Release 1.5.0

After half a year of hard work, VitastorFS is finally here ! :-)

New features:
- VitastorFS, a full-featured clustered (read-write-many) file system.
  Documentation: [VitastorFS](docs/usage/nfs.en.md)
- Embedded key-value database implementation based on Parallel Optimistic B-Tree
  algorithm and used for the metadata of VitastorFS
- Pool management commands in vitastor-cli (create-pool, list-pools, rm-pool, modify-pool).
  Thanks MIND Software (https://mindsw.io) for their contribution!
  [Documentation](docs/usage/cli.en.md#create-pool)

Bug fixes:
- Fix a very rare "infinite loop" in the client library
- Fix a rare OSD hang on during start when zeroing out bad metadata entries left from the previous run

.gitea/workflows/buildenv.Dockerfile

@@ -22,7 +22,7 @@ RUN apt-get update
 RUN apt-get -y install etcd qemu-system-x86 qemu-block-extra qemu-utils fio libasan5 \
     liburing1 liburing-dev libgoogle-perftools-dev devscripts libjerasure-dev cmake libibverbs-dev libisal-dev
 RUN apt-get -y build-dep fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`
-RUN apt-get -y install jq lp-solve sudo
+RUN apt-get -y install jq lp-solve sudo nfs-common
 RUN apt-get --download-only source fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`
 
 RUN set -ex; \

.gitea/workflows/test.yml

@@ -856,3 +856,21 @@ jobs:
           echo ""
         done
 
+  test_nfs:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /root/vitastor/tests/test_nfs.sh
+    - name: Print logs
+      if: always() && steps.test.outcome == 'failure'
+      run: |
+        for i in /root/vitastor/testdata/*.log /root/vitastor/testdata/*.txt; do
+          echo "-------- $i --------"
+          cat $i
+          echo ""
+        done
+

CMakeLists.txt

@@ -2,6 +2,6 @@ cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
-set(VERSION "1.4.5")
+set(VERSION "1.5.0")
 
 add_subdirectory(src)

README-ru.md

@@ -6,8 +6,8 @@
 
 Вернём былую скорость кластерному блочному хранилищу!
 
-Vitastor - распределённая блочная SDS (программная СХД), прямой аналог Ceph RBD и
-внутренних СХД популярных облачных провайдеров. Однако, в отличие от них, Vitastor
+Vitastor - распределённая блочная и файловая SDS (программная СХД), прямой аналог Ceph RBD и CephFS,
+а также внутренних СХД популярных облачных провайдеров. Однако, в отличие от них, Vitastor
 быстрый и при этом простой. Только пока маленький :-).
 
 Vitastor архитектурно похож на Ceph, что означает атомарность и строгую консистентность,
@@ -63,7 +63,7 @@ Vitastor поддерживает QEMU-драйвер, протоколы NBD и
   - [fio](docs/usage/fio.ru.md) для тестов производительности
   - [NBD](docs/usage/nbd.ru.md) для монтирования ядром
   - [QEMU и qemu-img](docs/usage/qemu.ru.md)
-  - [NFS](docs/usage/nfs.ru.md)-прокси для VMWare и подобных
+  - [NFS](docs/usage/nfs.ru.md) кластерная файловая система и псевдо-ФС прокси
 - Производительность
   - [Понимание сути производительности](docs/performance/understanding.ru.md)
   - [Теоретический максимум](docs/performance/theoretical.ru.md)

README.md

@@ -6,9 +6,9 @@
 
 Make Clustered Block Storage Fast Again.
 
-Vitastor is a distributed block SDS, direct replacement of Ceph RBD and internal SDS's
-of public clouds. However, in contrast to them, Vitastor is fast and simple at the same time.
-The only thing is it's slightly young :-).
+Vitastor is a distributed block and file SDS, direct replacement of Ceph RBD and CephFS,
+and also internal SDS's of public clouds. However, in contrast to them, Vitastor is fast
+and simple at the same time. The only thing is it's slightly young :-).
 
 Vitastor is architecturally similar to Ceph which means strong consistency,
 primary-replication, symmetric clustering and automatic data distribution over any
@@ -63,7 +63,7 @@ Read more details below in the documentation.
   - [fio](docs/usage/fio.en.md) for benchmarks
   - [NBD](docs/usage/nbd.en.md) for kernel mounts
   - [QEMU and qemu-img](docs/usage/qemu.en.md)
-  - [NFS](docs/usage/nfs.en.md) emulator for VMWare and similar
+  - [NFS](docs/usage/nfs.en.md) clustered file system and pseudo-FS proxy
 - Performance
   - [Understanding storage performance](docs/performance/understanding.en.md)
   - [Theoretical performance](docs/performance/theoretical.en.md)

csi/Makefile

@@ -1,4 +1,4 @@
-VERSION ?= v1.4.5
+VERSION ?= v1.5.0
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v1.4.5
+          image: vitalif/vitastor-csi:v1.5.0
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/deploy/007-csi-provisioner.yaml

@@ -121,7 +121,7 @@ spec:
             privileged: true
             capabilities:
               add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v1.4.5
+          image: vitalif/vitastor-csi:v1.5.0
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "1.4.5"
+    vitastorCSIDriverVersion = "1.5.0"
 )
 
 // Config struct fills the parameters of request or user input

debian/build-vitastor-bookworm.sh

@@ -3,5 +3,5 @@
 cat < vitastor.Dockerfile > ../Dockerfile
 cd ..
 mkdir -p packages
-sudo podman build --build-arg REL=bookworm -v `pwd`/packages:/root/packages -f Dockerfile .
+sudo podman build --build-arg DISTRO=debian --build-arg REL=bookworm -v `pwd`/packages:/root/packages -f Dockerfile .
 rm Dockerfile

debian/build-vitastor-bullseye.sh

@@ -3,5 +3,5 @@
 cat < vitastor.Dockerfile > ../Dockerfile
 cd ..
 mkdir -p packages
-sudo podman build --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f Dockerfile .
+sudo podman build --build-arg DISTRO=debian --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f Dockerfile .
 rm Dockerfile

debian/build-vitastor-buster.sh

@@ -3,5 +3,5 @@
 cat < vitastor.Dockerfile > ../Dockerfile
 cd ..
 mkdir -p packages
-sudo podman build --build-arg REL=buster -v `pwd`/packages:/root/packages -f Dockerfile .
+sudo podman build --build-arg DISTRO=debian --build-arg REL=buster -v `pwd`/packages:/root/packages -f Dockerfile .
 rm Dockerfile

debian/build-vitastor-ubuntu-jammy.sh

@@ -0,0 +1,7 @@
+#!/bin/bash
+
+cat < vitastor.Dockerfile > ../Dockerfile
+cd ..
+mkdir -p packages
+sudo podman build --build-arg DISTRO=ubuntu --build-arg REL=jammy -v `pwd`/packages:/root/packages -f Dockerfile .
+rm Dockerfile

debian/changelog

@@ -1,4 +1,4 @@
-vitastor (1.4.5-1) unstable; urgency=medium
+vitastor (1.5.0-1) unstable; urgency=medium
 
   * Bugfixes
 

debian/libvirt.Dockerfile

@@ -1,13 +1,14 @@
 # Build patched libvirt for Debian Buster or Bullseye/Sid inside a container
-# cd ..; podman build --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f debian/libvirt.Dockerfile .
+# cd ..; podman build --build-arg DISTRO=debian --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f debian/libvirt.Dockerfile .
 
+ARG DISTRO=
 ARG REL=
-FROM debian:$REL
+FROM $DISTRO:$REL
 ARG REL=
 
 WORKDIR /root
 
-RUN if [ "$REL" = "buster" -o "$REL" = "bullseye" ]; then \
+RUN if ([ "${DISTRO}" = "debian" ]) && ( [ "${REL}" = "buster" -o "${REL}" = "bullseye" ] ); then \
         echo "deb http://deb.debian.org/debian $REL-backports main" >> /etc/apt/sources.list; \
         echo >> /etc/apt/preferences; \
         echo 'Package: *' >> /etc/apt/preferences; \
@@ -23,7 +24,7 @@ RUN apt-get -y build-dep libvirt0
 RUN apt-get -y install libglusterfs-dev
 RUN apt-get --download-only source libvirt
 
-ADD patches/libvirt-5.0-vitastor.diff patches/libvirt-7.0-vitastor.diff patches/libvirt-7.5-vitastor.diff patches/libvirt-7.6-vitastor.diff /root
+ADD patches/libvirt-5.0-vitastor.diff patches/libvirt-7.0-vitastor.diff patches/libvirt-7.5-vitastor.diff patches/libvirt-7.6-vitastor.diff patches/libvirt-8.0-vitastor.diff /root
 RUN set -e; \
     mkdir -p /root/packages/libvirt-$REL; \
     rm -rf /root/packages/libvirt-$REL/*; \

debian/vitastor-client.install

@@ -3,4 +3,6 @@ usr/bin/vitastor-cli
 usr/bin/vitastor-rm
 usr/bin/vitastor-nbd
 usr/bin/vitastor-nfs
+usr/bin/vitastor-kv
+usr/bin/vitastor-kv-stress
 usr/lib/*/libvitastor*.so*

debian/vitastor.Dockerfile

@@ -1,8 +1,10 @@
 # Build Vitastor packages for Debian inside a container
-# cd ..; podman build --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f debian/vitastor.Dockerfile .
+# cd ..; podman build --build-arg DISTRO=debian --build-arg REL=bullseye -v `pwd`/packages:/root/packages -f debian/vitastor.Dockerfile .
 
+ARG DISTRO=debian
 ARG REL=
-FROM debian:$REL
+FROM $DISTRO:$REL
+ARG DISTRO=debian
 ARG REL=
 
 WORKDIR /root
@@ -35,8 +37,8 @@ RUN set -e -x; \
     mkdir -p /root/packages/vitastor-$REL; \
     rm -rf /root/packages/vitastor-$REL/*; \
     cd /root/packages/vitastor-$REL; \
-    cp -r /root/vitastor vitastor-1.4.5; \
-    cd vitastor-1.4.5; \
+    cp -r /root/vitastor vitastor-1.5.0; \
+    cd vitastor-1.5.0; \
     ln -s /root/fio-build/fio-*/ ./fio; \
     FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     ls /usr/include/linux/raw.h || cp ./debian/raw.h /usr/include/linux/raw.h; \
@@ -49,8 +51,8 @@ RUN set -e -x; \
     rm -rf a b; \
     echo "dep:fio=$FIO" > debian/fio_version; \
     cd /root/packages/vitastor-$REL; \
-    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_1.4.5.orig.tar.xz vitastor-1.4.5; \
-    cd vitastor-1.4.5; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_1.5.0.orig.tar.xz vitastor-1.5.0; \
+    cd vitastor-1.5.0; \
     V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
     DEBFULLNAME="Vitaliy Filippov <vitalif@yourcmc.ru>" dch -D $REL -v "$V""$REL" "Rebuild for $REL"; \
     DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage --jobs=auto -sa; \

docs/config/client.en.md

@@ -9,6 +9,8 @@
 These parameters apply only to Vitastor clients (QEMU, fio, NBD and so on) and
 affect their interaction with the cluster.
 
+- [client_retry_interval](#client_retry_interval)
+- [client_eio_retry_interval](#client_eio_retry_interval)
 - [client_max_dirty_bytes](#client_max_dirty_bytes)
 - [client_max_dirty_ops](#client_max_dirty_ops)
 - [client_enable_writeback](#client_enable_writeback)
@@ -19,6 +21,26 @@ affect their interaction with the cluster.
 - [nbd_max_devices](#nbd_max_devices)
 - [nbd_max_part](#nbd_max_part)
 
+## client_retry_interval
+
+- Type: milliseconds
+- Default: 50
+- Minimum: 10
+- Can be changed online: yes
+
+Retry time for I/O requests failed due to inactive PGs or network
+connectivity errors.
+
+## client_eio_retry_interval
+
+- Type: milliseconds
+- Default: 1000
+- Can be changed online: yes
+
+Retry time for I/O requests failed due to data corruption or unfinished
+EC object deletions (has_incomplete PG state). 0 disables such retries
+and clients are not blocked and just get EIO error code instead.
+
 ## client_max_dirty_bytes
 
 - Type: integer

docs/config/client.ru.md

@@ -9,6 +9,8 @@
 Данные параметры применяются только к клиентам Vitastor (QEMU, fio, NBD и т.п.) и
 затрагивают логику их работы с кластером.
 
+- [client_retry_interval](#client_retry_interval)
+- [client_eio_retry_interval](#client_eio_retry_interval)
 - [client_max_dirty_bytes](#client_max_dirty_bytes)
 - [client_max_dirty_ops](#client_max_dirty_ops)
 - [client_enable_writeback](#client_enable_writeback)
@@ -19,6 +21,27 @@
 - [nbd_max_devices](#nbd_max_devices)
 - [nbd_max_part](#nbd_max_part)
 
+## client_retry_interval
+
+- Тип: миллисекунды
+- Значение по умолчанию: 50
+- Минимальное значение: 10
+- Можно менять на лету: да
+
+Время повтора запросов ввода-вывода, неудачных из-за неактивных PG или
+ошибок сети.
+
+## client_eio_retry_interval
+
+- Тип: миллисекунды
+- Значение по умолчанию: 1000
+- Можно менять на лету: да
+
+Время повтора запросов ввода-вывода, неудачных из-за повреждения данных
+или незавершённых удалений EC-объектов (состояния PG has_incomplete).
+0 отключает повторы таких запросов и клиенты не блокируются, а вместо
+этого просто получают код ошибки EIO.
+
 ## client_max_dirty_bytes
 
 - Тип: целое число

docs/config/network.en.md

@@ -25,7 +25,6 @@ between clients, OSDs and etcd.
 - [peer_connect_timeout](#peer_connect_timeout)
 - [osd_idle_timeout](#osd_idle_timeout)
 - [osd_ping_timeout](#osd_ping_timeout)
-- [up_wait_retry_interval](#up_wait_retry_interval)
 - [max_etcd_attempts](#max_etcd_attempts)
 - [etcd_quick_timeout](#etcd_quick_timeout)
 - [etcd_slow_timeout](#etcd_slow_timeout)
@@ -212,17 +211,6 @@ Maximum time to wait for OSD keepalive responses. If an OSD doesn't respond
 within this time, the connection to it is dropped and a reconnection attempt
 is scheduled.
 
-## up_wait_retry_interval
-
-- Type: milliseconds
-- Default: 50
-- Minimum: 10
-- Can be changed online: yes
-
-OSDs respond to clients with a special error code when they receive I/O
-requests for a PG that's not synchronized and started. This parameter sets
-the time for the clients to wait before re-attempting such I/O requests.
-
 ## max_etcd_attempts
 
 - Type: integer

docs/config/network.ru.md

@@ -25,7 +25,6 @@
 - [peer_connect_timeout](#peer_connect_timeout)
 - [osd_idle_timeout](#osd_idle_timeout)
 - [osd_ping_timeout](#osd_ping_timeout)
-- [up_wait_retry_interval](#up_wait_retry_interval)
 - [max_etcd_attempts](#max_etcd_attempts)
 - [etcd_quick_timeout](#etcd_quick_timeout)
 - [etcd_slow_timeout](#etcd_slow_timeout)
@@ -221,19 +220,6 @@ OSD в любом случае согласовывают реальное зн
 Если OSD не отвечает за это время, соединение отключается и производится
 повторная попытка соединения.
 
-## up_wait_retry_interval
-
-- Тип: миллисекунды
-- Значение по умолчанию: 50
-- Минимальное значение: 10
-- Можно менять на лету: да
-
-Когда OSD получают от клиентов запросы ввода-вывода, относящиеся к не
-поднятым на данный момент на них PG, либо к PG в процессе синхронизации,
-они отвечают клиентам специальным кодом ошибки, означающим, что клиент
-должен некоторое время подождать перед повторением запроса. Именно это время
-ожидания задаёт данный параметр.
-
 ## max_etcd_attempts
 
 - Тип: целое число

docs/config/pool.en.md

@@ -41,6 +41,7 @@ Parameters:
 - [osd_tags](#osd_tags)
 - [primary_affinity_tags](#primary_affinity_tags)
 - [scrub_interval](#scrub_interval)
+- [used_for_fs](#used_for_fs)
 
 Examples:
 
@@ -154,8 +155,25 @@ That is, if it becomes impossible to place PG data on at least (pg_minsize)
 OSDs, PG is deactivated for both read and write. So you know that a fresh
 write always goes to at least (pg_minsize) OSDs (disks).
 
-That is, pg_size minus pg_minsize sets the number of disk failures to tolerate
-without temporary downtime (for [osd_out_time](monitor.en.md#osd_out_time)).
+For example, the difference between pg_minsize 2 and 1 in a 3-way replicated
+pool (pg_size=3) is:
+- If 2 hosts go down with pg_minsize=2, the pool becomes inactive and remains
+  inactive for [osd_out_time](monitor.en.md#osd_out_time) (10 minutes). After
+  this timeout, the monitor selects replacement hosts/OSDs and the pool comes
+  up and starts to heal. Therefore, if you don't have replacement OSDs, i.e.
+  if you only have 3 hosts with OSDs and 2 of them are down, the pool remains
+  inactive until you add or return at least 1 host (or change failure_domain
+  to "osd").
+- If 2 hosts go down with pg_minsize=1, the pool only experiences a short
+  I/O pause until the monitor notices that OSDs are down (5-10 seconds with
+  the default [etcd_report_interval](osd.en.md#etcd_report_interval)). After
+  this pause, I/O resumes, but new data is temporarily written in only 1 copy.
+  Then, after osd_out_time, the monitor also selects replacement OSDs and the
+  pool starts to heal.
+
+So, pg_minsize regulates the number of failures that a pool can tolerate
+without temporary downtime for [osd_out_time](monitor.en.md#osd_out_time),
+but at a cost of slightly reduced storage reliability.
 
 FIXME: pg_minsize behaviour may be changed in the future to only make PGs
 read-only instead of deactivating them.
@@ -168,8 +186,8 @@ read-only instead of deactivating them.
 Number of PGs for this pool. The value should be big enough for the monitor /
 LP solver to be able to optimize data placement.
 
-"Enough" is usually around 64-128 PGs per OSD, i.e. you set pg_count for pool
-to (total OSD count * 100 / pg_size). You can round it to the closest power of 2,
+"Enough" is usually around 10-100 PGs per OSD, i.e. you set pg_count for pool
+to (total OSD count * 10 / pg_size). You can round it to the closest power of 2,
 because it makes it easier to reduce or increase PG count later by dividing or
 multiplying it by 2.
 
@@ -282,6 +300,25 @@ of the OSDs containing a data chunk for a PG.
 Automatic scrubbing interval for this pool. Overrides
 [global scrub_interval setting](osd.en.md#scrub_interval).
 
+## used_for_fs
+
+- Type: string
+
+If non-empty, the pool is marked as used for VitastorFS with metadata stored
+in block image (regular Vitastor volume) named as the value of this pool parameter.
+
+When a pool is marked as used for VitastorFS, regular block volume creation in it
+is disabled (vitastor-cli refuses to create images without --force) to protect
+the user from block volume and FS file ID collisions and data loss.
+
+[vitastor-nfs](../usage/nfs.ru.md), in its turn, refuses to use pools not marked
+for the corresponding FS when starting. This also implies that you can use one
+pool only for one VitastorFS.
+
+The second thing that is disabled for VitastorFS pools is reporting per-inode space
+usage statistics in etcd because a FS pool may store a very large number of files
+and statistics for them all would take a lot of space in etcd.
+
 # Examples
 
 ## Replicated pool

docs/config/pool.ru.md

@@ -40,6 +40,7 @@
 - [osd_tags](#osd_tags)
 - [primary_affinity_tags](#primary_affinity_tags)
 - [scrub_interval](#scrub_interval)
+- [used_for_fs](#used_for_fs)
 
 Примеры:
 
@@ -157,9 +158,25 @@
 OSD, PG деактивируется на чтение и запись. Иными словами, всегда известно,
 что новые блоки данных всегда записываются как минимум на pg_minsize дисков.
 
-По сути, разница pg_size и pg_minsize задаёт число отказов дисков, которые пул
-может пережить без временной (на [osd_out_time](monitor.ru.md#osd_out_time))
-остановки обслуживания.
+Для примера, разница между pg_minsize 2 и 1 в реплицированном пуле с 3 копиями
+данных (pg_size=3), проявляется следующим образом:
+- Если 2 сервера отключаются при pg_minsize=2, пул становится неактивным и
+  остаётся неактивным в течение [osd_out_time](monitor.en.md#osd_out_time)
+  (10 минут), после чего монитор назначает другие OSD/серверы на замену, пул
+  поднимается и начинает восстанавливать недостающие копии данных. Соответственно,
+  если OSD на замену нет - то есть, если у вас всего 3 сервера с OSD и 2 из них
+  недоступны - пул так и остаётся недоступным до тех пор, пока вы не вернёте
+  или не добавите хотя бы 1 сервер (или не переключите failure_domain на "osd").
+- Если 2 сервера отключаются при pg_minsize=1, ввод-вывод лишь приостанавливается
+  на короткое время, до тех пор, пока монитор не поймёт, что OSD отключены
+  (что занимает 5-10 секунд при стандартном [etcd_report_interval](osd.en.md#etcd_report_interval)).
+  После этого ввод-вывод восстанавливается, но новые данные временно пишутся
+  всего в 1 копии. Когда же проходит osd_out_time, монитор точно так же назначает
+  другие OSD на замену выбывшим и пул начинает восстанавливать копии данных.
+
+То есть, pg_minsize регулирует число отказов, которые пул может пережить без
+временной остановки обслуживания на [osd_out_time](monitor.ru.md#osd_out_time),
+но ценой немного пониженных гарантий надёжности.
 
 FIXME: Поведение pg_minsize может быть изменено в будущем с полной деактивации
 PG на перевод их в режим только для чтения.
@@ -172,8 +189,8 @@ PG на перевод их в режим только для чтения.
 Число PG для данного пула. Число должно быть достаточно большим, чтобы монитор
 мог равномерно распределить по ним данные.
 
-Обычно это означает примерно 64-128 PG на 1 OSD, т.е. pg_count можно устанавливать
-равным (общему числу OSD * 100 / pg_size). Значение можно округлить до ближайшей
+Обычно это означает примерно 10-100 PG на 1 OSD, т.е. pg_count можно устанавливать
+равным (общему числу OSD * 10 / pg_size). Значение можно округлить до ближайшей
 степени 2, чтобы потом было легче уменьшать или увеличивать число PG, умножая
 или деля его на 2.
 
@@ -290,6 +307,27 @@ OSD с "all".
 Интервал скраба, то есть, автоматической фоновой проверки данных для данного пула.
 Переопределяет [глобальную настройку scrub_interval](osd.ru.md#scrub_interval).
 
+## used_for_fs
+
+- Type: string
+
+Если непусто, пул помечается как используемый для файловой системы VitastorFS с
+метаданными, хранимыми в блочном образе Vitastor с именем, равным значению
+этого параметра.
+
+Когда пул помечается как используемый для VitastorFS, создание обычных блочных
+образов в нём отключается (vitastor-cli отказывается создавать образы без --force),
+чтобы защитить пользователя от коллизий ID файлов и блочных образов и, таким
+образом, от потери данных.
+
+[vitastor-nfs](../usage/nfs.ru.md), в свою очередь, при запуске отказывается
+использовать для ФС пулы, не выделенные для неё. Это также означает, что один
+пул может использоваться только для одной VitastorFS.
+
+Также для ФС-пулов отключается передача статистики в etcd по отдельным инодам,
+так как ФС-пул может содержать очень много файлов и статистика по ним всем
+заняла бы очень много места в etcd.
+
 # Примеры
 
 ## Реплицированный пул

docs/config/src/client.yml

@@ -1,3 +1,27 @@
+- name: client_retry_interval
+  type: ms
+  min: 10
+  default: 50
+  online: true
+  info: |
+    Retry time for I/O requests failed due to inactive PGs or network
+    connectivity errors.
+  info_ru: |
+    Время повтора запросов ввода-вывода, неудачных из-за неактивных PG или
+    ошибок сети.
+- name: client_eio_retry_interval
+  type: ms
+  default: 1000
+  online: true
+  info: |
+    Retry time for I/O requests failed due to data corruption or unfinished
+    EC object deletions (has_incomplete PG state). 0 disables such retries
+    and clients are not blocked and just get EIO error code instead.
+  info_ru: |
+    Время повтора запросов ввода-вывода, неудачных из-за повреждения данных
+    или незавершённых удалений EC-объектов (состояния PG has_incomplete).
+    0 отключает повторы таких запросов и клиенты не блокируются, а вместо
+    этого просто получают код ошибки EIO.
 - name: client_max_dirty_bytes
   type: int
   default: 33554432

docs/config/src/network.yml

@@ -243,21 +243,6 @@
     Максимальное время ожидания ответа на запрос проверки состояния соединения.
     Если OSD не отвечает за это время, соединение отключается и производится
     повторная попытка соединения.
-- name: up_wait_retry_interval
-  type: ms
-  min: 10
-  default: 50
-  online: true
-  info: |
-    OSDs respond to clients with a special error code when they receive I/O
-    requests for a PG that's not synchronized and started. This parameter sets
-    the time for the clients to wait before re-attempting such I/O requests.
-  info_ru: |
-    Когда OSD получают от клиентов запросы ввода-вывода, относящиеся к не
-    поднятым на данный момент на них PG, либо к PG в процессе синхронизации,
-    они отвечают клиентам специальным кодом ошибки, означающим, что клиент
-    должен некоторое время подождать перед повторением запроса. Именно это время
-    ожидания задаёт данный параметр.
 - name: max_etcd_attempts
   type: int
   default: 5

docs/intro/features.en.md

@@ -33,6 +33,7 @@
 - [Checksums](../config/layout-osd.en.md#data_csum_type)
 - [Client write-back cache](../config/client.en.md#client_enable_writeback)
 - [Intelligent recovery auto-tuning](../config/osd.en.md#recovery_tune_interval)
+- [Clustered file system](../usage/nfs.en.md#vitastorfs)
 
 ## Plugins and tools
 
@@ -46,13 +47,12 @@
 - [CSI plugin for Kubernetes](../installation/kubernetes.en.md)
 - [OpenStack support: Cinder driver, Nova and libvirt patches](../installation/openstack.en.md)
 - [Proxmox storage plugin and packages](../installation/proxmox.en.md)
-- [Simplified NFS proxy for file-based image access emulation (suitable for VMWare)](../usage/nfs.en.md)
+- [Simplified NFS proxy for file-based image access emulation (suitable for VMWare)](../usage/nfs.en.md#pseudo-fs)
 
 ## Roadmap
 
 The following features are planned for the future:
 
-- File system
 - Control plane optimisation
 - Other administrative tools
 - Web GUI

docs/intro/features.ru.md

@@ -35,6 +35,7 @@
 - [Контрольные суммы](../config/layout-osd.ru.md#data_csum_type)
 - [Буферизация записи на стороне клиента](../config/client.ru.md#client_enable_writeback)
 - [Интеллектуальная автоподстройка скорости восстановления](../config/osd.ru.md#recovery_tune_interval)
+- [Кластерная файловая система](../usage/nfs.ru.md#vitastorfs)
 
 ## Драйверы и инструменты
 
@@ -48,11 +49,10 @@
 - [CSI-плагин для Kubernetes](../installation/kubernetes.ru.md)
 - [Базовая поддержка OpenStack: драйвер Cinder, патчи для Nova и libvirt](../installation/openstack.ru.md)
 - [Плагин для Proxmox](../installation/proxmox.ru.md)
-- [Упрощённая NFS-прокси для эмуляции файлового доступа к образам (подходит для VMWare)](../usage/nfs.ru.md)
+- [Упрощённая NFS-прокси для эмуляции файлового доступа к образам (подходит для VMWare)](../usage/nfs.ru.md#псевдо-фс)
 
 ## Планы развития
 
-- Файловая система
 - Оптимизация слоя управления
 - Другие инструменты администрирования
 - Web-интерфейс

docs/intro/quickstart.en.md

@@ -14,6 +14,7 @@
 - [Check cluster status](#check-cluster-status)
 - [Create an image](#create-an-image)
 - [Install plugins](#install-plugins)
+- [Create VitastorFS](#create-vitastorfs)
 
 ## Preparation
 
@@ -75,18 +76,16 @@ On the monitor hosts:
 
 ## Create a pool
 
-Create pool configuration in etcd:
+Create a pool using vitastor-cli:
 
 ```
-etcdctl --endpoints=... put /vitastor/config/pools '{"1":{"name":"testpool",
-  "scheme":"replicated","pg_size":2,"pg_minsize":1,"pg_count":256,"failure_domain":"host"}}'
+vitastor-cli create-pool testpool --pg_size 2 --pg_count 256
 ```
 
 For EC pools the configuration should look like the following:
 
 ```
-etcdctl --endpoints=... put /vitastor/config/pools '{"2":{"name":"ecpool",
-  "scheme":"ec","pg_size":4,"parity_chunks":2,"pg_minsize":2,"pg_count":256,"failure_domain":"host"}}'
+vitastor-cli create-pool testpool --ec 2+2 --pg_count 256
 ```
 
 After you do this, one of the monitors will configure PGs and OSDs will start them.
@@ -116,3 +115,9 @@ After that, you can [run benchmarks](../usage/fio.en.md) or [start QEMU manually
 - [Proxmox](../installation/proxmox.en.md)
 - [OpenStack](../installation/openstack.en.md)
 - [Kubernetes CSI](../installation/kubernetes.en.md)
+
+## Create VitastorFS
+
+If you want to use clustered file system in addition to VM or container images:
+
+- [Follow the instructions here](../usage/nfs.en.md#vitastorfs)

docs/intro/quickstart.ru.md

@@ -14,6 +14,7 @@
 - [Проверьте состояние кластера](#проверьте-состояние-кластера)
 - [Создайте образ](#создайте-образ)
 - [Установите плагины](#установите-плагины)
+- [Создайте VitastorFS](#создайте-vitastorfs)
 
 ## Подготовка
 
@@ -77,18 +78,16 @@
 
 ## Создайте пул
 
-Создайте конфигурацию пула с помощью etcdctl:
+Создайте пул с помощью vitastor-cli:
 
 ```
-etcdctl --endpoints=... put /vitastor/config/pools '{"1":{"name":"testpool",
-  "scheme":"replicated","pg_size":2,"pg_minsize":1,"pg_count":256,"failure_domain":"host"}}'
+vitastor-cli create-pool testpool --pg_size 2 --pg_count 256
 ```
 
 Для пулов с кодами коррекции ошибок конфигурация должна выглядеть примерно так:
 
 ```
-etcdctl --endpoints=... put /vitastor/config/pools '{"2":{"name":"ecpool",
-  "scheme":"ec","pg_size":4,"parity_chunks":2,"pg_minsize":2,"pg_count":256,"failure_domain":"host"}}'
+vitastor-cli create-pool testpool --ec 2+2 --pg_count 256
 ```
 
 После этого один из мониторов должен сконфигурировать PG, а OSD должны запустить их.
@@ -118,3 +117,10 @@ vitastor-cli create -s 10G testimg
 - [Proxmox](../installation/proxmox.ru.md)
 - [OpenStack](../installation/openstack.ru.md)
 - [Kubernetes CSI](../installation/kubernetes.ru.md)
+
+## Создайте VitastorFS
+
+Если вы хотите использовать не только блочные образы виртуальных машин или контейнеров,
+а также кластерную файловую систему, то:
+
+- [Следуйте инструкциям](../usage/nfs.en.md#vitastorfs)

docs/usage/cli.en.md

@@ -24,6 +24,10 @@ It supports the following commands:
 - [fix](#fix)
 - [alloc-osd](#alloc-osd)
 - [rm-osd](#rm-osd)
+- [create-pool](#create-pool)
+- [modify-pool](#modify-pool)
+- [ls-pools](#ls-pools)
+- [rm-pool](#rm-pool)
 
 Global options:
 
@@ -131,19 +135,18 @@ See also about [how to export snapshots](qemu.en.md#exporting-snapshots).
 
 ## modify
 
-`vitastor-cli modify <name> [--rename <new-name>] [--resize <size>] [--readonly | --readwrite] [-f|--force]`
+`vitastor-cli modify <name> [--rename <new-name>] [--resize <size>] [--readonly | --readwrite] [-f|--force] [--down-ok]`
 
 Rename, resize image or change its readonly status. Images with children can't be made read-write.
 If the new size is smaller than the old size, extra data will be purged.
 You should resize file system in the image, if present, before shrinking it.
 
-```
--f|--force  Proceed with shrinking or setting readwrite flag even if the image has children.
-```
+* `-f|--force` - Proceed with shrinking or setting readwrite flag even if the image has children.
+* `--down-ok` - Proceed with shrinking even if some data will be left on unavailable OSDs.
 
 ## rm
 
-`vitastor-cli rm <from> [<to>] [--writers-stopped]`
+`vitastor-cli rm <from> [<to>] [--writers-stopped] [--down-ok]`
 
 Remove `<from>` or all layers between `<from>` and `<to>` (`<to>` must be a child of `<from>`),
 rebasing all their children accordingly. --writers-stopped allows merging to be a bit
@@ -151,6 +154,10 @@ more effective in case of a single 'slim' read-write child and 'fat' removed par
 the child is merged into parent and parent is renamed to child in that case.
 In other cases parent layers are always merged into children.
 
+Other options:
+
+* `--down-ok` - Continue deletion/merging even if some data will be left on unavailable OSDs.
+
 ## flatten
 
 `vitastor-cli flatten <layer>`
@@ -238,3 +245,91 @@ Refuses to remove OSDs with data without `--force` and `--allow-data-loss`.
 
 With `--dry-run` only checks if deletion is possible without data loss and
 redundancy degradation.
+
+## create-pool
+
+`vitastor-cli create-pool|pool-create <name> (-s <pg_size>|--ec <N>+<K>) -n <pg_count> [OPTIONS]`
+
+Create a pool. Required parameters:
+
+| <!-- -->                 | <!-- -->                                                                              |
+|--------------------------|---------------------------------------------------------------------------------------|
+| `-s R` or `--pg_size R`  | Number of replicas for replicated pools                                               |
+| `--ec N+K`               | Number of data (N) and parity (K) chunks for erasure-coded pools                      |
+| `-n N` or `--pg_count N` | PG count for the new pool (start with 10*<OSD count>/pg_size rounded to a power of 2) |
+
+Optional parameters:
+
+| <!-- -->                       | <!-- -->                                                                   |
+|--------------------------------|----------------------------------------------------------------------------|
+| `--pg_minsize <number>`        | R or N+K minus number of failures to tolerate without downtime ([details](../config/pool.en.md#pg_minsize)) |
+| `--failure_domain host`        | Failure domain: host, osd or a level from placement_levels. Default: host  |
+| `--root_node <node>`           | Put pool only on child OSDs of this placement tree node                    |
+| `--osd_tags <tag>[,<tag>]...`  | Put pool only on OSDs tagged with all specified tags                       |
+| `--block_size 128k`            | Put pool only on OSDs with this data block size                            |
+| `--bitmap_granularity 4k`      | Put pool only on OSDs with this logical sector size                        |
+| `--immediate_commit none`      | Put pool only on OSDs with this or larger immediate_commit (none < small < all) |
+| `--primary_affinity_tags tags` | Prefer to put primary copies on OSDs with all specified tags               |
+| `--scrub_interval <time>`      | Enable regular scrubbing for this pool. Format: number + unit s/m/h/d/M/y  |
+| `--used_for_fs <name>`         | Mark pool as used for VitastorFS with metadata in image <name>             |
+| `--pg_stripe_size <number>`    | Increase object grouping stripe                                            |
+| `--max_osd_combinations 10000` | Maximum number of random combinations for LP solver input                  |
+| `--wait`                       | Wait for the new pool to come online                                       |
+| `-f` or `--force`              | Do not check that cluster has enough OSDs to create the pool               |
+
+See also [Pool configuration](../config/pool.en.md) for detailed parameter descriptions.
+
+Examples:
+
+`vitastor-cli create-pool test_x4 -s 4 -n 32`
+
+`vitastor-cli create-pool test_ec42 --ec 4+2 -n 32`
+
+## modify-pool
+
+`vitastor-cli modify-pool|pool-modify <id|name> [--name <new_name>] [PARAMETERS...]`
+
+Modify an existing pool. Modifiable parameters:
+
+```
+[-s|--pg_size <number>] [--pg_minsize <number>] [-n|--pg_count <count>]
+[--failure_domain <level>] [--root_node <node>] [--osd_tags <tags>] [--no_inode_stats 0|1]
+[--max_osd_combinations <number>] [--primary_affinity_tags <tags>] [--scrub_interval <time>]
+```
+
+Non-modifiable parameters (changing them WILL lead to data loss):
+
+```
+[--block_size <size>] [--bitmap_granularity <size>]
+[--immediate_commit <all|small|none>] [--pg_stripe_size <size>]
+```
+
+These, however, can still be modified with -f|--force.
+
+See [create-pool](#create-pool) for parameter descriptions.
+
+Examples:
+
+`vitastor-cli modify-pool pool_A --name pool_B`
+
+`vitastor-cli modify-pool 2 --pg_size 4 -n 128`
+
+## rm-pool
+
+`vitastor-cli rm-pool|pool-rm [--force] <id|name>`
+
+Remove a pool. Refuses to remove pools with images without `--force`.
+
+## ls-pools
+
+`vitastor-cli ls-pools|pool-ls|ls-pool|pools [-l] [--detail] [--sort FIELD] [-r] [-n N] [--stats] [<glob> ...]`
+
+List pools (only matching <glob> patterns if passed).
+
+| <!-- -->             | <!-- -->                                              |
+|----------------------|-------------------------------------------------------|
+| `-l` or `--long`     | Also report I/O statistics                            |
+| `--detail`           | Use list format (not table), show all details         |
+| `--sort FIELD`       | Sort by specified field (see fields in --json output) |
+| `-r` or `--reverse`  | Sort in descending order                              |
+| `-n` or `--count N`  | Only list first N items                               |

docs/usage/cli.ru.md

@@ -23,6 +23,10 @@ vitastor-cli - интерфейс командной строки для адм
 - [merge-data](#merge-data)
 - [alloc-osd](#alloc-osd)
 - [rm-osd](#rm-osd)
+- [create-pool](#create-pool)
+- [modify-pool](#modify-pool)
+- [ls-pools](#ls-pools)
+- [rm-pool](#rm-pool)
 
 Глобальные опции:
 
@@ -85,8 +89,8 @@ kaveri    2/1     32   0 B      10 G    0 B        100%    0%
 
 `vitastor-cli ls [-l] [-p POOL] [--sort FIELD] [-r] [-n N] [<glob> ...]`
 
-Показать список образов, если переданы шаблоны `<glob>`, то только с именами,
-соответствующими этим шаблонам (стандартные ФС-шаблоны с * и ?).
+Показать список образов, если передан(ы) шаблон(ы) `<glob>`, то только с именами,
+соответствующими одному из шаблонов (стандартные ФС-шаблоны с * и ?).
 
 Опции:
 
@@ -132,7 +136,7 @@ vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 
 ## modify
 
-`vitastor-cli modify <name> [--rename <new-name>] [--resize <size>] [--readonly | --readwrite] [-f|--force]`
+`vitastor-cli modify <name> [--rename <new-name>] [--resize <size>] [--readonly | --readwrite] [-f|--force] [--down-ok]`
 
 Изменить размер, имя образа или флаг "только для чтения". Снимать флаг "только для чтения"
 и уменьшать размер образов, у которых есть дочерние клоны, без `--force` нельзя.
@@ -140,13 +144,12 @@ vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 Если новый размер меньше старого, "лишние" данные будут удалены, поэтому перед уменьшением
 образа сначала уменьшите файловую систему в нём.
 
-```
--f|--force  Разрешить уменьшение или перевод в чтение-запись образа, у которого есть клоны.
-```
+* `-f|--force` - Разрешить уменьшение или перевод в чтение-запись образа, у которого есть клоны.
+* `--down-ok` - Разрешить уменьшение, даже если часть данных останется неудалённой на недоступных OSD.
 
 ## rm
 
-`vitastor-cli rm <from> [<to>] [--writers-stopped]`
+`vitastor-cli rm <from> [<to>] [--writers-stopped] [--down-ok]`
 
 Удалить образ `<from>` или все слои от `<from>` до `<to>` (`<to>` должен быть дочерним
 образом `<from>`), одновременно меняя родительские образы их клонов (если таковые есть).
@@ -158,6 +161,10 @@ vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 
 В других случаях родительские слои вливаются в дочерние.
 
+Другие опции:
+
+* `--down-ok` - Продолжать удаление/слияние, даже если часть данных останется неудалённой на недоступных OSD.
+
 ## flatten
 
 `vitastor-cli flatten <layer>`
@@ -255,3 +262,91 @@ vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 
 С опцией `--dry-run` только проверяет, возможно ли удаление без потери данных и деградации
 избыточности.
+
+## create-pool
+
+`vitastor-cli create-pool|pool-create <name> (-s <pg_size>|--ec <N>+<K>) -n <pg_count> [OPTIONS]`
+
+Создать пул. Обязательные параметры:
+
+| <!-- -->                  | <!-- -->                                                                                    |
+|---------------------------|---------------------------------------------------------------------------------------------|
+| `-s R` или `--pg_size R`  | Число копий данных для реплицированных пулов                                                |
+| `--ec N+K`                | Число частей данных (N) и чётности (K) для пулов с кодами коррекции ошибок                  |
+| `-n N` или `--pg_count N` | Число PG для нового пула (начните с 10*<число OSD>/pg_size, округлённого до степени двойки) |
+
+Необязательные параметры:
+
+| <!-- -->                       | <!-- -->                                                                   |
+|--------------------------------|----------------------------------------------------------------------------|
+| `--pg_minsize <number>`        | (R или N+K) минус число разрешённых отказов без остановки пула ([подробнее](../config/pool.ru.md#pg_minsize)) |
+| `--failure_domain host`        | Домен отказа: host, osd или другой из placement_levels. По умолчанию: host |
+| `--root_node <node>`           | Использовать для пула только дочерние OSD этого узла дерева размещения     |
+| `--osd_tags <tag>[,<tag>]...`  | ...только OSD со всеми заданными тегами                                    |
+| `--block_size 128k`            | ...только OSD с данным размером блока                                      |
+| `--bitmap_granularity 4k`      | ...только OSD с данным размером логического сектора                        |
+| `--immediate_commit none`      | ...только OSD с этим или большим immediate_commit (none < small < all)     |
+| `--primary_affinity_tags tags` | Предпочитать OSD со всеми данными тегами для роли первичных                |
+| `--scrub_interval <time>`      | Включить скрабы с заданным интервалом времени (число + единица s/m/h/d/M/y) |
+| `--pg_stripe_size <number>`    | Увеличить блок группировки объектов по PG                                  |
+| `--max_osd_combinations 10000` | Максимальное число случайных комбинаций OSD для ЛП-солвера                 |
+| `--wait`                       | Подождать, пока новый пул будет активирован                                |
+| `-f` или `--force`             | Не проверять, что в кластере достаточно доменов отказа для создания пула   |
+
+Подробно о параметрах см. [Конфигурация пулов](../config/pool.ru.md).
+
+Примеры:
+
+`vitastor-cli create-pool test_x4 -s 4 -n 32`
+
+`vitastor-cli create-pool test_ec42 --ec 4+2 -n 32`
+
+## modify-pool
+
+`vitastor-cli modify-pool|pool-modify <id|name> [--name <new_name>] [PARAMETERS...]`
+
+Изменить настройки существующего пула. Изменяемые параметры:
+
+```
+[-s|--pg_size <number>] [--pg_minsize <number>] [-n|--pg_count <count>]
+[--failure_domain <level>] [--root_node <node>] [--osd_tags <tags>]
+[--max_osd_combinations <number>] [--primary_affinity_tags <tags>] [--scrub_interval <time>]
+```
+
+Неизменяемые параметры (их изменение ПРИВЕДЁТ к потере данных):
+
+```
+[--block_size <size>] [--bitmap_granularity <size>]
+[--immediate_commit <all|small|none>] [--pg_stripe_size <size>]
+```
+
+Эти параметры можно изменить, только если явно передать опцию -f или --force.
+
+Описания параметров смотрите в [create-pool](#create-pool).
+
+Примеры:
+
+`vitastor-cli modify-pool pool_A --name pool_B`
+
+`vitastor-cli modify-pool 2 --pg_size 4 -n 128`
+
+## rm-pool
+
+`vitastor-cli rm-pool|pool-rm [--force] <id|name>`
+
+Удалить пул. Отказывается удалять пул, в котором ещё есть образы, без `--force`.
+
+## ls-pools
+
+`vitastor-cli ls-pools|pool-ls|ls-pool|pools [-l] [--detail] [--sort FIELD] [-r] [-n N] [--stats] [<glob> ...]`
+
+Показать список пулов. Если передан(ы) шаблон(ы) `<glob>`, то только с именами,
+соответствующими одному из шаблонов (стандартные ФС-шаблоны с * и ?).
+
+| <!-- -->              | <!-- -->                                                   |
+|-----------------------|------------------------------------------------------------|
+| `-l` или `--long`     | Вывести также статистику ввода-вывода                      |
+| `--detail`            | Максимально подробный вывод в виде списка (а не таблицы)   |
+| `--sort FIELD`        | Сортировать по заданному полю (поля см. в выводе с --json) |
+| `-r` или `--reverse`  | Сортировать в обратном порядке                             |
+| `-n` или `--count N`  | Выводить только первые N записей                           |

docs/usage/disk.en.md

@@ -261,7 +261,7 @@ Options (see also [Cluster-Wide Disk Layout Parameters](../config/layout-cluster
 ```
 --object_size 128k       Set blockstore block size
 --bitmap_granularity 4k  Set bitmap granularity
---journal_size 16M       Set journal size
+--journal_size 32M       Set journal size
 --data_csum_type none    Set data checksum type (crc32c or none)
 --csum_block_size 4k     Set data checksum block size
 --device_block_size 4k   Set device block size

docs/usage/disk.ru.md

@@ -267,7 +267,7 @@ OSD отключены fsync-и.
 ```
 --object_size 128k       Размер блока хранилища
 --bitmap_granularity 4k  Гранулярность битовых карт
---journal_size 16M       Размер журнала
+--journal_size 32M       Размер журнала
 --data_csum_type none    Задать тип контрольных сумм (crc32c или none)
 --csum_block_size 4k     Задать размер блока расчёта контрольных сумм
 --device_block_size 4k   Размер блока устройства

docs/usage/nfs.en.md

@@ -4,42 +4,150 @@
 
 [Читать на русском](nfs.ru.md)
 
-# NFS
+# VitastorFS and pseudo-FS
 
-Vitastor has a simplified NFS 3.0 proxy for file-based image access emulation. It's not
-suitable as a full-featured file system, at least because all file/image metadata is stored
-in etcd and kept in memory all the time - thus you can't put a lot of files in it.
+Vitastor has two file system implementations. Both can be used via `vitastor-nfs`.
 
-However, NFS proxy is totally fine as a method to provide VM image access and allows to
-plug Vitastor into, for example, VMWare. It's important to note that for VMWare it's a much
-better access method than iSCSI, because with iSCSI we'd have to put all VM images into one
-Vitastor image exported as a LUN to VMWare and formatted with VMFS. VMWare doesn't use VMFS
-over NFS.
+Commands:
+- [mount](#mount)
+- [start](#start)
 
-NFS proxy is stateless if you use immediate_commit=all mode (for SSD with capacitors or
-HDDs with disabled cache), so you can run multiple NFS proxies and use a network load
-balancer or any failover method you want to in that case.
+## Pseudo-FS
 
-vitastor-nfs usage:
+Simplified pseudo-FS proxy is used for file-based image access emulation. It's not
+suitable as a full-featured file system: it lacks a lot of FS features, it stores
+all file/image metadata in memory and in etcd. So it's fine for hundreds or thousands
+of large files/images, but not for millions.
+
+Pseudo-FS proxy is intended for environments where other block volume access methods
+can't be used or impose additional restrictions - for example, VMWare. NFS is better
+for VMWare than, for example, iSCSI, because with iSCSI, VMWare puts all VM images
+into one large shared block image in its own VMFS file system, and with NFS, VMWare
+doesn't use VMFS and puts each VM disk in a regular file which is equal to one
+Vitastor block image, just as originally intended.
+
+To use Vitastor pseudo-FS locally, run `vitastor-nfs mount --block /mnt/vita`.
+
+Also you can start the network server:
 
 ```
-vitastor-nfs [STANDARD OPTIONS] [OTHER OPTIONS]
-
---subdir <DIR>    export images prefixed <DIR>/ (default empty - export all images)
---portmap 0       do not listen on port 111 (portmap/rpcbind, requires root)
---bind <IP>       bind service to <IP> address (default 0.0.0.0)
---nfspath <PATH>  set NFS export path to <PATH> (default is /)
---port <PORT>     use port <PORT> for NFS services (default is 2049)
---pool <POOL>     use <POOL> as default pool for new files (images)
---foreground 1    stay in foreground, do not daemonize
+vitastor-nfs start --block --etcd_address 192.168.5.10:2379 --portmap 0 --port 2050 --pool testpool
 ```
 
-Example start and mount commands (etcd_address is optional):
+To mount the FS exported by this server, run:
 
 ```
-vitastor-nfs --etcd_address 192.168.5.10:2379 --portmap 0 --port 2050 --pool testpool
+mount server:/ /mnt/ -o port=2050,mountport=2050,nfsvers=3,soft,nolock,tcp
 ```
 
-```
-mount localhost:/ /mnt/ -o port=2050,mountport=2050,nfsvers=3,soft,nolock,tcp
-```
+## VitastorFS
+
+VitastorFS is a full-featured clustered (Read-Write-Many) file system. It supports most POSIX
+features like hierarchical organization, symbolic links, hard links, quick renames and so on.
+
+VitastorFS metadata is stored in a Parallel Optimistic B-Tree key-value database,
+implemented over a regular Vitastor block volume. Directory entries and inodes
+are stored in a simple human-readable JSON format in the B-Tree. `vitastor-kv` tool
+can be used to inspect the database.
+
+To use VitastorFS:
+
+1. Create a pool or choose an existing empty pool for FS data
+2. Create an image for FS metadata, preferably in a faster (SSD or replica-HDD) pool,
+   but you can create it in the data pool too if you want (image size doesn't matter):
+   `vitastor-cli create -s 10G -p fastpool testfs`
+3. Mark data pool as an FS pool: `vitastor-cli modify-pool --used-for-fs testfs data-pool`
+4. Either mount the FS: `vitastor-nfs mount --fs testfs --pool data-pool /mnt/vita`
+5. Or start the NFS server: `vitastor-nfs start --fs testfs --pool data-pool`
+
+### Supported POSIX features
+
+- Read-after-write semantics (read returns new data immediately after write)
+- Linear and random read and write
+- Writing outside current file size
+- Hierarchical structure, immediate rename of files and directories
+- File size change support (truncate)
+- Permissions (chmod/chown)
+- Flushing data to stable storage (if required) (fsync)
+- Symbolic links
+- Hard links
+- Special files (devices, sockets, named pipes)
+- File modification and attribute change time tracking (mtime and ctime)
+- Modification time (mtime) and last access time (atime) change support (utimes)
+- Correct handling of directory listing during file creation/deletion
+
+### Limitations
+
+POSIX features currently not implemented in VitastorFS:
+- File locking is not supported
+- Actually used space is not counted, so `du` always reports apparent file sizes
+  instead of actually allocated space
+- Access times (`atime`) are not tracked (like `-o noatime`)
+- Modification time (`mtime`) is updated lazily every second (like `-o lazytime`)
+
+Other notable missing features which should be addressed in the future:
+- Defragmentation of "shared" inodes. Files smaller than pool object size (block_size
+  multiplied by data part count if pool is EC) are internally stored in large block
+  volumes sequentially, one after another, and leave garbage after deleting or resizing.
+  Defragmentator will be implemented to collect this garbage.
+- Inode ID reuse. Currently inode IDs always grow, the limit is 2^48 inodes, so
+  in theory you may hit it if you create and delete a very large number of files
+- Compaction of the key-value B-Tree. Current implementation never merges or deletes
+  B-Tree blocks, so B-Tree may become bloated over time. Currently you can
+  use `vitastor-kv dumpjson` & `loadjson` commands to recreate the index in such
+  situations.
+- Filesystem check tool. VitastorFS doesn't have journal because it would impose a
+  severe performance hit, optimistic CAS-based transactions are used instead of it.
+  So, again, in theory an abnormal shutdown of the FS server may leave some garbage
+  in the DB. The FS is implemented is such way that this garbage doesn't affect its
+  function, but having a tool to clean it up still seems a right thing to do.
+
+## Horizontal scaling
+
+Linux NFS 3.0 client doesn't support built-in scaling or failover, i.e. you can't
+specify multiple server addresses when mounting the FS.
+
+However, you can use any regular TCP load balancing over multiple NFS servers.
+It's absolutely safe with `immediate_commit=all` and `client_enable_writeback=false`
+settings, because Vitastor NFS proxy doesn't keep uncommitted data in memory
+with these settings. But it may even work without `immediate_commit=all` because
+the Linux NFS client repeats all uncommitted writes if it loses the connection.
+
+## Commands
+
+### mount
+
+`vitastor-nfs (--fs <NAME> | --block) [-o <OPT>] mount <MOUNTPOINT>`
+
+Start local filesystem server and mount file system to <MOUNTPOINT>.
+
+Use regular `umount <MOUNTPOINT>` to unmount the FS.
+
+The server will be automatically stopped when the FS is unmounted.
+
+- `-o|--options <OPT>` - Pass additional NFS mount options (ex.: -o async).
+
+### start
+
+`vitastor-nfs (--fs <NAME> | --block) start`
+
+Start network NFS server. Options:
+
+| <!-- -->        | <!-- -->                                                   |
+|-----------------|------------------------------------------------------------|
+| `--bind <IP>`   | bind service to \<IP> address (default 0.0.0.0)            |
+| `--port <PORT>` | use port \<PORT> for NFS services (default is 2049)        |
+| `--portmap 0`   | do not listen on port 111 (portmap/rpcbind, requires root) |
+
+## Common options
+
+| <!-- -->           | <!-- -->                                                 |
+|--------------------|----------------------------------------------------------|
+| `--fs <NAME>`      | use VitastorFS with metadata in image \<NAME>            |
+| `--block`          | use pseudo-FS presenting images as files                 |
+| `--pool <POOL>`    | use \<POOL> as default pool for new files                |
+| `--subdir <DIR>`   | export \<DIR> instead of root directory (pseudo-FS only) |
+| `--nfspath <PATH>` | set NFS export path to \<PATH> (default is /)            |
+| `--pidfile <FILE>` | write process ID to the specified file                   |
+| `--logfile <FILE>` | log to the specified file                                |
+| `--foreground 1`   | stay in foreground, do not daemonize                     |

docs/usage/nfs.ru.md

@@ -4,41 +4,156 @@
 
 [Read in English](nfs.en.md)
 
-# NFS
+# VitastorFS и псевдо-ФС
 
-В Vitastor реализована упрощённая NFS 3.0 прокси для эмуляции файлового доступа к образам.
-Это не полноценная файловая система, т.к. метаданные всех файлов (образов) сохраняются
-в etcd и всё время хранятся в оперативной памяти - то есть, положить туда много файлов
-не получится.
+В Vitastor есть две реализации файловой системы. Обе используются через `vitastor-nfs`.
 
-Однако в качестве способа доступа к образам виртуальных машин NFS прокси прекрасно подходит
-и позволяет подключить Vitastor, например, к VMWare.
+Команды:
+- [mount](#mount)
+- [start](#start)
 
-При этом, если вы используете режим immediate_commit=all (для SSD с конденсаторами или HDD
-с отключённым кэшем), то NFS-сервер не имеет состояния и вы можете свободно поднять
-его в нескольких экземплярах и использовать поверх них сетевой балансировщик нагрузки или
-схему с отказоустойчивостью.
+## Псевдо-ФС
 
-Использование vitastor-nfs:
+Упрощённая реализация псевдо-ФС используется для эмуляции файлового доступа к блочным
+образам Vitastor. Это не полноценная файловая система - в ней отсутствуют многие функции
+POSIX ФС, а метаданные всех файлов (образов) сохраняются в etcd и всё время хранятся в
+оперативной памяти - то есть, псевдо-ФС подходит для сотен или тысяч файлов, но не миллионов.
+
+Псевдо-ФС предназначена для доступа к образам виртуальных машин в средах, где другие
+способы невозможны или неудобны - например, в VMWare. Для VMWare это лучшая опция, чем
+iSCSI, так как при использовании iSCSI VMWare размещает все виртуальные машины в одном
+большом блочном образе внутри собственной ФС VMFS, а с NFS VMFS не используется и каждый
+диск ВМ представляется в виде одного файла, то есть, соответствует одному блочному образу
+Vitastor, как это и задумано изначально.
+
+Чтобы подключить псевдо-ФС Vitastor, выполните команду `vitastor-nfs mount --block /mnt/vita`.
+
+Либо же запустите сетевой вариант сервера:
 
 ```
-vitastor-nfs [СТАНДАРТНЫЕ ОПЦИИ] [ДРУГИЕ ОПЦИИ]
-
---subdir <DIR>    экспортировать "поддиректорию" - образы с префиксом имени <DIR>/ (по умолчанию пусто - экспортировать все образы)
---portmap 0       отключить сервис portmap/rpcbind на порту 111 (по умолчанию включён и требует root привилегий)
---bind <IP>       принимать соединения по адресу <IP> (по умолчанию 0.0.0.0 - на всех)
---nfspath <PATH>  установить путь NFS-экспорта в <PATH> (по умолчанию /)
---port <PORT>     использовать порт <PORT> для NFS-сервисов (по умолчанию 2049)
---pool <POOL>     использовать пул <POOL> для новых образов (обязательно, если пул в кластере не один)
---foreground 1    не уходить в фон после запуска
+vitastor-nfs start --block --etcd_address 192.168.5.10:2379 --portmap 0 --port 2050 --pool testpool
 ```
 
-Пример монтирования Vitastor через NFS (etcd_address необязателен):
+Примонтировать ФС, запущенную с такими опциями, можно следующей командой:
 
 ```
-vitastor-nfs --etcd_address 192.168.5.10:2379 --portmap 0 --port 2050 --pool testpool
+mount server:/ /mnt/ -o port=2050,mountport=2050,nfsvers=3,soft,nolock,tcp
 ```
 
-```
-mount localhost:/ /mnt/ -o port=2050,mountport=2050,nfsvers=3,soft,nolock,tcp
-```
+## VitastorFS
+
+VitastorFS - полноценная кластерная (Read-Write-Many) файловая система. Она поддерживает
+большую часть функций POSIX - иерархическую организацию, символические ссылки, жёсткие
+ссылки, быстрые переименования и так далее.
+
+Метаданные VitastorFS хранятся в собственной реализации БД формата ключ-значения,
+основанной на Параллельном Оптимистичном Б-дереве поверх обычного блочного образа Vitastor.
+И записи каталогов, и иноды, как обычно в Vitastor, хранятся в простом человекочитаемом
+JSON-формате :-). Для инспекции содержимого БД можно использовать инструмент `vitastor-kv`.
+
+Чтобы использовать VitastorFS:
+
+1. Создайте пул для данных ФС или выберите существующий пустой пул
+2. Создайте блочный образ для метаданных ФС, желательно, в более быстром пуле (на SSD
+   или по крайней мере на HDD, но без EC), но можно и в том же пуле, что данные
+   (размер образа значения не имеет):
+   `vitastor-cli create -s 10G -p fastpool testfs`
+3. Пометьте пул данных как ФС-пул: `vitastor-cli modify-pool --used-for-fs testfs data-pool`
+4. Либо примонтируйте ФС: `vitastor-nfs mount --fs testfs --pool data-pool /mnt/vita`
+5. Либо запустите сетевой NFS-сервер: `vitastor-nfs start --fs testfs --pool data-pool`
+
+### Поддерживаемые функции POSIX
+
+- Чтение актуальной версии данных сразу после записи
+- Последовательное и произвольное чтение и запись
+- Запись за пределами текущего размера файла
+- Иерархическая организация, мгновенное переименование файлов и каталогов
+- Изменение размера файла (truncate)
+- Права на файлы (chmod/chown)
+- Фиксация данных на диски (когда необходимо) (fsync)
+- Символические ссылки
+- Жёсткие ссылки
+- Специальные файлы (устройства, сокеты, каналы)
+- Отслеживание времён модификации (mtime), изменения атрибутов (ctime)
+- Ручное изменение времён модификации (mtime), последнего доступа (atime)
+- Корректная обработка изменений списка файлов во время листинга
+
+### Ограничения
+
+Отсутствующие на данный момент в VitastorFS функции POSIX:
+- Блокировки файлов не поддерживаются
+- Фактически занятое файлами место не подсчитывается и не возвращается вызовами
+  stat(2), так что `du` всегда показывает сумму размеров файлов, а не фактически занятое место
+- Времена доступа (`atime`) не отслеживаются (как будто ФС смонтирована с `-o noatime`)
+- Времена модификации (`mtime`) отслеживаются асинхронно (как будто ФС смонтирована с `-o lazytime`)
+
+Другие недостающие функции, которые нужно добавить в будущем:
+- Дефрагментация "общих инодов". На уровне реализации ФС файлы, меньшие, чем размер
+  объекта пула (block_size умножить на число частей данных, если пул EC),
+  упаковываются друг за другом в большие "общие" иноды/тома. Если такие файлы удалять
+  или увеличивать, они перемещаются и оставляют за собой "мусор", вот тут-то и нужен
+  дефрагментатор.
+- Переиспользование номеров инодов. В текущей реализации номера инодов всё время
+  увеличиваются, так что в теории вы можете упереться в лимит, если насоздаёте
+  и наудаляете больше, чем 2^48 файлов.
+- Очистка места в Б-дереве метаданных. Текущая реализация никогда не сливает и не
+  удаляет блоки Б-дерева, так что в теории дерево может разростись и стать неоптимальным.
+  Если вы столкнётесь с такой ситуацией сейчас, вы можете решить её с помощью
+  команд `vitastor-kv dumpjson` и `loadjson` (т.е. пересоздав и загрузив обратно все метаданные ФС).
+- Инструмент проверки метаданных файловой системы. У VitastorFS нет журнала, так как
+  журнал бы сильно замедлил реализацию, вместо него используются оптимистичные
+  транзакции на основе CAS (сравнить-и-записать), и теоретически при нештатном
+  завершении сервера ФС в БД также могут оставаться неконсистентные "мусорные"
+  записи. ФС устроена так, что на работу они не влияют, но для порядка и их стоит
+  уметь подчищать.
+
+## Горизонтальное масштабирование
+
+Клиент Linux NFS 3.0 не поддерживает встроенное масштабирование или отказоустойчивость.
+То есть, вы не можете задать несколько адресов серверов при монтировании ФС.
+
+Однако вы можете использовать любые стандартные сетевые балансировщики нагрузки
+или схемы с отказоустойчивостью. Это точно безопасно при настройках `immediate_commit=all` и
+`client_enable_writeback=false`, так как с ними NFS-сервер Vitastor вообще не хранит
+в памяти ещё не зафиксированные на дисках данные; и вполне вероятно безопасно
+даже без `immediate_commit=all`, потому что NFS-клиент ядра Linux повторяет все
+незафиксированные запросы при потере соединения.
+
+## Команды
+
+### mount
+
+`vitastor-nfs (--fs <NAME> | --block) mount [-o <OPT>] <MOUNTPOINT>`
+
+Запустить локальный сервер и примонтировать ФС в директорию <MOUNTPOINT>.
+
+Чтобы отмонтировать ФС, используйте обычную команду `umount <MOUNTPOINT>`.
+
+Сервер автоматически останавливается при отмонтировании ФС.
+
+- `-o|--options <OPT>` - Передать дополнительные опции монтирования NFS (пример: -o async).
+
+### start
+
+`vitastor-nfs (--fs <NAME> | --block) start`
+
+Запустить сетевой NFS-сервер. Опции:
+
+| <!-- -->        | <!-- -->                                                              |
+|-----------------|-----------------------------------------------------------------------|
+| `--bind <IP>`   | принимать соединения по адресу \<IP> (по умолчанию 0.0.0.0 - на всех) |
+| `--port <PORT>` | использовать порт \<PORT> для NFS-сервисов (по умолчанию 2049)        |
+| `--portmap 0`   | отключить сервис portmap/rpcbind на порту 111 (по умолчанию включён и требует root привилегий) |
+
+## Общие опции
+
+| <!-- -->           | <!-- -->                                                |
+|--------------------|---------------------------------------------------------|
+| `--fs <NAME>`      | использовать VitastorFS с метаданными в образе \<NAME>  |
+| `--block`          | использовать псевдо-ФС для доступа к блочным образам    |
+| `--pool <POOL>`    | использовать пул \<POOL> для новых файлов (обязательно, если пул в кластере не один) |
+| `--subdir <DIR>`   | экспортировать подкаталог \<DIR>, а не корень (только для псевдо-ФС) |
+| `--nfspath <PATH>` | установить путь NFS-экспорта в \<PATH> (по умолчанию /) |
+| `--pidfile <FILE>` | записать ID процесса в заданный файл                    |
+| `--logfile <FILE>` | записывать логи в заданный файл                         |
+| `--foreground 1`   | не уходить в фон после запуска                          |

mon/mon.js

@@ -37,7 +37,7 @@ const etcd_allow = new RegExp('^'+[
     'pg/history/[1-9]\\d*/[1-9]\\d*',
     'pool/stats/[1-9]\\d*',
     'history/last_clean_pgs',
-    'inode/stats/[1-9]\\d*/[1-9]\\d*',
+    'inode/stats/[1-9]\\d*/\\d+',
     'pool/stats/[1-9]\\d*',
     'stats',
     'index/image/.*',
@@ -86,13 +86,14 @@ const etcd_tree = {
             client_max_buffered_bytes: 33554432,
             client_max_buffered_ops: 1024,
             client_max_writeback_iodepth: 256,
+            client_retry_interval: 50, // ms. min: 10
+            client_eio_retry_interval: 1000, // ms
             // client and osd - configurable online
             log_level: 0,
             peer_connect_interval: 5, // seconds. min: 1
             peer_connect_timeout: 5, // seconds. min: 1
             osd_idle_timeout: 5, // seconds. min: 1
             osd_ping_timeout: 5, // seconds. min: 1
-            up_wait_retry_interval: 50, // ms. min: 10
             max_etcd_attempts: 5,
             etcd_quick_timeout: 1000, // ms
             etcd_slow_timeout: 5000, // ms
@@ -1736,8 +1737,11 @@ class Mon
                 for (const inode_num in this.state.osd.space[osd_num][pool_id])
                 {
                     const u = BigInt(this.state.osd.space[osd_num][pool_id][inode_num]||0);
-                    inode_stats[pool_id][inode_num] = inode_stats[pool_id][inode_num] || inode_stub();
-                    inode_stats[pool_id][inode_num].raw_used += u;
+                    if (inode_num)
+                    {
+                        inode_stats[pool_id][inode_num] = inode_stats[pool_id][inode_num] || inode_stub();
+                        inode_stats[pool_id][inode_num].raw_used += u;
+                    }
                     this.state.pool.stats[pool_id].used_raw_tb += u;
                 }
             }

mon/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor-mon",
-  "version": "1.4.5",
+  "version": "1.5.0",
   "description": "Vitastor SDS monitor service",
   "main": "mon-main.js",
   "scripts": {

patches/cinder-vitastor.py

@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils
 
-VERSION = '1.4.5'
+VERSION = '1.5.0'
 
 LOG = logging.getLogger(__name__)
 

patches/libvirt-8.0-vitastor.diff

@@ -0,0 +1,692 @@
+commit d85024bd803b3b91f15578ed22de4ce31856626f
+Author: Vitaliy Filippov <vitalif@yourcmc.ru>
+Date:   Wed Jan 24 18:07:43 2024 +0300
+
+    Add Vitastor support
+
+diff --git a/docs/schemas/domaincommon.rng b/docs/schemas/domaincommon.rng
+index 7fa5c2b8b5..2d77f391e7 100644
+--- a/docs/schemas/domaincommon.rng
++++ b/docs/schemas/domaincommon.rng
+@@ -1898,6 +1898,35 @@
+     </element>
+   </define>
+ 
++  <define name="diskSourceNetworkProtocolVitastor">
++    <element name="source">
++      <interleave>
++        <attribute name="protocol">
++          <value>vitastor</value>
++        </attribute>
++        <ref name="diskSourceCommon"/>
++        <optional>
++          <attribute name="name"/>
++        </optional>
++        <optional>
++          <attribute name="query"/>
++        </optional>
++        <zeroOrMore>
++          <ref name="diskSourceNetworkHost"/>
++        </zeroOrMore>
++        <optional>
++          <element name="config">
++            <attribute name="file">
++              <ref name="absFilePath"/>
++            </attribute>
++            <empty/>
++          </element>
++        </optional>
++        <empty/>
++      </interleave>
++    </element>
++  </define>
++
+   <define name="diskSourceNetworkProtocolISCSI">
+     <element name="source">
+       <attribute name="protocol">
+@@ -2154,6 +2183,7 @@
+       <ref name="diskSourceNetworkProtocolSimple"/>
+       <ref name="diskSourceNetworkProtocolVxHS"/>
+       <ref name="diskSourceNetworkProtocolNFS"/>
++      <ref name="diskSourceNetworkProtocolVitastor"/>
+     </choice>
+   </define>
+ 
+diff --git a/include/libvirt/libvirt-storage.h b/include/libvirt/libvirt-storage.h
+index f89856b93e..a8cb9387e2 100644
+--- a/include/libvirt/libvirt-storage.h
++++ b/include/libvirt/libvirt-storage.h
+@@ -246,6 +246,7 @@ typedef enum {
+     VIR_CONNECT_LIST_STORAGE_POOLS_ZFS           = 1 << 17,
+     VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE      = 1 << 18,
+     VIR_CONNECT_LIST_STORAGE_POOLS_ISCSI_DIRECT  = 1 << 19,
++    VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR      = 1 << 20,
+ } virConnectListAllStoragePoolsFlags;
+ 
+ int                     virConnectListAllStoragePools(virConnectPtr conn,
+diff --git a/src/conf/domain_conf.c b/src/conf/domain_conf.c
+index 5691b8d2d5..6669e8451d 100644
+--- a/src/conf/domain_conf.c
++++ b/src/conf/domain_conf.c
+@@ -8293,7 +8293,8 @@ virDomainDiskSourceNetworkParse(xmlNodePtr node,
+     src->configFile = virXPathString("string(./config/@file)", ctxt);
+ 
+     if (src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTP ||
+-        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS)
++        src->protocol == VIR_STORAGE_NET_PROTOCOL_HTTPS ||
++        src->protocol == VIR_STORAGE_NET_PROTOCOL_VITASTOR)
+         src->query = virXMLPropString(node, "query");
+ 
+     if (virDomainStorageNetworkParseHosts(node, ctxt, &src->hosts, &src->nhosts) < 0)
+@@ -31267,6 +31268,7 @@ virDomainStorageSourceTranslateSourcePool(virStorageSource *src,
+ 
+     case VIR_STORAGE_POOL_MPATH:
+     case VIR_STORAGE_POOL_RBD:
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_SHEEPDOG:
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_LAST:
+diff --git a/src/conf/domain_validate.c b/src/conf/domain_validate.c
+index a4271f1247..621c1b7b31 100644
+--- a/src/conf/domain_validate.c
++++ b/src/conf/domain_validate.c
+@@ -508,7 +508,7 @@ virDomainDiskDefValidateSourceChainOne(const virStorageSource *src)
+         }
+     }
+ 
+-    /* internal snapshots and config files are currently supported only with rbd: */
++    /* internal snapshots are currently supported only with rbd: */
+     if (virStorageSourceGetActualType(src) != VIR_STORAGE_TYPE_NETWORK &&
+         src->protocol != VIR_STORAGE_NET_PROTOCOL_RBD) {
+         if (src->snapshot) {
+@@ -517,11 +517,15 @@ virDomainDiskDefValidateSourceChainOne(const virStorageSource *src)
+                              "only with 'rbd' disks"));
+             return -1;
+         }
+-
++    }
++    /* config files are currently supported only with rbd and vitastor: */
++    if (virStorageSourceGetActualType(src) != VIR_STORAGE_TYPE_NETWORK &&
++        src->protocol != VIR_STORAGE_NET_PROTOCOL_RBD &&
++        src->protocol != VIR_STORAGE_NET_PROTOCOL_VITASTOR) {
+         if (src->configFile) {
+             virReportError(VIR_ERR_XML_ERROR, "%s",
+                            _("<config> element is currently supported "
+-                             "only with 'rbd' disks"));
++                             "only with 'rbd' and 'vitastor' disks"));
+             return -1;
+         }
+     }
+diff --git a/src/conf/storage_conf.c b/src/conf/storage_conf.c
+index 6690d26ffd..2255df9d28 100644
+--- a/src/conf/storage_conf.c
++++ b/src/conf/storage_conf.c
+@@ -60,7 +60,7 @@ VIR_ENUM_IMPL(virStoragePool,
+               "logical", "disk", "iscsi",
+               "iscsi-direct", "scsi", "mpath",
+               "rbd", "sheepdog", "gluster",
+-              "zfs", "vstorage",
++              "zfs", "vstorage", "vitastor",
+ );
+ 
+ VIR_ENUM_IMPL(virStoragePoolFormatFileSystem,
+@@ -246,6 +246,18 @@ static virStoragePoolTypeInfo poolTypeInfo[] = {
+           .formatToString = virStorageFileFormatTypeToString,
+       }
+     },
++    {.poolType = VIR_STORAGE_POOL_VITASTOR,
++     .poolOptions = {
++         .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
++                   VIR_STORAGE_POOL_SOURCE_NETWORK |
++                   VIR_STORAGE_POOL_SOURCE_NAME),
++      },
++      .volOptions = {
++          .defaultFormat = VIR_STORAGE_FILE_RAW,
++          .formatFromString = virStorageVolumeFormatFromString,
++          .formatToString = virStorageFileFormatTypeToString,
++      }
++    },
+     {.poolType = VIR_STORAGE_POOL_SHEEPDOG,
+      .poolOptions = {
+          .flags = (VIR_STORAGE_POOL_SOURCE_HOST |
+@@ -546,6 +558,11 @@ virStoragePoolDefParseSource(xmlXPathContextPtr ctxt,
+                        _("element 'name' is mandatory for RBD pool"));
+         return -1;
+     }
++    if (pool_type == VIR_STORAGE_POOL_VITASTOR && source->name == NULL) {
++        virReportError(VIR_ERR_XML_ERROR, "%s",
++                       _("element 'name' is mandatory for Vitastor pool"));
++        return -1;
++    }
+ 
+     if (options->formatFromString) {
+         g_autofree char *format = NULL;
+@@ -1176,6 +1193,7 @@ virStoragePoolDefFormatBuf(virBuffer *buf,
+     /* RBD, Sheepdog, Gluster and Iscsi-direct devices are not local block devs nor
+      * files, so they don't have a target */
+     if (def->type != VIR_STORAGE_POOL_RBD &&
++        def->type != VIR_STORAGE_POOL_VITASTOR &&
+         def->type != VIR_STORAGE_POOL_SHEEPDOG &&
+         def->type != VIR_STORAGE_POOL_GLUSTER &&
+         def->type != VIR_STORAGE_POOL_ISCSI_DIRECT) {
+diff --git a/src/conf/storage_conf.h b/src/conf/storage_conf.h
+index aaecf138d6..97172db38b 100644
+--- a/src/conf/storage_conf.h
++++ b/src/conf/storage_conf.h
+@@ -106,6 +106,7 @@ typedef enum {
+     VIR_STORAGE_POOL_GLUSTER,  /* Gluster device */
+     VIR_STORAGE_POOL_ZFS,      /* ZFS */
+     VIR_STORAGE_POOL_VSTORAGE, /* Virtuozzo Storage */
++    VIR_STORAGE_POOL_VITASTOR, /* Vitastor */
+ 
+     VIR_STORAGE_POOL_LAST,
+ } virStoragePoolType;
+@@ -466,6 +467,7 @@ VIR_ENUM_DECL(virStoragePartedFs);
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SCSI     | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_MPATH    | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_RBD      | \
++                 VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER  | \
+                  VIR_CONNECT_LIST_STORAGE_POOLS_ZFS      | \
+diff --git a/src/conf/storage_source_conf.c b/src/conf/storage_source_conf.c
+index d42f715f26..29d8da3d10 100644
+--- a/src/conf/storage_source_conf.c
++++ b/src/conf/storage_source_conf.c
+@@ -86,6 +86,7 @@ VIR_ENUM_IMPL(virStorageNetProtocol,
+               "ssh",
+               "vxhs",
+               "nfs",
++              "vitastor",
+ );
+ 
+ 
+@@ -1265,6 +1266,7 @@ virStorageSourceNetworkDefaultPort(virStorageNetProtocol protocol)
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+             return 24007;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
+             /* we don't provide a default for RBD */
+             return 0;
+diff --git a/src/conf/storage_source_conf.h b/src/conf/storage_source_conf.h
+index c4a026881c..67568e9181 100644
+--- a/src/conf/storage_source_conf.h
++++ b/src/conf/storage_source_conf.h
+@@ -128,6 +128,7 @@ typedef enum {
+     VIR_STORAGE_NET_PROTOCOL_SSH,
+     VIR_STORAGE_NET_PROTOCOL_VXHS,
+     VIR_STORAGE_NET_PROTOCOL_NFS,
++    VIR_STORAGE_NET_PROTOCOL_VITASTOR,
+ 
+     VIR_STORAGE_NET_PROTOCOL_LAST
+ } virStorageNetProtocol;
+diff --git a/src/conf/virstorageobj.c b/src/conf/virstorageobj.c
+index 02903ac487..504df599fb 100644
+--- a/src/conf/virstorageobj.c
++++ b/src/conf/virstorageobj.c
+@@ -1481,6 +1481,7 @@ virStoragePoolObjSourceFindDuplicateCb(const void *payload,
+             return 1;
+         break;
+ 
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_RBD:
+     case VIR_STORAGE_POOL_LAST:
+         break;
+@@ -1978,6 +1979,8 @@ virStoragePoolObjMatch(virStoragePoolObj *obj,
+                (obj->def->type == VIR_STORAGE_POOL_MPATH))   ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_RBD) &&
+                (obj->def->type == VIR_STORAGE_POOL_RBD))     ||
++              (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR) &&
++               (obj->def->type == VIR_STORAGE_POOL_VITASTOR)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG) &&
+                (obj->def->type == VIR_STORAGE_POOL_SHEEPDOG)) ||
+               (MATCH(VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER) &&
+diff --git a/src/libvirt-storage.c b/src/libvirt-storage.c
+index cbc522b300..b4760fa58d 100644
+--- a/src/libvirt-storage.c
++++ b/src/libvirt-storage.c
+@@ -92,6 +92,7 @@ virStoragePoolGetConnect(virStoragePoolPtr pool)
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SCSI
+  * VIR_CONNECT_LIST_STORAGE_POOLS_MPATH
+  * VIR_CONNECT_LIST_STORAGE_POOLS_RBD
++ * VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR
+  * VIR_CONNECT_LIST_STORAGE_POOLS_SHEEPDOG
+  * VIR_CONNECT_LIST_STORAGE_POOLS_GLUSTER
+  * VIR_CONNECT_LIST_STORAGE_POOLS_ZFS
+diff --git a/src/libxl/libxl_conf.c b/src/libxl/libxl_conf.c
+index 1ac6253ad7..abe4587f94 100644
+--- a/src/libxl/libxl_conf.c
++++ b/src/libxl/libxl_conf.c
+@@ -962,6 +962,7 @@ libxlMakeNetworkDiskSrcStr(virStorageSource *src,
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
+     case VIR_STORAGE_NET_PROTOCOL_NFS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/src/libxl/xen_xl.c b/src/libxl/xen_xl.c
+index 7604e3d534..6453bb9776 100644
+--- a/src/libxl/xen_xl.c
++++ b/src/libxl/xen_xl.c
+@@ -1506,6 +1506,7 @@ xenFormatXLDiskSrcNet(virStorageSource *src)
+     case VIR_STORAGE_NET_PROTOCOL_SSH:
+     case VIR_STORAGE_NET_PROTOCOL_VXHS:
+     case VIR_STORAGE_NET_PROTOCOL_NFS:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+         virReportError(VIR_ERR_NO_SUPPORT,
+diff --git a/src/qemu/qemu_block.c b/src/qemu/qemu_block.c
+index e5ff653a60..884ecc79ea 100644
+--- a/src/qemu/qemu_block.c
++++ b/src/qemu/qemu_block.c
+@@ -943,6 +943,38 @@ qemuBlockStorageSourceGetRBDProps(virStorageSource *src,
+ }
+ 
+ 
++static virJSONValue *
++qemuBlockStorageSourceGetVitastorProps(virStorageSource *src)
++{
++    virJSONValue *ret = NULL;
++    virStorageNetHostDef *host;
++    size_t i;
++    g_auto(virBuffer) buf = VIR_BUFFER_INITIALIZER;
++    g_autofree char *etcd = NULL;
++
++    for (i = 0; i < src->nhosts; i++) {
++        host = src->hosts + i;
++        if ((virStorageNetHostTransport)host->transport != VIR_STORAGE_NET_HOST_TRANS_TCP) {
++            return NULL;
++        }
++        virBufferAsprintf(&buf, i > 0 ? ",%s:%u" : "%s:%u", host->name, host->port);
++    }
++    if (src->nhosts > 0) {
++        etcd = virBufferContentAndReset(&buf);
++    }
++
++    if (virJSONValueObjectCreate(&ret,
++                                 "S:etcd-host", etcd,
++                                 "S:etcd-prefix", src->query,
++                                 "S:config-path", src->configFile,
++                                 "s:image", src->path,
++                                 NULL) < 0)
++        return NULL;
++
++    return ret;
++}
++
++
+ static virJSONValue *
+ qemuBlockStorageSourceGetSheepdogProps(virStorageSource *src)
+ {
+@@ -1233,6 +1265,12 @@ qemuBlockStorageSourceGetBackendProps(virStorageSource *src,
+                 return NULL;
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            driver = "vitastor";
++            if (!(fileprops = qemuBlockStorageSourceGetVitastorProps(src)))
++                return NULL;
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             driver = "sheepdog";
+             if (!(fileprops = qemuBlockStorageSourceGetSheepdogProps(src)))
+@@ -2244,6 +2282,7 @@ qemuBlockGetBackingStoreString(virStorageSource *src,
+ 
+             case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             case VIR_STORAGE_NET_PROTOCOL_RBD:
++            case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+             case VIR_STORAGE_NET_PROTOCOL_VXHS:
+             case VIR_STORAGE_NET_PROTOCOL_NFS:
+             case VIR_STORAGE_NET_PROTOCOL_SSH:
+@@ -2626,6 +2665,12 @@ qemuBlockStorageSourceCreateGetStorageProps(virStorageSource *src,
+                 return -1;
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            driver = "vitastor";
++            if (!(location = qemuBlockStorageSourceGetVitastorProps(src)))
++                return -1;
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+             driver = "sheepdog";
+             if (!(location = qemuBlockStorageSourceGetSheepdogProps(src)))
+diff --git a/src/qemu/qemu_command.c b/src/qemu/qemu_command.c
+index d822533ccb..afe2087303 100644
+--- a/src/qemu/qemu_command.c
++++ b/src/qemu/qemu_command.c
+@@ -1723,6 +1723,43 @@ qemuBuildNetworkDriveStr(virStorageSource *src,
+             ret = virBufferContentAndReset(&buf);
+             break;
+ 
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++            if (strchr(src->path, ':')) {
++                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
++                               _("':' not allowed in Vitastor source volume name '%s'"),
++                               src->path);
++                return NULL;
++            }
++
++            virBufferStrcat(&buf, "vitastor:image=", src->path, NULL);
++
++            if (src->nhosts > 0) {
++                virBufferAddLit(&buf, ":etcd-host=");
++                for (i = 0; i < src->nhosts; i++) {
++                    if (i)
++                        virBufferAddLit(&buf, ",");
++
++                    /* assume host containing : is ipv6 */
++                    if (strchr(src->hosts[i].name, ':'))
++                        virBufferEscape(&buf, '\\', ":", "[%s]",
++                                        src->hosts[i].name);
++                    else
++                        virBufferAsprintf(&buf, "%s", src->hosts[i].name);
++
++                    if (src->hosts[i].port)
++                        virBufferAsprintf(&buf, "\\:%u", src->hosts[i].port);
++                }
++            }
++
++            if (src->configFile)
++                virBufferEscape(&buf, '\\', ":", ":config-path=%s", src->configFile);
++
++            if (src->query)
++                virBufferEscape(&buf, '\\', ":", ":etcd-prefix=%s", src->query);
++
++            ret = virBufferContentAndReset(&buf);
++            break;
++
+         case VIR_STORAGE_NET_PROTOCOL_VXHS:
+             virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                            _("VxHS protocol does not support URI syntax"));
+diff --git a/src/qemu/qemu_domain.c b/src/qemu/qemu_domain.c
+index a8401bac30..3dc1fe6db0 100644
+--- a/src/qemu/qemu_domain.c
++++ b/src/qemu/qemu_domain.c
+@@ -4731,7 +4731,8 @@ qemuDomainValidateStorageSource(virStorageSource *src,
+     if (src->query &&
+         (actualType != VIR_STORAGE_TYPE_NETWORK ||
+          (src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTPS &&
+-          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP))) {
++          src->protocol != VIR_STORAGE_NET_PROTOCOL_HTTP &&
++          src->protocol != VIR_STORAGE_NET_PROTOCOL_VITASTOR))) {
+         virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
+                        _("query is supported only with HTTP(S) protocols"));
+         return -1;
+@@ -9919,6 +9920,7 @@ qemuDomainPrepareStorageSourceTLS(virStorageSource *src,
+         break;
+ 
+     case VIR_STORAGE_NET_PROTOCOL_RBD:
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+     case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+diff --git a/src/qemu/qemu_snapshot.c b/src/qemu/qemu_snapshot.c
+index f92e00f9c0..854a3fbc90 100644
+--- a/src/qemu/qemu_snapshot.c
++++ b/src/qemu/qemu_snapshot.c
+@@ -393,6 +393,7 @@ qemuSnapshotPrepareDiskExternalInactive(virDomainSnapshotDiskDef *snapdisk,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+@@ -485,6 +486,7 @@ qemuSnapshotPrepareDiskExternalActive(virDomainObj *vm,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+         case VIR_STORAGE_NET_PROTOCOL_HTTP:
+@@ -638,6 +640,7 @@ qemuSnapshotPrepareDiskInternal(virDomainDiskDef *disk,
+         case VIR_STORAGE_NET_PROTOCOL_NONE:
+         case VIR_STORAGE_NET_PROTOCOL_NBD:
+         case VIR_STORAGE_NET_PROTOCOL_RBD:
++        case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
+         case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+         case VIR_STORAGE_NET_PROTOCOL_GLUSTER:
+         case VIR_STORAGE_NET_PROTOCOL_ISCSI:
+diff --git a/src/storage/storage_driver.c b/src/storage/storage_driver.c
+index 4df2c75a2b..5a5e48ef71 100644
+--- a/src/storage/storage_driver.c
++++ b/src/storage/storage_driver.c
+@@ -1643,6 +1643,7 @@ storageVolLookupByPathCallback(virStoragePoolObj *obj,
+ 
+         case VIR_STORAGE_POOL_GLUSTER:
+         case VIR_STORAGE_POOL_RBD:
++        case VIR_STORAGE_POOL_VITASTOR:
+         case VIR_STORAGE_POOL_SHEEPDOG:
+         case VIR_STORAGE_POOL_ZFS:
+         case VIR_STORAGE_POOL_LAST:
+diff --git a/src/storage_file/storage_source_backingstore.c b/src/storage_file/storage_source_backingstore.c
+index e48ae725ab..2017ccc88c 100644
+--- a/src/storage_file/storage_source_backingstore.c
++++ b/src/storage_file/storage_source_backingstore.c
+@@ -284,6 +284,75 @@ virStorageSourceParseRBDColonString(const char *rbdstr,
+ }
+ 
+ 
++static int
++virStorageSourceParseVitastorColonString(const char *colonstr,
++                                         virStorageSource *src)
++{
++    char *p, *e, *next;
++    g_autofree char *options = NULL;
++
++    /* optionally skip the "vitastor:" prefix if provided */
++    if (STRPREFIX(colonstr, "vitastor:"))
++        colonstr += strlen("vitastor:");
++
++    options = g_strdup(colonstr);
++
++    p = options;
++    while (*p) {
++        /* find : delimiter or end of string */
++        for (e = p; *e && *e != ':'; ++e) {
++            if (*e == '\\') {
++                e++;
++                if (*e == '\0')
++                    break;
++            }
++        }
++        if (*e == '\0') {
++            next = e;    /* last kv pair */
++        } else {
++            next = e + 1;
++            *e = '\0';
++        }
++
++        if (STRPREFIX(p, "image=")) {
++            src->path = g_strdup(p + strlen("image="));
++        } else if (STRPREFIX(p, "etcd-prefix=")) {
++            src->query = g_strdup(p + strlen("etcd-prefix="));
++        } else if (STRPREFIX(p, "config-path=")) {
++            src->configFile = g_strdup(p + strlen("config-path="));
++        } else if (STRPREFIX(p, "etcd-host=")) {
++            char *h, *sep;
++
++            h = p + strlen("etcd-host=");
++            while (h < e) {
++                for (sep = h; sep < e; ++sep) {
++                    if (*sep == '\\' && (sep[1] == ',' ||
++                                         sep[1] == ';' ||
++                                         sep[1] == ' ')) {
++                        *sep = '\0';
++                        sep += 2;
++                        break;
++                    }
++                }
++
++                if (virStorageSourceRBDAddHost(src, h) < 0)
++                    return -1;
++
++                h = sep;
++            }
++        }
++
++        p = next;
++    }
++
++    if (!src->path) {
++        return -1;
++    }
++
++    return 0;
++}
++
++
+ static int
+ virStorageSourceParseNBDColonString(const char *nbdstr,
+                                     virStorageSource *src)
+@@ -396,6 +465,11 @@ virStorageSourceParseBackingColon(virStorageSource *src,
+             return -1;
+         break;
+ 
++    case VIR_STORAGE_NET_PROTOCOL_VITASTOR:
++        if (virStorageSourceParseVitastorColonString(path, src) < 0)
++            return -1;
++        break;
++
+     case VIR_STORAGE_NET_PROTOCOL_SHEEPDOG:
+     case VIR_STORAGE_NET_PROTOCOL_LAST:
+     case VIR_STORAGE_NET_PROTOCOL_NONE:
+@@ -984,6 +1058,54 @@ virStorageSourceParseBackingJSONRBD(virStorageSource *src,
+     return 0;
+ }
+ 
++static int
++virStorageSourceParseBackingJSONVitastor(virStorageSource *src,
++                                         virJSONValue *json,
++                                         const char *jsonstr G_GNUC_UNUSED,
++                                         int opaque G_GNUC_UNUSED)
++{
++    const char *filename;
++    const char *image = virJSONValueObjectGetString(json, "image");
++    const char *conf = virJSONValueObjectGetString(json, "config-path");
++    const char *etcd_prefix = virJSONValueObjectGetString(json, "etcd-prefix");
++    virJSONValue *servers = virJSONValueObjectGetArray(json, "server");
++    size_t nservers;
++    size_t i;
++
++    src->type = VIR_STORAGE_TYPE_NETWORK;
++    src->protocol = VIR_STORAGE_NET_PROTOCOL_VITASTOR;
++
++    /* legacy syntax passed via 'filename' option */
++    if ((filename = virJSONValueObjectGetString(json, "filename")))
++        return virStorageSourceParseVitastorColonString(filename, src);
++
++    if (!image) {
++        virReportError(VIR_ERR_INVALID_ARG, "%s",
++                       _("missing image name in Vitastor backing volume "
++                         "JSON specification"));
++        return -1;
++    }
++
++    src->path = g_strdup(image);
++    src->configFile = g_strdup(conf);
++    src->query = g_strdup(etcd_prefix);
++
++    if (servers) {
++        nservers = virJSONValueArraySize(servers);
++
++        src->hosts = g_new0(virStorageNetHostDef, nservers);
++        src->nhosts = nservers;
++
++        for (i = 0; i < nservers; i++) {
++            if (virStorageSourceParseBackingJSONInetSocketAddress(src->hosts + i,
++                                                                  virJSONValueArrayGet(servers, i)) < 0)
++                return -1;
++        }
++    }
++
++    return 0;
++}
++
+ static int
+ virStorageSourceParseBackingJSONRaw(virStorageSource *src,
+                                     virJSONValue *json,
+@@ -1162,6 +1284,7 @@ static const struct virStorageSourceJSONDriverParser jsonParsers[] = {
+     {"sheepdog", false, virStorageSourceParseBackingJSONSheepdog, 0},
+     {"ssh", false, virStorageSourceParseBackingJSONSSH, 0},
+     {"rbd", false, virStorageSourceParseBackingJSONRBD, 0},
++    {"vitastor", false, virStorageSourceParseBackingJSONVitastor, 0},
+     {"raw", true, virStorageSourceParseBackingJSONRaw, 0},
+     {"nfs", false, virStorageSourceParseBackingJSONNFS, 0},
+     {"vxhs", false, virStorageSourceParseBackingJSONVxHS, 0},
+diff --git a/src/test/test_driver.c b/src/test/test_driver.c
+index 0e93b79922..b4d33f5f56 100644
+--- a/src/test/test_driver.c
++++ b/src/test/test_driver.c
+@@ -7367,6 +7367,7 @@ testStorageVolumeTypeForPool(int pooltype)
+     case VIR_STORAGE_POOL_ISCSI_DIRECT:
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_RBD:
++    case VIR_STORAGE_POOL_VITASTOR:
+         return VIR_STORAGE_VOL_NETWORK;
+     case VIR_STORAGE_POOL_LOGICAL:
+     case VIR_STORAGE_POOL_DISK:
+diff --git a/tests/storagepoolcapsschemadata/poolcaps-fs.xml b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
+index eee75af746..8bd0a57bdd 100644
+--- a/tests/storagepoolcapsschemadata/poolcaps-fs.xml
++++ b/tests/storagepoolcapsschemadata/poolcaps-fs.xml
+@@ -204,4 +204,11 @@
+       </enum>
+     </volOptions>
+   </pool>
++  <pool type='vitastor' supported='no'>
++    <volOptions>
++      <defaultFormat type='raw'/>
++      <enum name='targetFormatType'>
++      </enum>
++    </volOptions>
++  </pool>
+ </storagepoolCapabilities>
+diff --git a/tests/storagepoolcapsschemadata/poolcaps-full.xml b/tests/storagepoolcapsschemadata/poolcaps-full.xml
+index 805950a937..852df0de16 100644
+--- a/tests/storagepoolcapsschemadata/poolcaps-full.xml
++++ b/tests/storagepoolcapsschemadata/poolcaps-full.xml
+@@ -204,4 +204,11 @@
+       </enum>
+     </volOptions>
+   </pool>
++  <pool type='vitastor' supported='yes'>
++    <volOptions>
++      <defaultFormat type='raw'/>
++      <enum name='targetFormatType'>
++      </enum>
++    </volOptions>
++  </pool>
+ </storagepoolCapabilities>
+diff --git a/tests/storagepoolxml2argvtest.c b/tests/storagepoolxml2argvtest.c
+index 449b745519..7f95cc8e08 100644
+--- a/tests/storagepoolxml2argvtest.c
++++ b/tests/storagepoolxml2argvtest.c
+@@ -68,6 +68,7 @@ testCompareXMLToArgvFiles(bool shouldFail,
+     case VIR_STORAGE_POOL_GLUSTER:
+     case VIR_STORAGE_POOL_ZFS:
+     case VIR_STORAGE_POOL_VSTORAGE:
++    case VIR_STORAGE_POOL_VITASTOR:
+     case VIR_STORAGE_POOL_LAST:
+     default:
+         VIR_TEST_DEBUG("pool type '%s' has no xml2argv test", defTypeStr);
+diff --git a/tools/virsh-pool.c b/tools/virsh-pool.c
+index d391257f6e..46799c4a90 100644
+--- a/tools/virsh-pool.c
++++ b/tools/virsh-pool.c
+@@ -1213,6 +1213,9 @@ cmdPoolList(vshControl *ctl, const vshCmd *cmd G_GNUC_UNUSED)
+             case VIR_STORAGE_POOL_VSTORAGE:
+                 flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VSTORAGE;
+                 break;
++            case VIR_STORAGE_POOL_VITASTOR:
++                flags |= VIR_CONNECT_LIST_STORAGE_POOLS_VITASTOR;
++                break;
+             case VIR_STORAGE_POOL_LAST:
+                 break;
+             }

rpm/build-tarball.sh

@@ -24,4 +24,4 @@ rm fio
 mv fio-copy fio
 FIO=`rpm -qi fio | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Version[\s:]+(\S+)/ && print $1; /^Release[\s:]+(\S+)/ && print "-$1"; }'`
 perl -i -pe 's/(Requires:\s*fio)([^\n]+)?/$1 = '$FIO'/' $VITASTOR/rpm/vitastor-el$EL.spec
-tar --transform 's#^#vitastor-1.4.5/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-1.4.5$(rpm --eval '%dist').tar.gz *
+tar --transform 's#^#vitastor-1.5.0/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-1.5.0$(rpm --eval '%dist').tar.gz *

rpm/vitastor-el7.Dockerfile

@@ -36,7 +36,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-1.4.5.el7.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-1.5.0.el7.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el7.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el7.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        1.4.5
+Version:        1.5.0
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-1.4.5.el7.tar.gz
+Source0:        vitastor-1.5.0.el7.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -149,9 +149,12 @@ mkdir -p /etc/vitastor
 %_bindir/vitastor-nfs
 %_bindir/vitastor-cli
 %_bindir/vitastor-rm
+%_bindir/vitastor-kv
+%_bindir/vitastor-kv-stress
 %_bindir/vita
 %_libdir/libvitastor_blk.so*
 %_libdir/libvitastor_client.so*
+%_libdir/libvitastor_kv.so*
 
 
 %files -n vitastor-client-devel

rpm/vitastor-el8.Dockerfile

@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-1.4.5.el8.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-1.5.0.el8.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el8.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el8.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        1.4.5
+Version:        1.5.0
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-1.4.5.el8.tar.gz
+Source0:        vitastor-1.5.0.el8.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -146,9 +146,12 @@ mkdir -p /etc/vitastor
 %_bindir/vitastor-nfs
 %_bindir/vitastor-cli
 %_bindir/vitastor-rm
+%_bindir/vitastor-kv
+%_bindir/vitastor-kv-stress
 %_bindir/vita
 %_libdir/libvitastor_blk.so*
 %_libdir/libvitastor_client.so*
+%_libdir/libvitastor_kv.so*
 
 
 %files -n vitastor-client-devel

rpm/vitastor-el9.Dockerfile

@@ -18,7 +18,7 @@ ADD . /root/vitastor
 RUN set -e; \
     cd /root/vitastor/rpm; \
     sh build-tarball.sh; \
-    cp /root/vitastor-1.4.5.el9.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-1.5.0.el9.tar.gz ~/rpmbuild/SOURCES; \
     cp vitastor-el9.spec ~/rpmbuild/SPECS/vitastor.spec; \
     cd ~/rpmbuild/SPECS/; \
     rpmbuild -ba vitastor.spec; \

rpm/vitastor-el9.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        1.4.5
+Version:        1.5.0
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-1.4.5.el9.tar.gz
+Source0:        vitastor-1.5.0.el9.tar.gz
 
 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
@@ -139,9 +139,12 @@ mkdir -p /etc/vitastor
 %_bindir/vitastor-nfs
 %_bindir/vitastor-cli
 %_bindir/vitastor-rm
+%_bindir/vitastor-kv
+%_bindir/vitastor-kv-stress
 %_bindir/vita
 %_libdir/libvitastor_blk.so*
 %_libdir/libvitastor_client.so*
+%_libdir/libvitastor_kv.so*
 
 
 %files -n vitastor-client-devel

src/CMakeLists.txt

@@ -16,8 +16,8 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 	set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR}")
 endif()
 
-add_definitions(-DVERSION="1.4.5")
-add_definitions(-Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -fno-omit-frame-pointer -I ${CMAKE_SOURCE_DIR}/src)
+add_definitions(-DVERSION="1.5.0")
+add_definitions(-D_LARGEFILE64_SOURCE -D_FILE_OFFSET_BITS=64 -Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -fno-omit-frame-pointer -I ${CMAKE_SOURCE_DIR}/src)
 add_link_options(-fno-omit-frame-pointer)
 if (${WITH_ASAN})
 	add_definitions(-fsanitize=address)
@@ -145,7 +145,6 @@ add_library(vitastor_client SHARED
 	cli_status.cpp
 	cli_describe.cpp
 	cli_fix.cpp
-	cli_df.cpp
 	cli_ls.cpp
 	cli_create.cpp
 	cli_modify.cpp
@@ -154,6 +153,11 @@ add_library(vitastor_client SHARED
 	cli_rm_data.cpp
 	cli_rm.cpp
 	cli_rm_osd.cpp
+	cli_pool_cfg.cpp
+	cli_pool_create.cpp
+	cli_pool_ls.cpp
+	cli_pool_modify.cpp
+	cli_pool_rm.cpp
 )
 set_target_properties(vitastor_client PROPERTIES PUBLIC_HEADER "vitastor_c.h")
 target_link_libraries(vitastor_client
@@ -181,10 +185,48 @@ target_link_libraries(vitastor-nbd
 	vitastor_client
 )
 
+# libvitastor_kv.so
+add_library(vitastor_kv SHARED
+	kv_db.cpp
+	kv_db.h
+)
+target_link_libraries(vitastor_kv
+	vitastor_client
+)
+set_target_properties(vitastor_kv PROPERTIES VERSION ${VERSION} SOVERSION 0)
+
+# vitastor-kv
+add_executable(vitastor-kv
+	kv_cli.cpp
+)
+target_link_libraries(vitastor-kv
+	vitastor_kv
+)
+
+add_executable(vitastor-kv-stress
+	kv_stress.cpp
+)
+target_link_libraries(vitastor-kv-stress
+	vitastor_kv
+)
+
 # vitastor-nfs
 add_executable(vitastor-nfs
 	nfs_proxy.cpp
-	nfs_conn.cpp
+	nfs_block.cpp
+	nfs_kv.cpp
+	nfs_kv_create.cpp
+	nfs_kv_getattr.cpp
+	nfs_kv_link.cpp
+	nfs_kv_lookup.cpp
+	nfs_kv_read.cpp
+	nfs_kv_readdir.cpp
+	nfs_kv_remove.cpp
+	nfs_kv_rename.cpp
+	nfs_kv_setattr.cpp
+	nfs_kv_write.cpp
+	nfs_fsstat.cpp
+	nfs_mount.cpp
 	nfs_portmap.cpp
 	sha256.c
 	nfs/xdr_impl.cpp
@@ -194,6 +236,7 @@ add_executable(vitastor-nfs
 )
 target_link_libraries(vitastor-nfs
 	vitastor_client
+	vitastor_kv
 )
 
 # vitastor-cli
@@ -318,12 +361,12 @@ add_test(NAME test_cluster_client COMMAND test_cluster_client)
 
 ### Install
 
-install(TARGETS vitastor-osd vitastor-disk vitastor-nbd vitastor-nfs vitastor-cli RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR})
+install(TARGETS vitastor-osd vitastor-disk vitastor-nbd vitastor-nfs vitastor-cli vitastor-kv vitastor-kv-stress RUNTIME DESTINATION ${CMAKE_INSTALL_BINDIR})
 install_symlink(vitastor-disk ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_BINDIR}/vitastor-dump-journal)
 install_symlink(vitastor-cli ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_BINDIR}/vitastor-rm)
 install_symlink(vitastor-cli ${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_BINDIR}/vita)
 install(
-	TARGETS vitastor_blk vitastor_client
+	TARGETS vitastor_blk vitastor_client vitastor_kv
 	LIBRARY DESTINATION ${CMAKE_INSTALL_LIBDIR}
 	PUBLIC_HEADER DESTINATION ${CMAKE_INSTALL_INCLUDEDIR}
 )

src/blockstore.cpp

@@ -82,3 +82,8 @@ uint32_t blockstore_t::get_bitmap_granularity()
 {
     return impl->get_bitmap_granularity();
 }
+
+void blockstore_t::set_no_inode_stats(const std::vector<uint64_t> & pool_ids)
+{
+    impl->set_no_inode_stats(pool_ids);
+}

src/blockstore.h

@@ -216,6 +216,9 @@ public:
     // Get per-inode space usage statistics
     std::map<uint64_t, uint64_t> & get_inode_space_stats();
 
+    // Set per-pool no_inode_stats
+    void set_no_inode_stats(const std::vector<uint64_t> & pool_ids);
+
     // Print diagnostics to stdout
     void dump_diagnostics();
 

src/blockstore_disk.cpp

@@ -108,6 +108,10 @@ void blockstore_disk_t::parse_config(std::map<std::string, std::string> & config
     {
         throw std::runtime_error("journal_block_size must be a multiple of "+std::to_string(DIRECT_IO_ALIGNMENT));
     }
+    else if (journal_block_size > MAX_DATA_BLOCK_SIZE)
+    {
+        throw std::runtime_error("journal_block_size must not exceed "+std::to_string(MAX_DATA_BLOCK_SIZE));
+    }
     if (!meta_block_size)
     {
         meta_block_size = 4096;
@@ -116,6 +120,10 @@ void blockstore_disk_t::parse_config(std::map<std::string, std::string> & config
     {
         throw std::runtime_error("meta_block_size must be a multiple of "+std::to_string(DIRECT_IO_ALIGNMENT));
     }
+    else if (meta_block_size > MAX_DATA_BLOCK_SIZE)
+    {
+        throw std::runtime_error("meta_block_size must not exceed "+std::to_string(MAX_DATA_BLOCK_SIZE));
+    }
     if (data_offset % disk_alignment)
     {
         throw std::runtime_error("data_offset must be a multiple of disk_alignment = "+std::to_string(disk_alignment));

src/blockstore_flush.cpp

@@ -93,7 +93,7 @@ void journal_flusher_t::loop()
 void journal_flusher_t::enqueue_flush(obj_ver_id ov)
 {
 #ifdef BLOCKSTORE_DEBUG
-    printf("enqueue_flush %lx:%lx v%lu\n", ov.oid.inode, ov.oid.stripe, ov.version);
+    printf("enqueue_flush %jx:%jx v%ju\n", ov.oid.inode, ov.oid.stripe, ov.version);
 #endif
     auto it = flush_versions.find(ov.oid);
     if (it != flush_versions.end())
@@ -116,7 +116,7 @@ void journal_flusher_t::enqueue_flush(obj_ver_id ov)
 void journal_flusher_t::unshift_flush(obj_ver_id ov, bool force)
 {
 #ifdef BLOCKSTORE_DEBUG
-    printf("unshift_flush %lx:%lx v%lu\n", ov.oid.inode, ov.oid.stripe, ov.version);
+    printf("unshift_flush %jx:%jx v%ju\n", ov.oid.inode, ov.oid.stripe, ov.version);
 #endif
     auto it = flush_versions.find(ov.oid);
     if (it != flush_versions.end())
@@ -142,7 +142,7 @@ void journal_flusher_t::unshift_flush(obj_ver_id ov, bool force)
 void journal_flusher_t::remove_flush(object_id oid)
 {
 #ifdef BLOCKSTORE_DEBUG
-    printf("undo_flush %lx:%lx\n", oid.inode, oid.stripe);
+    printf("undo_flush %jx:%jx\n", oid.inode, oid.stripe);
 #endif
     auto v_it = flush_versions.find(oid);
     if (v_it != flush_versions.end())
@@ -233,7 +233,7 @@ void journal_flusher_t::dump_diagnostics()
         break;
     }
     printf(
-        "Flusher: queued=%ld first=%s%lx:%lx trim_wanted=%d dequeuing=%d trimming=%d cur=%d target=%d active=%d syncing=%d\n",
+        "Flusher: queued=%zd first=%s%jx:%jx trim_wanted=%d dequeuing=%d trimming=%d cur=%d target=%d active=%d syncing=%d\n",
         flush_queue.size(), unflushable_type, unflushable.oid.inode, unflushable.oid.stripe,
         trim_wanted, dequeuing, trimming, cur_flusher_count, target_flusher_count,
         active_flushers, syncing_flushers
@@ -266,7 +266,7 @@ bool journal_flusher_t::try_find_other(std::map<obj_ver_id, dirty_entry>::iterat
 {
     int search_left = flush_queue.size() - 1;
 #ifdef BLOCKSTORE_DEBUG
-    printf("Flusher overran writers (%lx:%lx v%lu, dirty_start=%08lx) - searching for older flushes (%d left)\n",
+    printf("Flusher overran writers (%jx:%jx v%ju, dirty_start=%08jx) - searching for older flushes (%d left)\n",
         cur.oid.inode, cur.oid.stripe, cur.version, bs->journal.dirty_start, search_left);
 #endif
     while (search_left > 0)
@@ -283,7 +283,7 @@ bool journal_flusher_t::try_find_other(std::map<obj_ver_id, dirty_entry>::iterat
                 dirty_end->second.journal_sector < bs->journal.used_start))
             {
 #ifdef BLOCKSTORE_DEBUG
-                printf("Write %lx:%lx v%lu is too new: offset=%08lx\n", cur.oid.inode, cur.oid.stripe, cur.version, dirty_end->second.journal_sector);
+                printf("Write %jx:%jx v%ju is too new: offset=%08jx\n", cur.oid.inode, cur.oid.stripe, cur.version, dirty_end->second.journal_sector);
 #endif
                 enqueue_flush(cur);
             }
@@ -386,7 +386,7 @@ stop_flusher:
         if (repeat_it != flusher->sync_to_repeat.end())
         {
 #ifdef BLOCKSTORE_DEBUG
-            printf("Postpone %lx:%lx v%lu\n", cur.oid.inode, cur.oid.stripe, cur.version);
+            printf("Postpone %jx:%jx v%ju\n", cur.oid.inode, cur.oid.stripe, cur.version);
 #endif
             // We don't flush different parts of history of the same object in parallel
             // So we check if someone is already flushing this object
@@ -415,12 +415,13 @@ stop_flusher:
                 flusher->sync_to_repeat.erase(cur.oid);
                 if (!flusher->try_find_other(dirty_end, cur))
                 {
+                    cur.oid = {};
                     goto stop_flusher;
                 }
             }
         }
 #ifdef BLOCKSTORE_DEBUG
-        printf("Flushing %lx:%lx v%lu\n", cur.oid.inode, cur.oid.stripe, cur.version);
+        printf("Flushing %jx:%jx v%ju\n", cur.oid.inode, cur.oid.stripe, cur.version);
 #endif
         flusher->active_flushers++;
         // Find it in clean_db
@@ -447,7 +448,7 @@ stop_flusher:
                 // Object not allocated. This is a bug.
                 char err[1024];
                 snprintf(
-                    err, 1024, "BUG: Object %lx:%lx v%lu that we are trying to flush is not allocated on the data device",
+                    err, 1024, "BUG: Object %jx:%jx v%ju that we are trying to flush is not allocated on the data device",
                     cur.oid.inode, cur.oid.stripe, cur.version
                 );
                 throw std::runtime_error(err);
@@ -537,7 +538,7 @@ resume_2:
                 clean_disk_entry *old_entry = (clean_disk_entry*)((uint8_t*)meta_old.buf + meta_old.pos*bs->dsk.clean_entry_size);
                 if (old_entry->oid.inode != 0 && old_entry->oid != cur.oid)
                 {
-                    printf("Fatal error (metadata corruption or bug): tried to wipe metadata entry %lu (%lx:%lx v%lu) as old location of %lx:%lx\n",
+                    printf("Fatal error (metadata corruption or bug): tried to wipe metadata entry %ju (%jx:%jx v%ju) as old location of %jx:%jx\n",
                         old_clean_loc >> bs->dsk.block_order, old_entry->oid.inode, old_entry->oid.stripe,
                         old_entry->version, cur.oid.inode, cur.oid.stripe);
                     exit(1);
@@ -570,7 +571,7 @@ resume_2:
         // Erase dirty_db entries
         bs->erase_dirty(dirty_start, std::next(dirty_end), clean_loc);
 #ifdef BLOCKSTORE_DEBUG
-        printf("Flushed %lx:%lx v%lu (%d copies, wr:%d, del:%d), %ld left\n", cur.oid.inode, cur.oid.stripe, cur.version,
+        printf("Flushed %jx:%jx v%ju (%d copies, wr:%d, del:%d), %jd left\n", cur.oid.inode, cur.oid.stripe, cur.version,
             copy_count, has_writes, has_delete, flusher->flush_queue.size());
 #endif
     release_oid:
@@ -609,8 +610,8 @@ void journal_flusher_co::update_metadata_entry()
     {
         printf(
             has_delete
-                ? "Fatal error (metadata corruption or bug): tried to delete metadata entry %lu (%lx:%lx v%lu) while deleting %lx:%lx v%lu\n"
-                : "Fatal error (metadata corruption or bug): tried to overwrite non-zero metadata entry %lu (%lx:%lx v%lu) with %lx:%lx v%lu\n",
+                ? "Fatal error (metadata corruption or bug): tried to delete metadata entry %ju (%jx:%jx v%ju) while deleting %jx:%jx v%ju\n"
+                : "Fatal error (metadata corruption or bug): tried to overwrite non-zero metadata entry %ju (%jx:%jx v%ju) with %jx:%jx v%ju\n",
             clean_loc >> bs->dsk.block_order, new_entry->oid.inode, new_entry->oid.stripe,
             new_entry->version, cur.oid.inode, cur.oid.stripe, cur.version
         );
@@ -710,7 +711,7 @@ bool journal_flusher_co::write_meta_block(flusher_meta_write_t & meta_block, int
     if (wait_state == wait_base)
         goto resume_0;
     await_sqe(0);
-    data->iov = (struct iovec){ meta_block.buf, bs->dsk.meta_block_size };
+    data->iov = (struct iovec){ meta_block.buf, (size_t)bs->dsk.meta_block_size };
     data->callback = simple_callback_w;
     my_uring_prep_writev(
         sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bs->dsk.meta_block_size + meta_block.sector
@@ -760,7 +761,7 @@ bool journal_flusher_co::clear_incomplete_csum_block_bits(int wait_base)
         {
             // If we encounter bad checksums during flush, we still update the bad block,
             // but intentionally mangle checksums to avoid hiding the corruption.
-            iovec iov = { .iov_base = v[i].buf, .iov_len = v[i].len };
+            iovec iov = { .iov_base = v[i].buf, .iov_len = (size_t)v[i].len };
             if (!(v[i].copy_flags & COPY_BUF_JOURNAL))
             {
                 assert(!(v[i].offset % bs->dsk.csum_block_size));
@@ -768,7 +769,7 @@ bool journal_flusher_co::clear_incomplete_csum_block_bits(int wait_base)
                 bs->verify_padded_checksums(new_clean_bitmap, new_clean_bitmap + 2*bs->dsk.clean_entry_bitmap_size,
                     v[i].offset, &iov, 1, [&](uint32_t bad_block, uint32_t calc_csum, uint32_t stored_csum)
                 {
-                    printf("Checksum mismatch in object %lx:%lx v%lu in data area at offset 0x%lx+0x%x: got %08x, expected %08x\n",
+                    printf("Checksum mismatch in object %jx:%jx v%ju in data area at offset 0x%jx+0x%x: got %08x, expected %08x\n",
                         cur.oid.inode, cur.oid.stripe, old_clean_ver, old_clean_loc, bad_block, calc_csum, stored_csum);
                     for (uint32_t j = 0; j < bs->dsk.csum_block_size; j += bs->dsk.bitmap_granularity)
                     {
@@ -781,7 +782,7 @@ bool journal_flusher_co::clear_incomplete_csum_block_bits(int wait_base)
             {
                 bs->verify_journal_checksums(v[i].csum_buf, v[i].offset, &iov, 1, [&](uint32_t bad_block, uint32_t calc_csum, uint32_t stored_csum)
                 {
-                    printf("Checksum mismatch in object %lx:%lx v%lu in journal at offset 0x%lx+0x%x (block offset 0x%lx): got %08x, expected %08x\n",
+                    printf("Checksum mismatch in object %jx:%jx v%ju in journal at offset 0x%jx+0x%x (block offset 0x%jx): got %08x, expected %08x\n",
                         cur.oid.inode, cur.oid.stripe, old_clean_ver,
                         v[i].disk_offset, bad_block, v[i].offset, calc_csum, stored_csum);
                     bad_block += (v[i].offset/bs->dsk.csum_block_size) * bs->dsk.csum_block_size;
@@ -805,7 +806,7 @@ bool journal_flusher_co::clear_incomplete_csum_block_bits(int wait_base)
             if (new_entry->oid != cur.oid)
             {
                 printf(
-                    "Fatal error (metadata corruption or bug): tried to make holes in %lu (%lx:%lx v%lu) with %lx:%lx v%lu\n",
+                    "Fatal error (metadata corruption or bug): tried to make holes in %ju (%jx:%jx v%ju) with %jx:%jx v%ju\n",
                     clean_loc >> bs->dsk.block_order, new_entry->oid.inode, new_entry->oid.stripe,
                     new_entry->version, cur.oid.inode, cur.oid.stripe, cur.version
                 );
@@ -925,7 +926,7 @@ void journal_flusher_co::scan_dirty()
         {
             char err[1024];
             snprintf(
-                err, 1024, "BUG: Unexpected dirty_entry %lx:%lx v%lu unstable state during flush: 0x%x",
+                err, 1024, "BUG: Unexpected dirty_entry %jx:%jx v%ju unstable state during flush: 0x%x",
                 dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version, dirty_it->second.state
             );
             throw std::runtime_error(err);
@@ -1021,7 +1022,7 @@ void journal_flusher_co::scan_dirty()
             // May happen if the metadata entry is corrupt, but journal isn't
             // FIXME: Report corrupted object to the upper layer (OSD)
             printf(
-                "Warning: object %lx:%lx has overwrites, but doesn't have a clean version."
+                "Warning: object %jx:%jx has overwrites, but doesn't have a clean version."
                 " Metadata is likely corrupted. Dropping object from the DB.\n",
                 cur.oid.inode, cur.oid.stripe
             );
@@ -1056,7 +1057,7 @@ void journal_flusher_co::scan_dirty()
         flusher->enqueue_flush(cur);
         cur.version = dirty_end->first.version;
 #ifdef BLOCKSTORE_DEBUG
-        printf("Partial checksum block overwrites found - rewinding flush back to %lx:%lx v%lu\n", cur.oid.inode, cur.oid.stripe, cur.version);
+        printf("Partial checksum block overwrites found - rewinding flush back to %jx:%jx v%ju\n", cur.oid.inode, cur.oid.stripe, cur.version);
 #endif
         v.clear();
         copy_count = 0;
@@ -1084,7 +1085,7 @@ bool journal_flusher_co::read_dirty(int wait_base)
         auto & vi = v[v.size()-i];
         assert(vi.len != 0);
         vi.buf = memalign_or_die(MEM_ALIGNMENT, vi.len);
-        data->iov = (struct iovec){ vi.buf, vi.len };
+        data->iov = (struct iovec){ vi.buf, (size_t)vi.len };
         data->callback = simple_callback_r;
         my_uring_prep_readv(
             sqe, bs->dsk.data_fd, &data->iov, 1, bs->dsk.data_offset + old_clean_loc + vi.offset
@@ -1208,7 +1209,7 @@ bool journal_flusher_co::modify_meta_read(uint64_t meta_loc, flusher_meta_write_
             .usage_count = 1,
         }).first;
         await_sqe(0);
-        data->iov = (struct iovec){ wr.it->second.buf, bs->dsk.meta_block_size };
+        data->iov = (struct iovec){ wr.it->second.buf, (size_t)bs->dsk.meta_block_size };
         data->callback = simple_callback_r;
         wr.submitted = true;
         my_uring_prep_readv(
@@ -1247,7 +1248,7 @@ void journal_flusher_co::free_data_blocks()
         auto uo_it = bs->used_clean_objects.find(old_clean_loc);
         bool used = uo_it != bs->used_clean_objects.end();
 #ifdef BLOCKSTORE_DEBUG
-        printf("%s block %lu from %lx:%lx v%lu (new location is %lu)\n",
+        printf("%s block %ju from %jx:%jx v%ju (new location is %ju)\n",
             used ? "Postpone free" : "Free",
             old_clean_loc >> bs->dsk.block_order,
             cur.oid.inode, cur.oid.stripe, cur.version,
@@ -1264,7 +1265,7 @@ void journal_flusher_co::free_data_blocks()
         auto uo_it = bs->used_clean_objects.find(old_clean_loc);
         bool used = uo_it != bs->used_clean_objects.end();
 #ifdef BLOCKSTORE_DEBUG
-        printf("%s block %lu from %lx:%lx v%lu (delete)\n",
+        printf("%s block %ju from %jx:%jx v%ju (delete)\n",
             used ? "Postpone free" : "Free",
             old_clean_loc >> bs->dsk.block_order,
             cur.oid.inode, cur.oid.stripe, cur.version);
@@ -1377,7 +1378,7 @@ bool journal_flusher_co::trim_journal(int wait_base)
                 .csum_block_size = bs->dsk.csum_block_size,
             };
             ((journal_entry_start*)flusher->journal_superblock)->crc32 = je_crc32((journal_entry*)flusher->journal_superblock);
-            data->iov = (struct iovec){ flusher->journal_superblock, bs->dsk.journal_block_size };
+            data->iov = (struct iovec){ flusher->journal_superblock, (size_t)bs->dsk.journal_block_size };
             data->callback = simple_callback_w;
             my_uring_prep_writev(sqe, bs->dsk.journal_fd, &data->iov, 1, bs->journal.offset);
             wait_count++;
@@ -1409,7 +1410,7 @@ bool journal_flusher_co::trim_journal(int wait_base)
             }
             bs->journal.used_start = new_trim_pos;
 #ifdef BLOCKSTORE_DEBUG
-            printf("Journal trimmed to %08lx (next_free=%08lx dirty_start=%08lx)\n", bs->journal.used_start, bs->journal.next_free, bs->journal.dirty_start);
+            printf("Journal trimmed to %08jx (next_free=%08jx dirty_start=%08jx)\n", bs->journal.used_start, bs->journal.next_free, bs->journal.dirty_start);
 #endif
             if (bs->journal.flush_journal && !flusher->flush_queue.size())
             {

src/blockstore_impl.cpp

@@ -269,7 +269,7 @@ void blockstore_impl_t::check_wait(blockstore_op_t *op)
         {
             // stop submission if there's still no free space
 #ifdef BLOCKSTORE_DEBUG
-            printf("Still waiting for %lu SQE(s)\n", PRIV(op)->wait_detail);
+            printf("Still waiting for %ju SQE(s)\n", PRIV(op)->wait_detail);
 #endif
             return;
         }
@@ -282,7 +282,7 @@ void blockstore_impl_t::check_wait(blockstore_op_t *op)
         {
             // do not submit
 #ifdef BLOCKSTORE_DEBUG
-            printf("Still waiting to flush journal offset %08lx\n", PRIV(op)->wait_detail);
+            printf("Still waiting to flush journal offset %08jx\n", PRIV(op)->wait_detail);
 #endif
             return;
         }
@@ -733,3 +733,86 @@ void blockstore_impl_t::disk_error_abort(const char *op, int retval, int expecte
     fprintf(stderr, "Disk %s failed: result is %d, expected %d. Can't continue, sorry :-(\n", op, retval, expected);
     exit(1);
 }
+
+void blockstore_impl_t::set_no_inode_stats(const std::vector<uint64_t> & pool_ids)
+{
+    for (auto & np: no_inode_stats)
+    {
+        np.second = 2;
+    }
+    for (auto pool_id: pool_ids)
+    {
+        if (!no_inode_stats[pool_id])
+            recalc_inode_space_stats(pool_id, false);
+        no_inode_stats[pool_id] = 1;
+    }
+    for (auto np_it = no_inode_stats.begin(); np_it != no_inode_stats.end(); )
+    {
+        if (np_it->second == 2)
+        {
+            recalc_inode_space_stats(np_it->first, true);
+            no_inode_stats.erase(np_it++);
+        }
+        else
+            np_it++;
+    }
+}
+
+void blockstore_impl_t::recalc_inode_space_stats(uint64_t pool_id, bool per_inode)
+{
+    auto sp_begin = inode_space_stats.lower_bound((pool_id << (64-POOL_ID_BITS)));
+    auto sp_end = inode_space_stats.lower_bound(((pool_id+1) << (64-POOL_ID_BITS)));
+    inode_space_stats.erase(sp_begin, sp_end);
+    auto sh_it = clean_db_shards.lower_bound((pool_id << (64-POOL_ID_BITS)));
+    while (sh_it != clean_db_shards.end() &&
+        (sh_it->first >> (64-POOL_ID_BITS)) == pool_id)
+    {
+        for (auto & pair: sh_it->second)
+        {
+            uint64_t space_id = per_inode ? pair.first.inode : (pool_id << (64-POOL_ID_BITS));
+            inode_space_stats[space_id] += dsk.data_block_size;
+        }
+        sh_it++;
+    }
+    object_id last_oid = {};
+    bool last_exists = false;
+    auto dirty_it = dirty_db.lower_bound((obj_ver_id){ .oid = { .inode = (pool_id << (64-POOL_ID_BITS)) } });
+    while (dirty_it != dirty_db.end() && (dirty_it->first.oid.inode >> (64-POOL_ID_BITS)) == pool_id)
+    {
+        if (IS_STABLE(dirty_it->second.state) && (IS_BIG_WRITE(dirty_it->second.state) || IS_DELETE(dirty_it->second.state)))
+        {
+            bool exists = false;
+            if (last_oid == dirty_it->first.oid)
+            {
+                exists = last_exists;
+            }
+            else
+            {
+                auto & clean_db = clean_db_shard(dirty_it->first.oid);
+                auto clean_it = clean_db.find(dirty_it->first.oid);
+                exists = clean_it != clean_db.end();
+            }
+            uint64_t space_id = per_inode ? dirty_it->first.oid.inode : (pool_id << (64-POOL_ID_BITS));
+            if (IS_BIG_WRITE(dirty_it->second.state))
+            {
+                if (!exists)
+                    inode_space_stats[space_id] += dsk.data_block_size;
+                last_exists = true;
+            }
+            else
+            {
+                if (exists)
+                {
+                    auto & sp = inode_space_stats[space_id];
+                    if (sp > dsk.data_block_size)
+                        sp -= dsk.data_block_size;
+                    else
+                        inode_space_stats.erase(space_id);
+                }
+                last_exists = false;
+            }
+            last_oid = dirty_it->first.oid;
+        }
+        dirty_it++;
+    }
+}

src/blockstore_impl.h

@@ -272,6 +272,7 @@ class blockstore_impl_t
 
     std::map<pool_id_t, pool_shard_settings_t> clean_db_settings;
     std::map<pool_pg_id_t, blockstore_clean_db_t> clean_db_shards;
+    std::map<uint64_t, int> no_inode_stats;
     uint8_t *clean_bitmaps = NULL;
     blockstore_dirty_db_t dirty_db;
     std::vector<blockstore_op_t*> submit_queue;
@@ -318,6 +319,7 @@ class blockstore_impl_t
 
     blockstore_clean_db_t& clean_db_shard(object_id oid);
     void reshard_clean_db(pool_id_t pool_id, uint32_t pg_count, uint32_t pg_stripe_size);
+    void recalc_inode_space_stats(uint64_t pool_id, bool per_inode);
 
     // Journaling
     void prepare_journal_sector_write(int sector, blockstore_op_t *op);
@@ -428,6 +430,9 @@ public:
     // Space usage statistics
     std::map<uint64_t, uint64_t> inode_space_stats;
 
+    // Set per-pool no_inode_stats
+    void set_no_inode_stats(const std::vector<uint64_t> & pool_ids);
+
     // Print diagnostics to stdout
     void dump_diagnostics();
 

src/blockstore_init.cpp

@@ -32,7 +32,7 @@ void blockstore_init_meta::handle_event(ring_data_t *data, int buf_num)
     if (data->res < 0)
     {
         throw std::runtime_error(
-            std::string("read metadata failed at offset ") + std::to_string(bufs[buf_num].offset) +
+            std::string("read metadata failed at offset ") + std::to_string(buf_num >= 0 ? bufs[buf_num].offset : last_read_offset) +
             std::string(": ") + strerror(-data->res)
         );
     }
@@ -63,7 +63,8 @@ int blockstore_init_meta::loop()
         throw std::runtime_error("Failed to allocate metadata read buffer");
     // Read superblock
     GET_SQE();
-    data->iov = { metadata_buffer, bs->dsk.meta_block_size };
+    last_read_offset = 0;
+    data->iov = { metadata_buffer, (size_t)bs->dsk.meta_block_size };
     data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
     my_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
     bs->ringloop->submit();
@@ -100,7 +101,8 @@ resume_1:
         {
             printf("Initializing metadata area\n");
             GET_SQE();
-            data->iov = (struct iovec){ metadata_buffer, bs->dsk.meta_block_size };
+            last_read_offset = 0;
+            data->iov = (struct iovec){ metadata_buffer, (size_t)bs->dsk.meta_block_size };
             data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
             my_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset);
             bs->ringloop->submit();
@@ -153,7 +155,7 @@ resume_1:
         else if (hdr->version > BLOCKSTORE_META_FORMAT_V2)
         {
             printf(
-                "Metadata format is too new for me (stored version is %lu, max supported %u).\n",
+                "Metadata format is too new for me (stored version is %ju, max supported %u).\n",
                 hdr->version, BLOCKSTORE_META_FORMAT_V2
             );
             exit(1);
@@ -167,7 +169,7 @@ resume_1:
             printf(
                 "Configuration stored in metadata superblock"
                 " (meta_block_size=%u, data_block_size=%u, bitmap_granularity=%u, data_csum_type=%u, csum_block_size=%u)"
-                " differs from OSD configuration (%lu/%u/%lu, %u/%u).\n",
+                " differs from OSD configuration (%ju/%u/%ju, %u/%u).\n",
                 hdr->meta_block_size, hdr->data_block_size, hdr->bitmap_granularity,
                 hdr->data_csum_type, hdr->csum_block_size,
                 bs->dsk.meta_block_size, bs->dsk.data_block_size, bs->dsk.bitmap_granularity,
@@ -199,7 +201,8 @@ resume_2:
                 submitted++;
                 next_offset += bufs[i].size;
                 GET_SQE();
-                data->iov = { bufs[i].buf, bufs[i].size };
+                assert(bufs[i].size <= 0x7fffffff);
+                data->iov = { bufs[i].buf, (size_t)bufs[i].size };
                 data->callback = [this, i](ring_data_t *data) { handle_event(data, i); };
                 if (!zero_on_init)
                     my_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bufs[i].offset);
@@ -231,9 +234,11 @@ resume_2:
             {
                 // write the modified buffer back
                 GET_SQE();
-                data->iov = { bufs[i].buf, bufs[i].size };
+                assert(bufs[i].size <= 0x7fffffff);
+                data->iov = { bufs[i].buf, (size_t)bufs[i].size };
                 data->callback = [this, i](ring_data_t *data) { handle_event(data, i); };
                 my_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + bufs[i].offset);
+                bs->ringloop->submit();
                 bufs[i].state = INIT_META_WRITING;
                 submitted++;
             }
@@ -257,9 +262,11 @@ resume_2:
             next_offset = entries_to_zero[i]/entries_per_block;
             for (j = i; j < entries_to_zero.size() && entries_to_zero[j]/entries_per_block == next_offset; j++) {}
             GET_SQE();
-            data->iov = { metadata_buffer, bs->dsk.meta_block_size };
+            last_read_offset = (1+next_offset)*bs->dsk.meta_block_size;
+            data->iov = { metadata_buffer, (size_t)bs->dsk.meta_block_size };
             data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
             my_uring_prep_readv(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + (1+next_offset)*bs->dsk.meta_block_size);
+            bs->ringloop->submit();
             submitted++;
 resume_5:
             if (submitted > 0)
@@ -273,9 +280,10 @@ resume_5:
                 memset((uint8_t*)metadata_buffer + pos*bs->dsk.clean_entry_size, 0, bs->dsk.clean_entry_size);
             }
             GET_SQE();
-            data->iov = { metadata_buffer, bs->dsk.meta_block_size };
+            data->iov = { metadata_buffer, (size_t)bs->dsk.meta_block_size };
             data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
             my_uring_prep_writev(sqe, bs->dsk.meta_fd, &data->iov, 1, bs->dsk.meta_offset + (1+next_offset)*bs->dsk.meta_block_size);
+            bs->ringloop->submit();
             submitted++;
 resume_6:
             if (submitted > 0)
@@ -287,7 +295,7 @@ resume_6:
         entries_to_zero.clear();
     }
     // metadata read finished
-    printf("Metadata entries loaded: %lu, free blocks: %lu / %lu\n", entries_loaded, bs->data_alloc->get_free_count(), bs->dsk.block_count);
+    printf("Metadata entries loaded: %ju, free blocks: %ju / %ju\n", entries_loaded, bs->data_alloc->get_free_count(), bs->dsk.block_count);
     if (!bs->inmemory_meta)
     {
         free(metadata_buffer);
@@ -297,6 +305,7 @@ resume_6:
     {
         GET_SQE();
         my_uring_prep_fsync(sqe, bs->dsk.meta_fd, IORING_FSYNC_DATASYNC);
+        last_read_offset = 0;
         data->iov = { 0 };
         data->callback = [this](ring_data_t *data) { handle_event(data, -1); };
         submitted++;
@@ -328,7 +337,7 @@ bool blockstore_init_meta::handle_meta_block(uint8_t *buf, uint64_t entries_per_
                 uint32_t *entry_csum = (uint32_t*)((uint8_t*)entry + bs->dsk.clean_entry_size - 4);
                 if (*entry_csum != crc32c(0, entry, bs->dsk.clean_entry_size - 4))
                 {
-                    printf("Metadata entry %lu is corrupt (checksum mismatch), skipping\n", done_cnt+i);
+                    printf("Metadata entry %ju is corrupt (checksum mismatch), skipping\n", done_cnt+i);
                     continue;
                 }
             }
@@ -366,7 +375,7 @@ bool blockstore_init_meta::handle_meta_block(uint8_t *buf, uint64_t entries_per_
                         entries_to_zero.push_back(clean_it->second.location >> bs->dsk.block_order);
                     }
 #ifdef BLOCKSTORE_DEBUG
-                    printf("Free block %lu from %lx:%lx v%lu (new location is %lu)\n",
+                    printf("Free block %ju from %jx:%jx v%ju (new location is %ju)\n",
                         old_clean_loc,
                         clean_it->first.inode, clean_it->first.stripe, clean_it->second.version,
                         done_cnt+i);
@@ -380,7 +389,7 @@ bool blockstore_init_meta::handle_meta_block(uint8_t *buf, uint64_t entries_per_
                 }
                 entries_loaded++;
 #ifdef BLOCKSTORE_DEBUG
-                printf("Allocate block (clean entry) %lu: %lx:%lx v%lu\n", done_cnt+i, entry->oid.inode, entry->oid.stripe, entry->version);
+                printf("Allocate block (clean entry) %ju: %jx:%jx v%ju\n", done_cnt+i, entry->oid.inode, entry->oid.stripe, entry->version);
 #endif
                 bs->data_alloc->set(done_cnt+i, true);
                 clean_db[entry->oid] = (struct clean_entry){
@@ -394,7 +403,7 @@ bool blockstore_init_meta::handle_meta_block(uint8_t *buf, uint64_t entries_per_
                 updated = true;
                 memset(entry, 0, bs->dsk.clean_entry_size);
 #ifdef BLOCKSTORE_DEBUG
-                printf("Old clean entry %lu: %lx:%lx v%lu\n", done_cnt+i, entry->oid.inode, entry->oid.stripe, entry->version);
+                printf("Old clean entry %ju: %jx:%jx v%ju\n", done_cnt+i, entry->oid.inode, entry->oid.stripe, entry->version);
 #endif
             }
         }
@@ -466,7 +475,7 @@ int blockstore_init_journal::loop()
     if (!sqe)
         throw std::runtime_error("io_uring is full while trying to read journal");
     data = ((ring_data_t*)sqe->user_data);
-    data->iov = { submitted_buf, bs->journal.block_size };
+    data->iov = { submitted_buf, (size_t)bs->journal.block_size };
     data->callback = simple_callback;
     my_uring_prep_readv(sqe, bs->dsk.journal_fd, &data->iov, 1, bs->journal.offset);
     bs->ringloop->submit();
@@ -507,7 +516,7 @@ resume_1:
             // FIXME: Randomize initial crc32. Track crc32 when trimming.
             printf("Resetting journal\n");
             GET_SQE();
-            data->iov = (struct iovec){ submitted_buf, 2*bs->journal.block_size };
+            data->iov = (struct iovec){ submitted_buf, (size_t)(2*bs->journal.block_size) };
             data->callback = simple_callback;
             my_uring_prep_writev(sqe, bs->dsk.journal_fd, &data->iov, 1, bs->journal.offset);
             wait_count++;
@@ -557,7 +566,7 @@ resume_1:
             (je_start->version != JOURNAL_VERSION_V2 || je_start->size != JE_START_V2_SIZE && je_start->size != JE_START_V1_SIZE))
         {
             fprintf(
-                stderr, "The code only supports journal versions 2 and 1, but it is %lu on disk."
+                stderr, "The code only supports journal versions 2 and 1, but it is %ju on disk."
                     " Please use vitastor-disk to rewrite the journal\n",
                 je_start->size == JE_START_V0_SIZE ? 0 : je_start->version
             );
@@ -606,7 +615,7 @@ resume_1:
                     submitted_buf = (uint8_t*)bs->journal.buffer + journal_pos;
                 data->iov = {
                     submitted_buf,
-                    end - journal_pos < JOURNAL_BUFFER_SIZE ? end - journal_pos : JOURNAL_BUFFER_SIZE,
+                    (size_t)(end - journal_pos < JOURNAL_BUFFER_SIZE ? end - journal_pos : JOURNAL_BUFFER_SIZE),
                 };
                 data->callback = [this](ring_data_t *data1) { handle_event(data1); };
                 my_uring_prep_readv(sqe, bs->dsk.journal_fd, &data->iov, 1, bs->journal.offset + journal_pos);
@@ -622,7 +631,7 @@ resume_1:
                     if (init_write_buf && !bs->readonly)
                     {
                         GET_SQE();
-                        data->iov = { init_write_buf, bs->journal.block_size };
+                        data->iov = { init_write_buf, (size_t)bs->journal.block_size };
                         data->callback = simple_callback;
                         my_uring_prep_writev(sqe, bs->dsk.journal_fd, &data->iov, 1, bs->journal.offset + init_write_sector);
                         wait_count++;
@@ -691,7 +700,7 @@ resume_1:
             IS_BIG_WRITE(dirty_it->second.state) &&
             dirty_it->second.location == UINT64_MAX)
         {
-            printf("Fatal error (bug): %lx:%lx v%lu big_write journal_entry was allocated over another object\n",
+            printf("Fatal error (bug): %jx:%jx v%ju big_write journal_entry was allocated over another object\n",
                 dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version);
             exit(1);
         }
@@ -699,7 +708,7 @@ resume_1:
     bs->flusher->mark_trim_possible();
     bs->journal.dirty_start = bs->journal.next_free;
     printf(
-        "Journal entries loaded: %lu, free journal space: %lu bytes (%08lx..%08lx is used), free blocks: %lu / %lu\n",
+        "Journal entries loaded: %ju, free journal space: %ju bytes (%08jx..%08jx is used), free blocks: %ju / %ju\n",
         entries_loaded,
         (bs->journal.next_free >= bs->journal.used_start
             ? bs->journal.len-bs->journal.block_size - (bs->journal.next_free-bs->journal.used_start)
@@ -754,7 +763,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
             {
 #ifdef BLOCKSTORE_DEBUG
                 printf(
-                    "je_small_write%s oid=%lx:%lx ver=%lu offset=%u len=%u\n",
+                    "je_small_write%s oid=%jx:%jx ver=%ju offset=%u len=%u\n",
                     je->type == JE_SMALL_WRITE_INSTANT ? "_instant" : "",
                     je->small_write.oid.inode, je->small_write.oid.stripe, je->small_write.version,
                     je->small_write.offset, je->small_write.len
@@ -776,7 +785,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                 if (location != je->small_write.data_offset)
                 {
                     char err[1024];
-                    snprintf(err, 1024, "BUG: calculated journal data offset (%08lx) != stored journal data offset (%08lx)", location, je->small_write.data_offset);
+                    snprintf(err, 1024, "BUG: calculated journal data offset (%08jx) != stored journal data offset (%08jx)", location, je->small_write.data_offset);
                     throw std::runtime_error(err);
                 }
                 small_write_data.clear();
@@ -803,7 +812,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                             covered += part_end - part_begin;
                             small_write_data.push_back((iovec){
                                 .iov_base = (uint8_t*)done[i].buf + part_begin - done[i].pos,
-                                .iov_len = part_end - part_begin,
+                                .iov_len = (size_t)(part_end - part_begin),
                             });
                         }
                     }
@@ -826,7 +835,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                     if (!data_csum_valid)
                     {
                         printf(
-                            "Journal entry data is corrupt for small_write%s oid=%lx:%lx ver=%lu offset=%u len=%u - data crc32 %x != %x\n",
+                            "Journal entry data is corrupt for small_write%s oid=%jx:%jx ver=%ju offset=%u len=%u - data crc32 %x != %x\n",
                             je->type == JE_SMALL_WRITE_INSTANT ? "_instant" : "",
                             je->small_write.oid.inode, je->small_write.oid.stripe, je->small_write.version,
                             je->small_write.offset, je->small_write.len,
@@ -845,7 +854,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                     if (je->size != required_size)
                     {
                         printf(
-                            "Journal entry data has invalid size for small_write%s oid=%lx:%lx ver=%lu offset=%u len=%u - should be %u bytes but is %u bytes\n",
+                            "Journal entry data has invalid size for small_write%s oid=%jx:%jx ver=%ju offset=%u len=%u - should be %u bytes but is %u bytes\n",
                             je->type == JE_SMALL_WRITE_INSTANT ? "_instant" : "",
                             je->small_write.oid.inode, je->small_write.oid.stripe, je->small_write.version,
                             je->small_write.offset, je->small_write.len,
@@ -893,7 +902,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                             if (block_crc32 != *block_csums)
                             {
                                 printf(
-                                    "Journal entry data is corrupt for small_write%s oid=%lx:%lx ver=%lu offset=%u len=%u - block %u crc32 %x != %x\n",
+                                    "Journal entry data is corrupt for small_write%s oid=%jx:%jx ver=%ju offset=%u len=%u - block %u crc32 %x != %x\n",
                                     je->type == JE_SMALL_WRITE_INSTANT ? "_instant" : "",
                                     je->small_write.oid.inode, je->small_write.oid.stripe, je->small_write.version,
                                     je->small_write.offset, je->small_write.len,
@@ -956,7 +965,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                     bs->journal.used_sectors[proc_pos]++;
 #ifdef BLOCKSTORE_DEBUG
                     printf(
-                        "journal offset %08lx is used by %lx:%lx v%lu (%lu refs)\n",
+                        "journal offset %08jx is used by %jx:%jx v%ju (%ju refs)\n",
                         proc_pos, ov.oid.inode, ov.oid.stripe, ov.version, bs->journal.used_sectors[proc_pos]
                     );
 #endif
@@ -972,7 +981,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
             {
 #ifdef BLOCKSTORE_DEBUG
                 printf(
-                    "je_big_write%s oid=%lx:%lx ver=%lu loc=%lu\n",
+                    "je_big_write%s oid=%jx:%jx ver=%ju loc=%ju\n",
                     je->type == JE_BIG_WRITE_INSTANT ? "_instant" : "",
                     je->big_write.oid.inode, je->big_write.oid.stripe, je->big_write.version, je->big_write.location >> bs->dsk.block_order
                 );
@@ -1049,7 +1058,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                     {
 #ifdef BLOCKSTORE_DEBUG
                         printf(
-                            "Allocate block (journal) %lu: %lx:%lx v%lu\n",
+                            "Allocate block (journal) %ju: %jx:%jx v%ju\n",
                             je->big_write.location >> bs->dsk.block_order,
                             ov.oid.inode, ov.oid.stripe, ov.version
                         );
@@ -1059,7 +1068,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
                     bs->journal.used_sectors[proc_pos]++;
 #ifdef BLOCKSTORE_DEBUG
                     printf(
-                        "journal offset %08lx is used by %lx:%lx v%lu (%lu refs)\n",
+                        "journal offset %08jx is used by %jx:%jx v%ju (%ju refs)\n",
                         proc_pos, ov.oid.inode, ov.oid.stripe, ov.version, bs->journal.used_sectors[proc_pos]
                     );
 #endif
@@ -1074,7 +1083,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
             else if (je->type == JE_STABLE)
             {
 #ifdef BLOCKSTORE_DEBUG
-                printf("je_stable oid=%lx:%lx ver=%lu\n", je->stable.oid.inode, je->stable.oid.stripe, je->stable.version);
+                printf("je_stable oid=%jx:%jx ver=%ju\n", je->stable.oid.inode, je->stable.oid.stripe, je->stable.version);
 #endif
                 // oid, version
                 obj_ver_id ov = {
@@ -1086,7 +1095,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
             else if (je->type == JE_ROLLBACK)
             {
 #ifdef BLOCKSTORE_DEBUG
-                printf("je_rollback oid=%lx:%lx ver=%lu\n", je->rollback.oid.inode, je->rollback.oid.stripe, je->rollback.version);
+                printf("je_rollback oid=%jx:%jx ver=%ju\n", je->rollback.oid.inode, je->rollback.oid.stripe, je->rollback.version);
 #endif
                 // rollback dirty writes of <oid> up to <version>
                 obj_ver_id ov = {
@@ -1098,7 +1107,7 @@ int blockstore_init_journal::handle_journal_part(void *buf, uint64_t done_pos, u
             else if (je->type == JE_DELETE)
             {
 #ifdef BLOCKSTORE_DEBUG
-                printf("je_delete oid=%lx:%lx ver=%lu\n", je->del.oid.inode, je->del.oid.stripe, je->del.version);
+                printf("je_delete oid=%jx:%jx ver=%ju\n", je->del.oid.inode, je->del.oid.stripe, je->del.version);
 #endif
                 bool dirty_exists = false;
                 auto dirty_it = bs->dirty_db.upper_bound((obj_ver_id){

src/blockstore_init.h

@@ -23,6 +23,7 @@ class blockstore_init_meta
     struct ring_data_t *data;
     uint64_t md_offset = 0;
     uint64_t next_offset = 0;
+    uint64_t last_read_offset = 0;
     uint64_t entries_loaded = 0;
     unsigned entries_per_block = 0;
     int i = 0, j = 0;

src/blockstore_journal.cpp

@@ -90,8 +90,8 @@ int blockstore_journal_check_t::check_available(blockstore_op_t *op, int entries
             }
             // In fact, it's even more rare than "ran out of journal space", so print a warning
             printf(
-                "Ran out of journal sector buffers: %d/%lu buffers used (%d dirty), next buffer (%ld)"
-                " is %s and flushed %lu times. Consider increasing \'journal_sector_buffer_count\'\n",
+                "Ran out of journal sector buffers: %d/%ju buffers used (%d dirty), next buffer (%jd)"
+                " is %s and flushed %ju times. Consider increasing \'journal_sector_buffer_count\'\n",
                 used, bs->journal.sector_count, dirty, next_sector,
                 bs->journal.sector_info[next_sector].dirty ? "dirty" : "not dirty",
                 bs->journal.sector_info[next_sector].flush_count
@@ -103,7 +103,7 @@ int blockstore_journal_check_t::check_available(blockstore_op_t *op, int entries
     if (data_after > 0)
     {
         next_pos = next_pos + data_after;
-        if (next_pos > bs->journal.len)
+        if (next_pos >= bs->journal.len)
         {
             if (right_dir)
                 next_pos = bs->journal.block_size + data_after;
@@ -114,7 +114,7 @@ int blockstore_journal_check_t::check_available(blockstore_op_t *op, int entries
     {
         // No space in the journal. Wait until used_start changes.
         printf(
-            "Ran out of journal space (used_start=%08lx, next_free=%08lx, dirty_start=%08lx)\n",
+            "Ran out of journal space (used_start=%08jx, next_free=%08jx, dirty_start=%08jx)\n",
             bs->journal.used_start, bs->journal.next_free, bs->journal.dirty_start
         );
         PRIV(op)->wait_for = WAIT_JOURNAL;
@@ -146,7 +146,7 @@ journal_entry* prefill_single_journal_entry(journal_t & journal, uint16_t type,
         journal.in_sector_pos = 0;
         auto next_next_free = (journal.next_free+journal.block_size) < journal.len ? journal.next_free + journal.block_size : journal.block_size;
         // double check that next_free doesn't cross used_start from the left
-        assert(journal.next_free >= journal.used_start || next_next_free < journal.used_start);
+        assert(journal.next_free >= journal.used_start && next_next_free >= journal.next_free || next_next_free < journal.used_start);
         journal.next_free = next_next_free;
         memset(journal.inmemory
             ? (uint8_t*)journal.buffer + journal.sector_info[journal.cur_sector].offset
@@ -183,7 +183,7 @@ void blockstore_impl_t::prepare_journal_sector_write(int cur_sector, blockstore_
             (journal.inmemory
                 ? (uint8_t*)journal.buffer + journal.sector_info[cur_sector].offset
                 : (uint8_t*)journal.sector_buf + journal.block_size*cur_sector),
-            journal.block_size
+            (size_t)journal.block_size
         };
         data->callback = [this, flush_id = journal.submit_id](ring_data_t *data) { handle_journal_write(data, flush_id); };
         my_uring_prep_writev(
@@ -263,7 +263,7 @@ uint64_t journal_t::get_trim_pos()
             // next_free does not need updating during trim
 #ifdef BLOCKSTORE_DEBUG
             printf(
-                "Trimming journal (used_start=%08lx, next_free=%08lx, dirty_start=%08lx, new_start=%08lx, new_refcount=%ld)\n",
+                "Trimming journal (used_start=%08jx, next_free=%08jx, dirty_start=%08jx, new_start=%08jx, new_refcount=%jd)\n",
                 used_start, next_free, dirty_start,
                 journal_used_it->first, journal_used_it->second
             );
@@ -276,7 +276,7 @@ uint64_t journal_t::get_trim_pos()
         // Journal is cleared up to <journal_used_it>
 #ifdef BLOCKSTORE_DEBUG
         printf(
-            "Trimming journal (used_start=%08lx, next_free=%08lx, dirty_start=%08lx, new_start=%08lx, new_refcount=%ld)\n",
+            "Trimming journal (used_start=%08jx, next_free=%08jx, dirty_start=%08jx, new_start=%08jx, new_refcount=%jd)\n",
             used_start, next_free, dirty_start,
             journal_used_it->first, journal_used_it->second
         );
@@ -296,7 +296,7 @@ void journal_t::dump_diagnostics()
         journal_used_it = used_sectors.begin();
     }
     printf(
-        "Journal: used_start=%08lx next_free=%08lx dirty_start=%08lx trim_to=%08lx trim_to_refs=%ld\n",
+        "Journal: used_start=%08jx next_free=%08jx dirty_start=%08jx trim_to=%08jx trim_to_refs=%jd\n",
         used_start, next_free, dirty_start,
         journal_used_it == used_sectors.end() ? 0 : journal_used_it->first,
         journal_used_it == used_sectors.end() ? 0 : journal_used_it->second

src/blockstore_read.cpp

@@ -25,7 +25,7 @@ int blockstore_impl_t::fulfill_read_push(blockstore_op_t *op, void *buf, uint64_
         return 1;
     }
     BS_SUBMIT_GET_SQE(sqe, data);
-    data->iov = (struct iovec){ buf, len };
+    data->iov = (struct iovec){ buf, (size_t)len };
     PRIV(op)->pending_ops++;
     my_uring_prep_readv(
         sqe,
@@ -505,7 +505,7 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *read_op)
         for (auto & rv: PRIV(read_op)->read_vec)
         {
             if (rv.journal_sector)
-                journal.used_sectors[rv.journal_sector-1]++;
+                journal.used_sectors.at(rv.journal_sector-1)++;
         }
     }
     read_op->retval = 0;
@@ -700,7 +700,7 @@ uint8_t* blockstore_impl_t::read_clean_meta_block(blockstore_op_t *op, uint64_t
         .buf = buf,
     });
     BS_SUBMIT_GET_SQE(sqe, data);
-    data->iov = (struct iovec){ buf, dsk.meta_block_size };
+    data->iov = (struct iovec){ buf, (size_t)dsk.meta_block_size };
     PRIV(op)->pending_ops++;
     my_uring_prep_readv(sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + dsk.meta_block_size + sector);
     data->callback = [this, op](ring_data_t *data) { handle_read_event(data, op); };
@@ -855,7 +855,7 @@ void blockstore_impl_t::handle_read_event(ring_data_t *data, blockstore_op_t *op
                             {
                                 ok = false;
                                 printf(
-                                    "Checksum mismatch in object %lx:%lx v%lu in journal at 0x%lx, checksum block #%u: got %08x, expected %08x\n",
+                                    "Checksum mismatch in object %jx:%jx v%ju in journal at 0x%jx, checksum block #%u: got %08x, expected %08x\n",
                                     op->oid.inode, op->oid.stripe, op->version,
                                     rv[i].disk_offset, bad_block / dsk.csum_block_size, calc_csum, stored_csum
                                 );
@@ -875,7 +875,7 @@ void blockstore_impl_t::handle_read_event(ring_data_t *data, blockstore_op_t *op
                                 {
                                     ok = false;
                                     printf(
-                                        "Checksum mismatch in object %lx:%lx v%lu in %s data at 0x%lx, checksum block #%u: got %08x, expected %08x\n",
+                                        "Checksum mismatch in object %jx:%jx v%ju in %s data at 0x%jx, checksum block #%u: got %08x, expected %08x\n",
                                         op->oid.inode, op->oid.stripe, op->version,
                                         (rv[i].copy_flags & COPY_BUF_JOURNALED_BIG ? "redirect-write" : "clean"),
                                         rv[i].disk_offset, bad_block / dsk.csum_block_size, calc_csum, stored_csum
@@ -918,7 +918,7 @@ void blockstore_impl_t::handle_read_event(ring_data_t *data, blockstore_op_t *op
                             {
                                 // checksum error
                                 printf(
-                                    "Checksum mismatch in object %lx:%lx v%lu in %s area at offset 0x%lx+0x%lx: %08x vs %08x\n",
+                                    "Checksum mismatch in object %jx:%jx v%ju in %s area at offset 0x%jx+0x%zx: %08x vs %08x\n",
                                     op->oid.inode, op->oid.stripe, op->version,
                                     (vec.copy_flags & COPY_BUF_JOURNAL) ? "journal" : "data", vec.disk_offset, p,
                                     crc32c(0, (uint8_t*)op->buf + vec.offset - op->offset + p, dsk.csum_block_size), *csum
@@ -966,7 +966,7 @@ void blockstore_impl_t::handle_read_event(ring_data_t *data, blockstore_op_t *op
             {
                 if (rv.journal_sector)
                 {
-                    auto used = --journal.used_sectors[rv.journal_sector-1];
+                    auto used = --journal.used_sectors.at(rv.journal_sector-1);
                     if (used == 0)
                     {
                         journal.used_sectors.erase(rv.journal_sector-1);

src/blockstore_rollback.cpp

@@ -179,7 +179,7 @@ void blockstore_impl_t::erase_dirty(blockstore_dirty_db_t::iterator dirty_start,
     {
         object_id oid = dirty_it->first.oid;
 #ifdef BLOCKSTORE_DEBUG
-        printf("Unblock writes-after-delete %lx:%lx v%lu\n", oid.inode, oid.stripe, dirty_it->first.version);
+        printf("Unblock writes-after-delete %jx:%jx v%ju\n", oid.inode, oid.stripe, dirty_it->first.version);
 #endif
         dirty_it = dirty_end;
         // Unblock operations blocked by delete flushing
@@ -210,21 +210,26 @@ void blockstore_impl_t::erase_dirty(blockstore_dirty_db_t::iterator dirty_start,
             dirty_it->second.location != UINT64_MAX)
         {
 #ifdef BLOCKSTORE_DEBUG
-            printf("Free block %lu from %lx:%lx v%lu\n", dirty_it->second.location >> dsk.block_order,
+            printf("Free block %ju from %jx:%jx v%ju\n", dirty_it->second.location >> dsk.block_order,
                 dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version);
 #endif
             data_alloc->set(dirty_it->second.location >> dsk.block_order, false);
         }
-        auto used = --journal.used_sectors[dirty_it->second.journal_sector];
+        auto used = --journal.used_sectors.at(dirty_it->second.journal_sector);
 #ifdef BLOCKSTORE_DEBUG
         printf(
-            "remove usage of journal offset %08lx by %lx:%lx v%lu (%lu refs)\n", dirty_it->second.journal_sector,
+            "remove usage of journal offset %08jx by %jx:%jx v%ju (%ju refs)\n", dirty_it->second.journal_sector,
             dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version, used
         );
 #endif
         if (used == 0)
         {
             journal.used_sectors.erase(dirty_it->second.journal_sector);
+            if (dirty_it->second.journal_sector == journal.sector_info[journal.cur_sector].offset)
+            {
+                // Mark current sector as "full" to select the new one
+                journal.in_sector_pos = dsk.journal_block_size;
+            }
             flusher->mark_trim_possible();
         }
         free_dirty_dyn_data(dirty_it->second);

src/blockstore_stable.cpp

@@ -298,7 +298,7 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
             if (clean_it == clean_db.end() || clean_it->second.version < ov.version)
             {
                 // No such object version
-                printf("Error: %lx:%lx v%lu not found while stabilizing\n", ov.oid.inode, ov.oid.stripe, ov.version);
+                printf("Error: %jx:%jx v%ju not found while stabilizing\n", ov.oid.inode, ov.oid.stripe, ov.version);
                 return -ENOENT;
             }
             else
@@ -307,35 +307,49 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
                 return STAB_SPLIT_DONE;
             }
         }
-        else if (IS_IN_FLIGHT(dirty_it->second.state))
-        {
-            // Object write is still in progress. Wait until the write request completes
-            return STAB_SPLIT_WAIT;
-        }
-        else if (!IS_SYNCED(dirty_it->second.state))
-        {
-            // Object not synced yet - sync it
-            // In previous versions we returned EBUSY here and required
-            // the caller (OSD) to issue a global sync first. But a global sync
-            // waits for all writes in the queue including inflight writes. And
-            // inflight writes may themselves be blocked by unstable writes being
-            // still present in the journal and not flushed away from it.
-            // So we must sync specific objects here.
-            //
-            // Even more, we have to process "stabilize" request in parts. That is,
-            // we must stabilize all objects which are already synced. Otherwise
-            // they may block objects which are NOT synced yet.
-            return STAB_SPLIT_SYNC;
-        }
         else if (IS_STABLE(dirty_it->second.state))
         {
             // Already stable
             return STAB_SPLIT_DONE;
         }
-        else
+        while (true)
         {
-            return STAB_SPLIT_TODO;
+            if (IS_IN_FLIGHT(dirty_it->second.state))
+            {
+                // Object write is still in progress. Wait until the write request completes
+                return STAB_SPLIT_WAIT;
+            }
+            else if (!IS_SYNCED(dirty_it->second.state))
+            {
+                // Object not synced yet - sync it
+                // In previous versions we returned EBUSY here and required
+                // the caller (OSD) to issue a global sync first. But a global sync
+                // waits for all writes in the queue including inflight writes. And
+                // inflight writes may themselves be blocked by unstable writes being
+                // still present in the journal and not flushed away from it.
+                // So we must sync specific objects here.
+                //
+                // Even more, we have to process "stabilize" request in parts. That is,
+                // we must stabilize all objects which are already synced. Otherwise
+                // they may block objects which are NOT synced yet.
+                return STAB_SPLIT_SYNC;
+            }
+            else if (IS_STABLE(dirty_it->second.state))
+            {
+                break;
+            }
+            // Check previous versions too
+            if (dirty_it == dirty_db.begin())
+            {
+                break;
+            }
+            dirty_it--;
+            if (dirty_it->first.oid != ov.oid)
+            {
+                break;
+            }
         }
+        return STAB_SPLIT_TODO;
     });
     if (r != 1)
     {
@@ -402,7 +416,7 @@ resume_4:
     {
         // Mark all dirty_db entries up to op->version as stable
 #ifdef BLOCKSTORE_DEBUG
-        printf("Stabilize %lx:%lx v%lu\n", v->oid.inode, v->oid.stripe, v->version);
+        printf("Stabilize %jx:%jx v%ju\n", v->oid.inode, v->oid.stripe, v->version);
 #endif
         mark_stable(*v);
     }
@@ -473,18 +487,24 @@ void blockstore_impl_t::mark_stable(obj_ver_id v, bool forget_dirty)
                     }
                     if (!exists)
                     {
-                        inode_space_stats[dirty_it->first.oid.inode] += dsk.data_block_size;
+                        uint64_t space_id = dirty_it->first.oid.inode;
+                        if (no_inode_stats[dirty_it->first.oid.inode >> (64-POOL_ID_BITS)])
+                            space_id = space_id & ~(((uint64_t)1 << (64-POOL_ID_BITS)) - 1);
+                        inode_space_stats[space_id] += dsk.data_block_size;
                         used_blocks++;
                     }
                     big_to_flush++;
                 }
                 else if (IS_DELETE(dirty_it->second.state))
                 {
-                    auto & sp = inode_space_stats[dirty_it->first.oid.inode];
+                    uint64_t space_id = dirty_it->first.oid.inode;
+                    if (no_inode_stats[dirty_it->first.oid.inode >> (64-POOL_ID_BITS)])
+                        space_id = space_id & ~(((uint64_t)1 << (64-POOL_ID_BITS)) - 1);
+                    auto & sp = inode_space_stats[space_id];
                     if (sp > dsk.data_block_size)
                         sp -= dsk.data_block_size;
                     else
-                        inode_space_stats.erase(dirty_it->first.oid.inode);
+                        inode_space_stats.erase(space_id);
                     used_blocks--;
                     big_to_flush++;
                 }
@@ -493,7 +513,7 @@ void blockstore_impl_t::mark_stable(obj_ver_id v, bool forget_dirty)
             {
                 // mark_stable should never be called for in-flight or submitted writes
                 printf(
-                    "BUG: Attempt to mark_stable object %lx:%lx v%lu state of which is %x\n",
+                    "BUG: Attempt to mark_stable object %jx:%jx v%ju state of which is %x\n",
                     dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version,
                     dirty_it->second.state
                 );

src/blockstore_sync.cpp

@@ -76,7 +76,6 @@ int blockstore_impl_t::continue_sync(blockstore_op_t *op)
         // 2nd step: Data device is synced, prepare & write journal entries
         // Check space in the journal and journal memory buffers
         blockstore_journal_check_t space_check(this);
-        auto reservation = (unstable_writes.size()+unstable_unsynced+PRIV(op)->sync_big_writes.size())*journal.block_size;
         if (dsk.csum_block_size)
         {
             // More complex check because all journal entries have different lengths
@@ -86,14 +85,14 @@ int blockstore_impl_t::continue_sync(blockstore_op_t *op)
                 left--;
                 auto & dirty_entry = dirty_db.at(sbw);
                 uint64_t dyn_size = dsk.dirty_dyn_size(dirty_entry.offset, dirty_entry.len);
-                if (!space_check.check_available(op, 1, sizeof(journal_entry_big_write) + dyn_size, left ? 0 : reservation))
+                if (!space_check.check_available(op, 1, sizeof(journal_entry_big_write) + dyn_size, 0))
                 {
                     return 0;
                 }
             }
         }
         else if (!space_check.check_available(op, PRIV(op)->sync_big_writes.size(),
-            sizeof(journal_entry_big_write) + dsk.clean_entry_bitmap_size, reservation))
+            sizeof(journal_entry_big_write) + dsk.clean_entry_bitmap_size, 0))
         {
             return 0;
         }
@@ -116,11 +115,14 @@ int blockstore_impl_t::continue_sync(blockstore_op_t *op)
                 journal, (dirty_entry.state & BS_ST_INSTANT) ? JE_BIG_WRITE_INSTANT : JE_BIG_WRITE,
                 sizeof(journal_entry_big_write) + dyn_size
             );
-            dirty_entry.journal_sector = journal.sector_info[journal.cur_sector].offset;
+            auto jsec = dirty_entry.journal_sector = journal.sector_info[journal.cur_sector].offset;
+            assert(journal.next_free >= journal.used_start
+                ? (jsec >= journal.used_start && jsec < journal.next_free)
+                : (jsec >= journal.used_start || jsec < journal.next_free));
             journal.used_sectors[journal.sector_info[journal.cur_sector].offset]++;
 #ifdef BLOCKSTORE_DEBUG
             printf(
-                "journal offset %08lx is used by %lx:%lx v%lu (%lu refs)\n",
+                "journal offset %08jx is used by %jx:%jx v%ju (%ju refs)\n",
                 dirty_entry.journal_sector, it->oid.inode, it->oid.stripe, it->version,
                 journal.used_sectors[journal.sector_info[journal.cur_sector].offset]
             );
@@ -174,7 +176,7 @@ void blockstore_impl_t::ack_sync(blockstore_op_t *op)
     for (auto it = PRIV(op)->sync_big_writes.begin(); it != PRIV(op)->sync_big_writes.end(); it++)
     {
 #ifdef BLOCKSTORE_DEBUG
-        printf("Ack sync big %lx:%lx v%lu\n", it->oid.inode, it->oid.stripe, it->version);
+        printf("Ack sync big %jx:%jx v%ju\n", it->oid.inode, it->oid.stripe, it->version);
 #endif
         auto & unstab = unstable_writes[it->oid];
         unstab = unstab < it->version ? it->version : unstab;
@@ -202,7 +204,7 @@ void blockstore_impl_t::ack_sync(blockstore_op_t *op)
     for (auto it = PRIV(op)->sync_small_writes.begin(); it != PRIV(op)->sync_small_writes.end(); it++)
     {
 #ifdef BLOCKSTORE_DEBUG
-        printf("Ack sync small %lx:%lx v%lu\n", it->oid.inode, it->oid.stripe, it->version);
+        printf("Ack sync small %jx:%jx v%ju\n", it->oid.inode, it->oid.stripe, it->version);
 #endif
         auto & unstab = unstable_writes[it->oid];
         unstab = unstab < it->version ? it->version : unstab;

src/blockstore_write.cpp

@@ -85,7 +85,7 @@ bool blockstore_impl_t::enqueue_write(blockstore_op_t *op)
             // It's allowed to write versions with low numbers over deletes
             // However, we have to flush those deletes first as we use version number for ordering
 #ifdef BLOCKSTORE_DEBUG
-            printf("Write %lx:%lx v%lu over delete (real v%lu) offset=%u len=%u\n", op->oid.inode, op->oid.stripe, version, op->version, op->offset, op->len);
+            printf("Write %jx:%jx v%ju over delete (real v%ju) offset=%u len=%u\n", op->oid.inode, op->oid.stripe, version, op->version, op->offset, op->len);
 #endif
             wait_del = true;
             PRIV(op)->real_version = op->version;
@@ -95,11 +95,13 @@ bool blockstore_impl_t::enqueue_write(blockstore_op_t *op)
                 // Issue an additional sync so the delete reaches the journal
                 blockstore_op_t *sync_op = new blockstore_op_t;
                 sync_op->opcode = BS_OP_SYNC;
-                sync_op->callback = [this, op](blockstore_op_t *sync_op)
+                sync_op->oid = op->oid;
+                sync_op->version = op->version;
+                sync_op->callback = [this](blockstore_op_t *sync_op)
                 {
                     flusher->unshift_flush((obj_ver_id){
-                        .oid = op->oid,
-                        .version = op->version-1,
+                        .oid = sync_op->oid,
+                        .version = sync_op->version-1,
                     }, true);
                     delete sync_op;
                 };
@@ -117,7 +119,7 @@ bool blockstore_impl_t::enqueue_write(blockstore_op_t *op)
         {
             // Invalid version requested
 #ifdef BLOCKSTORE_DEBUG
-            printf("Write %lx:%lx v%lu requested, but we already have v%lu\n", op->oid.inode, op->oid.stripe, op->version, version);
+            printf("Write %jx:%jx v%ju requested, but we already have v%ju\n", op->oid.inode, op->oid.stripe, op->version, version);
 #endif
             op->retval = -EEXIST;
             if (!is_del && alloc_dyn_data)
@@ -144,9 +146,9 @@ bool blockstore_impl_t::enqueue_write(blockstore_op_t *op)
         unsynced_queued_ops++;
 #ifdef BLOCKSTORE_DEBUG
     if (is_del)
-        printf("Delete %lx:%lx v%lu\n", op->oid.inode, op->oid.stripe, op->version);
+        printf("Delete %jx:%jx v%ju\n", op->oid.inode, op->oid.stripe, op->version);
     else if (!wait_del)
-        printf("Write %lx:%lx v%lu offset=%u len=%u\n", op->oid.inode, op->oid.stripe, op->version, op->offset, op->len);
+        printf("Write %jx:%jx v%ju offset=%u len=%u\n", op->oid.inode, op->oid.stripe, op->version, op->offset, op->len);
 #endif
     // No strict need to add it into dirty_db here except maybe for listings to return
     // correct data when there are inflight operations in the queue
@@ -286,7 +288,7 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
         }
         // Restore original low version number for unblocked operations
 #ifdef BLOCKSTORE_DEBUG
-        printf("Restoring %lx:%lx version: v%lu -> v%lu\n", op->oid.inode, op->oid.stripe, op->version, PRIV(op)->real_version);
+        printf("Restoring %jx:%jx version: v%ju -> v%ju\n", op->oid.inode, op->oid.stripe, op->version, PRIV(op)->real_version);
 #endif
         auto prev_it = dirty_it;
         if (prev_it != dirty_db.begin())
@@ -296,7 +298,7 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
             {
                 // Original version is still invalid
                 // All subsequent writes to the same object must be canceled too
-                printf("Tried to write %lx:%lx v%lu after delete (old version v%lu), but already have v%lu\n",
+                printf("Tried to write %jx:%jx v%ju after delete (old version v%ju), but already have v%ju\n",
                     op->oid.inode, op->oid.stripe, PRIV(op)->real_version, op->version, prev_it->first.version);
                 cancel_all_writes(op, dirty_it, -EEXIST);
                 return 2;
@@ -348,8 +350,8 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
             if (entry->oid.inode || entry->oid.stripe || entry->version)
             {
                 printf(
-                    "Fatal error (metadata corruption or bug): tried to write object %lx:%lx v%lu"
-                    " over a non-zero metadata entry %lu with %lx:%lx v%lu\n", op->oid.inode,
+                    "Fatal error (metadata corruption or bug): tried to write object %jx:%jx v%ju"
+                    " over a non-zero metadata entry %ju with %jx:%jx v%ju\n", op->oid.inode,
                     op->oid.stripe, op->version, loc, entry->oid.inode, entry->oid.stripe, entry->version
                 );
                 exit(1);
@@ -361,7 +363,7 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
         dirty_it->second.state = (dirty_it->second.state & ~BS_ST_WORKFLOW_MASK) | BS_ST_SUBMITTED;
 #ifdef BLOCKSTORE_DEBUG
         printf(
-            "Allocate block %lu for %lx:%lx v%lu\n",
+            "Allocate block %ju for %jx:%jx v%ju\n",
             loc, op->oid.inode, op->oid.stripe, op->version
         );
 #endif
@@ -372,13 +374,13 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
         int vcnt = 0;
         if (stripe_offset)
         {
-            PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ zero_object, stripe_offset };
+            PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ zero_object, (size_t)stripe_offset };
         }
         PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ op->buf, op->len };
         if (stripe_end)
         {
             stripe_end = dsk.bitmap_granularity - stripe_end;
-            PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ zero_object, stripe_end };
+            PRIV(op)->iov_zerofill[vcnt++] = (struct iovec){ zero_object, (size_t)stripe_end };
         }
         data->iov.iov_len = op->len + stripe_offset + stripe_end; // to check it in the callback
         data->callback = [this, op](ring_data_t *data) { handle_write_event(data, op); };
@@ -436,11 +438,23 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
             journal, op->opcode == BS_OP_WRITE_STABLE ? JE_SMALL_WRITE_INSTANT : JE_SMALL_WRITE,
             sizeof(journal_entry_small_write) + dyn_size
         );
-        dirty_it->second.journal_sector = journal.sector_info[journal.cur_sector].offset;
+        auto jsec = dirty_it->second.journal_sector = journal.sector_info[journal.cur_sector].offset;
+        if (!(journal.next_free >= journal.used_start
+            ? (jsec >= journal.used_start && jsec < journal.next_free)
+            : (jsec >= journal.used_start || jsec < journal.next_free)))
+        {
+            printf(
+                "BUG: journal offset %08jx is used by %jx:%jx v%ju (%ju refs) BUT used_start=%jx next_free=%jx\n",
+                dirty_it->second.journal_sector, dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version,
+                journal.used_sectors[journal.sector_info[journal.cur_sector].offset],
+                journal.used_start, journal.next_free
+            );
+            abort();
+        }
         journal.used_sectors[journal.sector_info[journal.cur_sector].offset]++;
 #ifdef BLOCKSTORE_DEBUG
         printf(
-            "journal offset %08lx is used by %lx:%lx v%lu (%lu refs)\n",
+            "journal offset %08jx is used by %jx:%jx v%ju (%ju refs)\n",
             dirty_it->second.journal_sector, dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version,
             journal.used_sectors[journal.sector_info[journal.cur_sector].offset]
         );
@@ -454,8 +468,8 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
                 journal_used_it->first < next_next_free + op->len)
             {
                 printf(
-                    "BUG: Attempt to overwrite used offset (%lx, %lu refs) of the journal with the object %lx:%lx v%lu: data at %lx, len %x!"
-                    " Journal used_start=%08lx (%lu refs), next_free=%08lx, dirty_start=%08lx\n",
+                    "BUG: Attempt to overwrite used offset (%jx, %ju refs) of the journal with the object %jx:%jx v%ju: data at %jx, len %x!"
+                    " Journal used_start=%08jx (%ju refs), next_free=%08jx, dirty_start=%08jx\n",
                     journal_used_it->first, journal_used_it->second, op->oid.inode, op->oid.stripe, op->version, next_next_free, op->len,
                     journal.used_start, journal.used_sectors[journal.used_start], journal.next_free, journal.dirty_start
                 );
@@ -463,7 +477,7 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
             }
         }
         // double check that next_free doesn't cross used_start from the left
-        assert(journal.next_free >= journal.used_start || next_next_free < journal.used_start);
+        assert(journal.next_free >= journal.used_start && next_next_free >= journal.next_free || next_next_free < journal.used_start);
         journal.next_free = next_next_free;
         je->oid = op->oid;
         je->version = op->version;
@@ -505,7 +519,7 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
         if (next_next_free >= journal.len)
             next_next_free = dsk.journal_block_size;
         // double check that next_free doesn't cross used_start from the left
-        assert(journal.next_free >= journal.used_start || next_next_free < journal.used_start);
+        assert(journal.next_free >= journal.used_start && next_next_free >= journal.next_free || next_next_free < journal.used_start);
         journal.next_free = next_next_free;
         if (!(dirty_it->second.state & BS_ST_INSTANT))
         {
@@ -558,11 +572,23 @@ resume_2:
             journal, op->opcode == BS_OP_WRITE_STABLE ? JE_BIG_WRITE_INSTANT : JE_BIG_WRITE,
             sizeof(journal_entry_big_write) + dyn_size
         );
-        dirty_it->second.journal_sector = journal.sector_info[journal.cur_sector].offset;
+        auto jsec = dirty_it->second.journal_sector = journal.sector_info[journal.cur_sector].offset;
+        if (!(journal.next_free >= journal.used_start
+            ? (jsec >= journal.used_start && jsec < journal.next_free)
+            : (jsec >= journal.used_start || jsec < journal.next_free)))
+        {
+            printf(
+                "BUG: journal offset %08jx is used by %jx:%jx v%ju (%ju refs) BUT used_start=%jx next_free=%jx\n",
+                dirty_it->second.journal_sector, dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version,
+                journal.used_sectors[journal.sector_info[journal.cur_sector].offset],
+                journal.used_start, journal.next_free
+            );
+            abort();
+        }
         journal.used_sectors[journal.sector_info[journal.cur_sector].offset]++;
 #ifdef BLOCKSTORE_DEBUG
         printf(
-            "journal offset %08lx is used by %lx:%lx v%lu (%lu refs)\n",
+            "journal offset %08jx is used by %jx:%jx v%ju (%ju refs)\n",
             journal.sector_info[journal.cur_sector].offset, op->oid.inode, op->oid.stripe, op->version,
             journal.used_sectors[journal.sector_info[journal.cur_sector].offset]
         );
@@ -589,7 +615,7 @@ resume_4:
         });
         assert(dirty_it != dirty_db.end());
 #ifdef BLOCKSTORE_DEBUG
-        printf("Ack write %lx:%lx v%lu = state 0x%x\n", op->oid.inode, op->oid.stripe, op->version, dirty_it->second.state);
+        printf("Ack write %jx:%jx v%ju = state 0x%x\n", op->oid.inode, op->oid.stripe, op->version, dirty_it->second.state);
 #endif
         bool is_big = (dirty_it->second.state & BS_ST_TYPE_MASK) == BS_ST_BIG_WRITE;
         bool imm = is_big ? (immediate_commit == IMMEDIATE_ALL) : (immediate_commit != IMMEDIATE_NONE);
@@ -782,7 +808,7 @@ int blockstore_impl_t::dequeue_del(blockstore_op_t *op)
     journal.used_sectors[journal.sector_info[journal.cur_sector].offset]++;
 #ifdef BLOCKSTORE_DEBUG
     printf(
-        "journal offset %08lx is used by %lx:%lx v%lu (%lu refs)\n",
+        "journal offset %08jx is used by %jx:%jx v%ju (%ju refs)\n",
         dirty_it->second.journal_sector, dirty_it->first.oid.inode, dirty_it->first.oid.stripe, dirty_it->first.version,
         journal.used_sectors[journal.sector_info[journal.cur_sector].offset]
     );

src/cli.cpp

@@ -46,18 +46,21 @@ static const char* help_text =
     "vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>\n"
     "  Create a snapshot of image <name>. May be used live if only a single writer is active.\n"
     "\n"
-    "vitastor-cli modify <name> [--rename <new-name>] [--resize <size>] [--readonly | --readwrite] [-f|--force]\n"
+    "vitastor-cli modify <name> [--rename <new-name>] [--resize <size>] [--readonly | --readwrite] [-f|--force] [--down-ok]\n"
     "  Rename, resize image or change its readonly status. Images with children can't be made read-write.\n"
     "  If the new size is smaller than the old size, extra data will be purged.\n"
     "  You should resize file system in the image, if present, before shrinking it.\n"
     "  -f|--force  Proceed with shrinking or setting readwrite flag even if the image has children.\n"
+    "  --down-ok   Proceed with shrinking even if some data will be left on unavailable OSDs.\n"
     "\n"
-    "vitastor-cli rm <from> [<to>] [--writers-stopped]\n"
+    "vitastor-cli rm <from> [<to>] [--writers-stopped] [--down-ok]\n"
     "  Remove <from> or all layers between <from> and <to> (<to> must be a child of <from>),\n"
     "  rebasing all their children accordingly. --writers-stopped allows merging to be a bit\n"
     "  more effective in case of a single 'slim' read-write child and 'fat' removed parent:\n"
     "  the child is merged into parent and parent is renamed to child in that case.\n"
     "  In other cases parent layers are always merged into children.\n"
+    "  Other options:\n"
+    "  --down-ok  Continue deletion/merging even if some data will be left on unavailable OSDs.\n"
     "\n"
     "vitastor-cli flatten <layer>\n"
     "  Flatten a layer, i.e. merge data and detach it from parents.\n"
@@ -113,6 +116,55 @@ static const char* help_text =
     "  With --dry-run only checks if deletion is possible without data loss and\n"
     "  redundancy degradation.\n"
     "\n"
+    "vitastor-cli create-pool|pool-create <name> (-s <pg_size>|--ec <N>+<K>) -n <pg_count> [OPTIONS]\n"
+    "  Create a pool. Required parameters:\n"
+    "    -s|--pg_size R   Number of replicas for replicated pools\n"
+    "    --ec N+K         Number of data (N) and parity (K) chunks for erasure-coded pools\n"
+    "    -n|--pg_count N  PG count for the new pool (start with 10*<OSD count>/pg_size rounded to a power of 2)\n"
+    "  Optional parameters:\n"
+    "    --pg_minsize <number>         R or N+K minus number of failures to tolerate without downtime\n"
+    "    --failure_domain host         Failure domain: host, osd or a level from placement_levels. Default: host\n"
+    "    --root_node <node>            Put pool only on child OSDs of this placement tree node\n"
+    "    --osd_tags <tag>[,<tag>]...   Put pool only on OSDs tagged with all specified tags\n"
+    "    --block_size 128k             Put pool only on OSDs with this data block size\n"
+    "    --bitmap_granularity 4k       Put pool only on OSDs with this logical sector size\n"
+    "    --immediate_commit none       Put pool only on OSDs with this or larger immediate_commit (none < small < all)\n"
+    "    --primary_affinity_tags tags  Prefer to put primary copies on OSDs with all specified tags\n"
+    "    --scrub_interval <time>       Enable regular scrubbing for this pool. Format: number + unit s/m/h/d/M/y\n"
+    "    --used_for_fs <name>          Mark pool as used for VitastorFS with metadata in image <name>\n"
+    "    --pg_stripe_size <number>     Increase object grouping stripe\n"
+    "    --max_osd_combinations 10000  Maximum number of random combinations for LP solver input\n"
+    "    --wait                        Wait for the new pool to come online\n"
+    "    -f|--force                    Do not check that cluster has enough OSDs to create the pool\n"
+    "  Examples:\n"
+    "    vitastor-cli create-pool test_x4 -s 4 -n 32\n"
+    "    vitastor-cli create-pool test_ec42 --ec 4+2 -n 32\n"
+    "\n"
+    "vitastor-cli modify-pool|pool-modify <id|name> [--name <new_name>] [PARAMETERS...]\n"
+    "  Modify an existing pool. Modifiable parameters:\n"
+    "    [-s|--pg_size <number>] [--pg_minsize <number>] [-n|--pg_count <count>]\n"
+    "    [--failure_domain <level>] [--root_node <node>] [--osd_tags <tags>] [--used_for_fs <name>]\n"
+    "    [--max_osd_combinations <number>] [--primary_affinity_tags <tags>] [--scrub_interval <time>]\n"
+    "  Non-modifiable parameters (changing them WILL lead to data loss):\n"
+    "    [--block_size <size>] [--bitmap_granularity <size>]\n"
+    "    [--immediate_commit <all|small|none>] [--pg_stripe_size <size>]\n"
+    "  These, however, can still be modified with -f|--force.\n"
+    "  See create-pool for parameter descriptions.\n"
+    "  Examples:\n"
+    "    vitastor-cli modify-pool pool_A --name pool_B\n"
+    "    vitastor-cli modify-pool 2 --pg_size 4 -n 128\n"
+    "\n"
+    "vitastor-cli rm-pool|pool-rm [--force] <id|name>\n"
+    "  Remove a pool. Refuses to remove pools with images without --force.\n"
+    "\n"
+    "vitastor-cli ls-pools|pool-ls|ls-pool|pools [-l] [--detail] [--sort FIELD] [-r] [-n N] [--stats] [<glob> ...]\n"
+    "  List pools (only matching <glob> patterns if passed).\n"
+    "  -l|--long       Also report I/O statistics\n"
+    "  --detail        Use list format (not table), show all details\n"
+    "  --sort FIELD    Sort by specified field (see fields in --json output)\n"
+    "  -r|--reverse    Sort in descending order\n"
+    "  -n|--count N    Only list first N items\n"
+    "\n"
     "Use vitastor-cli --help <command> for command details or vitastor-cli --help --all for all details.\n"
     "\n"
     "GLOBAL OPTIONS:\n"
@@ -122,7 +174,7 @@ static const char* help_text =
     "  --parallel_osds M   Work with M osds in parallel when possible (default 4)\n"
     "  --progress 1|0      Report progress (default 1)\n"
     "  --cas 1|0           Use CAS writes for flatten, merge, rm (default is decide automatically)\n"
-    "  --no-color          Disable colored output\n"
+    "  --color 1|0         Enable/disable colored output and CR symbols (default 1 if stdout is a terminal)\n"
     "  --json              JSON output\n"
 ;
 
@@ -133,6 +185,7 @@ static json11::Json::object parse_args(int narg, const char *args[])
     cfg["progress"] = "1";
     for (int i = 1; i < narg; i++)
     {
+        bool argHasValue = (!(i == narg-1) && (args[i+1][0] != '-'));
         if (args[i][0] == '-' && args[i][1] == 'h' && args[i][2] == 0)
         {
             cfg["help"] = "1";
@@ -143,15 +196,15 @@ static json11::Json::object parse_args(int narg, const char *args[])
         }
         else if (args[i][0] == '-' && args[i][1] == 'n' && args[i][2] == 0)
         {
-            cfg["count"] = args[++i];
+            cfg["count"] = argHasValue ? args[++i] : "";
         }
         else if (args[i][0] == '-' && args[i][1] == 'p' && args[i][2] == 0)
         {
-            cfg["pool"] = args[++i];
+            cfg["pool"] = argHasValue ? args[++i] : "";
         }
         else if (args[i][0] == '-' && args[i][1] == 's' && args[i][2] == 0)
         {
-            cfg["size"] = args[++i];
+            cfg["size"] = argHasValue ? args[++i] : "";
         }
         else if (args[i][0] == '-' && args[i][1] == 'r' && args[i][2] == 0)
         {
@@ -164,17 +217,24 @@ static json11::Json::object parse_args(int narg, const char *args[])
         else if (args[i][0] == '-' && args[i][1] == '-')
         {
             const char *opt = args[i]+2;
-            cfg[opt] = i == narg-1 || !strcmp(opt, "json") ||
+            if (!strcmp(opt, "json") || !strcmp(opt, "wait") ||
                 !strcmp(opt, "wait-list") || !strcmp(opt, "wait_list") ||
-                !strcmp(opt, "long") || !strcmp(opt, "del") ||
+                !strcmp(opt, "long") || !strcmp(opt, "detail") || !strcmp(opt, "del") ||
                 !strcmp(opt, "no-color") || !strcmp(opt, "no_color") ||
                 !strcmp(opt, "readonly") || !strcmp(opt, "readwrite") ||
                 !strcmp(opt, "force") || !strcmp(opt, "reverse") ||
                 !strcmp(opt, "allow-data-loss") || !strcmp(opt, "allow_data_loss") ||
+                !strcmp(opt, "down-ok") || !strcmp(opt, "down_ok") ||
                 !strcmp(opt, "dry-run") || !strcmp(opt, "dry_run") ||
                 !strcmp(opt, "help") || !strcmp(opt, "all") ||
-                (!strcmp(opt, "writers-stopped") || !strcmp(opt, "writers_stopped")) && strcmp("1", args[i+1]) != 0
-                ? "1" : args[++i];
+                !strcmp(opt, "writers-stopped") || !strcmp(opt, "writers_stopped"))
+            {
+                cfg[opt] = "1";
+            }
+            else
+            {
+                cfg[opt] = argHasValue ? args[++i] : "";
+            }
         }
         else
         {
@@ -217,7 +277,7 @@ static int run(cli_tool_t *p, json11::Json::object cfg)
     else if (cmd[0] == "df")
     {
         // Show pool space stats
-        action_cb = p->start_df(cfg);
+        action_cb = p->start_pool_ls(cfg);
     }
     else if (cmd[0] == "ls")
     {
@@ -324,6 +384,44 @@ static int run(cli_tool_t *p, json11::Json::object cfg)
         // Allocate a new OSD number
         action_cb = p->start_alloc_osd(cfg);
     }
+    else if (cmd[0] == "create-pool" || cmd[0] == "pool-create")
+    {
+        // Create a new pool
+        if (cmd.size() > 1 && cfg["name"].is_null())
+        {
+            cfg["name"] = cmd[1];
+        }
+        action_cb = p->start_pool_create(cfg);
+    }
+    else if (cmd[0] == "modify-pool" || cmd[0] == "pool-modify")
+    {
+        // Modify existing pool
+        if (cmd.size() > 1)
+        {
+            cfg["old_name"] = cmd[1];
+        }
+        action_cb = p->start_pool_modify(cfg);
+    }
+    else if (cmd[0] == "rm-pool" || cmd[0] == "pool-rm")
+    {
+        // Remove existing pool
+        if (cmd.size() > 1)
+        {
+            cfg["pool"] = cmd[1];
+        }
+        action_cb = p->start_pool_rm(cfg);
+    }
+    else if (cmd[0] == "ls-pool" || cmd[0] == "pool-ls" || cmd[0] == "ls-pools" || cmd[0] == "pools")
+    {
+        // Show pool list
+        cfg["show_recovery"] = 1;
+        if (cmd.size() > 1)
+        {
+            cmd.erase(cmd.begin(), cmd.begin()+1);
+            cfg["names"] = cmd;
+        }
+        action_cb = p->start_pool_ls(cfg);
+    }
     else
     {
         result = { .err = EINVAL, .text = "unknown command: "+cmd[0].string_value() };

src/cli.h

@@ -46,6 +46,7 @@ public:
     json11::Json etcd_result;
 
     void parse_config(json11::Json::object & cfg);
+    json11::Json parse_tags(std::string tags);
 
     void change_parent(inode_t cur, inode_t new_parent, cli_result_t *result);
     inode_config_t* get_inode_cfg(const std::string & name);
@@ -58,7 +59,6 @@ public:
     std::function<bool(cli_result_t &)> start_status(json11::Json);
     std::function<bool(cli_result_t &)> start_describe(json11::Json);
     std::function<bool(cli_result_t &)> start_fix(json11::Json);
-    std::function<bool(cli_result_t &)> start_df(json11::Json);
     std::function<bool(cli_result_t &)> start_ls(json11::Json);
     std::function<bool(cli_result_t &)> start_create(json11::Json);
     std::function<bool(cli_result_t &)> start_modify(json11::Json);
@@ -68,6 +68,10 @@ public:
     std::function<bool(cli_result_t &)> start_rm(json11::Json);
     std::function<bool(cli_result_t &)> start_rm_osd(json11::Json cfg);
     std::function<bool(cli_result_t &)> start_alloc_osd(json11::Json cfg);
+    std::function<bool(cli_result_t &)> start_pool_create(json11::Json);
+    std::function<bool(cli_result_t &)> start_pool_modify(json11::Json);
+    std::function<bool(cli_result_t &)> start_pool_rm(json11::Json);
+    std::function<bool(cli_result_t &)> start_pool_ls(json11::Json);
 
     // Should be called like loop_and_wait(start_status(), <completion callback>)
     void loop_and_wait(std::function<bool(cli_result_t &)> loop_cb, std::function<void(const cli_result_t &)> complete_cb);
@@ -77,8 +81,13 @@ public:
 
 std::string print_table(json11::Json items, json11::Json header, bool use_esc);
 
+size_t print_detail_title_len(json11::Json item, std::vector<std::pair<std::string, std::string>> names, size_t prev_len);
+std::string print_detail(json11::Json item, std::vector<std::pair<std::string, std::string>> names, size_t title_len, bool use_esc);
+
 std::string format_lat(uint64_t lat);
 
 std::string format_q(double depth);
 
 bool stupid_glob(const std::string str, const std::string glob);
+
+std::string implode(const std::string & sep, json11::Json array);

src/cli_alloc_osd.cpp

@@ -77,7 +77,7 @@ struct alloc_osd_t
                     std::string key = base64_decode(kv["key"].string_value());
                     osd_num_t cur_osd;
                     char null_byte = 0;
-                    int scanned = sscanf(key.c_str() + parent->cli->st_cli.etcd_prefix.length(), "/osd/stats/%lu%c", &cur_osd, &null_byte);
+                    int scanned = sscanf(key.c_str() + parent->cli->st_cli.etcd_prefix.length(), "/osd/stats/%ju%c", &cur_osd, &null_byte);
                     if (scanned != 1 || !cur_osd)
                     {
                         fprintf(stderr, "Invalid key in etcd: %s\n", key.c_str());

src/cli_common.cpp

@@ -1,6 +1,7 @@
 // Copyright (c) Vitaliy Filippov, 2019+
 // License: VNPL-1.1 (see README.md for details)
 
+#include <unistd.h>
 #include "str_util.h"
 #include "cluster_client.h"
 #include "cli.h"
@@ -11,7 +12,7 @@ void cli_tool_t::change_parent(inode_t cur, inode_t new_parent, cli_result_t *re
     if (cur_cfg_it == cli->st_cli.inode_config.end())
     {
         char buf[128];
-        snprintf(buf, 128, "Inode 0x%lx disappeared", cur);
+        snprintf(buf, 128, "Inode 0x%jx disappeared", cur);
         *result = (cli_result_t){ .err = EIO, .text = buf };
         return;
     }
@@ -113,7 +114,12 @@ void cli_tool_t::parse_config(json11::Json::object & cfg)
         else
             kv_it++;
     }
-    color = !cfg["no_color"].bool_value();
+    if (cfg.find("no_color") != cfg.end())
+        color = !cfg["no_color"].bool_value();
+    else if (cfg.find("color") != cfg.end())
+        color = cfg["color"].bool_value();
+    else
+        color = isatty(1);
     json_output = cfg["json"].bool_value();
     iodepth = cfg["iodepth"].uint64_value();
     if (!iodepth)
@@ -147,6 +153,7 @@ void cli_tool_t::loop_and_wait(std::function<bool(cli_result_t &)> loop_cb, std:
             ringloop->unregister_consumer(&looper->consumer);
             looper->loop_cb = NULL;
             looper->complete_cb(looper->result);
+            ringloop->submit();
             delete looper;
             return;
         }

src/cli_create.cpp

@@ -27,6 +27,7 @@ struct image_creator_t
     std::string image_name, new_snap, new_parent;
     json11::Json new_meta;
     uint64_t size;
+    bool force = false;
     bool force_size = false;
 
     pool_id_t old_pool_id = 0;
@@ -45,6 +46,7 @@ struct image_creator_t
 
     void loop()
     {
+        auto & pools = parent->cli->st_cli.pool_config;
         if (state >= 1)
             goto resume_1;
         if (image_name == "")
@@ -62,7 +64,6 @@ struct image_creator_t
         }
         if (new_pool_id)
         {
-            auto & pools = parent->cli->st_cli.pool_config;
             if (pools.find(new_pool_id) == pools.end())
             {
                 result = (cli_result_t){ .err = ENOENT, .text = "Pool "+std::to_string(new_pool_id)+" does not exist" };
@@ -72,7 +73,7 @@ struct image_creator_t
         }
         else if (new_pool_name != "")
         {
-            for (auto & ic: parent->cli->st_cli.pool_config)
+            for (auto & ic: pools)
             {
                 if (ic.second.name == new_pool_name)
                 {
@@ -87,10 +88,20 @@ struct image_creator_t
                 return;
             }
         }
-        else if (parent->cli->st_cli.pool_config.size() == 1)
+        else if (pools.size() == 1)
         {
-            auto it = parent->cli->st_cli.pool_config.begin();
-            new_pool_id = it->first;
+            new_pool_id = pools.begin()->first;
+        }
+        if (new_pool_id && !pools.at(new_pool_id).used_for_fs.empty() && !force)
+        {
+            result = (cli_result_t){
+                .err = EINVAL,
+                .text = "Pool "+pools.at(new_pool_id).name+
+                    " is used for VitastorFS "+pools.at(new_pool_id).used_for_fs+
+                    ". Use --force if you really know what you are doing",
+            };
+            state = 100;
+            return;
         }
         state = 1;
     resume_1:
@@ -183,7 +194,16 @@ resume_3:
         // Save into inode_config for library users to be able to take it from there immediately
         new_cfg.mod_revision = parent->etcd_result["responses"][0]["response_put"]["header"]["revision"].uint64_value();
         parent->cli->st_cli.insert_inode_config(new_cfg);
-        result = (cli_result_t){ .err = 0, .text = "Image "+image_name+" created" };
+        result = (cli_result_t){
+            .err = 0,
+            .text = "Image "+image_name+" created",
+            .data = json11::Json::object {
+                { "name", image_name },
+                { "pool", new_pool_name },
+                { "parent", new_parent },
+                { "size", size },
+            }
+        };
         state = 100;
     }
 
@@ -251,7 +271,16 @@ resume_4:
         // Save into inode_config for library users to be able to take it from there immediately
         new_cfg.mod_revision = parent->etcd_result["responses"][0]["response_put"]["header"]["revision"].uint64_value();
         parent->cli->st_cli.insert_inode_config(new_cfg);
-        result = (cli_result_t){ .err = 0, .text = "Snapshot "+image_name+"@"+new_snap+" created" };
+        result = (cli_result_t){
+            .err = 0,
+            .text = "Snapshot "+image_name+"@"+new_snap+" created",
+            .data = json11::Json::object {
+                { "name", image_name+"@"+new_snap },
+                { "pool", (uint64_t)new_pool_id },
+                { "parent", new_parent },
+                { "size", size },
+            }
+        };
         state = 100;
     }
 
@@ -514,6 +543,7 @@ std::function<bool(cli_result_t &)> cli_tool_t::start_create(json11::Json cfg)
     image_creator->image_name = cfg["image"].string_value();
     image_creator->new_pool_id = cfg["pool"].uint64_value();
     image_creator->new_pool_name = cfg["pool"].string_value();
+    image_creator->force = cfg["force"].bool_value();
     image_creator->force_size = cfg["force_size"].bool_value();
     if (cfg["image_meta"].is_object())
     {

src/cli_describe.cpp

@@ -160,14 +160,14 @@ struct cli_describe_t
                 if (op->reply.hdr.retval < 0)
                 {
                     fprintf(
-                        stderr, "Failed to describe objects on OSD %lu (retval=%ld)\n",
+                        stderr, "Failed to describe objects on OSD %ju (retval=%jd)\n",
                         osd_num, op->reply.hdr.retval
                     );
                 }
                 else if (op->reply.describe.result_bytes != op->reply.hdr.retval * sizeof(osd_reply_describe_item_t))
                 {
                     fprintf(
-                        stderr, "Invalid response size from OSD %lu (expected %lu bytes, got %lu bytes)\n",
+                        stderr, "Invalid response size from OSD %ju (expected %ju bytes, got %ju bytes)\n",
                         osd_num, op->reply.hdr.retval * sizeof(osd_reply_describe_item_t), op->reply.describe.result_bytes
                     );
                 }
@@ -178,11 +178,11 @@ struct cli_describe_t
                     {
                         if (!parent->json_output || parent->is_command_line)
                         {
-#define FMT "{\"inode\":\"0x%lx\",\"stripe\":\"0x%lx\",\"part\":%u,\"osd_num\":%lu%s%s%s}"
+#define FMT "{\"inode\":\"0x%jx\",\"stripe\":\"0x%jx\",\"part\":%u,\"osd_num\":%ju%s%s%s}"
                             printf(
                                 (parent->json_output
                                     ? (count > 0 ? ",\n  " FMT : "  " FMT)
-                                    : "%lx:%lx part %u on OSD %lu%s%s%s\n"),
+                                    : "%jx:%jx part %u on OSD %ju%s%s%s\n"),
 #undef FMT
                                 items[i].inode, items[i].stripe,
                                 items[i].role, items[i].osd_num,

src/cli_df.cpp

@@ -1,243 +0,0 @@
-// Copyright (c) Vitaliy Filippov, 2019+
-// License: VNPL-1.1 (see README.md for details)
-
-#include "cli.h"
-#include "cluster_client.h"
-#include "str_util.h"
-
-// List pools with space statistics
-struct pool_lister_t
-{
-    cli_tool_t *parent;
-
-    int state = 0;
-    json11::Json space_info;
-    cli_result_t result;
-    std::map<pool_id_t, json11::Json::object> pool_stats;
-
-    bool is_done()
-    {
-        return state == 100;
-    }
-
-    void get_stats()
-    {
-        if (state == 1)
-            goto resume_1;
-        // Space statistics - pool/stats/<pool>
-        parent->etcd_txn(json11::Json::object {
-            { "success", json11::Json::array {
-                json11::Json::object {
-                    { "request_range", json11::Json::object {
-                        { "key", base64_encode(
-                            parent->cli->st_cli.etcd_prefix+"/pool/stats/"
-                        ) },
-                        { "range_end", base64_encode(
-                            parent->cli->st_cli.etcd_prefix+"/pool/stats0"
-                        ) },
-                    } },
-                },
-                json11::Json::object {
-                    { "request_range", json11::Json::object {
-                        { "key", base64_encode(
-                            parent->cli->st_cli.etcd_prefix+"/osd/stats/"
-                        ) },
-                        { "range_end", base64_encode(
-                            parent->cli->st_cli.etcd_prefix+"/osd/stats0"
-                        ) },
-                    } },
-                },
-            } },
-        });
-        state = 1;
-resume_1:
-        if (parent->waiting > 0)
-            return;
-        if (parent->etcd_err.err)
-        {
-            result = parent->etcd_err;
-            state = 100;
-            return;
-        }
-        space_info = parent->etcd_result;
-        std::map<pool_id_t, uint64_t> osd_free;
-        for (auto & kv_item: space_info["responses"][0]["response_range"]["kvs"].array_items())
-        {
-            auto kv = parent->cli->st_cli.parse_etcd_kv(kv_item);
-            // pool ID
-            pool_id_t pool_id;
-            char null_byte = 0;
-            int scanned = sscanf(kv.key.substr(parent->cli->st_cli.etcd_prefix.length()).c_str(), "/pool/stats/%u%c", &pool_id, &null_byte);
-            if (scanned != 1 || !pool_id || pool_id >= POOL_ID_MAX)
-            {
-                fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
-                continue;
-            }
-            // pool/stats/<N>
-            pool_stats[pool_id] = kv.value.object_items();
-        }
-        for (auto & kv_item: space_info["responses"][1]["response_range"]["kvs"].array_items())
-        {
-            auto kv = parent->cli->st_cli.parse_etcd_kv(kv_item);
-            // osd ID
-            osd_num_t osd_num;
-            char null_byte = 0;
-            int scanned = sscanf(kv.key.substr(parent->cli->st_cli.etcd_prefix.length()).c_str(), "/osd/stats/%lu%c", &osd_num, &null_byte);
-            if (scanned != 1 || !osd_num || osd_num >= POOL_ID_MAX)
-            {
-                fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
-                continue;
-            }
-            // osd/stats/<N>::free
-            osd_free[osd_num] = kv.value["free"].uint64_value();
-        }
-        // Calculate max_avail for each pool
-        for (auto & pp: parent->cli->st_cli.pool_config)
-        {
-            auto & pool_cfg = pp.second;
-            uint64_t pool_avail = UINT64_MAX;
-            std::map<osd_num_t, uint64_t> pg_per_osd;
-            for (auto & pgp: pool_cfg.pg_config)
-            {
-                for (auto pg_osd: pgp.second.target_set)
-                {
-                    if (pg_osd != 0)
-                    {
-                        pg_per_osd[pg_osd]++;
-                    }
-                }
-            }
-            for (auto pg_per_pair: pg_per_osd)
-            {
-                uint64_t pg_free = osd_free[pg_per_pair.first] * pool_cfg.real_pg_count / pg_per_pair.second;
-                if (pool_avail > pg_free)
-                {
-                    pool_avail = pg_free;
-                }
-            }
-            if (pool_avail == UINT64_MAX)
-            {
-                pool_avail = 0;
-            }
-            if (pool_cfg.scheme != POOL_SCHEME_REPLICATED)
-            {
-                pool_avail *= (pool_cfg.pg_size - pool_cfg.parity_chunks);
-            }
-            pool_stats[pool_cfg.id] = json11::Json::object {
-                { "id", (uint64_t)pool_cfg.id },
-                { "name", pool_cfg.name },
-                { "pg_count", pool_cfg.pg_count },
-                { "real_pg_count", pool_cfg.real_pg_count },
-                { "scheme", pool_cfg.scheme == POOL_SCHEME_REPLICATED ? "replicated" : "ec" },
-                { "scheme_name", pool_cfg.scheme == POOL_SCHEME_REPLICATED
-                    ? std::to_string(pool_cfg.pg_size)+"/"+std::to_string(pool_cfg.pg_minsize)
-                    : "EC "+std::to_string(pool_cfg.pg_size-pool_cfg.parity_chunks)+"+"+std::to_string(pool_cfg.parity_chunks) },
-                { "used_raw", (uint64_t)(pool_stats[pool_cfg.id]["used_raw_tb"].number_value() * ((uint64_t)1<<40)) },
-                { "total_raw", (uint64_t)(pool_stats[pool_cfg.id]["total_raw_tb"].number_value() * ((uint64_t)1<<40)) },
-                { "max_available", pool_avail },
-                { "raw_to_usable", pool_stats[pool_cfg.id]["raw_to_usable"].number_value() },
-                { "space_efficiency", pool_stats[pool_cfg.id]["space_efficiency"].number_value() },
-                { "pg_real_size", pool_stats[pool_cfg.id]["pg_real_size"].uint64_value() },
-                { "failure_domain", pool_cfg.failure_domain },
-            };
-        }
-    }
-
-    json11::Json::array to_list()
-    {
-        json11::Json::array list;
-        for (auto & kv: pool_stats)
-        {
-            list.push_back(kv.second);
-        }
-        return list;
-    }
-
-    void loop()
-    {
-        get_stats();
-        if (parent->waiting > 0)
-            return;
-        if (state == 100)
-            return;
-        if (parent->json_output)
-        {
-            // JSON output
-            result.data = to_list();
-            state = 100;
-            return;
-        }
-        // Table output: name, scheme_name, pg_count, total, used, max_avail, used%, efficiency
-        json11::Json::array cols;
-        cols.push_back(json11::Json::object{
-            { "key", "name" },
-            { "title", "NAME" },
-        });
-        cols.push_back(json11::Json::object{
-            { "key", "scheme_name" },
-            { "title", "SCHEME" },
-        });
-        cols.push_back(json11::Json::object{
-            { "key", "pg_count_fmt" },
-            { "title", "PGS" },
-        });
-        cols.push_back(json11::Json::object{
-            { "key", "total_fmt" },
-            { "title", "TOTAL" },
-        });
-        cols.push_back(json11::Json::object{
-            { "key", "used_fmt" },
-            { "title", "USED" },
-        });
-        cols.push_back(json11::Json::object{
-            { "key", "max_avail_fmt" },
-            { "title", "AVAILABLE" },
-        });
-        cols.push_back(json11::Json::object{
-            { "key", "used_pct" },
-            { "title", "USED%" },
-        });
-        cols.push_back(json11::Json::object{
-            { "key", "eff_fmt" },
-            { "title", "EFFICIENCY" },
-        });
-        json11::Json::array list;
-        for (auto & kv: pool_stats)
-        {
-            double raw_to = kv.second["raw_to_usable"].number_value();
-            if (raw_to < 0.000001 && raw_to > -0.000001)
-                raw_to = 1;
-            kv.second["pg_count_fmt"] = kv.second["real_pg_count"] == kv.second["pg_count"]
-                ? kv.second["real_pg_count"].as_string()
-                : kv.second["real_pg_count"].as_string()+"->"+kv.second["pg_count"].as_string();
-            kv.second["total_fmt"] = format_size(kv.second["total_raw"].uint64_value() / raw_to);
-            kv.second["used_fmt"] = format_size(kv.second["used_raw"].uint64_value() / raw_to);
-            kv.second["max_avail_fmt"] = format_size(kv.second["max_available"].uint64_value());
-            kv.second["used_pct"] = format_q(kv.second["total_raw"].uint64_value()
-                ? (100 - 100*kv.second["max_available"].uint64_value() *
-                    kv.second["raw_to_usable"].number_value() / kv.second["total_raw"].uint64_value())
-                : 100)+"%";
-            kv.second["eff_fmt"] = format_q(kv.second["space_efficiency"].number_value()*100)+"%";
-        }
-        result.data = to_list();
-        result.text = print_table(result.data, cols, parent->color);
-        state = 100;
-    }
-};
-
-std::function<bool(cli_result_t &)> cli_tool_t::start_df(json11::Json cfg)
-{
-    auto lister = new pool_lister_t();
-    lister->parent = this;
-    return [lister](cli_result_t & result)
-    {
-        lister->loop();
-        if (lister->is_done())
-        {
-            result = lister->result;
-            delete lister;
-            return true;
-        }
-        return false;
-    };
-}

src/cli_fix.cpp

@@ -136,7 +136,7 @@ struct cli_fix_t
             auto pool_cfg_it = parent->cli->st_cli.pool_config.find(INODE_POOL(obj.inode));
             if (pool_cfg_it == parent->cli->st_cli.pool_config.end())
             {
-                fprintf(stderr, "Object %lx:%lx is from unknown pool\n", obj.inode, obj.stripe);
+                fprintf(stderr, "Object %jx:%jx is from unknown pool\n", obj.inode, obj.stripe);
                 continue;
             }
             auto & pool_cfg = pool_cfg_it->second;
@@ -146,7 +146,7 @@ struct cli_fix_t
                 !pg_it->second.cur_primary || !(pg_it->second.cur_state & PG_ACTIVE))
             {
                 fprintf(
-                    stderr, "Object %lx:%lx is from PG %u/%u which is not currently active\n",
+                    stderr, "Object %jx:%jx is from PG %u/%u which is not currently active\n",
                     obj.inode, obj.stripe, pool_cfg_it->first, pg_num
                 );
                 continue;
@@ -171,7 +171,7 @@ struct cli_fix_t
             {
                 if (op->reply.hdr.retval < 0 || op->reply.describe.result_bytes != op->reply.hdr.retval * sizeof(osd_reply_describe_item_t))
                 {
-                    fprintf(stderr, "Failed to describe objects on OSD %lu (retval=%ld)\n", primary_osd, op->reply.hdr.retval);
+                    fprintf(stderr, "Failed to describe objects on OSD %ju (retval=%jd)\n", primary_osd, op->reply.hdr.retval);
                     parent->waiting--;
                     loop();
                 }
@@ -209,7 +209,7 @@ struct cli_fix_t
                                 if (rm_op->reply.hdr.retval < 0)
                                 {
                                     fprintf(
-                                        stderr, "Failed to remove object %lx:%lx from OSD %lu (retval=%ld)\n",
+                                        stderr, "Failed to remove object %jx:%jx from OSD %ju (retval=%jd)\n",
                                         rm_op->req.sec_del.oid.inode, rm_op->req.sec_del.oid.stripe,
                                         rm_osd_num, rm_op->reply.hdr.retval
                                     );
@@ -226,7 +226,7 @@ struct cli_fix_t
                                 else
                                 {
                                     printf(
-                                        "Removed %lx:%lx (part %lu) from OSD %lu\n",
+                                        "Removed %jx:%jx (part %ju) from OSD %ju\n",
                                         rm_op->req.sec_del.oid.inode, rm_op->req.sec_del.oid.stripe & ~STRIPE_MASK,
                                         rm_op->req.sec_del.oid.stripe & STRIPE_MASK, rm_osd_num
                                     );
@@ -254,7 +254,7 @@ struct cli_fix_t
                                         if (scrub_op->reply.hdr.retval < 0 && scrub_op->reply.hdr.retval != -ENOENT)
                                         {
                                             fprintf(
-                                                stderr, "Failed to scrub %lx:%lx on OSD %lu (retval=%ld)\n",
+                                                stderr, "Failed to scrub %jx:%jx on OSD %ju (retval=%jd)\n",
                                                 obj.inode, obj.stripe, primary_osd, scrub_op->reply.hdr.retval
                                             );
                                         }

src/cli_ls.cpp

@@ -150,7 +150,7 @@ resume_1:
             inode_t only_inode_num;
             char null_byte = 0;
             int scanned = sscanf(kv.key.substr(parent->cli->st_cli.etcd_prefix.length()).c_str(),
-                "/inode/stats/%u/%lu%c", &pool_id, &only_inode_num, &null_byte);
+                "/inode/stats/%u/%ju%c", &pool_id, &only_inode_num, &null_byte);
             if (scanned != 2 || !pool_id || pool_id >= POOL_ID_MAX || INODE_POOL(only_inode_num) != 0)
             {
                 fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
@@ -456,7 +456,7 @@ std::string format_lat(uint64_t lat)
     char buf[256];
     int l = 0;
     if (lat < 100)
-        l = snprintf(buf, sizeof(buf), "%lu us", lat);
+        l = snprintf(buf, sizeof(buf), "%ju us", lat);
     else if (lat < 500000)
         l = snprintf(buf, sizeof(buf), "%.2f ms", (double)lat/1000);
     else

src/cli_merge.cpp

@@ -202,7 +202,7 @@ struct snap_merger_t
         if (parent->progress)
         {
             printf(
-                "Merging %ld layer(s) into target %s%s (inode %lu in pool %u)\n",
+                "Merging %zd layer(s) into target %s%s (inode %ju in pool %u)\n",
                 sources.size(), target_cfg->name.c_str(),
                 use_cas ? " online (with CAS)" : "", INODE_NO_POOL(target), INODE_POOL(target)
             );
@@ -275,7 +275,9 @@ struct snap_merger_t
                 processed++;
                 if (parent->progress && !(processed % 128))
                 {
-                    printf("\rFiltering target blocks: %lu/%lu", processed, to_process);
+                    fprintf(stderr, parent->color
+                        ? "\rFiltering target blocks: %ju/%ju"
+                        : "Filtering target blocks: %ju/%ju\n", processed, to_process);
                 }
             }
             if (in_flight > 0 || oit != merge_offsets.end())
@@ -285,7 +287,9 @@ struct snap_merger_t
             }
             if (parent->progress)
             {
-                printf("\r%lu full blocks of target filtered out\n", to_process-merge_offsets.size());
+                fprintf(stderr, parent->color
+                    ? "\r%ju full blocks of target filtered out\n"
+                    : "%ju full blocks of target filtered out\n", to_process-merge_offsets.size());
             }
         }
         state = 3;
@@ -320,7 +324,9 @@ struct snap_merger_t
             processed++;
             if (parent->progress && !(processed % 128))
             {
-                printf("\rOverwriting blocks: %lu/%lu", processed, to_process);
+                fprintf(stderr, parent->color
+                    ? "\rOverwriting blocks: %ju/%ju"
+                    : "Overwriting blocks: %ju/%ju\n", processed, to_process);
             }
         }
         if (in_flight == 0 && rwo_error.size())
@@ -339,10 +345,16 @@ struct snap_merger_t
         }
         if (parent->progress)
         {
-            printf("\rOverwriting blocks: %lu/%lu\n", to_process, to_process);
+            fprintf(stderr, parent->color
+                ? "\rOverwriting blocks: %ju/%ju\n"
+                : "Overwriting blocks: %ju/%ju\n", to_process, to_process);
         }
         // Done
-        result = (cli_result_t){ .text = "Done, layers from "+from_name+" to "+to_name+" merged into "+target_name };
+        result = (cli_result_t){ .text = "Done, layers from "+from_name+" to "+to_name+" merged into "+target_name, .data = json11::Json::object {
+            { "from", from_name },
+            { "to", to_name },
+            { "into", target_name },
+        }};
         state = 100;
     resume_100:
         return;
@@ -384,7 +396,7 @@ struct snap_merger_t
                         auto & name = parent->cli->st_cli.inode_config.at(src).name;
                         if (parent->progress)
                         {
-                            printf("Got listing of layer %s (inode %lu in pool %u)\n", name.c_str(), INODE_NO_POOL(src), INODE_POOL(src));
+                            printf("Got listing of layer %s (inode %ju in pool %u)\n", name.c_str(), INODE_NO_POOL(src), INODE_POOL(src));
                         }
                         if (delete_source)
                         {
@@ -416,7 +428,7 @@ struct snap_merger_t
         {
             if (op->retval < 0)
             {
-                fprintf(stderr, "error reading target bitmap at offset %lx: %s\n", op->offset, strerror(-op->retval));
+                fprintf(stderr, "error reading target bitmap at offset %jx: %s\n", op->offset, strerror(-op->retval));
             }
             else
             {
@@ -571,7 +583,7 @@ struct snap_merger_t
         {
             if (subop->retval != 0)
             {
-                fprintf(stderr, "error deleting from layer 0x%lx at offset %lx: %s", subop->inode, subop->offset, strerror(-subop->retval));
+                fprintf(stderr, "error deleting from layer 0x%jx at offset %jx: %s", subop->inode, subop->offset, strerror(-subop->retval));
             }
             delete subop;
         };
@@ -620,7 +632,7 @@ struct snap_merger_t
             if (rwo->error_code)
             {
                 char buf[1024];
-                snprintf(buf, 1024, "Error %s target at offset %lx: %s",
+                snprintf(buf, 1024, "Error %s target at offset %jx: %s",
                     rwo->error_read ? "reading" : "writing", rwo->error_offset, strerror(rwo->error_code));
                 rwo_error = std::string(buf);
             }

src/cli_modify.cpp

@@ -15,6 +15,7 @@ struct image_changer_t
     uint64_t new_size = 0;
     bool force_size = false, inc_size = false;
     bool set_readonly = false, set_readwrite = false, force = false;
+    bool down_ok = false;
     // interval between fsyncs
     int fsync_interval = 128;
 
@@ -84,7 +85,10 @@ struct image_changer_t
             (!new_size && !force_size || cfg.size == new_size || cfg.size >= new_size && inc_size) &&
             (new_name == "" || new_name == image_name))
         {
-            result = (cli_result_t){ .text = "No change" };
+            result = (cli_result_t){ .err = 0, .text = "No change", .data = json11::Json::object {
+                { "error_code", 0 },
+                { "error_text", "No change" },
+            }};
             state = 100;
             return;
         }
@@ -105,6 +109,7 @@ struct image_changer_t
                     { "pool", (uint64_t)INODE_POOL(inode_num) },
                     { "fsync-interval", fsync_interval },
                     { "min-offset", ((new_size+4095)/4096)*4096 },
+                    { "down-ok", down_ok },
                 });
 resume_1:
                 while (!cb(result))
@@ -220,7 +225,16 @@ resume_2:
             parent->cli->st_cli.inode_by_name.erase(image_name);
         }
         parent->cli->st_cli.insert_inode_config(cfg);
-        result = (cli_result_t){ .err = 0, .text = "Image "+image_name+" modified" };
+        result = (cli_result_t){
+            .err = 0,
+            .text = "Image "+image_name+" modified",
+            .data = json11::Json::object {
+                { "name", image_name },
+                { "inode", INODE_NO_POOL(inode_num) },
+                { "pool", (uint64_t)INODE_POOL(inode_num) },
+                { "size", new_size },
+            }
+        };
         state = 100;
     }
 };
@@ -240,6 +254,7 @@ std::function<bool(cli_result_t &)> cli_tool_t::start_modify(json11::Json cfg)
     changer->fsync_interval = cfg["fsync_interval"].uint64_value();
     if (!changer->fsync_interval)
         changer->fsync_interval = 128;
+    changer->down_ok = cfg["down_ok"].bool_value();
     // FIXME Check that the image doesn't have children when shrinking
     return [changer](cli_result_t & result)
     {

src/cli_pool_cfg.cpp

@@ -0,0 +1,270 @@
+// Copyright (c) Vitaliy Filippov, 2024
+// License: VNPL-1.1 (see README.md for details)
+
+#include "cli_pool_cfg.h"
+#include "etcd_state_client.h"
+#include "str_util.h"
+
+std::string validate_pool_config(json11::Json::object & new_cfg, json11::Json old_cfg,
+    uint64_t global_block_size, uint64_t global_bitmap_granularity, bool force)
+{
+    // short option names
+    if (new_cfg.find("count") != new_cfg.end())
+    {
+        new_cfg["pg_count"] = new_cfg["count"];
+        new_cfg.erase("count");
+    }
+    if (new_cfg.find("size") != new_cfg.end())
+    {
+        new_cfg["pg_size"] = new_cfg["size"];
+        new_cfg.erase("size");
+    }
+
+    // --ec shortcut
+    if (new_cfg.find("ec") != new_cfg.end())
+    {
+        if (new_cfg.find("scheme") != new_cfg.end() ||
+            new_cfg.find("pg_size") != new_cfg.end() ||
+            new_cfg.find("parity_chunks") != new_cfg.end())
+        {
+            return "--ec can't be used with --pg_size, --parity_chunks or --scheme";
+        }
+        // pg_size = N+K
+        // parity_chunks = K
+        uint64_t data_chunks = 0, parity_chunks = 0;
+        char null_byte = 0;
+        int ret = sscanf(new_cfg["ec"].string_value().c_str(), "%ju+%ju%c", &data_chunks, &parity_chunks, &null_byte);
+        if (ret != 2 || !data_chunks || !parity_chunks)
+        {
+            return "--ec should be <N>+<K> format (<N>, <K> - numbers)";
+        }
+        new_cfg.erase("ec");
+        new_cfg["scheme"] = "ec";
+        new_cfg["pg_size"] = data_chunks+parity_chunks;
+        new_cfg["parity_chunks"] = parity_chunks;
+    }
+
+    if (old_cfg.is_null() && new_cfg["scheme"].string_value() == "")
+    {
+        // Default scheme
+        new_cfg["scheme"] = "replicated";
+    }
+    if (new_cfg.find("pg_minsize") == new_cfg.end() && (old_cfg.is_null() || new_cfg.find("pg_size") != new_cfg.end()))
+    {
+        // Default pg_minsize
+        if (new_cfg["scheme"] == "replicated")
+        {
+            // pg_minsize = (N+K > 2) ? 2 : 1
+            new_cfg["pg_minsize"] = new_cfg["pg_size"].uint64_value() > 2 ? 2 : 1;
+        }
+        else // ec or xor
+        {
+            // pg_minsize = (K > 1) ? N + 1 : N
+            new_cfg["pg_minsize"] = new_cfg["pg_size"].uint64_value() - new_cfg["parity_chunks"].uint64_value() +
+                (new_cfg["parity_chunks"].uint64_value() > 1 ? 1 : 0);
+        }
+    }
+
+    // Check integer values and unknown keys
+    for (auto kv_it = new_cfg.begin(); kv_it != new_cfg.end(); )
+    {
+        auto & key = kv_it->first;
+        auto & value = kv_it->second;
+        if (key == "pg_size" || key == "parity_chunks" || key == "pg_minsize" ||
+            key == "pg_count" || key == "max_osd_combinations" || key == "block_size" ||
+            key == "bitmap_granularity" || key == "pg_stripe_size")
+        {
+            if (value.is_number() && value.uint64_value() != value.number_value() ||
+                value.is_string() && !value.uint64_value() && value.string_value() != "0")
+            {
+                return key+" must be a non-negative integer";
+            }
+            value = value.uint64_value();
+        }
+        else if (key == "name" || key == "scheme" || key == "immediate_commit" ||
+            key == "failure_domain" || key == "root_node" || key == "scrub_interval" || key == "used_for_fs")
+        {
+            // OK
+        }
+        else if (key == "osd_tags" || key == "primary_affinity_tags")
+        {
+            if (value.is_string())
+            {
+                value = explode(",", value.string_value(), true);
+            }
+        }
+        else
+        {
+            // Unknown parameter
+            new_cfg.erase(kv_it++);
+            continue;
+        }
+        kv_it++;
+    }
+
+    // Merge with the old config
+    if (!old_cfg.is_null())
+    {
+        for (auto & kv: old_cfg.object_items())
+        {
+            if (new_cfg.find(kv.first) == new_cfg.end())
+            {
+                new_cfg[kv.first] = kv.second;
+            }
+        }
+    }
+
+    // Check after merging
+    if (new_cfg["scheme"] != "ec")
+    {
+        new_cfg.erase("parity_chunks");
+    }
+    if (new_cfg.find("used_for_fs") != new_cfg.end() && new_cfg["used_for_fs"].string_value() == "")
+    {
+        new_cfg.erase("used_for_fs");
+    }
+
+    // Prevent autovivification of object keys. Now we don't modify the config, we just check it
+    json11::Json cfg = new_cfg;
+
+    // Validate changes
+    if (!old_cfg.is_null() && !force)
+    {
+        if (old_cfg["scheme"] != cfg["scheme"])
+        {
+            return "Changing scheme for an existing pool will lead to data loss. Use --force to proceed";
+        }
+        if (etcd_state_client_t::parse_scheme(old_cfg["scheme"].string_value()) == POOL_SCHEME_EC)
+        {
+            uint64_t old_data_chunks = old_cfg["pg_size"].uint64_value() - old_cfg["parity_chunks"].uint64_value();
+            uint64_t new_data_chunks = cfg["pg_size"].uint64_value() - cfg["parity_chunks"].uint64_value();
+            if (old_data_chunks != new_data_chunks)
+            {
+                return "Changing EC data chunk count for an existing pool will lead to data loss. Use --force to proceed";
+            }
+        }
+        if (old_cfg["block_size"] != cfg["block_size"] ||
+            old_cfg["bitmap_granularity"] != cfg["bitmap_granularity"] ||
+            old_cfg["immediate_commit"] != cfg["immediate_commit"])
+        {
+            return "Changing block_size, bitmap_granularity or immediate_commit"
+                " for an existing pool will lead to incomplete PGs. Use --force to proceed";
+        }
+        if (old_cfg["pg_stripe_size"] != cfg["pg_stripe_size"])
+        {
+            return "Changing pg_stripe_size for an existing pool will lead to data loss. Use --force to proceed";
+        }
+    }
+
+    // Validate values
+    if (cfg["name"].string_value() == "")
+    {
+        return "Non-empty pool name is required";
+    }
+
+    // scheme
+    auto scheme = etcd_state_client_t::parse_scheme(cfg["scheme"].string_value());
+    if (!scheme)
+    {
+        return "Scheme must be one of \"replicated\", \"ec\" or \"xor\"";
+    }
+
+    // pg_size
+    auto pg_size = cfg["pg_size"].uint64_value();
+    if (!pg_size)
+    {
+        return "Non-zero PG size is required";
+    }
+    if (scheme != POOL_SCHEME_REPLICATED && pg_size < 3)
+    {
+        return "PG size can't be smaller than 3 for EC/XOR pools";
+    }
+    if (pg_size > 256)
+    {
+        return "PG size can't be greater than 256";
+    }
+
+    // parity_chunks
+    uint64_t parity_chunks = 1;
+    if (scheme == POOL_SCHEME_EC)
+    {
+        parity_chunks = cfg["parity_chunks"].uint64_value();
+        if (!parity_chunks)
+        {
+            return "Non-zero parity_chunks is required";
+        }
+        if (parity_chunks > pg_size-2)
+        {
+            return "parity_chunks can't be greater than "+std::to_string(pg_size-2)+" (PG size - 2)";
+        }
+    }
+
+    // pg_minsize
+    auto pg_minsize = cfg["pg_minsize"].uint64_value();
+    if (!pg_minsize)
+    {
+        return "Non-zero pg_minsize is required";
+    }
+    else if (pg_minsize > pg_size)
+    {
+        return "pg_minsize can't be greater than "+std::to_string(pg_size)+" (PG size)";
+    }
+    else if (scheme != POOL_SCHEME_REPLICATED && pg_minsize < pg_size-parity_chunks)
+    {
+        return "pg_minsize can't be smaller than "+std::to_string(pg_size-parity_chunks)+
+            " (pg_size - parity_chunks) for XOR/EC pool";
+    }
+
+    // pg_count
+    if (!cfg["pg_count"].uint64_value())
+    {
+        return "Non-zero pg_count is required";
+    }
+
+    // max_osd_combinations
+    if (!cfg["max_osd_combinations"].is_null() && cfg["max_osd_combinations"].uint64_value() < 100)
+    {
+        return "max_osd_combinations must be at least 100, but it is "+cfg["max_osd_combinations"].as_string();
+    }
+
+    // block_size
+    auto block_size = cfg["block_size"].uint64_value();
+    if (!cfg["block_size"].is_null() && ((block_size & (block_size-1)) ||
+        block_size < MIN_DATA_BLOCK_SIZE || block_size > MAX_DATA_BLOCK_SIZE))
+    {
+        return "block_size must be a power of two between "+std::to_string(MIN_DATA_BLOCK_SIZE)+
+            " and "+std::to_string(MAX_DATA_BLOCK_SIZE)+", but it is "+std::to_string(block_size);
+    }
+    block_size = (block_size ? block_size : global_block_size);
+
+    // bitmap_granularity
+    auto bitmap_granularity = cfg["bitmap_granularity"].uint64_value();
+    if (!cfg["bitmap_granularity"].is_null() && (!bitmap_granularity || (bitmap_granularity % 512)))
+    {
+        return "bitmap_granularity must be a multiple of 512, but it is "+std::to_string(bitmap_granularity);
+    }
+    bitmap_granularity = (bitmap_granularity ? bitmap_granularity : global_bitmap_granularity);
+    if (block_size % bitmap_granularity)
+    {
+        return "bitmap_granularity must divide data block size ("+std::to_string(block_size)+"), but it is "+std::to_string(bitmap_granularity);
+    }
+
+    // immediate_commit
+    if (!cfg["immediate_commit"].is_null() && !etcd_state_client_t::parse_immediate_commit(cfg["immediate_commit"].string_value()))
+    {
+        return "immediate_commit must be one of \"all\", \"small\", or \"none\", but it is "+cfg["immediate_commit"].as_string();
+    }
+
+    // scrub_interval
+    if (!cfg["scrub_interval"].is_null())
+    {
+        bool ok;
+        parse_time(cfg["scrub_interval"].string_value(), &ok);
+        if (!ok)
+        {
+            return "scrub_interval must be a time interval (number + unit s/m/h/d/M/y), but it is "+cfg["scrub_interval"].as_string();
+        }
+    }
+
+    return "";
+}

src/cli_pool_cfg.h

@@ -0,0 +1,10 @@
+// Copyright (c) Vitaliy Filippov, 2024
+// License: VNPL-1.1 (see README.md for details)
+
+#pragma once
+
+#include "json11/json11.hpp"
+#include <stdint.h>
+
+std::string validate_pool_config(json11::Json::object & new_cfg, json11::Json old_cfg,
+    uint64_t global_block_size, uint64_t global_bitmap_granularity, bool force);

src/cli_pool_create.cpp

@@ -0,0 +1,622 @@
+// Copyright (c) MIND Software LLC, 2023 (info@mindsw.io)
+// I accept Vitastor CLA: see CLA-en.md for details
+// Copyright (c) Vitaliy Filippov, 2024
+// License: VNPL-1.1 (see README.md for details)
+
+#include <ctype.h>
+#include "cli.h"
+#include "cli_pool_cfg.h"
+#include "cluster_client.h"
+#include "epoll_manager.h"
+#include "pg_states.h"
+#include "str_util.h"
+
+struct pool_creator_t
+{
+    cli_tool_t *parent;
+    json11::Json::object cfg;
+
+    bool force = false;
+    bool wait = false;
+
+    int state = 0;
+    cli_result_t result;
+
+    struct {
+        uint32_t retries = 5;
+        uint32_t interval = 0;
+        bool passed = false;
+    } create_check;
+
+    uint64_t new_id = 1;
+    uint64_t new_pools_mod_rev;
+    json11::Json state_node_tree;
+    json11::Json new_pools;
+
+    bool is_done() { return state == 100; }
+
+    void loop()
+    {
+        if (state == 1)
+            goto resume_1;
+        else if (state == 2)
+            goto resume_2;
+        else if (state == 3)
+            goto resume_3;
+        else if (state == 4)
+            goto resume_4;
+        else if (state == 5)
+            goto resume_5;
+        else if (state == 6)
+            goto resume_6;
+        else if (state == 7)
+            goto resume_7;
+        else if (state == 8)
+            goto resume_8;
+
+        // Validate pool parameters
+        result.text = validate_pool_config(cfg, json11::Json(), parent->cli->st_cli.global_block_size,
+            parent->cli->st_cli.global_bitmap_granularity, force);
+        if (result.text != "")
+        {
+            result.err = EINVAL;
+            state = 100;
+            return;
+        }
+        state = 1;
+resume_1:
+        // If not forced, check that we have enough osds for pg_size
+        if (!force)
+        {
+            // Get node_placement configuration from etcd
+            parent->etcd_txn(json11::Json::object {
+                { "success", json11::Json::array {
+                    json11::Json::object {
+                        { "request_range", json11::Json::object {
+                            { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/node_placement") },
+                        } }
+                    },
+                } },
+            });
+
+            state = 2;
+resume_2:
+            if (parent->waiting > 0)
+                return;
+            if (parent->etcd_err.err)
+            {
+                result = parent->etcd_err;
+                state = 100;
+                return;
+            }
+
+            // Get state_node_tree based on node_placement and osd peer states
+            {
+                auto kv = parent->cli->st_cli.parse_etcd_kv(parent->etcd_result["responses"][0]["response_range"]["kvs"][0]);
+                state_node_tree = get_state_node_tree(kv.value.object_items());
+            }
+
+            // Skip tag checks, if pool has none
+            if (cfg["osd_tags"].array_items().size())
+            {
+                // Get osd configs (for tags) of osds in state_node_tree
+                {
+                    json11::Json::array osd_configs;
+                    for (auto osd_num: state_node_tree["osds"].array_items())
+                    {
+                        osd_configs.push_back(json11::Json::object {
+                            { "request_range", json11::Json::object {
+                                { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/osd/"+osd_num.as_string()) },
+                            } }
+                        });
+                    }
+                    parent->etcd_txn(json11::Json::object { { "success", osd_configs, }, });
+                }
+
+                state = 3;
+resume_3:
+                if (parent->waiting > 0)
+                    return;
+                if (parent->etcd_err.err)
+                {
+                    result = parent->etcd_err;
+                    state = 100;
+                    return;
+                }
+
+                // Filter out osds from state_node_tree based on pool/osd tags
+                {
+                    std::vector<json11::Json> osd_configs;
+                    for (auto & ocr: parent->etcd_result["responses"].array_items())
+                    {
+                        auto kv = parent->cli->st_cli.parse_etcd_kv(ocr["response_range"]["kvs"][0]);
+                        osd_configs.push_back(kv.value);
+                    }
+                    state_node_tree = filter_state_node_tree_by_tags(state_node_tree, osd_configs);
+                }
+            }
+
+            // Get stats (for block_size, bitmap_granularity, ...) of osds in state_node_tree
+            {
+                json11::Json::array osd_stats;
+
+                for (auto osd_num: state_node_tree["osds"].array_items())
+                {
+                    osd_stats.push_back(json11::Json::object {
+                        { "request_range", json11::Json::object {
+                            { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/osd/stats/"+osd_num.as_string()) },
+                        } }
+                    });
+                }
+
+                parent->etcd_txn(json11::Json::object { { "success", osd_stats, }, });
+            }
+
+            state = 4;
+resume_4:
+            if (parent->waiting > 0)
+                return;
+            if (parent->etcd_err.err)
+            {
+                result = parent->etcd_err;
+                state = 100;
+                return;
+            }
+
+            // Filter osds from state_node_tree based on pool parameters and osd stats
+            {
+                std::vector<json11::Json> osd_stats;
+                for (auto & ocr: parent->etcd_result["responses"].array_items())
+                {
+                    auto kv = parent->cli->st_cli.parse_etcd_kv(ocr["response_range"]["kvs"][0]);
+                    osd_stats.push_back(kv.value);
+                }
+                state_node_tree = filter_state_node_tree_by_stats(state_node_tree, osd_stats);
+            }
+
+            // Check that pg_size <= max_pg_size
+            {
+                auto failure_domain = cfg["failure_domain"].string_value() == ""
+                    ? "host" : cfg["failure_domain"].string_value();
+                uint64_t max_pg_size = get_max_pg_size(state_node_tree["nodes"].object_items(),
+                    failure_domain, cfg["root_node"].string_value());
+
+                if (cfg["pg_size"].uint64_value() > max_pg_size)
+                {
+                    result = (cli_result_t){
+                        .err = EINVAL,
+                        .text =
+                            "There are "+std::to_string(max_pg_size)+" \""+failure_domain+"\" failure domains with OSDs matching tags and"
+                            " block_size/bitmap_granularity/immediate_commit parameters, but you want to create a"
+                            " pool with "+cfg["pg_size"].as_string()+" OSDs from different failure domains in a PG."
+                            " Change parameters or add --force if you want to create a degraded pool and add OSDs later."
+                    };
+                    state = 100;
+                    return;
+                }
+            }
+        }
+        // Create pool
+        state = 5;
+resume_5:
+        // Get pools from etcd
+        parent->etcd_txn(json11::Json::object {
+            { "success", json11::Json::array {
+                json11::Json::object {
+                    { "request_range", json11::Json::object {
+                        { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                    } }
+                },
+            } },
+        });
+        state = 6;
+resume_6:
+        if (parent->waiting > 0)
+            return;
+        if (parent->etcd_err.err)
+        {
+            result = parent->etcd_err;
+            state = 100;
+            return;
+        }
+        {
+            // Add new pool
+            auto kv = parent->cli->st_cli.parse_etcd_kv(parent->etcd_result["responses"][0]["response_range"]["kvs"][0]);
+            new_pools = create_pool(kv);
+            if (new_pools.is_string())
+            {
+                result = (cli_result_t){ .err = EEXIST, .text = new_pools.string_value() };
+                state = 100;
+                return;
+            }
+            new_pools_mod_rev = kv.mod_revision;
+        }
+        // Update pools in etcd
+        parent->etcd_txn(json11::Json::object {
+            { "compare", json11::Json::array {
+                json11::Json::object {
+                    { "target", "MOD" },
+                    { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                    { "result", "LESS" },
+                    { "mod_revision", new_pools_mod_rev+1 },
+                }
+            } },
+            { "success", json11::Json::array {
+                json11::Json::object {
+                    { "request_put", json11::Json::object {
+                        { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                        { "value", base64_encode(new_pools.dump()) },
+                    } },
+                },
+            } },
+        });
+        state = 7;
+resume_7:
+        if (parent->waiting > 0)
+            return;
+        if (parent->etcd_err.err)
+        {
+            result = parent->etcd_err;
+            state = 100;
+            return;
+        }
+
+        // Perform final create-check
+        create_check.interval = parent->cli->config["mon_change_timeout"].uint64_value();
+        if (!create_check.interval)
+            create_check.interval = 1000;
+
+        state = 8;
+resume_8:
+        if (parent->waiting > 0)
+            return;
+
+        // Unless forced, check that pool was created and is active
+        if (!wait)
+        {
+            create_check.passed = true;
+        }
+        else if (create_check.retries)
+        {
+            create_check.retries--;
+            parent->waiting++;
+            parent->epmgr->tfd->set_timer(create_check.interval, false, [this](int timer_id)
+            {
+                if (parent->cli->st_cli.pool_config.find(new_id) != parent->cli->st_cli.pool_config.end())
+                {
+                    auto & pool_cfg = parent->cli->st_cli.pool_config[new_id];
+                    create_check.passed = pool_cfg.real_pg_count > 0;
+                    for (auto pg_it = pool_cfg.pg_config.begin(); pg_it != pool_cfg.pg_config.end(); pg_it++)
+                    {
+                        if (!(pg_it->second.cur_state & PG_ACTIVE))
+                        {
+                            create_check.passed = false;
+                            break;
+                        }
+                    }
+                    if (create_check.passed)
+                        create_check.retries = 0;
+                }
+                parent->waiting--;
+                parent->ringloop->wakeup();
+            });
+            return;
+        }
+
+        if (!create_check.passed)
+        {
+            result = (cli_result_t) {
+                .err = EAGAIN,
+                .text = "Pool "+cfg["name"].string_value()+" was created, but failed to become active."
+                    " This may indicate that cluster state has changed while the pool was being created."
+                    " Please check the current state and adjust the pool configuration if necessary.",
+            };
+        }
+        else
+        {
+            result = (cli_result_t){
+                .err = 0,
+                .text = "Pool "+cfg["name"].string_value()+" created",
+                .data = new_pools[std::to_string(new_id)],
+            };
+        }
+        state = 100;
+    }
+
+    // Returns a JSON object of form {"nodes": {...}, "osds": [...]} that
+    // contains: all nodes (osds, hosts, ...) based on node_placement config
+    // and current peer state, and a list of active peer osds.
+    json11::Json get_state_node_tree(json11::Json::object node_placement)
+    {
+        // Erase non-peer osd nodes from node_placement
+        for (auto np_it = node_placement.begin(); np_it != node_placement.end();)
+        {
+            // Numeric nodes are osds
+            osd_num_t osd_num = stoull_full(np_it->first);
+
+            // If node is osd and it is not in peer states, erase it
+            if (osd_num > 0 &&
+                parent->cli->st_cli.peer_states.find(osd_num) == parent->cli->st_cli.peer_states.end())
+            {
+                node_placement.erase(np_it++);
+            }
+            else
+                np_it++;
+        }
+
+        // List of peer osds
+        std::vector<std::string> peer_osds;
+
+        // Record peer osds and add missing osds/hosts to np
+        for (auto & ps: parent->cli->st_cli.peer_states)
+        {
+            std::string osd_num = std::to_string(ps.first);
+
+            // Record peer osd
+            peer_osds.push_back(osd_num);
+
+            // Add osd, if necessary
+            if (node_placement.find(osd_num) == node_placement.end())
+            {
+                std::string osd_host = ps.second["host"].as_string();
+
+                // Add host, if necessary
+                if (node_placement.find(osd_host) == node_placement.end())
+                {
+                    node_placement[osd_host] = json11::Json::object {
+                        { "level", "host" }
+                    };
+                }
+
+                node_placement[osd_num] = json11::Json::object {
+                    { "parent", osd_host }
+                };
+            }
+        }
+
+        return json11::Json::object { { "osds", peer_osds }, { "nodes", node_placement } };
+    }
+
+    // Returns new state_node_tree based on given state_node_tree with osds
+    // filtered out by tags in given osd_configs and current pool config.
+    // Requires: state_node_tree["osds"] must match osd_configs 1-1
+    json11::Json filter_state_node_tree_by_tags(const json11::Json & state_node_tree, std::vector<json11::Json> & osd_configs)
+    {
+        auto & osds = state_node_tree["osds"].array_items();
+
+        // Accepted state_node_tree nodes
+        auto accepted_nodes = state_node_tree["nodes"].object_items();
+
+        // List of accepted osds
+        std::vector<std::string> accepted_osds;
+
+        for (size_t i = 0; i < osd_configs.size(); i++)
+        {
+            auto & oc = osd_configs[i].object_items();
+
+            // Get osd number
+            auto osd_num = osds[i].as_string();
+
+            // We need tags in config to check against pool tags
+            if (oc.find("tags") == oc.end())
+            {
+                // Exclude osd from state_node_tree nodes
+                accepted_nodes.erase(osd_num);
+                continue;
+            }
+            else
+            {
+                // If all pool tags are in osd tags, accept osd
+                if (all_in_tags(osd_configs[i]["tags"], cfg["osd_tags"]))
+                {
+                    accepted_osds.push_back(osd_num);
+                }
+                // Otherwise, exclude osd
+                else
+                {
+                    // Exclude osd from state_node_tree nodes
+                    accepted_nodes.erase(osd_num);
+                }
+            }
+        }
+
+        return json11::Json::object { { "osds", accepted_osds }, { "nodes", accepted_nodes } };
+    }
+
+    // Returns new state_node_tree based on given state_node_tree with osds
+    // filtered out by stats parameters (block_size, bitmap_granularity) in
+    // given osd_stats and current pool config.
+    // Requires: state_node_tree["osds"] must match osd_stats 1-1
+    json11::Json filter_state_node_tree_by_stats(const json11::Json & state_node_tree, std::vector<json11::Json> & osd_stats)
+    {
+        auto & osds = state_node_tree["osds"].array_items();
+
+        // Accepted state_node_tree nodes
+        auto accepted_nodes = state_node_tree["nodes"].object_items();
+
+        // List of accepted osds
+        std::vector<std::string> accepted_osds;
+
+        uint64_t p_block_size = cfg["block_size"].uint64_value()
+            ? cfg["block_size"].uint64_value()
+            : parent->cli->st_cli.global_block_size;
+        uint64_t p_bitmap_granularity = cfg["bitmap_granularity"].uint64_value()
+            ? cfg["bitmap_granularity"].uint64_value()
+            : parent->cli->st_cli.global_bitmap_granularity;
+        uint32_t p_immediate_commit = cfg["immediate_commit"].is_string()
+            ? etcd_state_client_t::parse_immediate_commit(cfg["immediate_commit"].string_value())
+            : parent->cli->st_cli.global_immediate_commit;
+
+        for (size_t i = 0; i < osd_stats.size(); i++)
+        {
+            auto & os = osd_stats[i];
+            // Get osd number
+            auto osd_num = osds[i].as_string();
+            if (!os["data_block_size"].is_null() && os["data_block_size"] != p_block_size ||
+                !os["bitmap_granularity"].is_null() && os["bitmap_granularity"] != p_bitmap_granularity ||
+                !os["immediate_commit"].is_null() &&
+                etcd_state_client_t::parse_immediate_commit(os["immediate_commit"].string_value()) < p_immediate_commit)
+            {
+                accepted_nodes.erase(osd_num);
+            }
+            else
+            {
+                accepted_osds.push_back(osd_num);
+            }
+        }
+
+        return json11::Json::object { { "osds", accepted_osds }, { "nodes", accepted_nodes } };
+    }
+
+    // Returns maximum pg_size possible for given node_tree and failure_domain, starting at parent_node
+    uint64_t get_max_pg_size(json11::Json::object node_tree, const std::string & level, const std::string & parent_node)
+    {
+        uint64_t max_pg_sz = 0;
+
+        std::vector<std::string> nodes;
+
+        // Check if parent node is an osd (numeric)
+        if (parent_node != "" && stoull_full(parent_node))
+        {
+            // Add it to node list if osd is in node tree
+            if (node_tree.find(parent_node) != node_tree.end())
+                nodes.push_back(parent_node);
+        }
+        // If parent node given, ...
+        else if (parent_node != "")
+        {
+            // ... look for children nodes of this parent
+            for (auto & sn: node_tree)
+            {
+                auto & props = sn.second.object_items();
+
+                auto parent_prop = props.find("parent");
+                if (parent_prop != props.end() && (parent_prop->second.as_string() == parent_node))
+                {
+                    nodes.push_back(sn.first);
+
+                    // If we're not looking for all osds, we only need a single
+                    // child osd node
+                    if (level != "osd" && stoull_full(sn.first))
+                        break;
+                }
+            }
+        }
+        // No parent node given, and we're not looking for all osds
+        else if (level != "osd")
+        {
+            // ... look for all level nodes
+            for (auto & sn: node_tree)
+            {
+                auto & props = sn.second.object_items();
+
+                auto level_prop = props.find("level");
+                if (level_prop != props.end() && (level_prop->second.as_string() == level))
+                {
+                    nodes.push_back(sn.first);
+                }
+            }
+        }
+        // Otherwise, ...
+        else
+        {
+            // ... we're looking for osd nodes only
+            for (auto & sn: node_tree)
+            {
+                if (stoull_full(sn.first))
+                {
+                    nodes.push_back(sn.first);
+                }
+            }
+        }
+
+        // Process gathered nodes
+        for (auto & node: nodes)
+        {
+            // Check for osd node, return constant max size
+            if (stoull_full(node))
+            {
+                max_pg_sz += 1;
+            }
+            // Otherwise, ...
+            else
+            {
+                // ... exclude parent node from tree, and ...
+                node_tree.erase(parent_node);
+
+                // ... descend onto the resulting tree
+                max_pg_sz += get_max_pg_size(node_tree, level, node);
+            }
+        }
+
+        return max_pg_sz;
+    }
+
+    json11::Json create_pool(const etcd_kv_t & kv)
+    {
+        for (auto & p: kv.value.object_items())
+        {
+            // ID
+            uint64_t pool_id = stoull_full(p.first);
+            new_id = std::max(pool_id+1, new_id);
+            // Name
+            if (p.second["name"].string_value() == cfg["name"].string_value())
+            {
+                return "Pool with name \""+cfg["name"].string_value()+"\" already exists (ID "+std::to_string(pool_id)+")";
+            }
+        }
+        auto res = kv.value.object_items();
+        res[std::to_string(new_id)] = cfg;
+        return res;
+    }
+
+    // Checks whether tags2 tags are all in tags1 tags
+    bool all_in_tags(json11::Json tags1, json11::Json tags2)
+    {
+        if (!tags2.is_array())
+        {
+            tags2 = json11::Json::array{ tags2.string_value() };
+        }
+        if (!tags1.is_array())
+        {
+            tags1 = json11::Json::array{ tags1.string_value() };
+        }
+        for (auto & tag2: tags2.array_items())
+        {
+            bool found = false;
+            for (auto & tag1: tags1.array_items())
+            {
+                if (tag1 == tag2)
+                {
+                    found = true;
+                    break;
+                }
+            }
+            if (!found)
+            {
+                return false;
+            }
+        }
+        return true;
+    }
+};
+
+std::function<bool(cli_result_t &)> cli_tool_t::start_pool_create(json11::Json cfg)
+{
+    auto pool_creator = new pool_creator_t();
+    pool_creator->parent = this;
+    pool_creator->cfg = cfg.object_items();
+    pool_creator->force = cfg["force"].bool_value();
+    pool_creator->wait = cfg["wait"].bool_value();
+    return [pool_creator](cli_result_t & result)
+    {
+        pool_creator->loop();
+        if (pool_creator->is_done())
+        {
+            result = pool_creator->result;
+            delete pool_creator;
+            return true;
+        }
+        return false;
+    };
+}

src/cli_pool_ls.cpp

@@ -0,0 +1,723 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+
+#include <algorithm>
+#include "cli.h"
+#include "cluster_client.h"
+#include "str_util.h"
+#include "pg_states.h"
+
+// List pools with space statistics
+// - df - minimal list with % used space
+// - pool-ls - same but with PG state and recovery %
+// - pool-ls -l - same but also include I/O statistics
+// - pool-ls --detail - use list format, include PG states, I/O stats and all pool parameters
+struct pool_lister_t
+{
+    cli_tool_t *parent;
+    std::string sort_field;
+    std::set<std::string> only_names;
+    bool reverse = false;
+    int max_count = 0;
+    bool show_recovery = false;
+    bool show_stats = false;
+    bool detailed = false;
+
+    int state = 0;
+    cli_result_t result;
+    std::map<pool_id_t, json11::Json::object> pool_stats;
+    struct io_stats_t
+    {
+        uint64_t count = 0;
+        uint64_t read_iops = 0;
+        uint64_t read_bps = 0;
+        uint64_t read_lat = 0;
+        uint64_t write_iops = 0;
+        uint64_t write_bps = 0;
+        uint64_t write_lat = 0;
+        uint64_t delete_iops = 0;
+        uint64_t delete_bps = 0;
+        uint64_t delete_lat = 0;
+    };
+    struct object_counts_t
+    {
+        uint64_t object_count = 0;
+        uint64_t misplaced_count = 0;
+        uint64_t degraded_count = 0;
+        uint64_t incomplete_count = 0;
+    };
+
+    bool is_done()
+    {
+        return state == 100;
+    }
+
+    void get_pool_stats(int base_state)
+    {
+        if (state == base_state+1)
+            goto resume_1;
+        // Space statistics - pool/stats/<pool>
+        parent->etcd_txn(json11::Json::object {
+            { "success", json11::Json::array {
+                json11::Json::object {
+                    { "request_range", json11::Json::object {
+                        { "key", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/pool/stats/"
+                        ) },
+                        { "range_end", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/pool/stats0"
+                        ) },
+                    } },
+                },
+                json11::Json::object {
+                    { "request_range", json11::Json::object {
+                        { "key", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/osd/stats/"
+                        ) },
+                        { "range_end", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/osd/stats0"
+                        ) },
+                    } },
+                },
+                json11::Json::object {
+                    { "request_range", json11::Json::object {
+                        { "key", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/config/pools"
+                        ) },
+                    } },
+                },
+            } },
+        });
+        state = base_state+1;
+resume_1:
+        if (parent->waiting > 0)
+            return;
+        if (parent->etcd_err.err)
+        {
+            result = parent->etcd_err;
+            state = 100;
+            return;
+        }
+        auto space_info = parent->etcd_result;
+        auto config_pools = space_info["responses"][2]["response_range"]["kvs"][0];
+        if (!config_pools.is_null())
+        {
+            config_pools = parent->cli->st_cli.parse_etcd_kv(config_pools).value;
+        }
+        for (auto & kv_item: space_info["responses"][0]["response_range"]["kvs"].array_items())
+        {
+            auto kv = parent->cli->st_cli.parse_etcd_kv(kv_item);
+            // pool ID
+            pool_id_t pool_id;
+            char null_byte = 0;
+            int scanned = sscanf(kv.key.substr(parent->cli->st_cli.etcd_prefix.length()).c_str(), "/pool/stats/%u%c", &pool_id, &null_byte);
+            if (scanned != 1 || !pool_id || pool_id >= POOL_ID_MAX)
+            {
+                fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
+                continue;
+            }
+            // pool/stats/<N>
+            pool_stats[pool_id] = kv.value.object_items();
+        }
+        std::map<pool_id_t, uint64_t> osd_free;
+        for (auto & kv_item: space_info["responses"][1]["response_range"]["kvs"].array_items())
+        {
+            auto kv = parent->cli->st_cli.parse_etcd_kv(kv_item);
+            // osd ID
+            osd_num_t osd_num;
+            char null_byte = 0;
+            int scanned = sscanf(kv.key.substr(parent->cli->st_cli.etcd_prefix.length()).c_str(), "/osd/stats/%ju%c", &osd_num, &null_byte);
+            if (scanned != 1 || !osd_num || osd_num >= POOL_ID_MAX)
+            {
+                fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
+                continue;
+            }
+            // osd/stats/<N>::free
+            osd_free[osd_num] = kv.value["free"].uint64_value();
+        }
+        // Calculate max_avail for each pool
+        for (auto & pp: parent->cli->st_cli.pool_config)
+        {
+            auto & pool_cfg = pp.second;
+            uint64_t pool_avail = UINT64_MAX;
+            std::map<osd_num_t, uint64_t> pg_per_osd;
+            bool active = pool_cfg.real_pg_count > 0;
+            uint64_t pg_states = 0;
+            for (auto & pgp: pool_cfg.pg_config)
+            {
+                if (!(pgp.second.cur_state & PG_ACTIVE))
+                {
+                    active = false;
+                }
+                pg_states |= pgp.second.cur_state;
+                for (auto pg_osd: pgp.second.target_set)
+                {
+                    if (pg_osd != 0)
+                    {
+                        pg_per_osd[pg_osd]++;
+                    }
+                }
+            }
+            for (auto pg_per_pair: pg_per_osd)
+            {
+                uint64_t pg_free = osd_free[pg_per_pair.first] * pool_cfg.real_pg_count / pg_per_pair.second;
+                if (pool_avail > pg_free)
+                {
+                    pool_avail = pg_free;
+                }
+            }
+            if (pool_avail == UINT64_MAX)
+            {
+                pool_avail = 0;
+            }
+            if (pool_cfg.scheme != POOL_SCHEME_REPLICATED)
+            {
+                pool_avail *= (pool_cfg.pg_size - pool_cfg.parity_chunks);
+            }
+            // incomplete > has_incomplete > degraded > has_degraded > has_misplaced
+            std::string status;
+            if (!active)
+                status = "inactive";
+            else if (pg_states & PG_INCOMPLETE)
+                status = "incomplete";
+            else if (pg_states & PG_HAS_INCOMPLETE)
+                status = "has_incomplete";
+            else if (pg_states & PG_DEGRADED)
+                status = "degraded";
+            else if (pg_states & PG_HAS_DEGRADED)
+                status = "has_degraded";
+            else if (pg_states & PG_HAS_MISPLACED)
+                status = "has_misplaced";
+            else
+                status = "active";
+            pool_stats[pool_cfg.id] = json11::Json::object {
+                { "id", (uint64_t)pool_cfg.id },
+                { "name", pool_cfg.name },
+                { "status", status },
+                { "pg_count", pool_cfg.pg_count },
+                { "real_pg_count", pool_cfg.real_pg_count },
+                { "scheme_name", pool_cfg.scheme == POOL_SCHEME_REPLICATED
+                    ? std::to_string(pool_cfg.pg_size)+"/"+std::to_string(pool_cfg.pg_minsize)
+                    : "EC "+std::to_string(pool_cfg.pg_size-pool_cfg.parity_chunks)+"+"+std::to_string(pool_cfg.parity_chunks) },
+                { "used_raw", (uint64_t)(pool_stats[pool_cfg.id]["used_raw_tb"].number_value() * ((uint64_t)1<<40)) },
+                { "total_raw", (uint64_t)(pool_stats[pool_cfg.id]["total_raw_tb"].number_value() * ((uint64_t)1<<40)) },
+                { "max_available", pool_avail },
+                { "raw_to_usable", pool_stats[pool_cfg.id]["raw_to_usable"].number_value() },
+                { "space_efficiency", pool_stats[pool_cfg.id]["space_efficiency"].number_value() },
+                { "pg_real_size", pool_stats[pool_cfg.id]["pg_real_size"].uint64_value() },
+                { "osd_count", pg_per_osd.size() },
+            };
+        }
+        // Include full pool config
+        for (auto & pp: config_pools.object_items())
+        {
+            if (!pp.second.is_object())
+            {
+                continue;
+            }
+            auto pool_id = stoull_full(pp.first);
+            auto & st = pool_stats[pool_id];
+            for (auto & kv: pp.second.object_items())
+            {
+                if (st.find(kv.first) == st.end())
+                    st[kv.first] = kv.second;
+            }
+        }
+    }
+
+    void get_pg_stats(int base_state)
+    {
+        if (state == base_state+1)
+            goto resume_1;
+        // Space statistics - pool/stats/<pool>
+        parent->etcd_txn(json11::Json::object {
+            { "success", json11::Json::array {
+                json11::Json::object {
+                    { "request_range", json11::Json::object {
+                        { "key", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/pg/stats/"
+                        ) },
+                        { "range_end", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/pg/stats0"
+                        ) },
+                    } },
+                },
+            } },
+        });
+        state = base_state+1;
+resume_1:
+        if (parent->waiting > 0)
+            return;
+        if (parent->etcd_err.err)
+        {
+            result = parent->etcd_err;
+            state = 100;
+            return;
+        }
+        auto pg_stats = parent->etcd_result["responses"][0]["response_range"]["kvs"];
+        // Calculate recovery percent
+        std::map<pool_id_t, object_counts_t> counts;
+        for (auto & kv_item: pg_stats.array_items())
+        {
+            auto kv = parent->cli->st_cli.parse_etcd_kv(kv_item);
+            // pool ID & pg number
+            pool_id_t pool_id;
+            pg_num_t pg_num = 0;
+            char null_byte = 0;
+            int scanned = sscanf(kv.key.substr(parent->cli->st_cli.etcd_prefix.length()).c_str(),
+                "/pg/stats/%u/%u%c", &pool_id, &pg_num, &null_byte);
+            if (scanned != 2 || !pool_id || pool_id >= POOL_ID_MAX)
+            {
+                fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
+                continue;
+            }
+            auto & cnt = counts[pool_id];
+            cnt.object_count += kv.value["object_count"].uint64_value();
+            cnt.misplaced_count += kv.value["misplaced_count"].uint64_value();
+            cnt.degraded_count += kv.value["degraded_count"].uint64_value();
+            cnt.incomplete_count += kv.value["incomplete_count"].uint64_value();
+        }
+        for (auto & pp: pool_stats)
+        {
+            auto & cnt = counts[pp.first];
+            auto & st = pp.second;
+            st["object_count"] = cnt.object_count;
+            st["misplaced_count"] = cnt.misplaced_count;
+            st["degraded_count"] = cnt.degraded_count;
+            st["incomplete_count"] = cnt.incomplete_count;
+        }
+    }
+
+    void get_inode_stats(int base_state)
+    {
+        if (state == base_state+1)
+            goto resume_1;
+        // Space statistics - pool/stats/<pool>
+        parent->etcd_txn(json11::Json::object {
+            { "success", json11::Json::array {
+                json11::Json::object {
+                    { "request_range", json11::Json::object {
+                        { "key", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/inode/stats/"
+                        ) },
+                        { "range_end", base64_encode(
+                            parent->cli->st_cli.etcd_prefix+"/inode/stats0"
+                        ) },
+                    } },
+                },
+            } },
+        });
+        state = base_state+1;
+resume_1:
+        if (parent->waiting > 0)
+            return;
+        if (parent->etcd_err.err)
+        {
+            result = parent->etcd_err;
+            state = 100;
+            return;
+        }
+        auto inode_stats = parent->etcd_result["responses"][0]["response_range"]["kvs"];
+        // Performance statistics
+        std::map<pool_id_t, io_stats_t> pool_io;
+        for (auto & kv_item: inode_stats.array_items())
+        {
+            auto kv = parent->cli->st_cli.parse_etcd_kv(kv_item);
+            // pool ID & inode number
+            pool_id_t pool_id;
+            inode_t only_inode_num;
+            char null_byte = 0;
+            int scanned = sscanf(kv.key.substr(parent->cli->st_cli.etcd_prefix.length()).c_str(),
+                "/inode/stats/%u/%ju%c", &pool_id, &only_inode_num, &null_byte);
+            if (scanned != 2 || !pool_id || pool_id >= POOL_ID_MAX || INODE_POOL(only_inode_num) != 0)
+            {
+                fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
+                continue;
+            }
+            auto & io = pool_io[pool_id];
+            io.read_iops += kv.value["read"]["iops"].uint64_value();
+            io.read_bps += kv.value["read"]["bps"].uint64_value();
+            io.read_lat += kv.value["read"]["lat"].uint64_value();
+            io.write_iops += kv.value["write"]["iops"].uint64_value();
+            io.write_bps += kv.value["write"]["bps"].uint64_value();
+            io.write_lat += kv.value["write"]["lat"].uint64_value();
+            io.delete_iops += kv.value["delete"]["iops"].uint64_value();
+            io.delete_bps += kv.value["delete"]["bps"].uint64_value();
+            io.delete_lat += kv.value["delete"]["lat"].uint64_value();
+            io.count++;
+        }
+        for (auto & pp: pool_stats)
+        {
+            auto & io = pool_io[pp.first];
+            if (io.count > 0)
+            {
+                io.read_lat /= io.count;
+                io.write_lat /= io.count;
+                io.delete_lat /= io.count;
+            }
+            auto & st = pp.second;
+            st["read_iops"] = io.read_iops;
+            st["read_bps"] = io.read_bps;
+            st["read_lat"] = io.read_lat;
+            st["write_iops"] = io.write_iops;
+            st["write_bps"] = io.write_bps;
+            st["write_lat"] = io.write_lat;
+            st["delete_iops"] = io.delete_iops;
+            st["delete_bps"] = io.delete_bps;
+            st["delete_lat"] = io.delete_lat;
+        }
+    }
+
+    json11::Json::array to_list()
+    {
+        json11::Json::array list;
+        for (auto & kv: pool_stats)
+        {
+            if (!only_names.size())
+            {
+                list.push_back(kv.second);
+            }
+            else
+            {
+                for (auto glob: only_names)
+                {
+                    if (stupid_glob(kv.second["name"].string_value(), glob))
+                    {
+                        list.push_back(kv.second);
+                        break;
+                    }
+                }
+            }
+        }
+        if (sort_field == "name" || sort_field == "scheme" ||
+            sort_field == "scheme_name" || sort_field == "status")
+        {
+            std::sort(list.begin(), list.end(), [this](json11::Json a, json11::Json b)
+            {
+                auto av = a[sort_field].as_string();
+                auto bv = b[sort_field].as_string();
+                return reverse ? av > bv : av < bv;
+            });
+        }
+        else
+        {
+            std::sort(list.begin(), list.end(), [this](json11::Json a, json11::Json b)
+            {
+                auto av = a[sort_field].number_value();
+                auto bv = b[sort_field].number_value();
+                return reverse ? av > bv : av < bv;
+            });
+        }
+        if (max_count > 0 && list.size() > max_count)
+        {
+            list.resize(max_count);
+        }
+        return list;
+    }
+
+    void loop()
+    {
+        if (state == 1)
+            goto resume_1;
+        if (state == 2)
+            goto resume_2;
+        if (state == 3)
+            goto resume_3;
+        if (state == 100)
+            return;
+        show_stats = show_stats || detailed;
+        show_recovery = show_recovery || detailed;
+resume_1:
+        get_pool_stats(0);
+        if (parent->waiting > 0)
+            return;
+        if (show_stats)
+        {
+resume_2:
+            get_inode_stats(1);
+            if (parent->waiting > 0)
+                return;
+        }
+        if (show_recovery)
+        {
+resume_3:
+            get_pg_stats(2);
+            if (parent->waiting > 0)
+                return;
+        }
+        if (parent->json_output)
+        {
+            // JSON output
+            result.data = to_list();
+            state = 100;
+            return;
+        }
+        json11::Json::array list;
+        for (auto & kv: pool_stats)
+        {
+            auto & st = kv.second;
+            double raw_to = st["raw_to_usable"].number_value();
+            if (raw_to < 0.000001 && raw_to > -0.000001)
+                raw_to = 1;
+            st["pg_count_fmt"] = st["real_pg_count"] == st["pg_count"]
+                ? st["real_pg_count"].as_string()
+                : st["real_pg_count"].as_string()+"->"+st["pg_count"].as_string();
+            st["total_fmt"] = format_size(st["total_raw"].uint64_value() / raw_to);
+            st["used_fmt"] = format_size(st["used_raw"].uint64_value() / raw_to);
+            st["max_avail_fmt"] = format_size(st["max_available"].uint64_value());
+            st["used_pct"] = format_q(st["total_raw"].uint64_value()
+                ? (100 - 100*st["max_available"].uint64_value() *
+                    st["raw_to_usable"].number_value() / st["total_raw"].uint64_value())
+                : 100)+"%";
+            st["eff_fmt"] = format_q(st["space_efficiency"].number_value()*100)+"%";
+            if (show_stats)
+            {
+                st["read_bw"] = format_size(st["read_bps"].uint64_value())+"/s";
+                st["write_bw"] = format_size(st["write_bps"].uint64_value())+"/s";
+                st["delete_bw"] = format_size(st["delete_bps"].uint64_value())+"/s";
+                st["read_iops"] = format_q(st["read_iops"].number_value());
+                st["write_iops"] = format_q(st["write_iops"].number_value());
+                st["delete_iops"] = format_q(st["delete_iops"].number_value());
+                st["read_lat_f"] = format_lat(st["read_lat"].uint64_value());
+                st["write_lat_f"] = format_lat(st["write_lat"].uint64_value());
+                st["delete_lat_f"] = format_lat(st["delete_lat"].uint64_value());
+            }
+            if (show_recovery)
+            {
+                auto object_count = st["object_count"].uint64_value();
+                auto recovery_pct = 100.0 * (object_count - (st["misplaced_count"].uint64_value() +
+                    st["degraded_count"].uint64_value() + st["incomplete_count"].uint64_value())) /
+                    (object_count ? object_count : 1);
+                st["recovery_fmt"] = format_q(recovery_pct)+"%";
+            }
+        }
+        if (detailed)
+        {
+            for (auto & kv: pool_stats)
+            {
+                auto & st = kv.second;
+                auto total = st["object_count"].uint64_value();
+                auto obj_size = st["block_size"].uint64_value();
+                if (!obj_size)
+                    obj_size = parent->cli->st_cli.global_block_size;
+                if (st["scheme"] == "ec")
+                    obj_size *= st["pg_size"].uint64_value() - st["parity_chunks"].uint64_value();
+                else if (st["scheme"] == "xor")
+                    obj_size *= st["pg_size"].uint64_value() - 1;
+                auto n = st["misplaced_count"].uint64_value();
+                if (n > 0)
+                    st["misplaced_fmt"] = format_size(n * obj_size) + " / " + format_q(100.0 * n / total);
+                n = st["degraded_count"].uint64_value();
+                if (n > 0)
+                    st["degraded_fmt"] = format_size(n * obj_size) + " / " + format_q(100.0 * n / total);
+                n = st["incomplete_count"].uint64_value();
+                if (n > 0)
+                    st["incomplete_fmt"] = format_size(n * obj_size) + " / " + format_q(100.0 * n / total);
+                st["read_fmt"] = st["read_bw"].string_value()+", "+st["read_iops"].string_value()+" op/s, "+
+                    st["read_lat_f"].string_value()+" lat";
+                st["write_fmt"] = st["write_bw"].string_value()+", "+st["write_iops"].string_value()+" op/s, "+
+                    st["write_lat_f"].string_value()+" lat";
+                st["delete_fmt"] = st["delete_bw"].string_value()+", "+st["delete_iops"].string_value()+" op/s, "+
+                    st["delete_lat_f"].string_value()+" lat";
+                if (st["scheme"] == "replicated")
+                    st["scheme_name"] = "x"+st["pg_size"].as_string();
+                if (st["failure_domain"].string_value() == "")
+                    st["failure_domain"] = "host";
+                st["osd_tags_fmt"] = implode(", ", st["osd_tags"]);
+                st["primary_affinity_tags_fmt"] = implode(", ", st["primary_affinity_tags"]);
+                if (st["block_size"].uint64_value())
+                    st["block_size_fmt"] = format_size(st["block_size"].uint64_value());
+                if (st["bitmap_granularity"].uint64_value())
+                    st["bitmap_granularity_fmt"] = format_size(st["bitmap_granularity"].uint64_value());
+            }
+            // All pool parameters are only displayed in the "detailed" mode
+            // because there's too many of them to show them in table
+            auto cols = std::vector<std::pair<std::string, std::string>>{
+                { "name", "Name" },
+                { "id", "ID" },
+                { "scheme_name", "Scheme" },
+                { "used_for_fs", "Used for VitastorFS" },
+                { "status", "Status" },
+                { "pg_count_fmt", "PGs" },
+                { "pg_minsize", "PG minsize" },
+                { "failure_domain", "Failure domain" },
+                { "root_node", "Root node" },
+                { "osd_tags_fmt", "OSD tags" },
+                { "primary_affinity_tags_fmt", "Primary affinity" },
+                { "block_size_fmt", "Block size" },
+                { "bitmap_granularity_fmt", "Bitmap granularity" },
+                { "immediate_commit", "Immediate commit" },
+                { "scrub_interval", "Scrub interval" },
+                { "inode_stats_fmt", "Per-inode stats" },
+                { "pg_stripe_size", "PG stripe size" },
+                { "max_osd_combinations", "Max OSD combinations" },
+                { "total_fmt", "Total" },
+                { "used_fmt", "Used" },
+                { "max_avail_fmt", "Available" },
+                { "used_pct", "Used%" },
+                { "eff_fmt", "Efficiency" },
+                { "osd_count", "OSD count" },
+                { "misplaced_fmt", "Misplaced" },
+                { "degraded_fmt", "Degraded" },
+                { "incomplete_fmt", "Incomplete" },
+                { "read_fmt", "Read" },
+                { "write_fmt", "Write" },
+                { "delete_fmt", "Delete" },
+            };
+            auto list = to_list();
+            size_t title_len = 0;
+            for (auto & item: list)
+            {
+                title_len = print_detail_title_len(item, cols, title_len);
+            }
+            for (auto & item: list)
+            {
+                if (result.text != "")
+                    result.text += "\n";
+                result.text += print_detail(item, cols, title_len, parent->color);
+            }
+            state = 100;
+            return;
+        }
+        // Table output: name, scheme_name, pg_count, total, used, max_avail, used%, efficiency
+        json11::Json::array cols;
+        cols.push_back(json11::Json::object{
+            { "key", "name" },
+            { "title", "NAME" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "scheme_name" },
+            { "title", "SCHEME" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "status" },
+            { "title", "STATUS" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "pg_count_fmt" },
+            { "title", "PGS" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "total_fmt" },
+            { "title", "TOTAL" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "used_fmt" },
+            { "title", "USED" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "max_avail_fmt" },
+            { "title", "AVAILABLE" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "used_pct" },
+            { "title", "USED%" },
+        });
+        cols.push_back(json11::Json::object{
+            { "key", "eff_fmt" },
+            { "title", "EFFICIENCY" },
+        });
+        if (show_recovery)
+        {
+            cols.push_back(json11::Json::object{ { "key", "recovery_fmt" }, { "title", "RECOVERY" } });
+        }
+        if (show_stats)
+        {
+            cols.push_back(json11::Json::object{ { "key", "read_bw" }, { "title", "READ" } });
+            cols.push_back(json11::Json::object{ { "key", "read_iops" }, { "title", "IOPS" } });
+            cols.push_back(json11::Json::object{ { "key", "read_lat_f" }, { "title", "LAT" } });
+            cols.push_back(json11::Json::object{ { "key", "write_bw" }, { "title", "WRITE" } });
+            cols.push_back(json11::Json::object{ { "key", "write_iops" }, { "title", "IOPS" } });
+            cols.push_back(json11::Json::object{ { "key", "write_lat_f" }, { "title", "LAT" } });
+            cols.push_back(json11::Json::object{ { "key", "delete_bw" }, { "title", "DELETE" } });
+            cols.push_back(json11::Json::object{ { "key", "delete_iops" }, { "title", "IOPS" } });
+            cols.push_back(json11::Json::object{ { "key", "delete_lat_f" }, { "title", "LAT" } });
+        }
+        result.data = to_list();
+        result.text = print_table(result.data, cols, parent->color);
+        state = 100;
+    }
+};
+
+size_t print_detail_title_len(json11::Json item, std::vector<std::pair<std::string, std::string>> names, size_t prev_len)
+{
+    size_t title_len = prev_len;
+    for (auto & kv: names)
+    {
+        if (!item[kv.first].is_null() && (!item[kv.first].is_string() || item[kv.first].string_value() != ""))
+        {
+            size_t len = utf8_length(kv.second);
+            title_len = title_len < len ? len : title_len;
+        }
+    }
+    return title_len;
+}
+
+std::string print_detail(json11::Json item, std::vector<std::pair<std::string, std::string>> names, size_t title_len, bool use_esc)
+{
+    std::string str;
+    for (auto & kv: names)
+    {
+        if (!item[kv.first].is_null() && (!item[kv.first].is_string() || item[kv.first].string_value() != ""))
+        {
+            str += kv.second;
+            str += ": ";
+            size_t len = utf8_length(kv.second);
+            for (int j = 0; j < title_len-len; j++)
+                str += ' ';
+            if (use_esc)
+                str += "\033[1m";
+            str += item[kv.first].as_string();
+            if (use_esc)
+                str += "\033[0m";
+            str += "\n";
+        }
+    }
+    return str;
+}
+
+std::function<bool(cli_result_t &)> cli_tool_t::start_pool_ls(json11::Json cfg)
+{
+    auto lister = new pool_lister_t();
+    lister->parent = this;
+    lister->show_recovery = cfg["show_recovery"].bool_value();
+    lister->show_stats = cfg["long"].bool_value();
+    lister->detailed = cfg["detail"].bool_value();
+    lister->sort_field = cfg["sort"].string_value();
+    if ((lister->sort_field == "osd_tags") ||
+        (lister->sort_field == "primary_affinity_tags" ))
+        lister->sort_field = lister->sort_field + "_fmt";
+    lister->reverse = cfg["reverse"].bool_value();
+    lister->max_count = cfg["count"].uint64_value();
+    for (auto & item: cfg["names"].array_items())
+    {
+        lister->only_names.insert(item.string_value());
+    }
+    return [lister](cli_result_t & result)
+    {
+        lister->loop();
+        if (lister->is_done())
+        {
+            result = lister->result;
+            delete lister;
+            return true;
+        }
+        return false;
+    };
+}
+
+std::string implode(const std::string & sep, json11::Json array)
+{
+    if (array.is_number() || array.is_bool() || array.is_string())
+    {
+        return array.as_string();
+    }
+    std::string res;
+    bool first = true;
+    for (auto & item: array.array_items())
+    {
+        res += (first ? item.as_string() : sep+item.as_string());
+        first = false;
+    }
+    return res;
+}

src/cli_pool_modify.cpp

@@ -0,0 +1,203 @@
+// Copyright (c) MIND Software LLC, 2023 (info@mindsw.io)
+// I accept Vitastor CLA: see CLA-en.md for details
+// Copyright (c) Vitaliy Filippov, 2024
+// License: VNPL-1.1 (see README.md for details)
+
+#include <ctype.h>
+#include "cli.h"
+#include "cli_pool_cfg.h"
+#include "cluster_client.h"
+#include "str_util.h"
+
+struct pool_changer_t
+{
+    cli_tool_t *parent;
+
+    // Required parameters (id/name)
+    pool_id_t pool_id = 0;
+    std::string pool_name;
+    json11::Json::object cfg;
+    json11::Json::object new_cfg;
+    bool force = false;
+
+    json11::Json old_cfg;
+
+    int state = 0;
+    cli_result_t result;
+
+    // Updated pools
+    json11::Json new_pools;
+
+    // Expected pools mod revision
+    uint64_t pools_mod_rev;
+
+    bool is_done() { return state == 100; }
+
+    void loop()
+    {
+        if (state == 1)
+            goto resume_1;
+        else if (state == 2)
+            goto resume_2;
+        pool_id = stoull_full(cfg["old_name"].string_value());
+        if (!pool_id)
+        {
+            pool_name = cfg["old_name"].string_value();
+            if (pool_name == "")
+            {
+                result = (cli_result_t){ .err = ENOENT, .text = "Pool ID or name is required to modify it" };
+                state = 100;
+                return;
+            }
+        }
+resume_0:
+        // Get pools from etcd
+        parent->etcd_txn(json11::Json::object {
+            { "success", json11::Json::array {
+                json11::Json::object {
+                    { "request_range", json11::Json::object {
+                        { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                    } }
+                },
+            } },
+        });
+        state = 1;
+resume_1:
+        if (parent->waiting > 0)
+            return;
+        if (parent->etcd_err.err)
+        {
+            result = parent->etcd_err;
+            state = 100;
+            return;
+        }
+        {
+            // Parse received pools from etcd
+            auto kv = parent->cli->st_cli.parse_etcd_kv(parent->etcd_result["responses"][0]["response_range"]["kvs"][0]);
+
+            // Get pool by name or ID
+            old_cfg = json11::Json();
+            if (pool_name != "")
+            {
+                for (auto & pce: kv.value.object_items())
+                {
+                    if (pce.second["name"] == pool_name)
+                    {
+                        pool_id = stoull_full(pce.first);
+                        old_cfg = pce.second;
+                        break;
+                    }
+                }
+            }
+            else
+            {
+                pool_name = std::to_string(pool_id);
+                old_cfg = kv.value[pool_name];
+            }
+            if (!old_cfg.is_object())
+            {
+                result = (cli_result_t){ .err = ENOENT, .text = "Pool "+pool_name+" does not exist" };
+                state = 100;
+                return;
+            }
+
+            // Update pool
+            new_cfg = cfg;
+            result.text = validate_pool_config(new_cfg, old_cfg, parent->cli->st_cli.global_block_size,
+                parent->cli->st_cli.global_bitmap_granularity, force);
+            if (result.text != "")
+            {
+                result.err = EINVAL;
+                state = 100;
+                return;
+            }
+
+            if (new_cfg.find("used_for_fs") != new_cfg.end() && !force)
+            {
+                // Check that pool doesn't have images
+                auto img_it = parent->cli->st_cli.inode_config.lower_bound(INODE_WITH_POOL(pool_id, 0));
+                if (img_it != parent->cli->st_cli.inode_config.end() && INODE_POOL(img_it->first) == pool_id &&
+                    img_it->second.name == new_cfg["used_for_fs"].string_value())
+                {
+                    // Only allow metadata image to exist in the FS pool
+                    img_it++;
+                }
+                if (img_it != parent->cli->st_cli.inode_config.end() && INODE_POOL(img_it->first) == pool_id)
+                {
+                    result = (cli_result_t){ .err = ENOENT, .text = "Pool "+pool_name+" has block images, delete them before using it for VitastorFS" };
+                    state = 100;
+                    return;
+                }
+            }
+
+            // Update pool
+            auto pls = kv.value.object_items();
+            pls[std::to_string(pool_id)] = new_cfg;
+            new_pools = pls;
+
+            // Expected pools mod revision
+            pools_mod_rev = kv.mod_revision;
+        }
+        // Update pools in etcd
+        parent->etcd_txn(json11::Json::object {
+            { "compare", json11::Json::array {
+                json11::Json::object {
+                    { "target", "MOD" },
+                    { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                    { "result", "LESS" },
+                    { "mod_revision", pools_mod_rev+1 },
+                }
+            } },
+            { "success", json11::Json::array {
+                json11::Json::object {
+                    { "request_put", json11::Json::object {
+                        { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                        { "value", base64_encode(new_pools.dump()) },
+                    } },
+                },
+            } },
+        });
+        state = 2;
+resume_2:
+        if (parent->waiting > 0)
+            return;
+        if (parent->etcd_err.err)
+        {
+            result = parent->etcd_err;
+            state = 100;
+            return;
+        }
+        if (!parent->etcd_result["succeeded"].bool_value())
+        {
+            // CAS failure - retry
+            fprintf(stderr, "Warning: pool configuration was modified in the meantime by someone else\n");
+            goto resume_0;
+        }
+        // Successfully updated pool
+        result = (cli_result_t){
+            .err = 0,
+            .text = "Pool "+pool_name+" updated",
+            .data = new_pools,
+        };
+        state = 100;
+    }
+};
+
+std::function<bool(cli_result_t &)> cli_tool_t::start_pool_modify(json11::Json cfg)
+{
+    auto pool_changer = new pool_changer_t();
+    pool_changer->parent = this;
+    pool_changer->cfg = cfg.object_items();
+    pool_changer->force = cfg["force"].bool_value();
+    return [pool_changer](cli_result_t & result)
+    {
+        pool_changer->loop();
+        if (pool_changer->is_done())
+        {
+            result = pool_changer->result;
+            delete pool_changer;
+            return true;
+        }
+        return false;
+    };
+}

src/cli_pool_rm.cpp

@@ -0,0 +1,226 @@
+// Copyright (c) MIND Software LLC, 2023 (info@mindsw.io)
+// I accept Vitastor CLA: see CLA-en.md for details
+// Copyright (c) Vitaliy Filippov, 2024
+// License: VNPL-1.1 (see README.md for details)
+
+#include <ctype.h>
+#include "cli.h"
+#include "cluster_client.h"
+#include "str_util.h"
+
+struct pool_remover_t
+{
+    cli_tool_t *parent;
+
+    // Required parameters (id/name)
+
+    pool_id_t pool_id = 0;
+    std::string pool_name;
+
+    // Force removal
+    bool force;
+
+    int state = 0;
+    cli_result_t result;
+
+    // Is pool valid?
+    bool pool_valid = false;
+
+    // Updated pools
+    json11::Json new_pools;
+
+    // Expected pools mod revision
+    uint64_t pools_mod_rev;
+
+    bool is_done() { return state == 100; }
+
+    void loop()
+    {
+        if (state == 1)
+            goto resume_1;
+        else if (state == 2)
+            goto resume_2;
+        else if (state == 3)
+            goto resume_3;
+
+        // Pool name (or id) required
+        if (!pool_id && pool_name == "")
+        {
+            result = (cli_result_t){ .err = EINVAL, .text = "Pool name or id must be given" };
+            state = 100;
+            return;
+        }
+
+        // Validate pool name/id
+
+        // Get pool id by name (if name given)
+        if (pool_name != "")
+        {
+            for (auto & ic: parent->cli->st_cli.pool_config)
+            {
+                if (ic.second.name == pool_name)
+                {
+                    pool_id = ic.first;
+                    pool_valid = 1;
+                    break;
+                }
+            }
+        }
+        // Otherwise, check if given pool id is valid
+        else
+        {
+            // Set pool name from id (for easier logging)
+            pool_name = "id " + std::to_string(pool_id);
+
+            // Look-up pool id in pool_config
+            if (parent->cli->st_cli.pool_config.find(pool_id) != parent->cli->st_cli.pool_config.end())
+            {
+                pool_valid = 1;
+            }
+        }
+
+        // Need a valid pool to proceed
+        if (!pool_valid)
+        {
+            result = (cli_result_t){ .err = ENOENT, .text = "Pool "+pool_name+" does not exist" };
+            state = 100;
+            return;
+        }
+
+        // Unless forced, check if pool has associated Images/Snapshots
+        if (!force)
+        {
+            std::string images;
+
+            for (auto & ic: parent->cli->st_cli.inode_config)
+            {
+                if (pool_id && INODE_POOL(ic.second.num) != pool_id)
+                {
+                    continue;
+                }
+                images += ((images != "") ? ", " : "") + ic.second.name;
+            }
+
+            if (images != "")
+            {
+                result = (cli_result_t){
+                    .err = ENOTEMPTY,
+                    .text =
+                        "Pool "+pool_name+" cannot be removed as it still has the following "
+                        "images/snapshots associated with it: "+images
+                };
+                state = 100;
+                return;
+            }
+        }
+
+        // Proceed to deleting the pool
+        state = 1;
+        do
+        {
+resume_1:
+            // Get pools from etcd
+            parent->etcd_txn(json11::Json::object {
+                { "success", json11::Json::array {
+                    json11::Json::object {
+                        { "request_range", json11::Json::object {
+                            { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                        } }
+                    },
+                } },
+            });
+            state = 2;
+resume_2:
+            if (parent->waiting > 0)
+                return;
+            if (parent->etcd_err.err)
+            {
+                result = parent->etcd_err;
+                state = 100;
+                return;
+            }
+            {
+                // Parse received pools from etcd
+                auto kv = parent->cli->st_cli.parse_etcd_kv(parent->etcd_result["responses"][0]["response_range"]["kvs"][0]);
+
+                // Remove pool
+                auto p = kv.value.object_items();
+                if (p.erase(std::to_string(pool_id)) != 1)
+                {
+                    result = (cli_result_t){
+                        .err = ENOENT,
+                        .text = "Failed to erase pool "+pool_name+" from: "+kv.value.string_value()
+                    };
+                    state = 100;
+                    return;
+                }
+
+                // Record updated pools
+                new_pools = p;
+
+                // Expected pools mod revision
+                pools_mod_rev = kv.mod_revision;
+            }
+            // Update pools in etcd
+            parent->etcd_txn(json11::Json::object {
+                { "compare", json11::Json::array {
+                    json11::Json::object {
+                        { "target", "MOD" },
+                        { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                        { "result", "LESS" },
+                        { "mod_revision", pools_mod_rev+1 },
+                    }
+                } },
+                { "success", json11::Json::array {
+                    json11::Json::object {
+                        { "request_put", json11::Json::object {
+                            { "key", base64_encode(parent->cli->st_cli.etcd_prefix+"/config/pools") },
+                            { "value", base64_encode(new_pools.dump()) },
+                        } },
+                    },
+                } },
+            });
+            state = 3;
+resume_3:
+            if (parent->waiting > 0)
+                return;
+            if (parent->etcd_err.err)
+            {
+                result = parent->etcd_err;
+                state = 100;
+                return;
+            }
+        } while (!parent->etcd_result["succeeded"].bool_value());
+
+        // Successfully deleted pool
+        result = (cli_result_t){
+            .err = 0,
+            .text = "Pool "+pool_name+" deleted",
+            .data = new_pools
+        };
+        state = 100;
+    }
+};
+
+std::function<bool(cli_result_t &)> cli_tool_t::start_pool_rm(json11::Json cfg)
+{
+    auto pool_remover = new pool_remover_t();
+    pool_remover->parent = this;
+
+    pool_remover->pool_id = cfg["pool"].uint64_value();
+    pool_remover->pool_name = pool_remover->pool_id ? "" : cfg["pool"].as_string();
+
+    pool_remover->force = !cfg["force"].is_null();
+
+    return [pool_remover](cli_result_t & result)
+    {
+        pool_remover->loop();
+        if (pool_remover->is_done())
+        {
+            result = pool_remover->result;
+            delete pool_remover;
+            return true;
+        }
+        return false;
+    };
+}

src/cli_rm.cpp

@@ -53,6 +53,8 @@ struct snap_remover_t
     int use_cas = 1;
     // interval between fsyncs
     int fsync_interval = 128;
+    // ignore deletion errors
+    bool down_ok = false;
 
     std::map<inode_t,int> sources;
     std::map<inode_t,uint64_t> inode_used;
@@ -245,6 +247,7 @@ resume_8:
         }
         state = 100;
         result = (cli_result_t){
+            .err = 0,
             .text = "",
             .data = my_result(result.data),
         };
@@ -291,7 +294,7 @@ resume_100:
             if (it == parent->cli->st_cli.inode_config.end())
             {
                 char buf[1024];
-                snprintf(buf, 1024, "Parent inode of layer %s (id 0x%lx) not found", cur->name.c_str(), cur->parent_id);
+                snprintf(buf, 1024, "Parent inode of layer %s (id 0x%jx) not found", cur->name.c_str(), cur->parent_id);
                 state = 100;
                 return;
             }
@@ -384,7 +387,7 @@ resume_100:
                 pool_id_t pool_id = 0;
                 inode_t inode = 0;
                 char null_byte = 0;
-                int scanned = sscanf(kv.key.c_str() + parent->cli->st_cli.etcd_prefix.length()+13, "%u/%lu%c", &pool_id, &inode, &null_byte);
+                int scanned = sscanf(kv.key.c_str() + parent->cli->st_cli.etcd_prefix.length()+13, "%u/%ju%c", &pool_id, &inode, &null_byte);
                 if (scanned != 2 || !inode)
                 {
                     result = (cli_result_t){ .err = EIO, .text = "Bad key returned from etcd: "+kv.key };
@@ -439,7 +442,7 @@ resume_100:
         if (child_it == parent->cli->st_cli.inode_config.end())
         {
             char buf[1024];
-            snprintf(buf, 1024, "Inode 0x%lx disappeared", inverse_child);
+            snprintf(buf, 1024, "Inode 0x%jx disappeared", inverse_child);
             result = (cli_result_t){ .err = EIO, .text = std::string(buf) };
             state = 100;
             return;
@@ -448,7 +451,7 @@ resume_100:
         if (target_it == parent->cli->st_cli.inode_config.end())
         {
             char buf[1024];
-            snprintf(buf, 1024, "Inode 0x%lx disappeared", inverse_parent);
+            snprintf(buf, 1024, "Inode 0x%jx disappeared", inverse_parent);
             result = (cli_result_t){ .err = EIO, .text = std::string(buf) };
             state = 100;
             return;
@@ -576,7 +579,7 @@ resume_100:
         if (cur_cfg_it == parent->cli->st_cli.inode_config.end())
         {
             char buf[1024];
-            snprintf(buf, 1024, "Inode 0x%lx disappeared", cur);
+            snprintf(buf, 1024, "Inode 0x%jx disappeared", cur);
             result = (cli_result_t){ .err = EIO, .text = std::string(buf) };
             state = 100;
             return;
@@ -640,7 +643,7 @@ resume_100:
         if (child_it == parent->cli->st_cli.inode_config.end())
         {
             char buf[1024];
-            snprintf(buf, 1024, "Inode 0x%lx disappeared", child_inode);
+            snprintf(buf, 1024, "Inode 0x%jx disappeared", child_inode);
             result = (cli_result_t){ .err = EIO, .text = std::string(buf) };
             state = 100;
             return;
@@ -649,7 +652,7 @@ resume_100:
         if (target_it == parent->cli->st_cli.inode_config.end())
         {
             char buf[1024];
-            snprintf(buf, 1024, "Inode 0x%lx disappeared", target_inode);
+            snprintf(buf, 1024, "Inode 0x%jx disappeared", target_inode);
             result = (cli_result_t){ .err = EIO, .text = std::string(buf) };
             state = 100;
             return;
@@ -670,7 +673,7 @@ resume_100:
         if (source == parent->cli->st_cli.inode_config.end())
         {
             char buf[1024];
-            snprintf(buf, 1024, "Inode 0x%lx disappeared", inode);
+            snprintf(buf, 1024, "Inode 0x%jx disappeared", inode);
             result = (cli_result_t){ .err = EIO, .text = std::string(buf) };
             state = 100;
             return;
@@ -679,6 +682,7 @@ resume_100:
             { "inode", inode },
             { "pool", (uint64_t)INODE_POOL(inode) },
             { "fsync-interval", fsync_interval },
+            { "down-ok", down_ok },
         });
     }
 };
@@ -690,6 +694,7 @@ std::function<bool(cli_result_t &)> cli_tool_t::start_rm(json11::Json cfg)
     snap_remover->from_name = cfg["from"].string_value();
     snap_remover->to_name = cfg["to"].string_value();
     snap_remover->fsync_interval = cfg["fsync_interval"].uint64_value();
+    snap_remover->down_ok = cfg["down_ok"].bool_value();
     if (!snap_remover->fsync_interval)
         snap_remover->fsync_interval = 128;
     if (!cfg["cas"].is_null())

src/cli_rm_data.cpp

@@ -25,6 +25,7 @@ struct rm_inode_t
     uint64_t inode = 0;
     pool_id_t pool_id = 0;
     uint64_t min_offset = 0;
+    bool down_ok = false;
 
     cli_tool_t *parent = NULL;
     inode_list_t *lister = NULL;
@@ -95,7 +96,7 @@ struct rm_inode_t
             fprintf(stderr, "Some data may remain after delete on OSDs which are currently down: ");
             for (int i = 0; i < inactive_osds.size(); i++)
             {
-                fprintf(stderr, i > 0 ? ", %lu" : "%lu", inactive_osds[i]);
+                fprintf(stderr, i > 0 ? ", %ju" : "%ju", inactive_osds[i]);
             }
             fprintf(stderr, "\n");
         }
@@ -138,7 +139,7 @@ struct rm_inode_t
                     cur_list->in_flight--;
                     if (op->reply.hdr.retval < 0)
                     {
-                        fprintf(stderr, "Failed to remove object %lx:%lx from PG %u (OSD %lu) (retval=%ld)\n",
+                        fprintf(stderr, "Failed to remove object %jx:%jx from PG %u (OSD %ju) (retval=%jd)\n",
                             op->req.rw.inode, op->req.rw.offset,
                             cur_list->pg_num, cur_list->rm_osd_num, op->reply.hdr.retval);
                         error_count++;
@@ -174,7 +175,7 @@ struct rm_inode_t
                 cur_list->synced = true;
                 if (op->reply.hdr.retval < 0)
                 {
-                    fprintf(stderr, "Failed to sync OSD %lu (retval=%ld)\n",
+                    fprintf(stderr, "Failed to sync OSD %ju (retval=%jd)\n",
                         cur_list->rm_osd_num, op->reply.hdr.retval);
                     error_count++;
                 }
@@ -212,7 +213,9 @@ struct rm_inode_t
         }
         if (parent->progress && total_count > 0 && total_done*1000/total_count != total_prev_pct)
         {
-            fprintf(stderr, "\rRemoved %lu/%lu objects, %lu more PGs to list...", total_done, total_count, pgs_to_list);
+            fprintf(stderr, parent->color
+                ? "\rRemoved %ju/%ju objects, %ju more PGs to list..."
+                : "Removed %ju/%ju objects, %ju more PGs to list...\n", total_done, total_count, pgs_to_list);
             total_prev_pct = total_done*1000/total_count;
         }
         if (lists_done && !lists.size())
@@ -221,17 +224,18 @@ struct rm_inode_t
             {
                 fprintf(stderr, "\n");
             }
-            if (parent->progress && (total_done < total_count || inactive_osds.size() > 0 || error_count > 0))
+            bool is_error = (total_done < total_count || inactive_osds.size() > 0 || error_count > 0);
+            if (parent->progress && is_error)
             {
                 fprintf(
-                    stderr, "Warning: Pool:%u,ID:%lu inode data may not have been fully removed.\n"
-                    " Use `vitastor-cli rm-data --pool %u --inode %lu` if you encounter it in listings.\n",
+                    stderr, "Warning: Pool:%u,ID:%ju inode data may not have been fully removed.\n"
+                    "Use `vitastor-cli rm-data --pool %u --inode %ju` if you encounter it in listings.\n",
                     pool_id, INODE_NO_POOL(inode), pool_id, INODE_NO_POOL(inode)
                 );
             }
             result = (cli_result_t){
-                .err = error_count > 0 ? EIO : 0,
-                .text = error_count > 0 ? "Some blocks were not removed" : (
+                .err = is_error && !down_ok ? EIO : 0,
+                .text = is_error ? "Some blocks were not removed" : (
                     "Done, inode "+std::to_string(INODE_NO_POOL(inode))+" from pool "+
                     std::to_string(pool_id)+" removed"),
                 .data = json11::Json::object {
@@ -280,6 +284,7 @@ std::function<bool(cli_result_t &)> cli_tool_t::start_rm_data(json11::Json cfg)
     {
         remover->inode = (remover->inode & (((uint64_t)1 << (64-POOL_ID_BITS)) - 1)) | (((uint64_t)remover->pool_id) << (64-POOL_ID_BITS));
     }
+    remover->down_ok = cfg["down_ok"].bool_value();
     remover->pool_id = INODE_POOL(remover->inode);
     remover->min_offset = cfg["min_offset"].uint64_value();
     return [remover](cli_result_t & result)

src/cli_status.cpp

@@ -132,7 +132,7 @@ resume_2:
             auto kv = parent->cli->st_cli.parse_etcd_kv(osd_stats[i]);
             osd_num_t stat_osd_num = 0;
             char null_byte = 0;
-            int scanned = sscanf(kv.key.c_str() + parent->cli->st_cli.etcd_prefix.size(), "/osd/stats/%lu%c", &stat_osd_num, &null_byte);
+            int scanned = sscanf(kv.key.c_str() + parent->cli->st_cli.etcd_prefix.size(), "/osd/stats/%ju%c", &stat_osd_num, &null_byte);
             if (scanned != 1 || !stat_osd_num)
             {
                 fprintf(stderr, "Invalid key in etcd: %s\n", kv.key.c_str());
@@ -283,7 +283,7 @@ resume_2:
         }
         printf(
             "  cluster:\n"
-            "    etcd: %d / %ld up, %s database size\n"
+            "    etcd: %d / %zd up, %s database size\n"
             "    mon:  %d up%s\n"
             "    osd:  %d / %d up\n"
             "  \n"

src/cluster_client.cpp

@@ -6,7 +6,7 @@
 #include "cluster_client_impl.h"
 #include "http_client.h" // json_is_true
 
-cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json & config)
+cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config)
 {
     wb = new writeback_cache_t();
 
@@ -238,7 +238,8 @@ void cluster_client_t::erase_op(cluster_op_t *op)
         // which may continue following SYNCs, but these SYNCs
         // should know about the changed buffer state
         // This is ugly but this is the way we do it
-        std::function<void(cluster_op_t*)>(op->callback)(op);
+        auto cb = std::move(op->callback);
+        cb(op);
     }
     if (!(flags & OP_IMMEDIATE_COMMIT) || enable_writeback)
     {
@@ -248,7 +249,8 @@ void cluster_client_t::erase_op(cluster_op_t *op)
     {
         // Call callback at the end to avoid inconsistencies in prev_wait
         // if the callback adds more operations itself
-        std::function<void(cluster_op_t*)>(op->callback)(op);
+        auto cb = std::move(op->callback);
+        cb(op);
     }
     if (flags & OP_FLUSH_BUFFER)
     {
@@ -265,7 +267,7 @@ void cluster_client_t::erase_op(cluster_op_t *op)
     }
 }
 
-void cluster_client_t::continue_ops(bool up_retry)
+void cluster_client_t::continue_ops(int time_passed)
 {
     if (!pgs_loaded)
     {
@@ -277,22 +279,27 @@ void cluster_client_t::continue_ops(bool up_retry)
         // Attempt to reenter the function
         return;
     }
+    int reset_duration = 0;
 restart:
     continuing_ops = 1;
     for (auto op = op_queue_head; op; )
     {
         cluster_op_t *next_op = op->next;
-        if (!op->up_wait || up_retry)
+        if (op->retry_after && time_passed)
         {
-            op->up_wait = false;
-            if (!op->prev_wait)
+            op->retry_after = op->retry_after > time_passed ? op->retry_after-time_passed : 0;
+            if (op->retry_after && (!reset_duration || op->retry_after < reset_duration))
             {
-                if (op->opcode == OSD_OP_SYNC)
-                    continue_sync(op);
-                else
-                    continue_rw(op);
+                reset_duration = op->retry_after;
             }
         }
+        if (!op->retry_after && !op->prev_wait)
+        {
+            if (op->opcode == OSD_OP_SYNC)
+                continue_sync(op);
+            else
+                continue_rw(op);
+        }
         op = next_op;
         if (continuing_ops == 2)
         {
@@ -300,6 +307,27 @@ restart:
         }
     }
     continuing_ops = 0;
+    reset_retry_timer(reset_duration);
+}
+
+void cluster_client_t::reset_retry_timer(int new_duration)
+{
+    if (retry_timeout_duration && retry_timeout_duration <= new_duration || !new_duration)
+    {
+        return;
+    }
+    if (retry_timeout_id)
+    {
+        tfd->clear_timer(retry_timeout_id);
+    }
+    retry_timeout_duration = new_duration;
+    retry_timeout_id = tfd->set_timer(retry_timeout_duration, false, [this](int)
+    {
+        int time_passed = retry_timeout_duration;
+        retry_timeout_id = 0;
+        retry_timeout_duration = 0;
+        continue_ops(time_passed);
+    });
 }
 
 void cluster_client_t::on_load_config_hook(json11::Json::object & etcd_global_config)
@@ -349,15 +377,25 @@ void cluster_client_t::on_load_config_hook(json11::Json::object & etcd_global_co
     {
         client_max_writeback_iodepth = DEFAULT_CLIENT_MAX_WRITEBACK_IODEPTH;
     }
-    // up_wait_retry_interval
-    up_wait_retry_interval = config["up_wait_retry_interval"].uint64_value();
-    if (!up_wait_retry_interval)
+    // client_retry_interval
+    client_retry_interval = config["client_retry_interval"].uint64_value();
+    if (!client_retry_interval)
     {
-        up_wait_retry_interval = 50;
+        client_retry_interval = 50;
     }
-    else if (up_wait_retry_interval < 10)
+    else if (client_retry_interval < 10)
     {
-        up_wait_retry_interval = 10;
+        client_retry_interval = 10;
+    }
+    // client_eio_retry_interval
+    client_eio_retry_interval = 1000;
+    if (!config["client_eio_retry_interval"].is_null())
+    {
+        client_eio_retry_interval = config["client_eio_retry_interval"].uint64_value();
+        if (client_eio_retry_interval && client_eio_retry_interval < 10)
+        {
+            client_eio_retry_interval = 10;
+        }
     }
     // log_level
     log_level = config["log_level"].uint64_value();
@@ -512,7 +550,8 @@ void cluster_client_t::execute(cluster_op_t *op)
         op->opcode != OSD_OP_READ_BITMAP && op->opcode != OSD_OP_READ_CHAIN_BITMAP && op->opcode != OSD_OP_WRITE)
     {
         op->retval = -EINVAL;
-        std::function<void(cluster_op_t*)>(op->callback)(op);
+        auto cb = std::move(op->callback);
+        cb(op);
         return;
     }
     if (!pgs_loaded)
@@ -534,7 +573,7 @@ void cluster_client_t::execute_internal(cluster_op_t *op)
         return;
     }
     if (op->opcode == OSD_OP_WRITE && enable_writeback && !(op->flags & OP_FLUSH_BUFFER) &&
-        !op->version /* FIXME no CAS writeback */)
+        !op->version /* no CAS writeback */)
     {
         if (wb->writebacks_active >= client_max_writeback_iodepth)
         {
@@ -550,12 +589,13 @@ void cluster_client_t::execute_internal(cluster_op_t *op)
             wb->start_writebacks(this, 1);
         }
         op->retval = op->len;
-        std::function<void(cluster_op_t*)>(op->callback)(op);
+        auto cb = std::move(op->callback);
+        cb(op);
         return;
     }
     if (op->opcode == OSD_OP_WRITE && !(op->flags & OP_IMMEDIATE_COMMIT))
     {
-        if (!(op->flags & OP_FLUSH_BUFFER))
+        if (!(op->flags & OP_FLUSH_BUFFER) && !op->version /* no CAS write-repeat */)
         {
             wb->copy_write(op, CACHE_WRITTEN);
         }
@@ -619,7 +659,8 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
     if (!pool_id)
     {
         op->retval = -EINVAL;
-        std::function<void(cluster_op_t*)>(op->callback)(op);
+        auto cb = std::move(op->callback);
+        cb(op);
         return false;
     }
     auto pool_it = st_cli.pool_config.find(pool_id);
@@ -627,15 +668,17 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
     {
         // Pools are loaded, but this one is unknown
         op->retval = -EINVAL;
-        std::function<void(cluster_op_t*)>(op->callback)(op);
+        auto cb = std::move(op->callback);
+        cb(op);
         return false;
     }
     // Check alignment
-    if (!op->len && (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP || op->opcode == OSD_OP_WRITE) ||
+    if (!op->len && (op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP || op->opcode == OSD_OP_WRITE) ||
         op->offset % pool_it->second.bitmap_granularity || op->len % pool_it->second.bitmap_granularity)
     {
         op->retval = -EINVAL;
-        std::function<void(cluster_op_t*)>(op->callback)(op);
+        auto cb = std::move(op->callback);
+        cb(op);
         return false;
     }
     if (pool_it->second.immediate_commit == IMMEDIATE_ALL)
@@ -648,7 +691,8 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
         if (ino_it != st_cli.inode_config.end() && ino_it->second.readonly)
         {
             op->retval = -EROFS;
-            std::function<void(cluster_op_t*)>(op->callback)(op);
+            auto cb = std::move(op->callback);
+            cb(op);
             return false;
         }
     }
@@ -716,15 +760,8 @@ resume_1:
                 // We'll need to retry again
                 if (op->parts[i].flags & PART_RETRY)
                 {
-                    op->up_wait = true;
-                    if (!retry_timeout_id)
-                    {
-                        retry_timeout_id = tfd->set_timer(up_wait_retry_interval, false, [this](int)
-                        {
-                            retry_timeout_id = 0;
-                            continue_ops(true);
-                        });
-                    }
+                    op->retry_after = client_retry_interval;
+                    reset_retry_timer(client_retry_interval);
                 }
                 op->state = 1;
             }
@@ -780,10 +817,9 @@ resume_2:
         return 1;
     }
     else if (op->retval != 0 && !(op->flags & OP_FLUSH_BUFFER) &&
-        op->retval != -EPIPE && op->retval != -EIO && op->retval != -ENOSPC)
+        op->retval != -EPIPE && (op->retval != -EIO || !client_eio_retry_interval) && op->retval != -ENOSPC)
     {
         // Fatal error (neither -EPIPE, -EIO nor -ENOSPC)
-        // FIXME: Add a parameter to allow to not wait for EIOs (incomplete or corrupted objects) to heal
         erase_op(op);
         return 1;
     }
@@ -1138,7 +1174,6 @@ static inline void mem_or(void *res, const void *r2, unsigned int len)
 void cluster_client_t::handle_op_part(cluster_op_part_t *part)
 {
     cluster_op_t *op = part->parent;
-    op->inflight_count--;
     int expected = part->op.req.hdr.opcode == OSD_OP_SYNC ? 0 : part->op.req.rw.len;
     if (part->op.reply.hdr.retval != expected)
     {
@@ -1156,31 +1191,32 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
             if (op->retval != -EPIPE || log_level > 0)
             {
                 fprintf(
-                    stderr, "%s operation failed on OSD %lu: retval=%ld (expected %d), dropping connection\n",
+                    stderr, "%s operation failed on OSD %ju: retval=%jd (expected %d), dropping connection\n",
                     osd_op_names[part->op.req.hdr.opcode], part->osd_num, part->op.reply.hdr.retval, expected
                 );
             }
         }
-        else
+        else if (log_level > 0)
         {
             fprintf(
-                stderr, "%s operation failed on OSD %lu: retval=%ld (expected %d)\n",
+                stderr, "%s operation failed on OSD %ju: retval=%jd (expected %d)\n",
                 osd_op_names[part->op.req.hdr.opcode], part->osd_num, part->op.reply.hdr.retval, expected
             );
         }
         // All next things like timer, continue_sync/rw and stop_client may affect the operation again
         // So do all these things after modifying operation state, otherwise we may hit reenterability bugs
         // FIXME postpone such things to set_immediate here to avoid bugs
-        // Mark op->up_wait = true to retry operation after a short pause (not immediately)
-        op->up_wait = true;
-        if (!retry_timeout_id)
+        // Set op->retry_after to retry operation after a short pause (not immediately)
+        if (!op->retry_after)
         {
-            retry_timeout_id = tfd->set_timer(up_wait_retry_interval, false, [this](int)
-            {
-                retry_timeout_id = 0;
-                continue_ops(true);
-            });
+            op->retry_after = op->retval == -EIO ? client_eio_retry_interval : client_retry_interval;
         }
+        reset_retry_timer(op->retry_after);
+        if (stop_fd >= 0)
+        {
+            msgr.stop_client(stop_fd);
+        }
+        op->inflight_count--;
         if (op->inflight_count == 0)
         {
             if (op->opcode == OSD_OP_SYNC)
@@ -1188,14 +1224,11 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
             else
                 continue_rw(op);
         }
-        if (stop_fd >= 0)
-        {
-            msgr.stop_client(stop_fd);
-        }
     }
     else
     {
         // OK
+        op->inflight_count--;
         if ((op->opcode == OSD_OP_WRITE || op->opcode == OSD_OP_DELETE) && !(op->flags & OP_IMMEDIATE_COMMIT))
             dirty_osds.insert(part->osd_num);
         part->flags |= PART_DONE;

src/cluster_client.h

@@ -59,7 +59,7 @@ protected:
     void *buf = NULL;
     cluster_op_t *orig_op = NULL;
     bool needs_reslice = false;
-    bool up_wait = false;
+    int retry_after = 0;
     int inflight_count = 0, done_count = 0;
     std::vector<cluster_op_part_t> parts;
     void *part_bitmaps = NULL;
@@ -92,9 +92,11 @@ class cluster_client_t
     uint64_t client_max_writeback_iodepth = 0;
 
     int log_level = 0;
-    int up_wait_retry_interval = 500; // ms
+    int client_retry_interval = 50; // ms
+    int client_eio_retry_interval = 1000; // ms
 
     int retry_timeout_id = 0;
+    int retry_timeout_duration = 0;
     std::vector<cluster_op_t*> offline_ops;
     cluster_op_t *op_queue_head = NULL, *op_queue_tail = NULL;
     writeback_cache_t *wb = NULL;
@@ -121,7 +123,7 @@ public:
     json11::Json::object cli_config, file_config, etcd_global_config;
     json11::Json::object config;
 
-    cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json & config);
+    cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd, json11::Json config);
     ~cluster_client_t();
     void execute(cluster_op_t *op);
     void execute_raw(osd_num_t osd_num, osd_op_t *op);
@@ -131,7 +133,7 @@ public:
 
     bool get_immediate_commit(uint64_t inode);
 
-    void continue_ops(bool up_retry = false);
+    void continue_ops(int time_passed = 0);
     inode_list_t *list_inode_start(inode_t inode,
         std::function<void(inode_list_t* lst, std::set<object_id>&& objects, pg_num_t pg_num, osd_num_t primary_osd, int status)> callback);
     int list_pg_count(inode_list_t *lst);
@@ -152,6 +154,7 @@ protected:
     int continue_rw(cluster_op_t *op);
     bool check_rw(cluster_op_t *op);
     void slice_rw(cluster_op_t *op);
+    void reset_retry_timer(int new_duration);
     bool try_send(cluster_op_t *op, int i);
     int continue_sync(cluster_op_t *op);
     void send_sync(cluster_op_t *op, cluster_op_part_t *part);

src/cluster_client_list.cpp

@@ -226,7 +226,7 @@ void cluster_client_t::send_list(inode_list_osd_t *cur_list)
     {
         if (op->reply.hdr.retval < 0)
         {
-            fprintf(stderr, "Failed to get PG %u/%u object list from OSD %lu (retval=%ld), skipping\n",
+            fprintf(stderr, "Failed to get PG %u/%u object list from OSD %ju (retval=%jd), skipping\n",
                 cur_list->pg->lst->pool_id, cur_list->pg->pg_num, cur_list->osd_num, op->reply.hdr.retval);
         }
         else
@@ -236,7 +236,7 @@ void cluster_client_t::send_list(inode_list_osd_t *cur_list)
                 // Unstable objects, if present, mean that someone still writes into the inode. Warn the user about it.
                 cur_list->pg->has_unstable = true;
                 fprintf(
-                    stderr, "[PG %u/%u] Inode still has %lu unstable object versions out of total %lu - is it still open?\n",
+                    stderr, "[PG %u/%u] Inode still has %ju unstable object versions out of total %ju - is it still open?\n",
                     cur_list->pg->lst->pool_id, cur_list->pg->pg_num, op->reply.hdr.retval - op->reply.sec_list.stable_count,
                     op->reply.hdr.retval
                 );
@@ -244,7 +244,7 @@ void cluster_client_t::send_list(inode_list_osd_t *cur_list)
             if (log_level > 0)
             {
                 fprintf(
-                    stderr, "[PG %u/%u] Got inode object list from OSD %lu: %ld object versions\n",
+                    stderr, "[PG %u/%u] Got inode object list from OSD %ju: %jd object versions\n",
                     cur_list->pg->lst->pool_id, cur_list->pg->pg_num, cur_list->osd_num, op->reply.hdr.retval
                 );
             }

src/disk_simple_offsets.cpp

@@ -16,11 +16,6 @@
 void disk_tool_simple_offsets(json11::Json cfg, bool json_output)
 {
     std::string device = cfg["device"].string_value();
-    if (device == "")
-    {
-        fprintf(stderr, "Device path is missing\n");
-        exit(1);
-    }
     uint64_t data_block_size = parse_size(cfg["object_size"].string_value());
     uint64_t bitmap_granularity = parse_size(cfg["bitmap_granularity"].string_value());
     uint64_t journal_size = parse_size(cfg["journal_size"].string_value());
@@ -47,7 +42,7 @@ void disk_tool_simple_offsets(json11::Json cfg, bool json_output)
     if (!bitmap_granularity)
         bitmap_granularity = DEFAULT_BITMAP_GRANULARITY;
     if (!journal_size)
-        journal_size = 16*1024*1024;
+        journal_size = 32*1024*1024;
     if (!device_block_size)
         device_block_size = 4096;
     if (!data_csum_type)
@@ -57,6 +52,11 @@ void disk_tool_simple_offsets(json11::Json cfg, bool json_output)
     uint64_t orig_device_size = device_size;
     if (!device_size)
     {
+        if (device == "")
+        {
+            fprintf(stderr, "Device path is missing\n");
+            exit(1);
+        }
         struct stat st;
         if (stat(device.c_str(), &st) < 0)
         {
@@ -75,9 +75,9 @@ void disk_tool_simple_offsets(json11::Json cfg, bool json_output)
             if (st.st_blksize < device_block_size)
             {
                 fprintf(
-                    stderr, "Warning: %s reports %lu byte blocks, but we use %lu."
-                    " Set --device_block_size=%lu if you're sure it works well with %lu byte blocks.\n",
-                    device.c_str(), st.st_blksize, device_block_size, st.st_blksize, st.st_blksize
+                    stderr, "Warning: %s reports %ju byte blocks, but we use %ju."
+                    " Set --device_block_size=%ju if you're sure it works well with %ju byte blocks.\n",
+                    device.c_str(), (uint64_t)st.st_blksize, device_block_size, (uint64_t)st.st_blksize, (uint64_t)st.st_blksize
                 );
             }
         }
@@ -99,19 +99,19 @@ void disk_tool_simple_offsets(json11::Json cfg, bool json_output)
     if (device_block_size < 512 || device_block_size > 1048576 ||
         device_block_size & (device_block_size-1) != 0)
     {
-        fprintf(stderr, "Invalid device block size specified: %lu\n", device_block_size);
+        fprintf(stderr, "Invalid device block size specified: %ju\n", device_block_size);
         exit(1);
     }
     if (data_block_size < device_block_size || data_block_size > MAX_DATA_BLOCK_SIZE ||
         data_block_size & (data_block_size-1) != 0)
     {
-        fprintf(stderr, "Invalid object size specified: %lu\n", data_block_size);
+        fprintf(stderr, "Invalid object size specified: %ju\n", data_block_size);
         exit(1);
     }
     if (bitmap_granularity < device_block_size || bitmap_granularity > data_block_size ||
         bitmap_granularity & (bitmap_granularity-1) != 0)
     {
-        fprintf(stderr, "Invalid bitmap granularity specified: %lu\n", bitmap_granularity);
+        fprintf(stderr, "Invalid bitmap granularity specified: %ju\n", bitmap_granularity);
         exit(1);
     }
     if (csum_block_size && (data_block_size % csum_block_size))
@@ -145,8 +145,8 @@ void disk_tool_simple_offsets(json11::Json cfg, bool json_output)
     {
         // Env
         printf(
-            "meta_block_size=%lu\njournal_block_size=%lu\ndata_size=%lu\n"
-            "data_device=%s\njournal_offset=%lu\nmeta_offset=%lu\ndata_offset=%lu\n",
+            "meta_block_size=%ju\njournal_block_size=%ju\ndata_size=%ju\n"
+            "data_device=%s\njournal_offset=%ju\nmeta_offset=%ju\ndata_offset=%ju\n",
             device_block_size, device_block_size, device_size-data_offset,
             device.c_str(), journal_offset, meta_offset, data_offset
         );
@@ -160,14 +160,14 @@ void disk_tool_simple_offsets(json11::Json cfg, bool json_output)
         }
         if (device_block_size != 4096)
         {
-            printf("--meta_block_size %lu\n--journal_block_size %lu\n", device_block_size, device_block_size);
+            printf("--meta_block_size %ju\n--journal_block_size %ju\n", device_block_size, device_block_size);
         }
         if (orig_device_size)
         {
-            printf("--data_size %lu\n", device_size-data_offset);
+            printf("--data_size %ju\n", device_size-data_offset);
         }
         printf(
-            "--data_device %s\n--journal_offset %lu\n--meta_offset %lu\n--data_offset %lu\n",
+            "--data_device %s\n--journal_offset %ju\n--meta_offset %ju\n--data_offset %ju\n",
             device.c_str(), journal_offset, meta_offset, data_offset
         );
     }

src/disk_tool.cpp

@@ -167,7 +167,7 @@ static const char *help_text =
     "  Calculate offsets for old simple&stupid (no superblock) OSD deployment. Options:\n"
     "    --object_size 128k       Set blockstore block size\n"
     "    --bitmap_granularity 4k  Set bitmap granularity\n"
-    "    --journal_size 16M       Set journal size\n"
+    "    --journal_size 32M       Set journal size\n"
     "    --data_csum_type none    Set data checksum type (crc32c or none)\n"
     "    --csum_block_size 4k     Set data checksum block size\n"
     "    --device_block_size 4k   Set device block size\n"

src/disk_tool.h

@@ -4,7 +4,7 @@
 #pragma once
 
 #ifndef _LARGEFILE64_SOURCE
-#define _LARGEFILE64_SOURCE 1
+#define _LARGEFILE64_SOURCE
 #endif
 
 #include <map>
@@ -132,9 +132,6 @@ void disk_tool_simple_offsets(json11::Json cfg, bool json_output);
 
 uint64_t sscanf_json(const char *fmt, const json11::Json & str);
 void fromhexstr(const std::string & from, int bytes, uint8_t *to);
-std::string realpath_str(std::string path, bool nofail = true);
-std::string read_all_fd(int fd);
-std::string read_file(std::string file, bool allow_enoent = false);
 int disable_cache(std::string dev);
 std::string get_parent_device(std::string dev);
 bool json_is_true(const json11::Json & val);

src/disk_tool_journal.cpp

@@ -38,7 +38,7 @@ int disk_tool_t::dump_journal()
             }
             if (json)
             {
-                printf("%s{\"offset\":\"0x%lx\"", first_block ? "" : ",\n", journal_pos);
+                printf("%s{\"offset\":\"0x%jx\"", first_block ? "" : ",\n", journal_pos);
                 first_block = false;
             }
             if (s == dsk.journal_block_size)
@@ -46,13 +46,13 @@ int disk_tool_t::dump_journal()
                 if (json)
                     printf(",\"type\":\"zero\"}");
                 else
-                    printf("offset %08lx: zeroes\n", journal_pos);
+                    printf("offset %08jx: zeroes\n", journal_pos);
                 journal_pos += dsk.journal_block_size;
             }
             else if (((journal_entry*)journal_buf)->magic == JOURNAL_MAGIC)
             {
                 if (!json)
-                    printf("offset %08lx:\n", journal_pos);
+                    printf("offset %08jx:\n", journal_pos);
                 else
                     printf(",\"entries\":[\n");
                 if (journal_pos == 0)
@@ -80,9 +80,9 @@ int disk_tool_t::dump_journal()
             else
             {
                 if (json)
-                    printf(",\"type\":\"data\",\"pattern\":\"%08lx\"}", *((uint64_t*)journal_buf));
+                    printf(",\"type\":\"data\",\"pattern\":\"%08jx\"}", *((uint64_t*)journal_buf));
                 else
-                    printf("offset %08lx: no magic in the beginning, looks like random data (pattern=%08lx)\n", journal_pos, *((uint64_t*)journal_buf));
+                    printf("offset %08jx: no magic in the beginning, looks like random data (pattern=%08jx)\n", journal_pos, *((uint64_t*)journal_buf));
                 journal_pos += dsk.journal_block_size;
             }
         }
@@ -98,12 +98,12 @@ int disk_tool_t::dump_journal()
             if (json && dump_with_blocks)
                 first_entry = true;
             if (!json)
-                printf("offset %08lx:\n", journal_pos);
+                printf("offset %08jx:\n", journal_pos);
             auto pos = journal_pos;
             int r = process_journal_block(data, [this, pos](int num, journal_entry *je)
             {
                 if (json && dump_with_blocks && first_entry)
-                    printf("%s{\"offset\":\"0x%lx\",\"entries\":[\n", first_block ? "" : ",\n", pos);
+                    printf("%s{\"offset\":\"0x%jx\",\"entries\":[\n", first_block ? "" : ",\n", pos);
                 dump_journal_entry(num, je, json);
                 first_block = false;
             });
@@ -134,12 +134,12 @@ int disk_tool_t::process_journal(std::function<int(void*)> block_fn)
     journal_entry *je = (journal_entry*)(data);
     if (je->magic != JOURNAL_MAGIC || je->type != JE_START || je_crc32(je) != je->crc32)
     {
-        fprintf(stderr, "offset %08lx: journal superblock is invalid\n", journal_pos);
+        fprintf(stderr, "offset %08jx: journal superblock is invalid\n", journal_pos);
         r = 1;
     }
     else if (je->start.size != JE_START_V0_SIZE && je->start.version != JOURNAL_VERSION_V1 && je->start.version != JOURNAL_VERSION_V2)
     {
-        fprintf(stderr, "offset %08lx: journal superblock contains version %lu, but I only understand 0, 1 and 2\n",
+        fprintf(stderr, "offset %08jx: journal superblock contains version %ju, but I only understand 0, 1 and 2\n",
             journal_pos, je->start.size == JE_START_V0_SIZE ? 0 : je->start.version);
         r = 1;
     }
@@ -296,7 +296,7 @@ void disk_tool_t::dump_journal_entry(int num, journal_entry *je, bool json)
     if (je->type == JE_START)
     {
         printf(
-            json ? ",\"type\":\"start\",\"start\":\"0x%lx\"" : "je_start start=%08lx",
+            json ? ",\"type\":\"start\",\"start\":\"0x%jx\"" : "je_start start=%08jx",
             je->start.journal_start
         );
         if (je->start.data_csum_type)
@@ -312,15 +312,15 @@ void disk_tool_t::dump_journal_entry(int num, journal_entry *je, bool json)
     {
         auto & sw = je->small_write;
         printf(
-            json ? ",\"type\":\"small_write%s\",\"inode\":\"0x%lx\",\"stripe\":\"0x%lx\",\"ver\":\"%lu\",\"offset\":%u,\"len\":%u,\"loc\":\"0x%lx\""
-                : "je_small_write%s oid=%lx:%lx ver=%lu offset=%u len=%u loc=%08lx",
+            json ? ",\"type\":\"small_write%s\",\"inode\":\"0x%jx\",\"stripe\":\"0x%jx\",\"ver\":\"%ju\",\"offset\":%u,\"len\":%u,\"loc\":\"0x%jx\""
+                : "je_small_write%s oid=%jx:%jx ver=%ju offset=%u len=%u loc=%08jx",
             je->type == JE_SMALL_WRITE_INSTANT ? "_instant" : "",
             sw.oid.inode, sw.oid.stripe, sw.version, sw.offset, sw.len, sw.data_offset
         );
         if (journal_calc_data_pos != sw.data_offset)
         {
-            printf(json ? ",\"bad_loc\":true,\"calc_loc\":\"0x%lx\""
-                : " (mismatched, calculated = %08lx)", journal_pos);
+            printf(json ? ",\"bad_loc\":true,\"calc_loc\":\"0x%jx\""
+                : " (mismatched, calculated = %08jx)", journal_pos);
         }
         uint32_t data_csum_size = (!je_start.csum_block_size
             ? 0
@@ -367,8 +367,8 @@ void disk_tool_t::dump_journal_entry(int num, journal_entry *je, bool json)
     {
         auto & bw = je->big_write;
         printf(
-            json ? ",\"type\":\"big_write%s\",\"inode\":\"0x%lx\",\"stripe\":\"0x%lx\",\"ver\":\"%lu\",\"offset\":%u,\"len\":%u,\"loc\":\"0x%lx\""
-                : "je_big_write%s oid=%lx:%lx ver=%lu offset=%u len=%u loc=%08lx",
+            json ? ",\"type\":\"big_write%s\",\"inode\":\"0x%jx\",\"stripe\":\"0x%jx\",\"ver\":\"%ju\",\"offset\":%u,\"len\":%u,\"loc\":\"0x%jx\""
+                : "je_big_write%s oid=%jx:%jx ver=%ju offset=%u len=%u loc=%08jx",
             je->type == JE_BIG_WRITE_INSTANT ? "_instant" : "",
             bw.oid.inode, bw.oid.stripe, bw.version, bw.offset, bw.len, bw.location
         );
@@ -398,24 +398,24 @@ void disk_tool_t::dump_journal_entry(int num, journal_entry *je, bool json)
     else if (je->type == JE_STABLE)
     {
         printf(
-            json ? ",\"type\":\"stable\",\"inode\":\"0x%lx\",\"stripe\":\"0x%lx\",\"ver\":\"%lu\"}"
-                : "je_stable oid=%lx:%lx ver=%lu\n",
+            json ? ",\"type\":\"stable\",\"inode\":\"0x%jx\",\"stripe\":\"0x%jx\",\"ver\":\"%ju\"}"
+                : "je_stable oid=%jx:%jx ver=%ju\n",
             je->stable.oid.inode, je->stable.oid.stripe, je->stable.version
         );
     }
     else if (je->type == JE_ROLLBACK)
     {
         printf(
-            json ? ",\"type\":\"rollback\",\"inode\":\"0x%lx\",\"stripe\":\"0x%lx\",\"ver\":\"%lu\"}"
-                : "je_rollback oid=%lx:%lx ver=%lu\n",
+            json ? ",\"type\":\"rollback\",\"inode\":\"0x%jx\",\"stripe\":\"0x%jx\",\"ver\":\"%ju\"}"
+                : "je_rollback oid=%jx:%jx ver=%ju\n",
             je->rollback.oid.inode, je->rollback.oid.stripe, je->rollback.version
         );
     }
     else if (je->type == JE_DELETE)
     {
         printf(
-            json ? ",\"type\":\"delete\",\"inode\":\"0x%lx\",\"stripe\":\"0x%lx\",\"ver\":\"%lu\"}"
-                : "je_delete oid=%lx:%lx ver=%lu\n",
+            json ? ",\"type\":\"delete\",\"inode\":\"0x%jx\",\"stripe\":\"0x%jx\",\"ver\":\"%ju\"}"
+                : "je_delete oid=%jx:%jx ver=%ju\n",
             je->del.oid.inode, je->del.oid.stripe, je->del.version
         );
     }

src/disk_tool_meta.cpp

@@ -54,7 +54,7 @@ int disk_tool_t::process_meta(std::function<void(blockstore_meta_header_v2_t *)>
         else
         {
             // Unsupported version
-            fprintf(stderr, "Metadata format is too new for me (stored version is %lu, max supported %u).\n", hdr->version, BLOCKSTORE_META_FORMAT_V2);
+            fprintf(stderr, "Metadata format is too new for me (stored version is %ju, max supported %u).\n", hdr->version, BLOCKSTORE_META_FORMAT_V2);
             free(data);
             close(dsk.meta_fd);
             dsk.meta_fd = -1;
@@ -108,7 +108,7 @@ int disk_tool_t::process_meta(std::function<void(blockstore_meta_header_v2_t *)>
                             uint32_t *entry_csum = (uint32_t*)((uint8_t*)entry + dsk.clean_entry_size - 4);
                             if (*entry_csum != crc32c(0, entry, dsk.clean_entry_size - 4))
                             {
-                                fprintf(stderr, "Metadata entry %lu is corrupt (checksum mismatch), skipping\n", block_num);
+                                fprintf(stderr, "Metadata entry %ju is corrupt (checksum mismatch), skipping\n", block_num);
                                 continue;
                             }
                         }
@@ -184,7 +184,7 @@ void disk_tool_t::dump_meta_header(blockstore_meta_header_v2_t *hdr)
     }
     else
     {
-        printf("{\"version\":\"0.5\",\"meta_block_size\":%lu,\"entries\":[\n", dsk.meta_block_size);
+        printf("{\"version\":\"0.5\",\"meta_block_size\":%ju,\"entries\":[\n", dsk.meta_block_size);
     }
     first_entry = true;
 }
@@ -192,7 +192,7 @@ void disk_tool_t::dump_meta_header(blockstore_meta_header_v2_t *hdr)
 void disk_tool_t::dump_meta_entry(uint64_t block_num, clean_disk_entry *entry, uint8_t *bitmap)
 {
     printf(
-#define ENTRY_FMT "{\"block\":%lu,\"pool\":%u,\"inode\":\"0x%lx\",\"stripe\":\"0x%lx\",\"version\":%lu"
+#define ENTRY_FMT "{\"block\":%ju,\"pool\":%u,\"inode\":\"0x%jx\",\"stripe\":\"0x%jx\",\"version\":%ju"
         (first_entry ? ENTRY_FMT : (",\n" ENTRY_FMT)),
 #undef ENTRY_FMT
         block_num, INODE_POOL(entry->oid.inode), INODE_NO_POOL(entry->oid.inode),
@@ -265,7 +265,7 @@ int disk_tool_t::write_json_meta(json11::Json meta)
         {
             free(new_meta_buf);
             new_meta_buf = NULL;
-            fprintf(stderr, "Metadata (data block %lu) doesn't fit into the new area\n", data_block);
+            fprintf(stderr, "Metadata (data block %ju) doesn't fit into the new area\n", data_block);
             return 1;
         }
         clean_disk_entry *new_entry = (clean_disk_entry*)(new_meta_buf +

src/disk_tool_prepare.cpp

@@ -210,10 +210,10 @@ int disk_tool_t::prepare_one(std::map<std::string, std::string> options, int is_
         desc += " with metadata on "+realpath_str(options["meta_device"]);
     if (sep_j)
         desc += (sep_m ? " and journal on " : " with journal on ") + realpath_str(options["journal_device"]);
-    fprintf(stderr, "Initialized OSD %lu on %s\n", osd_num, desc.c_str());
+    fprintf(stderr, "Initialized OSD %ju on %s\n", osd_num, desc.c_str());
     if (shell_exec({ "systemctl", "enable", "--now", "vitastor-osd@"+std::to_string(osd_num) }, "", NULL, NULL) != 0)
     {
-        fprintf(stderr, "Failed to enable systemd unit vitastor-osd@%lu\n", osd_num);
+        fprintf(stderr, "Failed to enable systemd unit vitastor-osd@%ju\n", osd_num);
         return 1;
     }
     return 0;
@@ -337,7 +337,7 @@ json11::Json disk_tool_t::add_partitions(vitastor_dev_info_t & devinfo, std::vec
     std::string out;
     if (shell_exec({ "sfdisk", "--no-reread", "--force", devinfo.path }, script, &out, NULL) != 0)
     {
-        fprintf(stderr, "Failed to add %lu partition(s) with sfdisk\n", sizes.size());
+        fprintf(stderr, "Failed to add %zu partition(s) with sfdisk\n", sizes.size());
         return {};
     }
     // Get new partition table and find created partitions
@@ -352,7 +352,7 @@ json11::Json disk_tool_t::add_partitions(vitastor_dev_info_t & devinfo, std::vec
     }
     if (new_parts.size() != sizes.size())
     {
-        fprintf(stderr, "Failed to add %lu partition(s) with sfdisk: new partitions not found in table\n", sizes.size());
+        fprintf(stderr, "Failed to add %zu partition(s) with sfdisk: new partitions not found in table\n", sizes.size());
         return {};
     }
     // Check if new nodes exist and run partprobe if not
@@ -456,7 +456,7 @@ std::vector<std::string> disk_tool_t::get_new_data_parts(vitastor_dev_info_t & d
                     bool is_journal = sb["params"]["journal_device"].string_value() == part_path;
                     bool is_data = sb["params"]["data_device"].string_value() == part_path;
                     fprintf(
-                        stderr, "%s is already initialized for OSD %lu%s, skipping\n",
+                        stderr, "%s is already initialized for OSD %ju%s, skipping\n",
                         part["node"].string_value().c_str(), sb["params"]["osd_num"].uint64_value(),
                         (is_data ? " data" : (is_meta ? " meta" : (is_journal ? " journal" : "")))
                     );
@@ -539,7 +539,7 @@ int disk_tool_t::get_meta_partition(std::vector<vitastor_dev_info_t> & ssds, std
     if (sel < 0)
     {
         fprintf(
-            stderr, "Could not find free space for new SSD journal and metadata (need %lu + %lu MiB)\n",
+            stderr, "Could not find free space for new SSD journal and metadata (need %ju + %ju MiB)\n",
             meta_size/1024/1024, journal_size/1024/1024
         );
         return 1;

src/disk_tool_resize.cpp

@@ -184,7 +184,7 @@ void disk_tool_t::resize_init(blockstore_meta_header_v2_t *hdr)
     }
     if (new_meta_len < dsk.meta_block_size*new_meta_blocks)
     {
-        fprintf(stderr, "New metadata area size is too small, should be at least %lu bytes\n", dsk.meta_block_size*new_meta_blocks);
+        fprintf(stderr, "New metadata area size is too small, should be at least %ju bytes\n", dsk.meta_block_size*new_meta_blocks);
         exit(1);
     }
     // Check that new metadata, journal and data areas don't overlap
@@ -289,7 +289,7 @@ int disk_tool_t::resize_copy_data()
                         if (data->res != dsk.data_block_size)
                         {
                             fprintf(
-                                stderr, "Failed to read %u bytes at %lu from %s: %s\n", dsk.data_block_size,
+                                stderr, "Failed to read %u bytes at %ju from %s: %s\n", dsk.data_block_size,
                                 dsk.data_offset + moving_blocks[i].old_loc*dsk.data_block_size, dsk.data_device.c_str(),
                                 data->res < 0 ? strerror(-data->res) : "short read"
                             );
@@ -314,7 +314,7 @@ int disk_tool_t::resize_copy_data()
                         if (data->res != dsk.data_block_size)
                         {
                             fprintf(
-                                stderr, "Failed to write %u bytes at %lu to %s: %s\n", dsk.data_block_size,
+                                stderr, "Failed to write %u bytes at %ju to %s: %s\n", dsk.data_block_size,
                                 dsk.data_offset + moving_blocks[i].new_loc*dsk.data_block_size, dsk.data_device.c_str(),
                                 data->res < 0 ? strerror(-data->res) : "short write"
                             );

src/disk_tool_udev.cpp

@@ -43,8 +43,8 @@ int disk_tool_t::udev_import(std::string device)
     }
     uint64_t osd_num = sb["params"]["osd_num"].uint64_value();
     // Print variables for udev
-    printf("VITASTOR_OSD_NUM=%lu\n", osd_num);
-    printf("VITASTOR_ALIAS=osd%lu-%s\n", osd_num, sb["device_type"].string_value().c_str());
+    printf("VITASTOR_OSD_NUM=%ju\n", osd_num);
+    printf("VITASTOR_ALIAS=osd%ju-%s\n", osd_num, sb["device_type"].string_value().c_str());
     printf("VITASTOR_DATA_DEVICE=%s\n", udev_escape(sb["params"]["data_device"].string_value()).c_str());
     if (sb["real_meta_device"].string_value() != "" && sb["real_meta_device"] != sb["real_data_device"])
         printf("VITASTOR_META_DEVICE=%s\n", udev_escape(sb["params"]["meta_device"].string_value()).c_str());
@@ -466,12 +466,12 @@ int disk_tool_t::purge_devices(const std::vector<std::string> & devices)
                     close(fd);
                 if (r != 0)
                 {
-                    fprintf(stderr, "Failed to clear OSD %lu %s device %s superblock: %s\n",
+                    fprintf(stderr, "Failed to clear OSD %ju %s device %s superblock: %s\n",
                         sb["params"]["osd_num"].uint64_value(), dev_type.c_str(), dev.c_str(), strerror(errno));
                 }
                 else
                 {
-                    fprintf(stderr, "OSD %lu %s device %s superblock cleared\n",
+                    fprintf(stderr, "OSD %ju %s device %s superblock cleared\n",
                         sb["params"]["osd_num"].uint64_value(), dev_type.c_str(), dev.c_str());
                 }
                 if (sb["params"][dev_type+"_device"].string_value().substr(0, 22) == "/dev/disk/by-partuuid/")

src/disk_tool_utils.cpp

@@ -12,9 +12,9 @@ uint64_t sscanf_json(const char *fmt, const json11::Json & str)
 {
     uint64_t value = 0;
     if (fmt)
-        sscanf(str.string_value().c_str(), "%lx", &value);
+        sscanf(str.string_value().c_str(), "%jx", &value);
     else if (str.string_value().size() > 2 && (str.string_value()[0] == '0' && str.string_value()[1] == 'x'))
-        sscanf(str.string_value().c_str(), "0x%lx", &value);
+        sscanf(str.string_value().c_str(), "0x%jx", &value);
     else
         value = str.uint64_value();
     return value;
@@ -42,36 +42,6 @@ void fromhexstr(const std::string & from, int bytes, uint8_t *to)
     }
 }
 
-std::string realpath_str(std::string path, bool nofail)
-{
-    char *p = realpath((char*)path.c_str(), NULL);
-    if (!p)
-    {
-        fprintf(stderr, "Failed to resolve %s: %s\n", path.c_str(), strerror(errno));
-        return nofail ? path : "";
-    }
-    std::string rp(p);
-    free(p);
-    return rp;
-}
-
-std::string read_file(std::string file, bool allow_enoent)
-{
-    std::string res;
-    int fd = open(file.c_str(), O_RDONLY);
-    if (fd < 0 || (res = read_all_fd(fd)) == "")
-    {
-        int err = errno;
-        if (fd >= 0)
-            close(fd);
-        if (!allow_enoent || err != ENOENT)
-            fprintf(stderr, "Can't read %s: %s\n", file.c_str(), strerror(err));
-        return "";
-    }
-    close(fd);
-    return res;
-}
-
 // returns 1 = check error, 0 = write through, -1 = write back
 // (similar to 1 = warning, -1 = error, 0 = success in disable_cache)
 static int check_queue_cache(std::string dev, std::string parent_dev)

src/epoll_manager.cpp

@@ -101,7 +101,7 @@ void epoll_manager_t::handle_uring_event()
     my_uring_prep_poll_add(sqe, epoll_fd, POLLIN);
     data->callback = [this](ring_data_t *data)
     {
-        if (data->res < 0)
+        if (data->res < 0 && data->res != -ECANCELED)
         {
             throw std::runtime_error(std::string("epoll failed: ") + strerror(-data->res));
         }

src/etcd_state_client.cpp

@@ -333,7 +333,7 @@ void etcd_state_client_t::start_etcd_watcher()
         etcd_watch_ws = NULL;
     }
     if (this->log_level > 1)
-        fprintf(stderr, "Trying to connect to etcd websocket at %s, watch from revision %lu\n", etcd_address.c_str(), etcd_watch_revision);
+        fprintf(stderr, "Trying to connect to etcd websocket at %s, watch from revision %ju\n", etcd_address.c_str(), etcd_watch_revision);
     etcd_watch_ws = open_websocket(tfd, etcd_address, etcd_api_path+"/watch", etcd_slow_timeout,
         [this, cur_addr = selected_etcd_address](const http_response_t *msg)
     {
@@ -357,7 +357,7 @@ void etcd_state_client_t::start_etcd_watcher()
                         watch_id == ETCD_OSD_STATE_WATCH_ID)
                         etcd_watches_initialised++;
                     if (etcd_watches_initialised == ETCD_TOTAL_WATCHES && this->log_level > 0)
-                        fprintf(stderr, "Successfully subscribed to etcd at %s, revision %lu\n", cur_addr.c_str(), etcd_watch_revision);
+                        fprintf(stderr, "Successfully subscribed to etcd at %s, revision %ju\n", cur_addr.c_str(), etcd_watch_revision);
                 }
                 if (data["result"]["canceled"].bool_value())
                 {
@@ -371,7 +371,7 @@ void etcd_state_client_t::start_etcd_watcher()
                             // check to not trigger on_reload_hook multiple times
                             if (etcd_watch_ws != NULL)
                             {
-                                fprintf(stderr, "Revisions before %lu were compacted by etcd, reloading state\n",
+                                fprintf(stderr, "Revisions before %ju were compacted by etcd, reloading state\n",
                                     data["result"]["compact_revision"].uint64_value());
                                 http_close(etcd_watch_ws);
                                 etcd_watch_ws = NULL;
@@ -382,7 +382,7 @@ void etcd_state_client_t::start_etcd_watcher()
                         }
                         else
                         {
-                            fprintf(stderr, "Revisions before %lu were compacted by etcd, exiting\n",
+                            fprintf(stderr, "Revisions before %ju were compacted by etcd, exiting\n",
                                 data["result"]["compact_revision"].uint64_value());
                             exit(1);
                         }
@@ -573,8 +573,7 @@ void etcd_state_client_t::load_global_config()
         {
             global_bitmap_granularity = DEFAULT_BITMAP_GRANULARITY;
         }
-        global_immediate_commit = global_config["immediate_commit"].string_value() == "all"
-            ? IMMEDIATE_ALL : (global_config["immediate_commit"].string_value() == "small" ? IMMEDIATE_SMALL : IMMEDIATE_NONE);
+        global_immediate_commit = parse_immediate_commit(global_config["immediate_commit"].string_value());
         on_load_config_hook(global_config);
     });
 }
@@ -646,7 +645,7 @@ void etcd_state_client_t::load_pgs()
             etcd_watch_revision = data["header"]["revision"].uint64_value()+1;
             if (this->log_level > 3)
             {
-                fprintf(stderr, "Loaded revision %lu of PG configuration\n", etcd_watch_revision-1);
+                fprintf(stderr, "Loaded revision %ju of PG configuration\n", etcd_watch_revision-1);
             }
         }
         for (auto & res: data["responses"].array_items())
@@ -740,7 +739,7 @@ void etcd_state_client_t::clean_nonexistent_pgs()
     {
         if (seen_peers.find(peer_item.first) == seen_peers.end())
         {
-            fprintf(stderr, "OSD %lu state disappeared after reload, forgetting it\n", peer_item.first);
+            fprintf(stderr, "OSD %ju state disappeared after reload, forgetting it\n", peer_item.first);
             parse_state((etcd_kv_t){
                 .key = etcd_prefix+"/osd/state/"+std::to_string(peer_item.first),
             });
@@ -782,13 +781,8 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
             // Failure Domain
             pc.failure_domain = pool_item.second["failure_domain"].string_value();
             // Coding Scheme
-            if (pool_item.second["scheme"] == "replicated")
-                pc.scheme = POOL_SCHEME_REPLICATED;
-            else if (pool_item.second["scheme"] == "xor")
-                pc.scheme = POOL_SCHEME_XOR;
-            else if (pool_item.second["scheme"] == "ec" || pool_item.second["scheme"] == "jerasure")
-                pc.scheme = POOL_SCHEME_EC;
-            else
+            pc.scheme = parse_scheme(pool_item.second["scheme"].string_value());
+            if (!pc.scheme)
             {
                 fprintf(stderr, "Pool %u has invalid coding scheme (one of \"xor\", \"replicated\", \"ec\" or \"jerasure\" required), skipping pool\n", pool_id);
                 continue;
@@ -869,11 +863,11 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
             pc.scrub_interval = parse_time(pool_item.second["scrub_interval"].string_value());
             if (!pc.scrub_interval)
                 pc.scrub_interval = 0;
+            // Mark pool as VitastorFS pool (disable per-inode stats and block volume creation)
+            pc.used_for_fs = pool_item.second["used_for_fs"].as_string();
             // Immediate Commit Mode
             pc.immediate_commit = pool_item.second["immediate_commit"].is_string()
-                ? (pool_item.second["immediate_commit"].string_value() == "all"
-                    ? IMMEDIATE_ALL : (pool_item.second["immediate_commit"].string_value() == "small"
-                        ? IMMEDIATE_SMALL : IMMEDIATE_NONE))
+                ? parse_immediate_commit(pool_item.second["immediate_commit"].string_value())
                 : global_immediate_commit;
             // PG Stripe Size
             pc.pg_stripe_size = pool_item.second["pg_stripe_size"].uint64_value();
@@ -890,7 +884,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
             {
                 if (pg_item.second.target_set.size() != parsed_cfg.pg_size)
                 {
-                    fprintf(stderr, "Pool %u PG %u configuration is invalid: osd_set size %lu != pool pg_size %lu\n",
+                    fprintf(stderr, "Pool %u PG %u configuration is invalid: osd_set size %zu != pool pg_size %ju\n",
                         pool_id, pg_item.first, pg_item.second.target_set.size(), parsed_cfg.pg_size);
                     pg_item.second.pause = true;
                 }
@@ -936,7 +930,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
                 }
                 if (parsed_cfg.target_set.size() != pool_config[pool_id].pg_size)
                 {
-                    fprintf(stderr, "Pool %u PG %u configuration is invalid: osd_set size %lu != pool pg_size %lu\n",
+                    fprintf(stderr, "Pool %u PG %u configuration is invalid: osd_set size %zu != pool pg_size %ju\n",
                         pool_id, pg_num, parsed_cfg.target_set.size(), pool_config[pool_id].pg_size);
                     parsed_cfg.pause = true;
                 }
@@ -950,7 +944,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
                 if (pg_it->second.config_exists && pg_it->first != ++n)
                 {
                     fprintf(
-                        stderr, "Invalid pool %u PG configuration: PG numbers don't cover whole 1..%lu range\n",
+                        stderr, "Invalid pool %u PG configuration: PG numbers don't cover whole 1..%zu range\n",
                         pool_item.second.id, pool_item.second.pg_config.size()
                     );
                     for (pg_it = pool_item.second.pg_config.begin(); pg_it != pool_item.second.pg_config.end(); pg_it++)
@@ -1066,7 +1060,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
                 (state & PG_PEERING) && state != PG_PEERING ||
                 (state & PG_INCOMPLETE) && state != PG_INCOMPLETE)
             {
-                fprintf(stderr, "Unexpected pool %u PG %u state in etcd: primary=%lu, state=%s\n", pool_id, pg_num, cur_primary, value["state"].dump().c_str());
+                fprintf(stderr, "Unexpected pool %u PG %u state in etcd: primary=%ju, state=%s\n", pool_id, pg_num, cur_primary, value["state"].dump().c_str());
                 return;
             }
             pg_cfg.cur_primary = cur_primary;
@@ -1102,7 +1096,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
         uint64_t pool_id = 0;
         uint64_t inode_num = 0;
         char null_byte = 0;
-        int scanned = sscanf(key.c_str() + etcd_prefix.length()+14, "%lu/%lu%c", &pool_id, &inode_num, &null_byte);
+        int scanned = sscanf(key.c_str() + etcd_prefix.length()+14, "%ju/%ju%c", &pool_id, &inode_num, &null_byte);
         if (scanned != 2 || !pool_id || pool_id >= POOL_ID_MAX || !inode_num || (inode_num >> (64-POOL_ID_BITS)))
         {
             fprintf(stderr, "Bad etcd key %s, ignoring\n", key.c_str());
@@ -1145,7 +1139,7 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
                     else if (parent_pool_id >= POOL_ID_MAX)
                     {
                         fprintf(
-                            stderr, "Inode %lu/%lu parent_pool value is invalid, ignoring parent setting\n",
+                            stderr, "Inode %ju/%ju parent_pool value is invalid, ignoring parent setting\n",
                             inode_num >> (64-POOL_ID_BITS), inode_num & (((uint64_t)1 << (64-POOL_ID_BITS)) - 1)
                         );
                         parent_inode_num = 0;
@@ -1167,6 +1161,23 @@ void etcd_state_client_t::parse_state(const etcd_kv_t & kv)
     }
 }
 
+uint32_t etcd_state_client_t::parse_immediate_commit(const std::string & immediate_commit_str)
+{
+    return immediate_commit_str == "all" ? IMMEDIATE_ALL :
+        (immediate_commit_str == "small" ? IMMEDIATE_SMALL : IMMEDIATE_NONE);
+}
+
+uint32_t etcd_state_client_t::parse_scheme(const std::string & scheme)
+{
+    if (scheme == "replicated")
+        return POOL_SCHEME_REPLICATED;
+    else if (scheme == "xor")
+        return POOL_SCHEME_XOR;
+    else if (scheme == "ec" || scheme == "jerasure")
+        return POOL_SCHEME_EC;
+    return 0;
+}
+
 void etcd_state_client_t::insert_inode_config(const inode_config_t & cfg)
 {
     this->inode_config[cfg.num] = cfg;

src/etcd_state_client.h

@@ -60,6 +60,7 @@ struct pool_config_t
     uint64_t pg_stripe_size;
     std::map<pg_num_t, pg_config_t> pg_config;
     uint64_t scrub_interval;
+    std::string used_for_fs;
 };
 
 struct inode_config_t
@@ -151,4 +152,7 @@ public:
     void close_watch(inode_watch_t* watch);
     int address_count();
     ~etcd_state_client_t();
+
+    static uint32_t parse_immediate_commit(const std::string & immediate_commit_str);
+    static uint32_t parse_scheme(const std::string & scheme_str);
 };

src/fio_cluster.cpp

@@ -377,7 +377,7 @@ static void io_callback(void *opaque, long retval)
     bsd->completed.push_back(io);
     if (bsd->trace)
     {
-        printf("--- %s 0x%lx retval=%ld\n", io->ddir == DDIR_READ ? "READ" :
+        printf("--- %s 0x%jx retval=%ld\n", io->ddir == DDIR_READ ? "READ" :
             (io->ddir == DDIR_WRITE ? "WRITE" : "SYNC"), (uint64_t)io, retval);
     }
 }
@@ -405,10 +405,11 @@ static enum fio_q_status sec_queue(struct thread_data *td, struct io_u *io)
     bsd->inflight++;
 
     uint64_t inode = opt->image ? vitastor_c_inode_get_num(bsd->watch) : opt->inode;
+    assert(io->xfer_buflen < (size_t)-1);
     switch (io->ddir)
     {
     case DDIR_READ:
-        iov = { .iov_base = io->xfer_buf, .iov_len = io->xfer_buflen };
+        iov = { .iov_base = io->xfer_buf, .iov_len = (size_t)io->xfer_buflen };
         vitastor_c_read(bsd->cli, inode, io->offset, io->xfer_buflen, &iov, 1, read_callback, io);
         bsd->last_sync = false;
         break;
@@ -436,7 +437,7 @@ static enum fio_q_status sec_queue(struct thread_data *td, struct io_u *io)
             io->error = EROFS;
             return FIO_Q_COMPLETED;
         }
-        iov = { .iov_base = io->xfer_buf, .iov_len = io->xfer_buflen };
+        iov = { .iov_base = io->xfer_buf, .iov_len = (size_t)io->xfer_buflen };
         vitastor_c_write(bsd->cli, inode, io->offset, io->xfer_buflen, 0, &iov, 1, io_callback, io);
         bsd->last_sync = false;
         break;
@@ -453,11 +454,11 @@ static enum fio_q_status sec_queue(struct thread_data *td, struct io_u *io)
     {
         if (io->ddir == DDIR_SYNC)
         {
-            printf("+++ SYNC 0x%lx\n", (uint64_t)io);
+            printf("+++ SYNC 0x%jx\n", (uint64_t)io);
         }
         else
         {
-            printf("+++ %s 0x%lx 0x%llx+%lx\n",
+            printf("+++ %s 0x%jx 0x%llx+%jx\n",
                 io->ddir == DDIR_READ ? "READ" : "WRITE",
                 (uint64_t)io, io->offset, (uint64_t)io->xfer_buflen);
         }

src/fio_sec_osd.cpp

@@ -310,7 +310,8 @@ static enum fio_q_status sec_queue(struct thread_data *td, struct io_u *io)
     int iovcnt = 1, wtotal = OSD_PACKET_SIZE;
     if (io->ddir == DDIR_WRITE)
     {
-        iov[iovcnt++] = { .iov_base = io->xfer_buf, .iov_len = io->xfer_buflen };
+        assert(io->xfer_buflen <= 0x7fffffff);
+        iov[iovcnt++] = { .iov_base = io->xfer_buf, .iov_len = (size_t)io->xfer_buflen };
         wtotal += io->xfer_buflen;
     }
     if (sendv_blocking(bsd->connect_fd, iov, iovcnt,
@@ -341,13 +342,13 @@ static int sec_getevents(struct thread_data *td, unsigned int min, unsigned int
         read_blocking(bsd->connect_fd, reply.buf, OSD_PACKET_SIZE);
         if (reply.hdr.magic != SECONDARY_OSD_REPLY_MAGIC)
         {
-            fprintf(stderr, "bad reply: magic = %lx instead of %lx\n", reply.hdr.magic, SECONDARY_OSD_REPLY_MAGIC);
+            fprintf(stderr, "bad reply: magic = %jx instead of %jx\n", reply.hdr.magic, SECONDARY_OSD_REPLY_MAGIC);
             exit(1);
         }
         auto it = bsd->queue.find(reply.hdr.id);
         if (it == bsd->queue.end())
         {
-            fprintf(stderr, "bad reply: op id %lx missing in local queue\n", reply.hdr.id);
+            fprintf(stderr, "bad reply: op id %jx missing in local queue\n", reply.hdr.id);
             exit(1);
         }
         io_u* io = it->second->fio_op;
@@ -357,7 +358,7 @@ static int sec_getevents(struct thread_data *td, unsigned int min, unsigned int
         {
             if (reply.hdr.retval != io->xfer_buflen)
             {
-                fprintf(stderr, "Short read: retval = %ld instead of %lu\n", reply.hdr.retval, (uint64_t)io->xfer_buflen);
+                fprintf(stderr, "Short read: retval = %jd instead of %ju\n", reply.hdr.retval, (uint64_t)io->xfer_buflen);
                 exit(1);
             }
             // Support bitmap
@@ -371,7 +372,8 @@ static int sec_getevents(struct thread_data *td, unsigned int min, unsigned int
                 else
                     iov[iovcnt++] = { .iov_base = (void*)(bitmap = (uint64_t)malloc(reply.sec_rw.attr_len)), .iov_len = reply.sec_rw.attr_len };
             }
-            iov[iovcnt++] = { .iov_base = io->xfer_buf, .iov_len = io->xfer_buflen };
+            assert(io->xfer_buflen <= 0x7FFFFFFF);
+            iov[iovcnt++] = { .iov_base = io->xfer_buf, .iov_len = (size_t)io->xfer_buflen };
             readv_blocking(bsd->connect_fd, iov, iovcnt);
             if (reply.sec_rw.attr_len > 8)
             {
@@ -382,7 +384,7 @@ static int sec_getevents(struct thread_data *td, unsigned int min, unsigned int
         {
             if (reply.hdr.retval != io->xfer_buflen)
             {
-                fprintf(stderr, "Short write: retval = %ld instead of %lu\n", reply.hdr.retval, (uint64_t)io->xfer_buflen);
+                fprintf(stderr, "Short write: retval = %jd instead of %ju\n", reply.hdr.retval, (uint64_t)io->xfer_buflen);
                 exit(1);
             }
         }
@@ -390,13 +392,13 @@ static int sec_getevents(struct thread_data *td, unsigned int min, unsigned int
         {
             if (reply.hdr.retval != 0)
             {
-                fprintf(stderr, "Sync failed: retval = %ld\n", reply.hdr.retval);
+                fprintf(stderr, "Sync failed: retval = %jd\n", reply.hdr.retval);
                 exit(1);
             }
         }
         if (opt->trace)
         {
-            printf("--- %s # %ld\n", io->ddir == DDIR_READ ? "READ" :
+            printf("--- %s # %ju\n", io->ddir == DDIR_READ ? "READ" :
                 (io->ddir == DDIR_WRITE ? "WRITE" : "SYNC"), reply.hdr.id);
         }
         bsd->completed.push_back(io);

src/kv_cli.cpp

@@ -0,0 +1,673 @@
+// Copyright (c) Vitaliy Filippov, 2019+
+// License: VNPL-1.1 (see README.md for details)
+//
+// Vitastor shared key/value database test CLI
+
+#define _XOPEN_SOURCE
+#include <limits.h>
+
+#include <netinet/tcp.h>
+#include <sys/epoll.h>
+#include <unistd.h>
+#include <fcntl.h>
+//#include <signal.h>
+
+#include "epoll_manager.h"
+#include "str_util.h"
+#include "kv_db.h"
+
+const char *exe_name = NULL;
+
+class kv_cli_t
+{
+public:
+    json11::Json::object cfg;
+    std::vector<std::string> cli_cmd;
+
+    kv_dbw_t *db = NULL;
+    ring_loop_t *ringloop = NULL;
+    epoll_manager_t *epmgr = NULL;
+    cluster_client_t *cli = NULL;
+    int load_parallelism = 16;
+    bool opened = false;
+    bool interactive = false, is_file = false;
+    int in_progress = 0;
+    char *cur_cmd = NULL;
+    int cur_cmd_size = 0, cur_cmd_alloc = 0;
+    bool finished = false, eof = false;
+
+    std::function<void(int)> load_cb;
+    bool loading_json = false, in_loadjson = false;
+    int load_state = 0;
+    std::string load_key;
+
+    ~kv_cli_t();
+
+    void parse_args(int narg, const char *args[]);
+    void run();
+    void read_cmd();
+    void next_cmd();
+    std::vector<std::string> parse_cmd(const std::string & cmdstr);
+    void handle_cmd(const std::vector<std::string> & cmd, std::function<void(int)> cb);
+    void loadjson();
+};
+
+kv_cli_t::~kv_cli_t()
+{
+    if (cur_cmd)
+    {
+        free(cur_cmd);
+        cur_cmd = NULL;
+    }
+    cur_cmd_alloc = 0;
+    if (db)
+        delete db;
+    if (cli)
+    {
+        cli->flush();
+        delete cli;
+    }
+    if (epmgr)
+        delete epmgr;
+    if (ringloop)
+        delete ringloop;
+}
+
+void kv_cli_t::parse_args(int narg, const char *args[])
+{
+    bool db = false;
+    for (int i = 1; i < narg; i++)
+    {
+        if (!strcmp(args[i], "-h") || !strcmp(args[i], "--help"))
+        {
+            printf(
+                "Vitastor Key/Value CLI\n"
+                "(c) Vitaliy Filippov, 2023+ (VNPL-1.1)\n"
+                "\n"
+                "USAGE: %s [OPTIONS] [<IMAGE> [<COMMAND>]]\n"
+                "\n"
+                "COMMANDS:\n"
+                "  get <key>\n"
+                "  set <key> <value>\n"
+                "  del <key>\n"
+                "  list [<start> [end]]\n"
+                "  dump [<start> [end]]\n"
+                "  dumpjson [<start> [end]]\n"
+                "  loadjson\n"
+                "\n"
+                "<IMAGE> should be the name of Vitastor image with the DB.\n"
+                "Without <COMMAND>, you get an interactive DB shell.\n"
+                "\n"
+                "OPTIONS:\n"
+                "  --kv_block_size 4k\n"
+                "    Key-value B-Tree block size\n"
+                "  --kv_memory_limit 128M\n"
+                "    Maximum memory to use for vitastor-kv index cache\n"
+                "  --kv_allocate_blocks 4\n"
+                "    Number of PG blocks used for new tree block allocation in parallel\n"
+                "  --kv_evict_max_misses 10\n"
+                "    Eviction algorithm parameter: retry eviction from another random spot\n"
+                "    if this number of keys is used currently or was used recently\n"
+                "  --kv_evict_attempts_per_level 3\n"
+                "    Retry eviction at most this number of times per tree level, starting\n"
+                "    with bottom-most levels\n"
+                "  --kv_evict_unused_age 1000\n"
+                "    Evict only keys unused during this number of last operations\n"
+                "  --kv_log_level 1\n"
+                "    Log level. 0 = errors, 1 = warnings, 10 = trace operations\n"
+                ,
+                exe_name
+            );
+            exit(0);
+        }
+        else if (args[i][0] == '-' && args[i][1] == '-')
+        {
+            const char *opt = args[i]+2;
+            cfg[opt] = !strcmp(opt, "json") || i == narg-1 ? "1" : args[++i];
+        }
+        else if (!db)
+        {
+            cfg["db"] = args[i];
+            db = true;
+        }
+        else
+        {
+            cli_cmd.push_back(args[i]);
+        }
+    }
+}
+
+void kv_cli_t::run()
+{
+    // Create client
+    ringloop = new ring_loop_t(512);
+    epmgr = new epoll_manager_t(ringloop);
+    cli = new cluster_client_t(ringloop, epmgr->tfd, cfg);
+    db = new kv_dbw_t(cli);
+    // Load image metadata
+    while (!cli->is_ready())
+    {
+        ringloop->loop();
+        if (cli->is_ready())
+            break;
+        ringloop->wait();
+    }
+    // Open if DB is set in options
+    if (cfg.find("db") != cfg.end())
+    {
+        bool done = false;
+        handle_cmd({ "open", cfg.at("db").string_value() }, [&done](int res) { if (res != 0) exit(1); done = true; });
+        while (!done)
+        {
+            ringloop->loop();
+            if (done)
+                break;
+            ringloop->wait();
+        }
+    }
+    // Run single command from CLI
+    if (cli_cmd.size())
+    {
+        bool done = false;
+        handle_cmd(cli_cmd, [&done](int res) { if (res != 0) exit(1); done = true; });
+        while (!done)
+        {
+            ringloop->loop();
+            if (done)
+                break;
+            ringloop->wait();
+        }
+    }
+    else
+    {
+        // Run interactive shell
+        fcntl(0, F_SETFL, fcntl(0, F_GETFL, 0) | O_NONBLOCK);
+        try
+        {
+            epmgr->tfd->set_fd_handler(0, false, [this](int fd, int events)
+            {
+                if (events & EPOLLIN)
+                {
+                    read_cmd();
+                }
+                if (events & EPOLLRDHUP)
+                {
+                    epmgr->tfd->set_fd_handler(0, false, NULL);
+                    finished = true;
+                }
+            });
+            interactive = isatty(0);
+            if (interactive)
+                printf("> ");
+        }
+        catch (std::exception & e)
+        {
+            // Can't add to epoll, STDIN is probably a file
+            is_file = true;
+            read_cmd();
+        }
+        while (!finished)
+        {
+            ringloop->loop();
+            if (!finished)
+                ringloop->wait();
+        }
+    }
+    // Destroy the client
+    delete db;
+    db = NULL;
+    cli->flush();
+    delete cli;
+    delete epmgr;
+    delete ringloop;
+    cli = NULL;
+    epmgr = NULL;
+    ringloop = NULL;
+}
+
+void kv_cli_t::read_cmd()
+{
+    if (!cur_cmd_alloc)
+    {
+        cur_cmd_alloc = 65536;
+        cur_cmd = (char*)malloc_or_die(cur_cmd_alloc);
+    }
+    while (cur_cmd_size < cur_cmd_alloc)
+    {
+        int r = read(0, cur_cmd+cur_cmd_size, cur_cmd_alloc-cur_cmd_size);
+        if (r < 0 && errno != EAGAIN)
+            fprintf(stderr, "Error reading from stdin: %s\n", strerror(errno));
+        if (r > 0)
+            cur_cmd_size += r;
+        if (r == 0)
+            eof = true;
+        if (r <= 0)
+            break;
+    }
+    next_cmd();
+}
+
+void kv_cli_t::next_cmd()
+{
+    if (loading_json)
+    {
+        loadjson();
+        return;
+    }
+    if (in_progress > 0)
+    {
+        return;
+    }
+    int pos = 0;
+    for (; pos < cur_cmd_size; pos++)
+    {
+        if (cur_cmd[pos] == '\n' || cur_cmd[pos] == '\r')
+        {
+            auto cmd = trim(std::string(cur_cmd, pos));
+            pos++;
+            memmove(cur_cmd, cur_cmd+pos, cur_cmd_size-pos);
+            cur_cmd_size -= pos;
+            in_progress++;
+            handle_cmd(parse_cmd(cmd), [this](int res)
+            {
+                in_progress--;
+                if (interactive)
+                    printf("> ");
+                next_cmd();
+                if (!in_progress)
+                    read_cmd();
+            });
+            break;
+        }
+    }
+    if (eof && !in_progress)
+    {
+        finished = true;
+    }
+}
+
+struct kv_cli_list_t
+{
+    kv_dbw_t *db = NULL;
+    void *handle = NULL;
+    int format = 0;
+    int n = 0;
+    std::function<void(int)> cb;
+};
+
+std::vector<std::string> kv_cli_t::parse_cmd(const std::string & str)
+{
+    std::vector<std::string> res;
+    size_t pos = 0;
+    auto cmd = scan_escaped(str, pos);
+    if (cmd.empty())
+        return res;
+    res.push_back(cmd);
+    int max_args = (cmd == "set" || cmd == "config" ||
+        cmd == "list" || cmd == "dump" || cmd == "dumpjson" ? 3 :
+        (cmd == "open" || cmd == "get" || cmd == "del" ? 2 : 1));
+    while (pos < str.size() && res.size() < max_args)
+    {
+        if (res.size() == max_args-1)
+        {
+            // Allow unquoted last argument
+            pos = str.find_first_not_of(" \t\r\n", pos);
+            if (pos == std::string::npos)
+                break;
+            if (str[pos] != '"' && str[pos] != '\'')
+            {
+                res.push_back(trim(str.substr(pos)));
+                break;
+            }
+        }
+        auto arg = scan_escaped(str, pos);
+        if (arg.size())
+            res.push_back(arg);
+    }
+    return res;
+}
+
+void kv_cli_t::loadjson()
+{
+    // simple streaming json parser
+    if (in_progress >= load_parallelism || in_loadjson)
+    {
+        return;
+    }
+    in_loadjson = true;
+    if (load_state == 5)
+    {
+st_5:
+        if (!in_progress)
+        {
+            loading_json = false;
+            auto cb = std::move(load_cb);
+            cb(0);
+        }
+        in_loadjson = false;
+        return;
+    }
+    do
+    {
+        read_cmd();
+        size_t pos = 0;
+        while (true)
+        {
+            while (pos < cur_cmd_size && is_white(cur_cmd[pos]))
+            {
+                pos++;
+            }
+            if (pos >= cur_cmd_size)
+            {
+                break;
+            }
+            if (load_state == 0 || load_state == 2)
+            {
+                char expected = "{ :"[load_state];
+                if (cur_cmd[pos] != expected)
+                {
+                    fprintf(stderr, "Unexpected %c, expected %c\n", cur_cmd[pos], expected);
+                    exit(1);
+                }
+                pos++;
+                load_state++;
+            }
+            else if (load_state == 1 || load_state == 3)
+            {
+                if (cur_cmd[pos] != '"')
+                {
+                    fprintf(stderr, "Unexpected %c, expected \"\n", cur_cmd[pos]);
+                    exit(1);
+                }
+                size_t prev = pos;
+                auto str = scan_escaped(cur_cmd, cur_cmd_size, pos, false);
+                if (pos == prev)
+                {
+                    break;
+                }
+                load_state++;
+                if (load_state == 2)
+                {
+                    load_key = str;
+                }
+                else
+                {
+                    in_progress++;
+                    handle_cmd({ "set", load_key, str }, [this](int res)
+                    {
+                        in_progress--;
+                        next_cmd();
+                    });
+                    if (in_progress >= load_parallelism)
+                    {
+                        break;
+                    }
+                }
+            }
+            else if (load_state == 4)
+            {
+                if (cur_cmd[pos] == ',')
+                {
+                    pos++;
+                    load_state = 1;
+                }
+                else if (cur_cmd[pos] == '}')
+                {
+                    pos++;
+                    load_state = 5;
+                    goto st_5;
+                }
+                else
+                {
+                    fprintf(stderr, "Unexpected %c, expected , or }\n", cur_cmd[pos]);
+                    exit(1);
+                }
+            }
+        }
+        if (pos < cur_cmd_size)
+        {
+            memmove(cur_cmd, cur_cmd+pos, cur_cmd_size-pos);
+        }
+        cur_cmd_size -= pos;
+    } while (loading_json && is_file);
+    in_loadjson = false;
+}
+
+void kv_cli_t::handle_cmd(const std::vector<std::string> & cmd, std::function<void(int)> cb)
+{
+    if (!cmd.size())
+    {
+        cb(-EINVAL);
+        return;
+    }
+    auto & opname = cmd[0];
+    if (!opened && opname != "open" && opname != "config" && opname != "quit" && opname != "q")
+    {
+        fprintf(stderr, "Error: database not opened\n");
+        cb(-EINVAL);
+        return;
+    }
+    if (opname == "open")
+    {
+        auto name = cmd.size() > 1 ? cmd[1] : "";
+        uint64_t pool_id = 0;
+        inode_t inode_id = 0;
+        int scanned = sscanf(name.c_str(), "%lu %lu", &pool_id, &inode_id);
+        if (scanned < 2 || !pool_id || !inode_id)
+        {
+            inode_id = 0;
+            name = trim(name);
+            for (auto & ic: cli->st_cli.inode_config)
+            {
+                if (ic.second.name == name)
+                {
+                    inode_id = ic.first;
+                    break;
+                }
+            }
+            if (!inode_id)
+            {
+                fprintf(stderr, "Usage: open <image> OR open <pool_id> <inode_id>\n");
+                cb(-EINVAL);
+                return;
+            }
+        }
+        else
+            inode_id = INODE_WITH_POOL(pool_id, inode_id);
+        db->open(inode_id, cfg, [=](int res)
+        {
+            if (res < 0)
+            {
+                fprintf(stderr, "Error opening index: %s (code %d)\n", strerror(-res), res);
+            }
+            else
+            {
+                opened = true;
+                fprintf(interactive ? stdout : stderr, "Index opened. Current size: %lu bytes\n", db->get_size());
+            }
+            cb(res);
+        });
+    }
+    else if (opname == "config")
+    {
+        if (cmd.size() < 3)
+        {
+            fprintf(stderr, "Usage: config <property> <value>\n");
+            cb(-EINVAL);
+            return;
+        }
+        auto & key = cmd[1];
+        auto & value = cmd[2];
+        if (key != "kv_memory_limit" &&
+            key != "kv_allocate_blocks" &&
+            key != "kv_evict_max_misses" &&
+            key != "kv_evict_attempts_per_level" &&
+            key != "kv_evict_unused_age" &&
+            key != "kv_log_level" &&
+            key != "kv_block_size")
+        {
+            fprintf(
+                stderr, "Allowed properties: kv_block_size, kv_memory_limit, kv_allocate_blocks,"
+                " kv_evict_max_misses, kv_evict_attempts_per_level, kv_evict_unused_age, kv_log_level\n"
+            );
+            cb(-EINVAL);
+        }
+        else if (key == "kv_block_size")
+        {
+            if (opened)
+            {
+                fprintf(stderr, "kv_block_size can't be set after opening DB\n");
+                cb(-EINVAL);
+            }
+            else
+            {
+                cfg[key] = value;
+                cb(0);
+            }
+        }
+        else
+        {
+            cfg[key] = value;
+            db->set_config(cfg);
+            cb(0);
+        }
+    }
+    else if (opname == "get" || opname == "set" || opname == "del")
+    {
+        if (opname == "get" || opname == "del")
+        {
+            if (cmd.size() < 2)
+            {
+                fprintf(stderr, "Usage: %s <key>\n", opname.c_str());
+                cb(-EINVAL);
+                return;
+            }
+            auto & key = cmd[1];
+            if (opname == "get")
+            {
+                db->get(key, [this, cb](int res, const std::string & value)
+                {
+                    if (res < 0)
+                        fprintf(stderr, "Error: %s (code %d)\n", strerror(-res), res);
+                    else
+                    {
+                        if (write(1, value.c_str(), value.size()) < 0 || write(1, "\n", 1) < 0)
+                            exit(1);
+                    }
+                    cb(res);
+                });
+            }
+            else
+            {
+                db->del(key, [this, cb](int res)
+                {
+                    if (res < 0)
+                        fprintf(stderr, "Error: %s (code %d)\n", strerror(-res), res);
+                    else
+                        fprintf(interactive ? stdout : stderr, "OK\n");
+                    cb(res);
+                });
+            }
+        }
+        else
+        {
+            if (cmd.size() < 3)
+            {
+                fprintf(stderr, "Usage: set <key> <value>\n");
+                cb(-EINVAL);
+                return;
+            }
+            auto & key = cmd[1];
+            auto & value = cmd[2];
+            db->set(key, value, [this, cb, l = loading_json](int res)
+            {
+                if (res < 0)
+                    fprintf(stderr, "Error: %s (code %d)\n", strerror(-res), res);
+                else if (!l)
+                    fprintf(interactive ? stdout : stderr, "OK\n");
+                cb(res);
+            });
+        }
+    }
+    else if (opname == "list" || opname == "dump" || opname == "dumpjson")
+    {
+        kv_cli_list_t *lst = new kv_cli_list_t;
+        std::string start = cmd.size() >= 2 ? cmd[1] : "";
+        std::string end = cmd.size() >= 3 ? cmd[2] : "";
+        lst->handle = db->list_start(start);
+        lst->db = db;
+        lst->format = opname == "dump" ? 1 : (opname == "dumpjson" ? 2 : 0);
+        lst->cb = std::move(cb);
+        db->list_next(lst->handle, [lst](int res, const std::string & key, const std::string & value)
+        {
+            if (res < 0)
+            {
+                if (res != -ENOENT)
+                {
+                    fprintf(stderr, "Error: %s (code %d)\n", strerror(-res), res);
+                }
+                if (lst->format == 2)
+                    printf("\n}\n");
+                lst->db->list_close(lst->handle);
+                lst->cb(res == -ENOENT ? 0 : res);
+                delete lst;
+            }
+            else
+            {
+                if (lst->format == 2)
+                    printf(lst->n ? ",\n  %s: %s" : "{\n  %s: %s", addslashes(key).c_str(), addslashes(value).c_str());
+                else if (lst->format == 1)
+                    printf("set %s %s\n", auto_addslashes(key).c_str(), value.c_str());
+                else
+                    printf("%s = %s\n", key.c_str(), value.c_str());
+                lst->n++;
+                lst->db->list_next(lst->handle, NULL);
+            }
+        });
+    }
+    else if (opname == "loadjson")
+    {
+        loading_json = true;
+        load_state = 0;
+        load_cb = cb;
+        loadjson();
+    }
+    else if (opname == "close")
+    {
+        db->close([=]()
+        {
+            fprintf(interactive ? stdout : stderr, "Index closed\n");
+            opened = false;
+            cb(0);
+        });
+    }
+    else if (opname == "quit" || opname == "q")
+    {
+        ::close(0);
+        finished = true;
+    }
+    else
+    {
+        fprintf(
+            stderr, "Unknown operation: %s. Supported operations:\n"
+            "open <image>\nopen <pool_id> <inode_id>\n"
+            "config <property> <value>\n"
+            "get <key>\nset <key> <value>\ndel <key>\n"
+            "list [<start> [end]]\ndump [<start> [end]]\ndumpjson [<start> [end]]\nloadjson\n"
+            "close\nquit\n", opname.c_str()
+        );
+        cb(-EINVAL);
+    }
+}
+
+int main(int narg, const char *args[])
+{
+    setvbuf(stdout, NULL, _IONBF, 0);
+    setvbuf(stderr, NULL, _IONBF, 0);
+    exe_name = args[0];
+    kv_cli_t *p = new kv_cli_t();
+    p->parse_args(narg, args);
+    p->run();
+    delete p;
+    return 0;
+}