Release 3.0.6

- Fix one more bug in old store checksums O_o - checksum verification wasn't working correctly on reads quickly after small initial offsetted writes (when a BIG_WRITE was read from the journal). - Fix clients possibly not destroyed after being switched to RDMA. Not 100% verified, but in theory it could rarely lead to PGs hanging in peering/stopping states with RDMA enabled. - Use vitastor-cli in the Cinder driver for modifications instead of direct etcd interaction. - Fix mon_https_ca parsing. - Fix Docker install.sh, add instructions for Podman.
Fix str_replace
2026-03-22 22:49:51 +03:00 · 2026-03-19 01:19:29 +03:00 · 2026-03-17 22:26:38 +00:00 · 2026-03-17 01:04:17 +03:00 · 2026-03-15 01:44:30 +03:00 · 2026-03-15 01:01:41 +03:00
96 changed files with 2009 additions and 901 deletions
@@ -1,28 +1,29 @@
-FROM node:16-bullseye
+FROM node:16-bookworm

 WORKDIR /root

-ADD ./docker/vitastor.gpg /etc/apt/trusted.gpg.d
+ADD ./docker/etc/apt/trusted.gpg.d /etc/apt/trusted.gpg.d

-RUN echo 'deb http://deb.debian.org/debian bullseye-backports main' >> /etc/apt/sources.list; \
-    echo 'deb http://vitastor.io/debian bullseye main' >> /etc/apt/sources.list; \
+RUN echo 'deb http://deb.debian.org/debian bookworm-backports main' >> /etc/apt/sources.list; \
+    echo 'deb http://vitastor.io/debian bookworm main' >> /etc/apt/sources.list; \
    echo >> /etc/apt/preferences; \
    echo 'Package: *' >> /etc/apt/preferences; \
-    echo 'Pin: release a=bullseye-backports' >> /etc/apt/preferences; \
+    echo 'Pin: release n=bookworm-backports' >> /etc/apt/preferences; \
    echo 'Pin-Priority: 500' >> /etc/apt/preferences; \
    echo >> /etc/apt/preferences; \
    echo 'Package: *' >> /etc/apt/preferences; \
    echo 'Pin: origin "vitastor.io"' >> /etc/apt/preferences; \
    echo 'Pin-Priority: 1000' >> /etc/apt/preferences; \
+    perl -i -pe 's/Types: deb$/Types: deb deb-src/' /etc/apt/sources.list.d/debian.sources; \
    grep '^deb ' /etc/apt/sources.list | perl -pe 's/^deb/deb-src/' >> /etc/apt/sources.list; \
    echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf; \
    echo 'APT::Install-Suggests false;' >> /etc/apt/apt.conf

 RUN apt-get update
-RUN apt-get -y install etcd qemu-system-x86 qemu-block-extra qemu-utils fio libasan5 \
+RUN apt-get -y install etcd qemu-system-x86 qemu-block-extra qemu-utils fio libasan8 \
    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 update && apt-get -y install jq lp-solve sudo nfs-common fdisk parted
+RUN apt-get update && apt-get -y install jq lp-solve sudo nfs-common fdisk parted libc-ares-dev udev
 RUN apt-get --download-only source fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`

 RUN set -ex; \
@@ -1206,6 +1206,78 @@ jobs:
          echo ""
        done

+  test_checksum:
+    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_checksum.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
+
+  test_old_checksum:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: OLD=1 /root/vitastor/tests/test_checksum.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
+
+  test_corrupt_all:
+    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_corrupt_all.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
+
+  test_old_corrupt_all:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: OLD=1 /root/vitastor/tests/test_corrupt_all.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
+
  test_reweight_half:
    runs-on: ubuntu-latest
    needs: build
@@ -2,7 +2,7 @@ cmake_minimum_required(VERSION 2.8.12)

 project(vitastor)

-set(VITASTOR_VERSION "3.0.2")
+set(VITASTOR_VERSION "3.0.6")

 include(CTest)

@@ -1,5 +1,5 @@
 # Compile stage
-FROM golang:bookworm AS build
+FROM golang:trixie AS build

 ADD go.sum go.mod /app/
 RUN cd /app; CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go mod download -x
@@ -9,7 +9,7 @@ RUN perl -i -e '$/ = undef; while(<>) { s/\n\s*(\{\s*\n)/$1\n/g; s/\}(\s*\n\s*)e
    CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go build -o vitastor-csi

 # Final stage
-FROM debian:bookworm
+FROM debian:trixie

 LABEL maintainers="Vitaliy Filippov <vitalif@yourcmc.ru>"
 LABEL description="Vitastor CSI Driver"
@@ -25,20 +25,20 @@ RUN apt-get update && \
        # NFS mount dependencies
        nfs-common netbase \
        # dependencies of qemu-storage-daemon
-        libnuma1 liburing2 libglib2.0-0 libfuse3-3 libaio1 libzstd1 libnettle8 \
-        libgmp10 libhogweed6 libp11-kit0 libidn2-0 libunistring2 libtasn1-6 libpcre2-8-0 libffi8 && \
+        libaio1t64 libc6 libfuse3-4 libglib2.0-0t64 libgmp10 libgnutls30t64 \
+        libhogweed6t64 libnettle8t64 libnuma1 libselinux1 liburing2 libzstd1 zlib1g && \
    apt-get clean && \
    (echo options nbd nbds_max=128 > /etc/modprobe.d/nbd.conf)

 COPY --from=build /app/vitastor-csi /bin/

-RUN (echo deb http://vitastor.io/debian bookworm main > /etc/apt/sources.list.d/vitastor.list) && \
+RUN (echo deb http://vitastor.io/debian trixie main > /etc/apt/sources.list.d/vitastor.list) && \
    ((echo 'Package: *'; echo 'Pin: origin "vitastor.io"'; echo 'Pin-Priority: 1000') > /etc/apt/preferences.d/vitastor.pref) && \
    wget -q -O /etc/apt/trusted.gpg.d/vitastor.gpg https://vitastor.io/debian/pubkey.gpg && \
    apt-get update && \
    apt-get install -y vitastor-client ibverbs-providers && \
-    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-utils_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
-    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-block-extra_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-trixie-10.0.2%2Bds-2%2Bvitastor1/qemu-utils_10.0.2%2Bds-2%2Bvitastor1_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-trixie-10.0.2%2Bds-2%2Bvitastor1/qemu-block-extra_10.0.2%2Bds-2%2Bvitastor1_amd64.deb && \
    dpkg -x qemu-utils*.deb tmp1 && \
    dpkg -x qemu-block-extra*.deb tmp1 && \
    cp -a tmp1/usr/bin/qemu-storage-daemon /usr/bin/ && \
@@ -1,5 +1,5 @@
 # Compile stage
-FROM golang:bookworm AS build
+FROM golang:trixie AS build

 ADD go.sum go.mod /app/
 RUN cd /app; CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go mod download -x
@@ -9,7 +9,7 @@ RUN perl -i -e '$/ = undef; while(<>) { s/\n\s*(\{\s*\n)/$1\n/g; s/\}(\s*\n\s*)e
    CGO_ENABLED=1 GOOS=linux GOARCH=amd64 go build -o vitastor-csi

 # Final stage
-FROM debian:bookworm
+FROM debian:trixie

 LABEL maintainers="Vitaliy Filippov <vitalif@yourcmc.ru>"
 LABEL description="Vitastor CSI Driver"
@@ -36,8 +36,8 @@ ADD deb /deb

 RUN apt-get update && \
    apt-get -y install /deb/vitastor-client_*.deb && \
-    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-utils_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
-    wget https://vitastor.io/archive/qemu/qemu-bookworm-9.2.2%2Bds-1%2Bvitastor4/qemu-block-extra_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-trixie-9.2.2%2Bds-1%2Bvitastor4/qemu-utils_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
+    wget https://vitastor.io/archive/qemu/qemu-trixie-9.2.2%2Bds-1%2Bvitastor4/qemu-block-extra_9.2.2%2Bds-1%2Bvitastor4_amd64.deb && \
    dpkg -x qemu-utils*.deb tmp1 && \
    dpkg -x qemu-block-extra*.deb tmp1 && \
    cp -a tmp1/usr/bin/qemu-storage-daemon /usr/bin/ && \
@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.2
+VITASTOR_VERSION ?= v3.0.6

 all: build push

@@ -49,7 +49,7 @@ spec:
            capabilities:
              add: ["SYS_ADMIN"]
            allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v3.0.2
+          image: vitalif/vitastor-csi:v3.0.6
          args:
            - "--node=$(NODE_ID)"
            - "--endpoint=$(CSI_ENDPOINT)"
@@ -121,7 +121,7 @@ spec:
            privileged: true
            capabilities:
              add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v3.0.2
+          image: vitalif/vitastor-csi:v3.0.6
          args:
            - "--node=$(NODE_ID)"
            - "--endpoint=$(CSI_ENDPOINT)"
@@ -5,7 +5,7 @@ package vitastor

 const (
    vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "3.0.2"
+    vitastorCSIDriverVersion = "3.0.6"
 )

 // Config struct fills the parameters of request or user input
@@ -0,0 +1,5 @@
+#!/bin/bash
+# 26.04 Resolute Raccoon
+
+docker build --build-arg DISTRO=ubuntu --build-arg REL=resolute -t vitastor-buildenv:resolute -f vitastor-buildenv.Dockerfile .
+docker run -it --rm -e REL=resolute -v `dirname $0`/../:/root/vitastor vitastor-buildenv:resolute /root/vitastor/debian/vitastor-build.sh
@@ -1,4 +1,4 @@
-vitastor (3.0.2-1) unstable; urgency=medium
+vitastor (3.0.6-1) unstable; urgency=medium

  * Bugfixes

@@ -1,6 +1,6 @@
 # Build Docker image with Vitastor packages

-FROM debian:bookworm
+FROM debian:trixie

 ADD etc/apt /etc/apt/
 RUN apt-get update && apt-get -y install vitastor ibverbs-providers udev systemd qemu-system-x86 qemu-system-common qemu-block-extra qemu-utils jq nfs-common && apt-get clean
@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.2
+VITASTOR_VERSION ?= v3.0.6

 all: build push

@@ -1,3 +1,3 @@
 Package: *
-Pin: release n=bookworm-backports
+Pin: release n=trixie-backports
 Pin-Priority: 500
@@ -1,2 +1,2 @@
-deb http://vitastor.io/debian bookworm main
-deb http://http.debian.net/debian/ bookworm-backports main
+deb http://vitastor.io/debian trixie main
+#deb http://http.debian.net/debian/ trixie-backports main
@@ -7,7 +7,7 @@ PartOf=vitastor.target
 [Service]
 Restart=always
 EnvironmentFile=/etc/vitastor/docker.conf
-ExecStart=bash -c 'docker run --rm -i -v /etc/vitastor:/etc/vitastor -v /dev:/dev -v /run:/run \
+ExecStart=bash -c 'docker run --rm -i -v /etc/vitastor:/etc/vitastor -v /dev:/dev -v /run:/run -e SYSTEMD_IN_CHROOT=0 \
    --security-opt seccomp=unconfined --privileged --pid=host --log-driver none --network host --name vitastor vitastor:$VITASTOR_VERSION \
    sleep.sh'
 ExecStartPost=udevadm trigger
@@ -4,7 +4,7 @@
 #

 # Desired Vitastor version
-VITASTOR_VERSION=v3.0.2
+VITASTOR_VERSION=v3.0.6

 # Additional arguments for all containers
 # For example, you may want to specify a custom logging driver here
@@ -2,8 +2,7 @@

 set -e

-cp -urv /etc/default /host-etc/
-cp -urv /etc/systemd /host-etc/
-cp -urv /etc/udev /host-etc/
+cp -urv /etc/systemd/system/vitastor* /host-etc/systemd/system/
+cp -urv /etc/udev/rules.d /host-etc/udev/
 cp -urnv /etc/vitastor /host-etc/
 cp -urnv /opt/scripts/* /host-bin/
@@ -22,7 +22,6 @@ between clients, OSDs and etcd.
 - [rdma_max_msg](#rdma_max_msg)
 - [rdma_max_recv](#rdma_max_recv)
 - [rdma_max_send](#rdma_max_send)
- [rdma_odp](#rdma_odp)
 - [peer_connect_interval](#peer_connect_interval)
 - [peer_connect_timeout](#peer_connect_timeout)
 - [osd_idle_timeout](#osd_idle_timeout)
@@ -102,11 +101,6 @@ found or if `osd_network` is not specified. Auto-selection is also
 unsupported with old libibverbs < v32, like in Debian 10 Buster or
 CentOS 7.

-Vitastor supports all adapters, even ones without ODP support, like
-Mellanox ConnectX-3 and non-Mellanox cards. Versions up to Vitastor
-1.2.0 required ODP which is only present in Mellanox ConnectX >= 4.
-See also [rdma_odp](#rdma_odp).
-
 Run `ibv_devinfo -v` as root to list available RDMA devices and their
 features.

@@ -116,6 +110,23 @@ the manual of your network vendor for details about setting up the switch
 for RoCEv2 correctly. Usually it means setting up Lossless Ethernet with
 PFC (Priority Flow Control) and ECN (Explicit Congestion Notification).

+Vitastor supports all adapters, even ones without ODP (On-Demand Paging)
+support, like Mellanox ConnectX-3 and non-Mellanox cards. ODP is only present
+in Mellanox ConnectX >= 4 adapters and allows to skip memory registration
+for RDMA and thus, in theory, avoid memory copying.
+
+Versions up to Vitastor 1.2.0 required ODP, then it was disabled by default,
+but it was still supported up to 3.0.3. Now ODP support is removed because it
+actually only hurts performance: an example 3-node cluster with 8 NVMe in each
+node and 2*25 GBit/s ConnectX-6 RDMA network pushed 3950000 read iops without
+ODP, but only 239000 iops with ODP.
+
+This happens because Mellanox ODP implementation seems to be based on
+message retransmissions when the adapter doesn't know about the buffer yet -
+it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
+which is generally slow in RDMA/RoCE networks. Here's a presentation about
+it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
+
 ## rdma_port_num

 - Type: integer
@@ -187,28 +198,6 @@ less than `rdma_max_recv` so the receiving side doesn't run out of buffers.
 Doesn't affect memory usage - additional memory isn't allocated for send
 operations.

-## rdma_odp
-
- Type: boolean
- Default: false
-
-Use RDMA with On-Demand Paging. ODP is currently only available on Mellanox
-ConnectX-4 and newer adapters. ODP allows to not register memory explicitly
-for RDMA adapter to be able to use it. This, in turn, allows to skip memory
-copying during sending. One would think this should improve performance, but
-**in reality** RDMA performance with ODP is **drastically** worse. Example
-3-node cluster with 8 NVMe in each node and 2*25 GBit/s ConnectX-6 RDMA network
-without ODP pushes 3950000 read iops, but only 239000 iops with ODP...
-
-This happens because Mellanox ODP implementation seems to be based on
-message retransmissions when the adapter doesn't know about the buffer yet -
-it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
-which is generally slow in RDMA/RoCE networks. Here's a presentation about
-it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
-
-ODP support is retained in the code just in case a good ODP implementation
-appears one day.
-
 ## peer_connect_interval

 - Type: seconds
@@ -22,7 +22,6 @@
 - [rdma_max_msg](#rdma_max_msg)
 - [rdma_max_recv](#rdma_max_recv)
 - [rdma_max_send](#rdma_max_send)
- [rdma_odp](#rdma_odp)
 - [peer_connect_interval](#peer_connect_interval)
 - [peer_connect_timeout](#peer_connect_timeout)
 - [osd_idle_timeout](#osd_idle_timeout)
@@ -101,12 +100,6 @@ RoCEv1/RoCEv2, и даже позволяет полностью отключи
 не задана. Также автовыбор не поддерживается со старыми версиями библиотеки
 libibverbs < v32, например в Debian 10 Buster или CentOS 7.

-Vitastor поддерживает все модели адаптеров, включая те, у которых
-нет поддержки ODP, то есть вы можете использовать RDMA с ConnectX-3 и
-картами производства не Mellanox. Версии Vitastor до 1.2.0 включительно
-требовали ODP, который есть только на Mellanox ConnectX 4 и более новых.
-См. также [rdma_odp](#rdma_odp).
-
 Запустите `ibv_devinfo -v` от имени суперпользователя, чтобы посмотреть
 список доступных RDMA-устройств, их параметры и возможности.

@@ -117,6 +110,24 @@ Vitastor поддерживает все модели адаптеров, вкл
 подразумевает настройку сети без потерь на основе PFC (Priority Flow
 Control) и ECN (Explicit Congestion Notification).

+Vitastor поддерживает все модели адаптеров, включая те, у которых нет
+поддержки ODP (On-Demand Paging), например, ConnectX-3 и карты производства
+не Mellanox. Функция ODP доступна только на адаптерах Mellanox ConnectX-4 и
+более новых и позволяет не регистрировать память для её использования RDMA-картой,
+благодаря чему в теории можно избежать лишних копирований памяти.
+
+Версии Vitastor до 1.2.0 включительно требовали ODP, потом функция был отключена
+по умолчанию, но поддерживалась вплоть до версии 3.0.3. Сейчас поддержка ODP
+полностью удалена, так как на самом деле она только портит производительность:
+например, на 3-узловом кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на
+чтение с RDMA без ODP удаётся снять 3950000 iops, а с ODP - всего 239000 iops.
+
+Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
+основана на повторной передаче сообщений, когда карте не известен буфер -
+вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
+А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
+Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
+
 ## rdma_port_num

 - Тип: целое число
@@ -192,29 +203,6 @@ OSD в любом случае согласовывают реальное зн
 Не влияет на потребление памяти - дополнительная память на операции отправки
 не выделяется.

-## rdma_odp
-
- Тип: булево (да/нет)
- Значение по умолчанию: false
-
-Использовать RDMA с On-Demand Paging. ODP - функция, доступная пока что
-исключительно на адаптерах Mellanox ConnectX-4 и более новых. ODP позволяет
-не регистрировать память для её использования RDMA-картой. Благодаря этому
-можно не копировать данные при отправке их в сеть и, казалось бы, это должно
-улучшать производительность - но **по факту** получается так, что
-производительность только ухудшается, причём сильно. Пример - на 3-узловом
-кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на чтение с RDMA без ODP
-удаётся снять 3950000 iops, а с ODP - всего 239000 iops...
-
-Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
-основана на повторной передаче сообщений, когда карте не известен буфер -
-вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
-А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
-Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
-
-Возможность использования ODP сохранена в коде на случай, если вдруг в один
-прекрасный день появится хорошая реализация ODP.
-
 ## peer_connect_interval

 - Тип: секунды
@@ -38,6 +38,7 @@ with an OSD restart or, for some of them, even without restarting by updating co
 - [journal_io](#journal_io)
 - [journal_sector_buffer_count](#journal_sector_buffer_count)
 - [journal_no_same_sector_overwrites](#journal_no_same_sector_overwrites)
+- [skip_corrupted_meta_entries](#skip_corrupted_meta_entries)
 - [throttle_small_writes](#throttle_small_writes)
 - [throttle_target_iops](#throttle_target_iops)
 - [throttle_target_mbs](#throttle_target_mbs)
@@ -279,13 +280,19 @@ Maximum number of journal flushers (see above min_flusher_count).
 - Type: boolean
 - Default: true

-This parameter makes Vitastor always keep metadata area of the block device
-in memory. It's required for good performance because it allows to avoid
-additional read-modify-write cycles during metadata modifications. Metadata
-area size is currently roughly 224 MB per 1 TB of data. You can turn it off
-to reduce memory usage by this value, but it will hurt performance. This
-restriction is likely to be removed in the future along with the upgrade
-of the metadata storage scheme.
+Only for the old store ([meta_format](layout-osd.en.md#meta_format) 2).
+
+This parameter makes Vitastor keep a copy of metadata area in memory as it is
+on disk, in addition to the metadata database. When the option is enabled, every
+metadata entry is effectively stored in RAM twice. It's required for good performance
+because it allows to avoid additional read-modify-write cycles during metadata
+modifications. Metadata area size with the old store is roughly 224 MB per 1 TB
+of data. You can turn the option off to reduce memory usage by this value, but
+it will reduce performance.
+
+For the new store ([meta_format](layout-osd.en.md#meta_format) 3), the option
+may be changed in the future to support operation without loading full metadata
+database in memory.

 ## inmemory_journal

@@ -364,6 +371,8 @@ blocks. The only situation when you should increase it to a larger value
 is when you enable journal_no_same_sector_overwrites. In this case set
 it to, for example, 1024.

+Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
+
 ## journal_no_same_sector_overwrites

 - Type: boolean
@@ -377,6 +386,17 @@ journal after writing it instead of possibly overwriting it the second time.

 Most (99%) other SSDs don't need this option.

+Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
+
+## skip_corrupted_meta_entries
+
+- Type: boolean
+- Default: false
+
+Only for the new store ([meta_format](layout-osd.en.md#meta_format) 3).
+Allow OSD to start when some metadata entries or blocks are corrupted by
+skipping them. Should be only used as an emergency measure.
+
 ## throttle_small_writes

 - Type: boolean
@@ -39,6 +39,7 @@
 - [journal_io](#journal_io)
 - [journal_sector_buffer_count](#journal_sector_buffer_count)
 - [journal_no_same_sector_overwrites](#journal_no_same_sector_overwrites)
+- [skip_corrupted_meta_entries](#skip_corrupted_meta_entries)
 - [throttle_small_writes](#throttle_small_writes)
 - [throttle_target_iops](#throttle_target_iops)
 - [throttle_target_mbs](#throttle_target_mbs)
@@ -287,13 +288,19 @@ Flusher - это микро-поток (корутина), которая коп
 - Тип: булево (да/нет)
 - Значение по умолчанию: true

-Данный параметр заставляет Vitastor всегда держать область метаданных диска
-в памяти. Это нужно, чтобы избегать дополнительных операций чтения с диска
-при записи. Размер области метаданных на данный момент составляет примерно
-224 МБ на 1 ТБ данных. При включении потребление памяти снизится примерно
-на эту величину, но при этом также снизится и производительность. В будущем,
-после обновления схемы хранения метаданных, это ограничение, скорее всего,
-будет ликвидировано.
+Только для старого хранилища ([meta_format](layout-osd.en.md#meta_format) 2).
+
+Данный параметр заставляет Vitastor всегда держать копию области метаданных
+в памяти в том же виде, как она лежит на диске, в дополнение к БД метаданных.
+То есть, с включённой опцией каждая запись метаданных хранится в памяти дважды.
+Это нужно, чтобы избегать дополнительных операций чтения с диска при записи.
+Размер области метаданных в старом хранилище составляет примерно 224 МБ на
+1 ТБ данных. Вы можете отключить опцию, чтобы снизить потребление памяти
+примерно на эту величину, но при этом также снизится и производительность.
+
+Для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3) опция,
+возможно, будет переработана в будущем для поддержки работы без полной
+загрузки метаданных в памяти.

 ## inmemory_journal

@@ -376,6 +383,8 @@ fsync небезопасным даже с режимом "directsync".
 нужно менять - это если вы включаете journal_no_same_sector_overwrites. В
 этом случае установите данный параметр, например, в 1024.

+Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
+
 ## journal_no_same_sector_overwrites

 - Тип: булево (да/нет)
@@ -391,6 +400,18 @@ fsync небезопасным даже с режимом "directsync".

 Почти все другие SSD (99% моделей) не требуют данной опции.

+Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
+
+## skip_corrupted_meta_entries
+
+- Тип: булево (да/нет)
+- Значение по умолчанию: false
+
+Только для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3).
+Разрешить OSD запускаться, даже если часть блоков или записей метаданных
+повреждена, пропуская их. Опция предназначена для использования только в
+целях аварийного восстановления.
+
 ## throttle_small_writes

 - Тип: булево (да/нет)
@@ -84,11 +84,6 @@
    unsupported with old libibverbs < v32, like in Debian 10 Buster or
    CentOS 7.

-    Vitastor supports all adapters, even ones without ODP support, like
-    Mellanox ConnectX-3 and non-Mellanox cards. Versions up to Vitastor
-    1.2.0 required ODP which is only present in Mellanox ConnectX >= 4.
-    See also [rdma_odp](#rdma_odp).
-
    Run `ibv_devinfo -v` as root to list available RDMA devices and their
    features.

@@ -97,6 +92,23 @@
    the manual of your network vendor for details about setting up the switch
    for RoCEv2 correctly. Usually it means setting up Lossless Ethernet with
    PFC (Priority Flow Control) and ECN (Explicit Congestion Notification).
+
+    Vitastor supports all adapters, even ones without ODP (On-Demand Paging)
+    support, like Mellanox ConnectX-3 and non-Mellanox cards. ODP is only present
+    in Mellanox ConnectX >= 4 adapters and allows to skip memory registration
+    for RDMA and thus, in theory, avoid memory copying.
+
+    Versions up to Vitastor 1.2.0 required ODP, then it was disabled by default,
+    but it was still supported up to 3.0.3. Now ODP support is removed because it
+    actually only hurts performance: an example 3-node cluster with 8 NVMe in each
+    node and 2*25 GBit/s ConnectX-6 RDMA network pushed 3950000 read iops without
+    ODP, but only 239000 iops with ODP.
+
+    This happens because Mellanox ODP implementation seems to be based on
+    message retransmissions when the adapter doesn't know about the buffer yet -
+    it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
+    which is generally slow in RDMA/RoCE networks. Here's a presentation about
+    it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
  info_ru: |
    Название RDMA-устройства для связи с Vitastor OSD (например, "rocep5s0f0").
    Если не указано, Vitastor попробует найти RoCE-устройство, соответствующее
@@ -105,12 +117,6 @@
    не задана. Также автовыбор не поддерживается со старыми версиями библиотеки
    libibverbs < v32, например в Debian 10 Buster или CentOS 7.

-    Vitastor поддерживает все модели адаптеров, включая те, у которых
-    нет поддержки ODP, то есть вы можете использовать RDMA с ConnectX-3 и
-    картами производства не Mellanox. Версии Vitastor до 1.2.0 включительно
-    требовали ODP, который есть только на Mellanox ConnectX 4 и более новых.
-    См. также [rdma_odp](#rdma_odp).
-
    Запустите `ibv_devinfo -v` от имени суперпользователя, чтобы посмотреть
    список доступных RDMA-устройств, их параметры и возможности.

@@ -120,6 +126,24 @@
    коммутатора для RoCEv2 ищите в документации производителя. Обычно это
    подразумевает настройку сети без потерь на основе PFC (Priority Flow
    Control) и ECN (Explicit Congestion Notification).
+
+    Vitastor поддерживает все модели адаптеров, включая те, у которых нет
+    поддержки ODP (On-Demand Paging), например, ConnectX-3 и карты производства
+    не Mellanox. Функция ODP доступна только на адаптерах Mellanox ConnectX-4 и
+    более новых и позволяет не регистрировать память для её использования RDMA-картой,
+    благодаря чему в теории можно избежать лишних копирований памяти.
+
+    Версии Vitastor до 1.2.0 включительно требовали ODP, потом функция был отключена
+    по умолчанию, но поддерживалась вплоть до версии 3.0.3. Сейчас поддержка ODP
+    полностью удалена, так как на самом деле она только портит производительность:
+    например, на 3-узловом кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на
+    чтение с RDMA без ODP удаётся снять 3950000 iops, а с ODP - всего 239000 iops.
+
+    Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
+    основана на повторной передаче сообщений, когда карте не известен буфер -
+    вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
+    А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
+    Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
 - name: rdma_port_num
  type: int
  info: |
@@ -218,45 +242,6 @@
    у принимающей стороны в процессе работы не заканчивались буферы на приём.
    Не влияет на потребление памяти - дополнительная память на операции отправки
    не выделяется.
- name: rdma_odp
-  type: bool
-  default: false
-  online: false
-  info: |
-    Use RDMA with On-Demand Paging. ODP is currently only available on Mellanox
-    ConnectX-4 and newer adapters. ODP allows to not register memory explicitly
-    for RDMA adapter to be able to use it. This, in turn, allows to skip memory
-    copying during sending. One would think this should improve performance, but
-    **in reality** RDMA performance with ODP is **drastically** worse. Example
-    3-node cluster with 8 NVMe in each node and 2*25 GBit/s ConnectX-6 RDMA network
-    without ODP pushes 3950000 read iops, but only 239000 iops with ODP...
-
-    This happens because Mellanox ODP implementation seems to be based on
-    message retransmissions when the adapter doesn't know about the buffer yet -
-    it likely uses standard "RNR retransmissions" (RNR = receiver not ready)
-    which is generally slow in RDMA/RoCE networks. Here's a presentation about
-    it from ISPASS-2021 conference: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
-
-    ODP support is retained in the code just in case a good ODP implementation
-    appears one day.
-  info_ru: |
-    Использовать RDMA с On-Demand Paging. ODP - функция, доступная пока что
-    исключительно на адаптерах Mellanox ConnectX-4 и более новых. ODP позволяет
-    не регистрировать память для её использования RDMA-картой. Благодаря этому
-    можно не копировать данные при отправке их в сеть и, казалось бы, это должно
-    улучшать производительность - но **по факту** получается так, что
-    производительность только ухудшается, причём сильно. Пример - на 3-узловом
-    кластере с 8 NVMe в каждом узле и сетью 2*25 Гбит/с на чтение с RDMA без ODP
-    удаётся снять 3950000 iops, а с ODP - всего 239000 iops...
-
-    Это происходит из-за того, что реализация ODP у Mellanox неоптимальная и
-    основана на повторной передаче сообщений, когда карте не известен буфер -
-    вероятно, на стандартных "RNR retransmission" (RNR = receiver not ready).
-    А данные повторные передачи в RDMA/RoCE - всегда очень медленная штука.
-    Презентация на эту тему с конференции ISPASS-2021: https://tkygtr6.github.io/pub/ISPASS21_slides.pdf
-
-    Возможность использования ODP сохранена в коде на случай, если вдруг в один
-    прекрасный день появится хорошая реализация ODP.
 - name: peer_connect_interval
  type: sec
  min: 1
@@ -253,21 +253,33 @@
  type: bool
  default: true
  info: |
-    This parameter makes Vitastor always keep metadata area of the block device
-    in memory. It's required for good performance because it allows to avoid
-    additional read-modify-write cycles during metadata modifications. Metadata
-    area size is currently roughly 224 MB per 1 TB of data. You can turn it off
-    to reduce memory usage by this value, but it will hurt performance. This
-    restriction is likely to be removed in the future along with the upgrade
-    of the metadata storage scheme.
+    Only for the old store ([meta_format](layout-osd.en.md#meta_format) 2).
+
+    This parameter makes Vitastor keep a copy of metadata area in memory as it is
+    on disk, in addition to the metadata database. When the option is enabled, every
+    metadata entry is effectively stored in RAM twice. It's required for good performance
+    because it allows to avoid additional read-modify-write cycles during metadata
+    modifications. Metadata area size with the old store is roughly 224 MB per 1 TB
+    of data. You can turn the option off to reduce memory usage by this value, but
+    it will reduce performance.
+
+    For the new store ([meta_format](layout-osd.en.md#meta_format) 3), the option
+    may be changed in the future to support operation without loading full metadata
+    database in memory.
  info_ru: |
-    Данный параметр заставляет Vitastor всегда держать область метаданных диска
-    в памяти. Это нужно, чтобы избегать дополнительных операций чтения с диска
-    при записи. Размер области метаданных на данный момент составляет примерно
-    224 МБ на 1 ТБ данных. При включении потребление памяти снизится примерно
-    на эту величину, но при этом также снизится и производительность. В будущем,
-    после обновления схемы хранения метаданных, это ограничение, скорее всего,
-    будет ликвидировано.
+    Только для старого хранилища ([meta_format](layout-osd.en.md#meta_format) 2).
+
+    Данный параметр заставляет Vitastor всегда держать копию области метаданных
+    в памяти в том же виде, как она лежит на диске, в дополнение к БД метаданных.
+    То есть, с включённой опцией каждая запись метаданных хранится в памяти дважды.
+    Это нужно, чтобы избегать дополнительных операций чтения с диска при записи.
+    Размер области метаданных в старом хранилище составляет примерно 224 МБ на
+    1 ТБ данных. Вы можете отключить опцию, чтобы снизить потребление памяти
+    примерно на эту величину, но при этом также снизится и производительность.
+
+    Для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3) опция,
+    возможно, будет переработана в будущем для поддержки работы без полной
+    загрузки метаданных в памяти.
 - name: inmemory_journal
  type: bool
  default: true
@@ -386,11 +398,15 @@
    blocks. The only situation when you should increase it to a larger value
    is when you enable journal_no_same_sector_overwrites. In this case set
    it to, for example, 1024.
+
+    Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
  info_ru: |
    Максимальное число буферов, разрешённых для использования под записываемые
    в журнал блоки метаданных. Единственная ситуация, в которой этот параметр
    нужно менять - это если вы включаете journal_no_same_sector_overwrites. В
    этом случае установите данный параметр, например, в 1024.
+
+    Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
 - name: journal_no_same_sector_overwrites
  type: bool
  default: false
@@ -402,6 +418,8 @@
    journal after writing it instead of possibly overwriting it the second time.

    Most (99%) other SSDs don't need this option.
+
+    Not applicable to the new store ([meta_format](layout-osd.en.md#meta_format) 3).
  info_ru: |
    Включайте данную опцию для SSD вроде Intel D3-S4510 и D3-S4610, которые
    ОЧЕНЬ не любят, когда ПО перезаписывает один и тот же сектор несколько раз
@@ -412,6 +430,20 @@
    самого сектора.

    Почти все другие SSD (99% моделей) не требуют данной опции.
+
+    Неприменимо к новому хранилищу ([meta_format](layout-osd.en.md#meta_format) 3).
+- name: skip_corrupted_meta_entries
+  type: bool
+  default: false
+  info: |
+    Only for the new store ([meta_format](layout-osd.en.md#meta_format) 3).
+    Allow OSD to start when some metadata entries or blocks are corrupted by
+    skipping them. Should be only used as an emergency measure.
+  info_ru: |
+    Только для нового хранилища ([meta_format](layout-osd.en.md#meta_format) 3).
+    Разрешить OSD запускаться, даже если часть блоков или записей метаданных
+    повреждена, пропуская их. Опция предназначена для использования только в
+    целях аварийного восстановления.
 - name: throttle_small_writes
  type: bool
  default: false
@@ -26,13 +26,37 @@ at Vitastor Kubernetes operator: https://github.com/Antilles7227/vitastor-operat
 The instruction is very simple.

 1. Download a Docker image of the desired version: \
-   `docker pull vitalif/vitastor:v3.0.2`
+   `docker pull vitalif/vitastor:v3.0.6`
 2. Install scripts to the host system: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.2 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.6 install.sh`
 3. Reload udev rules: \
   `udevadm control --reload-rules`
+4. Enable the vitastor-host service: \
+   `systemctl enable --now vitastor-host`

-And you can return to [Quick Start](../intro/quickstart.en.md).
+After these steps, you can return to [Quick Start](../intro/quickstart.en.md).
+
+## Podman
+
+If you use Podman, run the following commands as root before installing Vitastor containers:
+
+```
+ln -s podman /usr/bin/docker
+
+mkdir -p /etc/systemd/system/systemd-udevd.service.d
+
+cat >/etc/systemd/system/systemd-udevd.service.d/override.conf <<EOF
+[Service]
+CapabilityBoundingSet=~
+SystemCallFilter=@mount capset
+EOF
+
+systemctl daemon-reload
+
+systemctl restart systemd-udevd
+```
+
+Without it, udev fails to do calls into a Podman container and Vitastor disk detection doesn't work.

 ## Upgrading Containers

@@ -25,14 +25,39 @@ Vitastor можно установить в Docker/Podman. При этом etcd,
 Инструкция по установке максимально простая.

 1. Скачайте Docker-образ желаемой версии: \
-   `docker pull vitalif/vitastor:v3.0.2`
+   `docker pull vitalif/vitastor:v3.0.6`
 2. Установите скрипты в хост-систему командой: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.2 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.6 install.sh`
 3. Перезагрузите правила udev: \
   `udevadm control --reload-rules`
+4. Включите сервис vitastor-host: \
+   `systemctl enable --now vitastor-host`

 После этого вы можете возвращаться к разделу [Быстрый старт](../intro/quickstart.ru.md).

+## Podman
+
+Если вы используете Podman, перед установкой контейнеров Vitastor выполните следующие
+команды от имени суперпользователя:
+
+```
+ln -s podman /usr/bin/docker
+
+mkdir -p /etc/systemd/system/systemd-udevd.service.d
+
+cat >/etc/systemd/system/systemd-udevd.service.d/override.conf <<EOF
+[Service]
+CapabilityBoundingSet=~
+SystemCallFilter=@mount capset
+EOF
+
+systemctl daemon-reload
+
+systemctl restart systemd-udevd
+```
+
+Без этих настроек udev не может делать вызовы внутрь Podman-контейнеров и определение дисков Vitastor не работает.
+
 ## Обновление контейнеров

 Сначала обязательно проверьте раздел [Обновление Vitastor](../usage/admin.ru.md#обновление-vitastor),
@@ -33,15 +33,17 @@
  - CentOS 7: `yum install https://vitastor.io/rpms/centos/7/vitastor-release.rpm`
  - CentOS 8: `dnf install https://vitastor.io/rpms/centos/8/vitastor-release.rpm`
  - AlmaLinux 9 and other RHEL 9 clones (Rocky, Oracle...): `dnf install https://vitastor.io/rpms/centos/9/vitastor-release.rpm`
+  - AlmaLinux 10 and other RHEL 10 clones: `dnf install https://vitastor.io/rpms/centos/10/vitastor-release.rpm`
 - Enable EPEL: `yum/dnf install epel-release`
 - Enable additional CentOS repositories:
  - CentOS 7: `yum install centos-release-scl`
  - CentOS 8: `dnf install centos-release-advanced-virtualization`
-  - RHEL 9 clones: not required
+  - RHEL 9/10 clones: not required
 - Enable elrepo-kernel:
  - CentOS 7: `yum install https://www.elrepo.org/elrepo-release-7.el7.elrepo.noarch.rpm`
  - CentOS 8: `dnf install https://www.elrepo.org/elrepo-release-8.el8.elrepo.noarch.rpm`
  - RHEL 9 clones: `dnf install https://www.elrepo.org/elrepo-release-9.el9.elrepo.noarch.rpm`
+  - RHEL 10 clones: not required
 - Install packages: `yum/dnf install vitastor lpsolve etcd kernel-ml qemu-kvm`

 ## Installation requirements
@@ -33,15 +33,17 @@
  - CentOS 7: `yum install https://vitastor.io/rpms/centos/7/vitastor-release.rpm`
  - CentOS 8: `dnf install https://vitastor.io/rpms/centos/8/vitastor-release.rpm`
  - AlmaLinux 9 и другие клоны RHEL 9 (Rocky, Oracle...): `dnf install https://vitastor.io/rpms/centos/9/vitastor-release.rpm`
+  - AlmaLinux 10 и другие клоны RHEL 10: `dnf install https://vitastor.io/rpms/centos/10/vitastor-release.rpm`
 - Включите EPEL: `yum/dnf install epel-release`
 - Включите дополнительные репозитории CentOS:
  - CentOS 7: `yum install centos-release-scl`
  - CentOS 8: `dnf install centos-release-advanced-virtualization`
-  - Клоны RHEL 9: не нужно
+  - Клоны RHEL 9/10: не нужно
 - Включите elrepo-kernel:
  - CentOS 7: `yum install https://www.elrepo.org/elrepo-release-7.el7.elrepo.noarch.rpm`
  - CentOS 8: `dnf install https://www.elrepo.org/elrepo-release-8.el8.elrepo.noarch.rpm`
  - Клоны RHEL 9: `dnf install https://www.elrepo.org/elrepo-release-9.el9.elrepo.noarch.rpm`
+  - Клоны RHEL 10: не нужно
 - Установите пакеты: `yum/dnf install vitastor lpsolve etcd kernel-ml qemu-kvm`

 ## Установочные требования
@@ -16,7 +16,7 @@ async function create_http_server(cfg, handler)
        };
        if (cfg.mon_https_ca)
        {
-            tls.mon_https_ca = await fsp.readFile(cfg.mon_https_ca);
+            tls.ca = await fsp.readFile(cfg.mon_https_ca);
        }
        if (cfg.mon_https_client_auth)
        {
@@ -10,16 +10,19 @@ const NO_OSD = 'Z';
 async function lp_solve(text)
 {
    const cp = child_process.spawn('lp_solve');
-    let stdout = '', stderr = '', finish_cb;
+    let stdout = '', stderr = '', finish_cb, finished = 0;
    cp.stdout.on('data', buf => stdout += buf.toString());
    cp.stderr.on('data', buf => stderr += buf.toString());
-    cp.on('exit', () => finish_cb && finish_cb());
+    cp.stdout.on('end', () => finish_cb());
+    cp.stderr.on('end', () => finish_cb());
    cp.stdin.write(text);
    cp.stdin.end();
-    if (cp.exitCode == null)
+    await new Promise(ok => (finish_cb = () =>
    {
-        await new Promise(ok => finish_cb = ok);
-    }
+        finished++;
+        if (finished == 2)
+            ok();
+    }));
    if (!stdout.trim())
    {
        return null;
@@ -1,6 +1,6 @@
 {
  "name": "vitastor-mon",
-  "version": "3.0.2",
+  "version": "3.0.6",
  "description": "Vitastor SDS monitor service",
  "main": "mon-main.js",
  "scripts": {
@@ -9,7 +9,7 @@
  "author": "Vitaliy Filippov",
  "license": "UNLICENSED",
  "dependencies": {
-    "antietcd": "^1.2.2",
+    "antietcd": "^1.2.4",
    "sprintf-js": "^1.1.2",
    "ws": "^7.2.5"
  },
@@ -195,7 +195,7 @@ async function generate_pool_pgs(state, global_config, pool_id, osd_tree, levels
    const folded_tree = make_hier_tree(global_config, folded.nodes.reduce((a, c) => { a[c.id] = c; return a; }, {}));
    const old_pg_count = prev_pgs.length;
    const optimize_cfg = {
-        osd_weights: folded.nodes.reduce((a, c) => { if (Number(c.id)) { a[c.id] = c.size; } return a; }, {}),
+        osd_weights: folded.nodes.reduce((a, c) => { if (/^\d+$/.exec(c.id) && c.size != null) { a[c.id] = c.size||0; } return a; }, {}),
        combinator: use_rules
            // new algorithm:
            ? new RuleCombinator(folded_tree, rules, pool_cfg.max_osd_combinations)
@@ -1,6 +1,6 @@
 {
  "name": "vitastor",
-  "version": "3.0.2",
+  "version": "3.0.6",
  "description": "Low-level native bindings to Vitastor client library",
  "main": "index.js",
  "keywords": [
@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils

-VITASTOR_VERSION = '3.0.2'
+VITASTOR_VERSION = '3.0.6'

 LOG = logging.getLogger(__name__)

@@ -275,7 +275,7 @@ class VitastorDriver(driver.CloneableImageVD,
            LOG.exception('error getting vitastor pool stats: '+str(e))

        self._stats = stats
-        
+
    def get_volume_stats(self, refresh=False):
        """Get volume stats.
        If 'refresh' is True, run update the stats first.
@@ -291,6 +291,14 @@ class VitastorDriver(driver.CloneableImageVD,
        else:
            return (1 + resp['kvs'][0]['value'], resp['kvs'][0]['mod_revision'])

+    def _cli(self, descr, *args):
+        args = [ 'vitastor-cli', *args, *(self._vitastor_args()) ]
+        try:
+            self._execute(*args)
+        except processutils.ProcessExecutionError as exc:
+            LOG.error("Failed to "+descr+": "+exc)
+            raise exception.VolumeBackendAPIException(data = exc.stderr)
+
    def create_volume(self, volume):
        """Creates a logical volume."""

@@ -302,7 +310,7 @@ class VitastorDriver(driver.CloneableImageVD,

        LOG.debug("creating volume '%s'", vol_name)

-        self._create_image(vol_name, { 'size': size })
+        self._cli('create volume', 'create', vol_name, '--size', size)

        if volume.encryption_key_id:
            self._create_encrypted_volume(volume, volume.obj_context)
@@ -346,7 +354,7 @@ class VitastorDriver(driver.CloneableImageVD,
        snap_name = utils.convert_str(snapshot.name)
        if snap_name.find('@') >= 0 or snap_name.find('/') >= 0:
            raise exception.VolumeBackendAPIException(data = '@ and / are forbidden in volume and snapshot names')
-        self._create_snapshot(vol_name, vol_name+'@'+snap_name)
+        self._cli('create snapshot', 'snap-create', vol_name+'@'+snap_name)

    def snapshot_revert_use_temp_snapshot(self):
        """Disable the use of a temporary snapshot on revert."""
@@ -359,21 +367,8 @@ class VitastorDriver(driver.CloneableImageVD,
        snap_name = utils.convert_str(snapshot.name)

        # Delete the image and recreate it from the snapshot
-        args = [ 'vitastor-cli', 'rm', vol_name, *(self._vitastor_args()) ]
-        try:
-            self._execute(*args)
-        except processutils.ProcessExecutionError as exc:
-            LOG.error("Failed to delete image "+vol_name+": "+exc)
-            raise exception.VolumeBackendAPIException(data = exc.stderr)
-        args = [
-            'vitastor-cli', 'create', '--parent', vol_name+'@'+snap_name,
-            vol_name, *(self._vitastor_args())
-        ]
-        try:
-            self._execute(*args)
-        except processutils.ProcessExecutionError as exc:
-            LOG.error("Failed to recreate image "+vol_name+" from "+vol_name+"@"+snap_name+": "+exc)
-            raise exception.VolumeBackendAPIException(data = exc.stderr)
+        self._cli('delete image', 'rm', vol_name)
+        self._cli('recreate image', 'create', '--parent', vol_name+'@'+snap_name, vol_name)

    def delete_snapshot(self, snapshot):
        """Deletes a snapshot."""
@@ -381,15 +376,7 @@ class VitastorDriver(driver.CloneableImageVD,
        vol_name = utils.convert_str(snapshot.volume_name)
        snap_name = utils.convert_str(snapshot.name)

-        args = [
-            'vitastor-cli', 'rm', vol_name+'@'+snap_name,
-            *(self._vitastor_args())
-        ]
-        try:
-            self._execute(*args)
-        except processutils.ProcessExecutionError as exc:
-            LOG.error("Failed to remove snapshot "+vol_name+'@'+snap_name+": "+exc)
-            raise exception.VolumeBackendAPIException(data = exc.stderr)
+        self._cli('remove snapshot', 'rm', vol_name+'@'+snap_name)

    def _child_count(self, parents):
        children = 0
@@ -427,13 +414,7 @@ class VitastorDriver(driver.CloneableImageVD,
        if src_vref.admin_metadata.get('readonly') == 'True':
            # source volume is a volume-image cache entry or other readonly volume
            # clone without intermediate snapshot
-            src = self._get_image(src_name)
-            LOG.debug("creating image '%s' from '%s'", dest_name, src_name)
-            new_cfg = self._create_image(dest_name, {
-                'size': size,
-                'parent_id': src['idx']['id'],
-                'parent_pool_id': src['idx']['pool_id'],
-            })
+            self._cli('create clone', 'create', '--parent', src_name, '--size', size, dest_name)
            return {}

        clone_snap = "%s@%s.clone_snap" % (src_name, dest_name)
@@ -446,15 +427,12 @@ class VitastorDriver(driver.CloneableImageVD,
            clone_snap = dest_name
            make_img = False

-        LOG.debug("creating layer '%s' under '%s'", clone_snap, src_name)
-        new_cfg = self._create_snapshot(src_name, clone_snap, True)
+        LOG.debug("creating snapshot '%s'", clone_snap)
+        self._cli('create base snapshot', 'snap-create', '--allow-existing', '1', clone_snap)
+
        if make_img:
            # Then create a clone from it
-            new_cfg = self._create_image(dest_name, {
-                'size': size,
-                'parent_id': new_cfg['parent_id'],
-                'parent_pool_id': new_cfg['parent_pool_id'],
-            })
+            self._cli('create clone', 'create', '--parent', clone_snap, '--size', size, dest_name)

        return {}

@@ -464,7 +442,8 @@ class VitastorDriver(driver.CloneableImageVD,
        vol_name = utils.convert_str(volume.name)
        snap_name = utils.convert_str(snapshot.name)

-        snap = self._get_image('volume-'+snapshot.volume_id+'@'+snap_name)
+        src_snap = 'volume-'+snapshot.volume_id+'@'+snap_name
+        snap = self._get_image(src_snap)
        if not snap:
            raise exception.SnapshotNotFound(snapshot_id = snap_name)
        snap_inode_id = int(resp['responses'][0]['kvs'][0]['value']['id'])
@@ -473,12 +452,8 @@ class VitastorDriver(driver.CloneableImageVD,
        size = snap['cfg']['size']
        if int(volume.size):
            size = int(volume.size) * units.Gi
-        new_cfg = self._create_image(vol_name, {
-            'size': size,
-            'parent_id': snap['idx']['id'],
-            'parent_pool_id': snap['idx']['pool_id'],
-        })

+        self._cli('create clone', 'create', vol_name, '--size', size, '--parent', src_snap)
        return {}

    def _vitastor_args(self):
@@ -505,49 +480,7 @@ class VitastorDriver(driver.CloneableImageVD,
        """Deletes a logical volume."""

        vol_name = utils.convert_str(volume.name)
-
-        # Find the volume and all its snapshots
-        range_end = b'index/image/' + vol_name.encode('utf-8')
-        range_end = range_end[0 : len(range_end)-1] + six.int2byte(range_end[len(range_end)-1] + 1)
-        resp = self._etcd_txn({ 'success': [
-            { 'request_range': { 'key': 'index/image/'+vol_name, 'range_end': range_end } },
-        ] })
-        if len(resp['responses'][0]['kvs']) == 0:
-            # already deleted
-            LOG.info("volume %s no longer exists in backend", vol_name)
-            return
-        layers = resp['responses'][0]['kvs']
-        layer_ids = {}
-        for kv in layers:
-            inode_id = int(kv['value']['id'])
-            pool_id = int(kv['value']['pool_id'])
-            inode_pool_id = (pool_id << 48) | (inode_id & 0xffffffffffff)
-            layer_ids[inode_pool_id] = True
-
-        # Check if the volume has clones and raise 'busy' if so
-        children = self._child_count(layer_ids)
-        if children > 0:
-            raise exception.VolumeIsBusy(volume_name = vol_name)
-
-        # Clear data
-        for kv in layers:
-            args = [
-                'vitastor-cli', 'rm-data', '--pool', str(kv['value']['pool_id']),
-                '--inode', str(kv['value']['id']), '--progress', '0',
-                *(self._vitastor_args())
-            ]
-            try:
-                self._execute(*args)
-            except processutils.ProcessExecutionError as exc:
-                LOG.error("Failed to remove layer "+kv['key']+": "+exc)
-                raise exception.VolumeBackendAPIException(data = exc.stderr)
-
-        # Delete all layers from etcd
-        requests = []
-        for kv in layers:
-            requests.append({ 'request_delete_range': { 'key': kv['key'] } })
-            requests.append({ 'request_delete_range': { 'key': 'config/inode/'+str(kv['value']['pool_id'])+'/'+str(kv['value']['id']) } })
-        self._etcd_txn({ 'success': requests })
+        self._cli('delete volume', 'rm', '--matching', vol_name, vol_name+'@*', '--progress', '0')

    def retype(self, context, volume, new_type, diff, host):
        """Change extra type specifications for a volume."""
@@ -567,98 +500,6 @@ class VitastorDriver(driver.CloneableImageVD,
        """Removes an export for a logical volume."""
        pass

-    def _create_image(self, vol_name, cfg):
-        pool_s = str(self.cfg['pool_id'])
-        image_id = 0
-        while image_id == 0:
-            # check if the image already exists and find a free ID
-            resp = self._etcd_txn({ 'success': [
-                { 'request_range': { 'key': 'index/image/'+vol_name } },
-                { 'request_range': { 'key': 'index/maxid/'+pool_s } },
-            ] })
-            if len(resp['responses'][0]['kvs']) > 0:
-                # already exists
-                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' already exists')
-            image_id, id_mod = self._next_id(resp['responses'][1])
-            # try to create the image
-            resp = self._etcd_txn({ 'compare': [
-                { 'target': 'MOD', 'mod_revision': id_mod, 'key': 'index/maxid/'+pool_s },
-                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+vol_name },
-                { 'target': 'VERSION', 'version': 0, 'key': 'config/inode/'+pool_s+'/'+str(image_id) },
-            ], 'success': [
-                { 'request_put': { 'key': 'index/maxid/'+pool_s, 'value': image_id } },
-                { 'request_put': { 'key': 'index/image/'+vol_name, 'value': json.dumps({
-                    'id': image_id, 'pool_id': self.cfg['pool_id']
-                }) } },
-                { 'request_put': { 'key': 'config/inode/'+pool_s+'/'+str(image_id), 'value': json.dumps({
-                    **cfg, 'name': vol_name,
-                }) } },
-            ] })
-            if not resp.get('succeeded'):
-                # repeat
-                image_id = 0
-
-    def _create_snapshot(self, vol_name, snap_vol_name, allow_existing = False):
-        while True:
-            # check if the image already exists and snapshot doesn't
-            resp = self._etcd_txn({ 'success': [
-                { 'request_range': { 'key': 'index/image/'+vol_name } },
-                { 'request_range': { 'key': 'index/image/'+snap_vol_name } },
-            ] })
-            if len(resp['responses'][0]['kvs']) == 0:
-                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
-            if len(resp['responses'][1]['kvs']) > 0:
-                if allow_existing:
-                    snap_idx = resp['responses'][1]['kvs'][0]['value']
-                    resp = self._etcd_txn({ 'success': [
-                        { 'request_range': { 'key': 'config/inode/'+str(snap_idx['pool_id'])+'/'+str(snap_idx['id']) } },
-                    ] })
-                    if len(resp['responses'][0]['kvs']) == 0:
-                        raise exception.VolumeBackendAPIException(data =
-                            'Volume '+snap_vol_name+' is already indexed, but does not exist'
-                        )
-                    return resp['responses'][0]['kvs'][0]['value']
-                raise exception.VolumeBackendAPIException(
-                    data = 'Volume '+snap_vol_name+' already exists'
-                )
-            vol_idx = resp['responses'][0]['kvs'][0]['value']
-            vol_idx_mod = resp['responses'][0]['kvs'][0]['mod_revision']
-            # get image inode config and find a new ID
-            resp = self._etcd_txn({ 'success': [
-                { 'request_range': { 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']) } },
-                { 'request_range': { 'key': 'index/maxid/'+str(self.cfg['pool_id']) } },
-            ] })
-            if len(resp['responses'][0]['kvs']) == 0:
-                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
-            vol_cfg = resp['responses'][0]['kvs'][0]['value']
-            vol_mod = resp['responses'][0]['kvs'][0]['mod_revision']
-            new_id, id_mod = self._next_id(resp['responses'][1])
-            # try to redirect image to the new inode
-            new_cfg = {
-                **vol_cfg, 'name': vol_name, 'parent_id': vol_idx['id'], 'parent_pool_id': vol_idx['pool_id']
-            }
-            resp = self._etcd_txn({ 'compare': [
-                { 'target': 'MOD', 'mod_revision': vol_idx_mod, 'key': 'index/image/'+vol_name },
-                { 'target': 'MOD', 'mod_revision': vol_mod, 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']) },
-                { 'target': 'MOD', 'mod_revision': id_mod, 'key': 'index/maxid/'+str(self.cfg['pool_id']) },
-                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+snap_vol_name },
-                { 'target': 'VERSION', 'version': 0, 'key': 'config/inode/'+str(self.cfg['pool_id'])+'/'+str(new_id) },
-            ], 'success': [
-                { 'request_put': { 'key': 'index/maxid/'+str(self.cfg['pool_id']), 'value': new_id } },
-                { 'request_put': { 'key': 'index/image/'+vol_name, 'value': json.dumps({
-                    'id': new_id, 'pool_id': self.cfg['pool_id']
-                }) } },
-                { 'request_put': { 'key': 'config/inode/'+str(self.cfg['pool_id'])+'/'+str(new_id), 'value': json.dumps(new_cfg) } },
-                { 'request_put': { 'key': 'index/image/'+snap_vol_name, 'value': json.dumps({
-                    'id': vol_idx['id'], 'pool_id': vol_idx['pool_id']
-                }) } },
-                { 'request_put': { 'key': 'config/inode/'+str(vol_idx['pool_id'])+'/'+str(vol_idx['id']), 'value': json.dumps({
-                    **vol_cfg, 'name': snap_vol_name, 'readonly': True
-                }) } }
-            ] })
-            if resp.get('succeeded'):
-                return new_cfg
-
    def initialize_connection(self, volume, connector):
        data = {
            'driver_volume_type': 'vitastor',
@@ -697,13 +538,9 @@ class VitastorDriver(driver.CloneableImageVD,
                    size = int(volume.size) * units.Gi
                    dest_name = utils.convert_str(volume.name)
                    # Find or create the base snapshot
-                    snap_cfg = self._create_snapshot(base_vol.name, base_vol.name+'@.clone_snap', True)
+                    self._cli('create base snapshot', 'create', '--allow-existing', '1', base_vol.name+'@.clone_snap')
                    # Then create a clone from it
-                    new_cfg = self._create_image(dest_name, {
-                        'size': size,
-                        'parent_id': snap_cfg['parent_id'],
-                        'parent_pool_id': snap_cfg['parent_pool_id'],
-                    })
+                    self._cli('create clone', 'create', dest_name, '--size', size, '--parent', base_vol.name+'@.clone_snap')
                    return ({}, True)
        return ({}, False)

@@ -770,26 +607,8 @@ class VitastorDriver(driver.CloneableImageVD,
    def extend_volume(self, volume, new_size):
        """Extend an existing volume."""
        vol_name = utils.convert_str(volume.name)
-        while True:
-            vol = self._get_image(vol_name)
-            if not vol:
-                raise exception.VolumeBackendAPIException(data = 'Volume '+vol_name+' does not exist')
-            # change size
-            size = int(new_size) * units.Gi
-            if size == vol['cfg']['size']:
-                break
-            resp = self._etcd_txn({ 'compare': [ {
-                'target': 'MOD',
-                'mod_revision': vol['cfg_mod'],
-                'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
-            } ], 'success': [
-                { 'request_put': {
-                    'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
-                    'value': json.dumps({ **vol['cfg'], 'size': size }),
-                } },
-            ] })
-            if resp.get('succeeded'):
-                break
+        size = int(new_size) * units.Gi
+        self._cli('extend volume', 'modify', vol_name, '--resize', new_size)
        LOG.debug(
            "Extend volume from %(old_size)s GB to %(new_size)s GB.",
            {'old_size': volume.size, 'new_size': new_size}
@@ -862,28 +681,7 @@ class VitastorDriver(driver.CloneableImageVD,
        """
        from_name = self._get_existing_name(existing_ref)
        to_name = utils.convert_str(volume.name)
-        self._rename(from_name, to_name)
-
-    def _rename(self, from_name, to_name):
-        while True:
-            vol = self._get_image(from_name)
-            if not vol:
-                raise exception.VolumeBackendAPIException(data = 'Volume '+from_name+' does not exist')
-            to = self._get_image(to_name)
-            if to:
-                raise exception.VolumeBackendAPIException(data = 'Volume '+to_name+' already exists')
-            resp = self._etcd_txn({ 'compare': [
-                { 'target': 'MOD', 'mod_revision': vol['idx_mod'], 'key': 'index/image/'+vol['cfg']['name'] },
-                { 'target': 'MOD', 'mod_revision': vol['cfg_mod'], 'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']) },
-                { 'target': 'VERSION', 'version': 0, 'key': 'index/image/'+to_name },
-            ], 'success': [
-                { 'request_delete_range': { 'key': 'index/image/'+vol['cfg']['name'] } },
-                { 'request_put': { 'key': 'index/image/'+to_name, 'value': json.dumps(vol['idx']) } },
-                { 'request_put': { 'key': 'config/inode/'+str(vol['idx']['pool_id'])+'/'+str(vol['idx']['id']),
-                    'value': json.dumps({ **vol['cfg'], 'name': to_name }) } },
-            ] })
-            if resp.get('succeeded'):
-                break
+        self._cli('rename', 'modify', from_name, '--rename', to_name)

    def unmanage(self, volume):
        pass
@@ -956,7 +754,7 @@ class VitastorDriver(driver.CloneableImageVD,
        snap_name = self._get_existing_name(existing_ref)
        from_name = vol_name+'@'+snap_name
        to_name = vol_name+'@'+utils.convert_str(snapshot.name)
-        self._rename(from_name, to_name)
+        self._cli('rename', 'modify', from_name, '--rename', to_name)

    def unmanage_snapshot(self, snapshot):
        """Removes the specified snapshot from Cinder management."""
@@ -21,7 +21,7 @@ rpmbuild -bp fio.spec
 cd $VITASTOR
 VER=$(grep ^Version: rpm/vitastor-$REL.spec | awk '{print $2}')
 rm -rf fio
-ln -s ~/rpmbuild/BUILD/fio*/ fio
+ln -s $(ls -d ~/rpmbuild/BUILD/fio*/ | grep -v SPECPARTS) fio
 sh copy-fio-includes.sh
 rm fio
 mv fio-copy fio
@@ -0,0 +1,17 @@
+# Build packages for AlmaLinux 10 inside a container
+# cd ..
+# docker pull --platform=linux/amd64/v2 quay.io/almalinuxorg/almalinux:10
+# docker build -t vitastor-buildenv:el10 -f rpm/vitastor-el10.Dockerfile .
+# docker run -i --rm -v ./:/root/vitastor vitastor-buildenv:el10 /root/vitastor/rpm/vitastor-build.sh
+
+FROM quay.io/almalinuxorg/almalinux:10
+
+WORKDIR /root
+
+RUN sed -i 's/enabled=0/enabled=1/' /etc/yum.repos.d/*.repo
+RUN dnf -y install epel-release dnf-plugins-core
+RUN dnf -y install https://vitastor.io/rpms/centos/10/vitastor-release-1.0-1.el10.noarch.rpm
+RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel isa-l-devel gf-complete-devel rdma-core-devel cmake libnl3-devel
+RUN dnf download --source fio
+RUN rpm --nomd5 -i fio*.src.rpm
+RUN cd ~/rpmbuild/SPECS && dnf builddep -y --spec fio.spec
@@ -0,0 +1,198 @@
+Name:           vitastor
+Version:        3.0.6
+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-3.0.6.el10.tar.gz
+
+BuildRequires:  gperftools-devel
+BuildRequires:  gcc-c++
+BuildRequires:  nodejs >= 10
+BuildRequires:  jerasure-devel
+BuildRequires:  isa-l-devel
+BuildRequires:  gf-complete-devel
+BuildRequires:  rdma-core-devel
+BuildRequires:  cmake
+BuildRequires:  libnl3-devel
+Requires:       vitastor-osd = %{version}-%{release}
+Requires:       vitastor-mon = %{version}-%{release}
+Requires:       vitastor-client = %{version}-%{release}
+Requires:       vitastor-client-devel = %{version}-%{release}
+Requires:       vitastor-fio = %{version}-%{release}
+
+%description
+Vitastor is a small, simple and fast clustered block storage (storage for VM drives),
+architecturally similar to Ceph which means strong consistency, primary-replication,
+symmetric clustering and automatic data distribution over any number of drives of any
+size with configurable redundancy (replication or erasure codes/XOR).
+
+
+%package -n vitastor-osd
+Summary:        Vitastor - OSD
+Requires:       vitastor-client = %{version}-%{release}
+Requires:       util-linux
+Requires:       parted
+
+
+%description -n vitastor-osd
+Vitastor object storage daemon, i.e. server program that stores data.
+
+
+%package -n vitastor-mon
+Summary:        Vitastor - monitor
+Requires:       nodejs >= 10
+Requires:       lpsolve
+
+
+%description -n vitastor-mon
+Vitastor monitor, i.e. server program responsible for watching cluster state and
+scheduling cluster-level operations.
+
+
+%package -n vitastor-client
+Summary:        Vitastor - client
+
+
+%description -n vitastor-client
+Vitastor client library and command-line interface.
+
+
+%package -n vitastor-client-devel
+Summary:        Vitastor - development files
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+
+
+%description -n vitastor-client-devel
+Vitastor library headers for development.
+
+
+%package -n vitastor-fio
+Summary:        Vitastor - fio drivers
+Group:          Development/Libraries
+Requires:       vitastor-client = %{version}-%{release}
+Requires:       fio = 3.36-5.el10
+
+
+%description -n vitastor-fio
+Vitastor fio drivers for benchmarking.
+
+
+%package -n vitastor-opennebula
+Summary:        Vitastor for OpenNebula
+Group:          Development/Libraries
+Requires:       vitastor-client
+Requires:       jq
+Requires:       python3-lxml
+Requires:       patch
+Requires:       qemu-kvm-block-vitastor
+
+
+%description -n vitastor-opennebula
+Vitastor storage plugin for OpenNebula.
+
+
+%prep
+%setup -q
+
+
+%build
+%cmake
+%cmake_build
+
+
+%install
+rm -rf $RPM_BUILD_ROOT
+%cmake_install
+cd mon
+npm install --production
+cd ..
+mkdir -p %buildroot/usr/lib/vitastor
+cp -r mon %buildroot/usr/lib/vitastor
+mv %buildroot/usr/lib/vitastor/mon/scripts/make-etcd %buildroot/usr/lib/vitastor/mon/
+mkdir -p %buildroot/lib/systemd/system
+cp mon/scripts/vitastor.target mon/scripts/vitastor-mon.service mon/scripts/vitastor-osd@.service %buildroot/lib/systemd/system
+mkdir -p %buildroot/lib/udev/rules.d
+cp mon/scripts/90-vitastor.rules %buildroot/lib/udev/rules.d
+mkdir -p %buildroot/var/lib/one
+cp -r opennebula/remotes %buildroot/var/lib/one
+cp opennebula/install.sh %buildroot/var/lib/one/remotes/datastore/vitastor/
+mkdir -p %buildroot/etc/
+cp -r opennebula/sudoers.d %buildroot/etc/
+
+
+%files
+%doc GPL-2.0.txt VNPL-1.1.txt README.md README-ru.md
+
+
+%files -n vitastor-osd
+%_bindir/vitastor-osd
+%_bindir/vitastor-disk
+%_bindir/vitastor-dump-journal
+/lib/systemd/system/vitastor-osd@.service
+/lib/systemd/system/vitastor.target
+/lib/udev/rules.d/90-vitastor.rules
+
+
+%pre -n vitastor-osd
+groupadd -r -f vitastor 2>/dev/null ||:
+useradd -r -g vitastor -s /sbin/nologin -c "Vitastor daemons" -M -d /nonexistent vitastor 2>/dev/null ||:
+install -o vitastor -g vitastor -d /var/log/vitastor
+mkdir -p /etc/vitastor
+
+
+%files -n vitastor-mon
+/usr/lib/vitastor/mon
+/lib/systemd/system/vitastor-mon.service
+
+
+%pre -n vitastor-mon
+groupadd -r -f vitastor 2>/dev/null ||:
+useradd -r -g vitastor -s /sbin/nologin -c "Vitastor daemons" -M -d /nonexistent vitastor 2>/dev/null ||:
+mkdir -p /etc/vitastor
+mkdir -p /var/lib/vitastor
+chown vitastor:vitastor /var/lib/vitastor
+
+
+%files -n vitastor-client
+%_bindir/vitastor-nbd
+%_bindir/vitastor-ublk
+%_bindir/vitastor-nfs
+%_bindir/vitastor-cli
+%_bindir/vitastor-rm
+%_bindir/vitastor-kv
+%_bindir/vitastor-kv-stress
+%_bindir/vita
+%_libdir/libvitastor_client.so*
+%_libdir/libvitastor_kv.so*
+
+
+%files -n vitastor-client-devel
+%_includedir/vitastor_c.h
+%_includedir/vitastor_kv.h
+%_libdir/pkgconfig
+
+
+%files -n vitastor-fio
+%_libdir/libfio_vitastor.so
+%_libdir/libfio_vitastor_blk.so
+%_libdir/libfio_vitastor_sec.so
+
+
+%files -n vitastor-opennebula
+/var/lib/one
+/etc/sudoers.d/opennebula-vitastor
+
+
+%triggerin -n vitastor-opennebula -- opennebula
+[ $2 = 0 ] || exit 0
+/var/lib/one/remotes/datastore/vitastor/install.sh
+
+
+# Turn off the brp-python-bytecompile script
+%global __os_install_post %(echo '%{__os_install_post}' | sed -e 's!/usr/lib[^[:space:]]*/brp-python-bytecompile[[:space:]].*$!!g')
+
+
+%changelog
@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.2
+Version:        3.0.6
 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-3.0.2.el7.tar.gz
+Source0:        vitastor-3.0.6.el7.tar.gz

 BuildRequires:  gperftools-devel
 BuildRequires:  devtoolset-9-gcc-c++
@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.2
+Version:        3.0.6
 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-3.0.2.el8.tar.gz
+Source0:        vitastor-3.0.6.el8.tar.gz

 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-toolset-9-gcc-c++
@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.2
+Version:        3.0.6
 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-3.0.2.el9.tar.gz
+Source0:        vitastor-3.0.6.el9.tar.gz

 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-c++
@@ -21,7 +21,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 endif()
 set(ENABLE_COVERAGE false CACHE BOOL "Enable code coverage")

-add_definitions(-DVITASTOR_VERSION="3.0.2")
+add_definitions(-DVITASTOR_VERSION="3.0.6")
 add_definitions(-D_GNU_SOURCE -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 -fvisibility=hidden -I ${CMAKE_SOURCE_DIR}/src)
 add_link_options(-fno-omit-frame-pointer)
 if (${WITH_ASAN})
@@ -187,7 +187,6 @@ public:
    // MUST be called only when nobody makes any modifications to the DB for this pool
    virtual void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit) = 0;
    virtual bool reshard_continue(void *reshard_state, uint64_t chunk_limit) = 0;
-    virtual void reshard_abort(void *reshard_state) = 0;

    // Event loop
    virtual void loop() = 0;
@@ -171,6 +171,10 @@ void blockstore_disk_t::parse_config(std::map<std::string, std::string> & config
    {
        throw std::runtime_error("Data block size must be a multiple of sparse write tracking granularity");
    }
+    if (data_block_size / bitmap_granularity < 8)
+    {
+        throw std::runtime_error("Data block size must be at least bitmap_granularity*8");
+    }
    if (!data_csum_type)
    {
        csum_block_size = 0;
@@ -259,7 +263,7 @@ void blockstore_disk_t::calc_lengths(bool skip_meta_check)
    }
    // required metadata size
    block_count = data_len / data_block_size;
-    clean_entry_bitmap_size = data_block_size / bitmap_granularity / 8;
+    clean_entry_bitmap_size = (data_block_size / bitmap_granularity + 7) / 8;
    clean_dyn_size = clean_entry_bitmap_size*2 + (csum_block_size
        ? data_block_size/csum_block_size*(data_csum_type & 0xFF) : 0);
 recalc:
@@ -26,14 +26,14 @@ class allocator_t;
 struct blockstore_disk_t
 {
    std::string data_device, meta_device, journal_device;
-    uint32_t data_block_size;
+    uint64_t data_block_size;
    uint64_t cfg_journal_size, cfg_data_size;
    // Required write alignment and journal/metadata/data areas' location alignment
    uint32_t disk_alignment = 4096;
    // Journal block size - minimum_io_size of the journal device is the best choice
-    uint32_t journal_block_size = 4096;
+    uint64_t journal_block_size = 4096;
    // Metadata block size - minimum_io_size of the metadata device is the best choice
-    uint32_t meta_block_size = 4096;
+    uint64_t meta_block_size = 4096;
    // Atomic write size of the data block device
    uint32_t atomic_write_size = 4096;
    // Whether we should set RWF_ATOMIC on atomic writes
@@ -58,7 +58,6 @@ class journal_flusher_co
    int i, res;
    bool read_to_fill_incomplete;
    int copy_count;
-    bool do_repeat = false;

    friend class journal_flusher_t;

@@ -301,12 +301,13 @@ int blockstore_heap_t::read_blocks(uint64_t disk_offset, uint64_t disk_size, uin
            heap_entry_t *wr = (heap_entry_t*)data;
            if (wr->size > dsk->meta_block_size-block_offset)
            {
-                fprintf(stderr, "Error: entry is too large in metadata block %u at %u (%u > max %u bytes). ",
+                fprintf(stderr, "Error: entry is too large in metadata block %u at %u (%u > max %ju bytes). ",
                    block_num, block_offset, wr->size, dsk->meta_block_size-block_offset);
 corrupted_block:
                if (allow_corrupted)
                {
                    fprintf(stderr, "Metadata block is corrupted, skipping\n");
+                    recheck_modified_blocks.insert(block_num);
                    break;
                }
                else
@@ -331,7 +332,19 @@ corrupted_block:
                    block_num, block_offset, wr->size, sizeof(heap_entry_t));
                goto corrupted_block;
            }
-            wr->entry_type &= ~BS_HEAP_GARBAGE;
+            if (wr->is_garbage())
+            {
+                // Garbage collection is only performed when writing new entries into the block
+                // because it needs a fake LSN and modified blocks require consecutive modified LSNs
+                // That's why garbage entries may persist on disk
+                if (log_level > 5)
+                {
+                    fprintf(stderr, "Notice: skipping garbage entry %jx:%jx v%ju l%ju in metadata block %u at %u\n",
+                        wr->inode, wr->stripe, wr->version, wr->lsn, block_num, block_offset);
+                }
+                block_offset += wr->size;
+                continue;
+            }
            if ((wr->entry_type & BS_HEAP_TYPE) < BS_HEAP_BIG_WRITE ||
                (wr->entry_type & BS_HEAP_TYPE) > BS_HEAP_ROLLBACK ||
                (wr->entry_type & ~(BS_HEAP_TYPE|BS_HEAP_STABLE)) ||
@@ -345,6 +358,7 @@ corrupted_object:
                if (allow_corrupted)
                {
                    fprintf(stderr, "Entry is corrupted, skipping\n");
+                    recheck_modified_blocks.insert(block_num);
                    block_offset += wr->size;
                    continue;
                }
@@ -360,7 +374,7 @@ corrupted_object:
            {
                // Small writes require accessing offset & len to calculate correct length,
                // so require at least sizeof(heap_small_write_t) for them
-                fprintf(stderr, "Error: entry %jx:%jx v%ju has invalid size in metadata block %u at %u (%u < min %zu bytes). Metadata is corrupted, aborting\n",
+                fprintf(stderr, "Error: entry %jx:%jx v%ju has invalid size in metadata block %u at %u (%u < min %zu bytes)\n",
                    wr->inode, wr->stripe, wr->version, block_num, block_offset, wr->size, sizeof(heap_small_write_t));
                goto corrupted_object;
            }
@@ -374,8 +388,8 @@ corrupted_object:
            uint32_t expected_crc32c = wr->calc_crc32c();
            if (wr->crc32c != expected_crc32c)
            {
-                fprintf(stderr, "Error: entry %jx:%jx v%ju in metadata block %u at %u is corrupt (crc32c mismatch: expected %08x, got %08x). Metadata is corrupted, aborting\n",
-                    wr->inode, wr->stripe, wr->version,
+                fprintf(stderr, "Error: entry %jx:%jx v%ju l%ju in metadata block %u at %u is corrupt (crc32c mismatch: expected %08x, got %08x). ",
+                    wr->inode, wr->stripe, wr->version, wr->lsn,
                    block_num, block_offset, expected_crc32c, wr->crc32c);
                goto corrupted_object;
            }
@@ -385,7 +399,7 @@ corrupted_object:
                wr->small().offset % dsk->bitmap_granularity ||
                wr->small().len % dsk->bitmap_granularity))
            {
-                fprintf(stderr, "Error: %s entry %jx:%jx v%ju has invalid offset/length: %u/%u. Metadata is incompatible with current parameters, aborting\n",
+                fprintf(stderr, "Error: %s entry %jx:%jx v%ju has invalid offset/length: %u/%u. Metadata is incompatible with current parameters. ",
                    wr->type() == BS_HEAP_SMALL_WRITE ? "small_write" : "intent_write",
                    wr->inode, wr->stripe, wr->version, wr->small().offset, wr->small().len);
                goto corrupted_object;
@@ -395,7 +409,7 @@ corrupted_object:
                wr->big_intent().offset % dsk->bitmap_granularity ||
                wr->big_intent().len % dsk->bitmap_granularity))
            {
-                fprintf(stderr, "Error: big_intent entry %jx:%jx v%ju has invalid offset/length: %u/%u. Metadata is incompatible with current parameters, aborting\n",
+                fprintf(stderr, "Error: big_intent entry %jx:%jx v%ju has invalid offset/length: %u/%u. Metadata is incompatible with current parameters. ",
                    wr->inode, wr->stripe, wr->version, wr->big_intent().offset, wr->big_intent().len);
                goto corrupted_object;
            }
@@ -422,7 +436,7 @@ int blockstore_heap_t::load_blocks(uint64_t disk_offset, uint64_t size, uint8_t
            next_lsn = wr->lsn;
        }
        entries_loaded++;
-        insert_list_item(li);
+        loaded_list_items.push_back(li);
        modify_alloc(block_num, [&](heap_block_info_t & inf)
        {
            if (!inf.entries.size())
@@ -458,19 +472,22 @@ bool blockstore_heap_t::validate_object(heap_entry_t *obj)
        next_wr = wr;
        if (wr->type() == BS_HEAP_ROLLBACK)
        {
-            if (commit_wr && wr->version > commit_wr->version)
-            {
-                // rollback may not come before commit with a smaller version
-                fprintf(stderr, "Error: rollback entry %jx:%jx v%ju l%ju comes before a commit entry v%ju l%ju\n",
-                    wr->inode, wr->stripe, wr->version, wr->lsn, commit_wr->version, commit_wr->lsn);
-                return false;
-            }
            rollback_wr = wr;
            continue;
        }
        if (wr->type() == BS_HEAP_COMMIT)
        {
-            commit_wr = wr;
+            if (commit_wr && wr->version > commit_wr->version)
+            {
+                // commit may not come before commit with a smaller version
+                fprintf(stderr, "Error: commit entry %jx:%jx v%ju l%ju comes before a commit entry v%ju l%ju\n",
+                    wr->inode, wr->stripe, wr->version, wr->lsn, commit_wr->version, commit_wr->lsn);
+                return false;
+            }
+            if (!commit_wr)
+            {
+                commit_wr = wr;
+            }
            continue;
        }
        if (wr->entry_type & BS_HEAP_STABLE)
@@ -523,6 +540,23 @@ bool blockstore_heap_t::validate_object(heap_entry_t *obj)
    return true;
 }

+void blockstore_heap_t::finish_load()
+{
+    if (loaded_list_items.size())
+    {
+        // Sort everything and load in correct order
+        std::sort(loaded_list_items.begin(), loaded_list_items.end(), [this](const heap_list_item_t* a, const heap_list_item_t* b)
+        {
+            return a->entry.lsn < b->entry.lsn;
+        });
+        for (auto & li: loaded_list_items)
+        {
+            insert_list_item(li);
+        }
+        loaded_list_items.clear();
+    }
+}
+
 void blockstore_heap_t::fill_recheck_queue()
 {
    for (auto & pgp: block_index)
@@ -709,6 +743,7 @@ bool blockstore_heap_t::recheck_small_writes(std::function<void(bool is_data, ui
    }
    if (!recheck_queue_filled)
    {
+        finish_load();
        fill_recheck_queue();
        recheck_queue_filled = true;
    }
@@ -800,7 +835,7 @@ std::vector<uint32_t> blockstore_heap_t::get_recheck_modified_blocks()
    return modified;
 }

-int blockstore_heap_t::finish_load(bool allow_corrupted)
+int blockstore_heap_t::finish_recheck()
 {
    if (!marked_used_blocks)
    {
@@ -837,7 +872,6 @@ bool blockstore_heap_t::calc_checksums(heap_entry_t *wr, uint8_t *data, bool set
        {
            return true;
        }
-        uint32_t len = 0;
        if (wr->type() == BS_HEAP_SMALL_WRITE || wr->type() == BS_HEAP_INTENT_WRITE)
            len = wr->small().len;
        else if (wr->type() == BS_HEAP_BIG_INTENT)
@@ -854,13 +888,14 @@ bool blockstore_heap_t::calc_checksums(heap_entry_t *wr, uint8_t *data, bool set
    }
    if (wr->type() == BS_HEAP_BIG_WRITE)
    {
+        assert(offset != UINT32_MAX && len != UINT32_MAX);
        return calc_block_checksums((uint32_t*)(wr->get_checksums(this) + offset/dsk->csum_block_size * (dsk->data_csum_type & 0xFF)),
            data, wr->get_int_bitmap(this), offset, offset+len, set, NULL);
    }
    if (wr->type() == BS_HEAP_BIG_INTENT)
    {
        auto & bi = wr->big_intent();
-        return calc_block_checksums((uint32_t*)(wr->get_checksums(this) + offset/dsk->csum_block_size * (dsk->data_csum_type & 0xFF)),
+        return calc_block_checksums((uint32_t*)(wr->get_checksums(this) + bi.offset/dsk->csum_block_size * (dsk->data_csum_type & 0xFF)),
            data, wr->get_int_bitmap(this), bi.offset, bi.offset+bi.len, set, NULL);
    }
    assert(wr->type() == BS_HEAP_SMALL_WRITE || wr->type() == BS_HEAP_INTENT_WRITE);
@@ -1077,16 +1112,6 @@ bool blockstore_heap_t::reshard_check(pool_id_t pool, uint32_t pg_count, uint32_
        set_it->second.pg_stripe_size == pg_stripe_size);
 }

-void blockstore_heap_t::reshard_abort(void* reshard_state)
-{
-    heap_reshard_state_t *st = (heap_reshard_state_t*)reshard_state;
-    for (auto sh_it = st->old_shards.begin(); sh_it != st->old_shards.end(); sh_it++)
-    {
-        block_index[sh_it->first] = std::move(sh_it->second);
-    }
-    delete st;
-}
-
 heap_entry_t *blockstore_heap_t::lock_and_read_entry(object_id oid)
 {
    auto obj = read_entry(oid);
@@ -1099,27 +1124,6 @@ heap_entry_t *blockstore_heap_t::lock_and_read_entry(object_id oid)
    return obj;
 }

-heap_entry_t *blockstore_heap_t::read_locked_entry(object_id oid, uint64_t lsn)
-{
-    auto obj = read_entry(oid);
-    assert(obj);
-    for (auto wr = obj; wr; wr = prev(wr))
-    {
-        if (wr->is_overwrite())
-        {
-            if (lsn == wr->lsn)
-            {
-                return obj;
-            }
-            else
-            {
-                obj = prev(wr);
-            }
-        }
-    }
-    return NULL;
-}
-
 bool blockstore_heap_t::unlock_entry(object_id oid)
 {
    auto mvcc_it = object_mvcc.find(oid);
@@ -1156,6 +1160,35 @@ heap_entry_t *blockstore_heap_t::read_entry(object_id oid)
    return &li->entry;
 }

+void blockstore_heap_t::gc_block(heap_block_info_t & inf)
+{
+    if (inf.has_garbage)
+    {
+        size_t i = 0, j = 0;
+        for (; i < inf.entries.size(); i++)
+        {
+            if (inf.entries[i]->entry.is_garbage())
+            {
+                // old entry invalidated by a newer one, mark it as freeable on block write
+                // assign a 'virtual' LSN to track GC completion
+                assert(!inf.mod_lsn_to || inf.mod_lsn_to == next_lsn);
+                uint64_t gc_lsn = ++next_lsn;
+                inf.mod_lsn = inf.mod_lsn ? inf.mod_lsn : gc_lsn;
+                inf.mod_lsn_to = gc_lsn;
+                push_inflight_lsn(gc_lsn, &inf.entries[i]->entry, HEAP_INFLIGHT_GC);
+            }
+            else
+            {
+                if (j != i)
+                    inf.entries[j] = inf.entries[i];
+                j++;
+            }
+        }
+        inf.entries.resize(j);
+        inf.has_garbage = false;
+    }
+}
+
 int blockstore_heap_t::allocate_entry(uint32_t entry_size, uint32_t *block_num, bool allow_last_free)
 {
    if (last_allocated_block != UINT32_MAX)
@@ -1222,31 +1255,7 @@ int blockstore_heap_t::allocate_entry(uint32_t entry_size, uint32_t *block_num,
    }
    // Write into the same block
    auto & inf = block_info.at(last_allocated_block);
-    if (inf.has_garbage)
-    {
-        size_t i = 0, j = 0;
-        for (; i < inf.entries.size(); i++)
-        {
-            if (inf.entries[i]->entry.is_garbage())
-            {
-                // old entry invalidated by a newer one, mark it as freeable on block write
-                // assign a 'virtual' LSN to track GC completion
-                assert(!inf.mod_lsn_to || inf.mod_lsn_to == next_lsn);
-                uint64_t gc_lsn = ++next_lsn;
-                inf.mod_lsn = inf.mod_lsn ? inf.mod_lsn : gc_lsn;
-                inf.mod_lsn_to = gc_lsn;
-                push_inflight_lsn(gc_lsn, &inf.entries[i]->entry, HEAP_INFLIGHT_GC);
-            }
-            else
-            {
-                if (j != i)
-                    inf.entries[j] = inf.entries[i];
-                j++;
-            }
-        }
-        inf.entries.resize(j);
-        inf.has_garbage = false;
-    }
+    gc_block(inf);
    *block_num = last_allocated_block;
    modify_alloc(last_allocated_block, [&](heap_block_info_t & inf)
    {
@@ -1418,7 +1427,7 @@ int blockstore_heap_t::add_redirect_intent(object_id oid, heap_entry_t **obj_ptr
        bitmap_set(wr->get_int_bitmap(this), offset, len, dsk->bitmap_granularity);
        if (dsk->data_csum_type)
            memset(wr->get_checksums(this), 0, get_csum_size(wr));
-        calc_checksums(wr, (uint8_t*)data, true, offset, len);
+        calc_checksums(wr, (uint8_t*)data, true);
        *obj_ptr = wr;
    });
 }
@@ -1457,11 +1466,11 @@ int blockstore_heap_t::add_big_intent(object_id oid, heap_entry_t **obj_ptr, uin
        if (dsk->data_csum_type)
        {
            if (checksums)
-                memcpy(wr->get_checksums(this), checksums, dsk->clean_entry_bitmap_size);
+                memcpy(wr->get_checksums(this), checksums, get_csum_size(wr));
            else
            {
-                memcpy(wr->get_checksums(this), obj->get_checksums(this), dsk->clean_entry_bitmap_size);
-                calc_checksums(wr, (uint8_t*)data, true, offset, len);
+                memcpy(wr->get_checksums(this), obj->get_checksums(this), get_csum_size(wr));
+                calc_checksums(wr, (uint8_t*)data, true);
            }
        }
        else
@@ -1582,7 +1591,7 @@ int blockstore_heap_t::add_commit(heap_entry_t *obj, uint64_t version, uint32_t
    }
    if (!uncommitted)
    {
-        return EBUSY;
+        return 0;
    }
    return add_simple(obj, version, modified_block, BS_HEAP_COMMIT);
 }
@@ -1592,23 +1601,32 @@ int blockstore_heap_t::add_rollback(heap_entry_t *obj, uint64_t version, uint32_
    heap_entry_t *wr = obj;
    bool found_uncommitted = false;
    uint64_t commit_version = 0;
-    while (wr && !wr->is_overwrite())
+    uint64_t rollback_version = UINT64_MAX;
+    while (wr)
    {
        if (wr->type() == BS_HEAP_ROLLBACK)
        {
-            auto rollback_version = wr->version;
-            wr = prev(wr);
-            while (wr->version > rollback_version)
+            if (wr->version <= version)
            {
-                assert(!(wr->entry_type & BS_HEAP_STABLE));
-                wr = prev(wr);
+                // All previous writes are already rolled back, stop
+                break;
            }
+            rollback_version = wr->version;
+            wr = prev(wr);
            continue;
        }
        if (wr->type() == BS_HEAP_COMMIT)
        {
            if (commit_version < wr->version)
+            {
                commit_version = wr->version;
+            }
+            wr = prev(wr);
+            continue;
+        }
+        if (wr->version > rollback_version)
+        {
+            // Already rolled back, skip
            wr = prev(wr);
            continue;
        }
@@ -1619,14 +1637,10 @@ int blockstore_heap_t::add_rollback(heap_entry_t *obj, uint64_t version, uint32_
            {
                return EBUSY;
            }
-            else if (wr->version == version)
+            else
            {
                break;
            }
-            else if (wr->version < version)
-            {
-                return ENOENT;
-            }
        }
        else if (wr->version > version)
        {
@@ -1807,9 +1821,9 @@ void blockstore_heap_t::iterate_with_stable(heap_entry_t *obj, uint64_t max_lsn,
        }
        else
        {
-            // 1) 1 2 3 ROLLBACK(2) COMMIT(3) -> impossible
+            // 1) 1 2 3 ROLLBACK(2) COMMIT(3) -> 3 is unstable
            // 2) 1 2 3 4 ROLLBACK(3) COMMIT(2) -> OK
-            // 3) 1 2 3 ROLLBACK(2) 3 COMMIT(3) -> first 3 shouldn't be treated as stable
+            // 3) 1 2 3 ROLLBACK(2) 3 COMMIT(3) -> first 3 is unstable
            // 4) 1 2 3 COMMIT(3) ROLLBACK(2) -> impossible
            //    I.e. a rollback always has version >= previous commit
            // 5) 1 2 3 4 5 ROLLBACK(4) 5 ROLLBACK(3)
@@ -1966,24 +1980,21 @@ int blockstore_heap_t::list_objects(uint32_t pg_num, object_id min_oid, object_i
                return;
            }
            uint64_t stable_version = 0;
-            auto first_wr = obj;
-            for (auto wr = first_wr; wr; wr = prev(wr))
+            iterate_with_stable(obj, UINT64_MAX, [&](heap_entry_t* wr, bool stable)
            {
-                if ((wr->entry_type & BS_HEAP_STABLE) || wr->type() == BS_HEAP_COMMIT || wr->type() == BS_HEAP_ROLLBACK)
+                if (stable)
                {
                    stable_version = wr->version;
-                    break;
+                    return false;
                }
-                else
+                if (unstable_size >= unstable_alloc)
                {
-                    if (unstable_size >= unstable_alloc)
-                    {
-                        unstable_alloc = (!unstable_alloc ? 128 : unstable_alloc*2);
-                        unstable = (obj_ver_id*)realloc_or_die(unstable, sizeof(obj_ver_id) * unstable_alloc);
-                    }
-                    unstable[unstable_size++] = (obj_ver_id){ .oid = oid, .version = wr->version };
+                    unstable_alloc = (!unstable_alloc ? 128 : unstable_alloc*2);
+                    unstable = (obj_ver_id*)realloc_or_die(unstable, sizeof(obj_ver_id) * unstable_alloc);
                }
-            }
+                unstable[unstable_size++] = (obj_ver_id){ .oid = oid, .version = wr->version };
+                return true;
+            });
            if (stable_version)
            {
                if (res_size >= res_alloc)
@@ -2045,7 +2056,13 @@ void blockstore_heap_t::free_data(inode_t inode, uint64_t location)
        inode = (INODE_POOL(inode) << POOL_ID_BITS);
    assert(data_alloc->get(location / dsk->data_block_size));
    data_alloc->set(location / dsk->data_block_size, false);
-    inode_space_stats[inode] -= dsk->data_block_size;
+    auto sp_it = inode_space_stats.find(inode);
+    if (sp_it != inode_space_stats.end())
+    {
+        sp_it->second -= dsk->data_block_size;
+        if (sp_it->second == 0)
+            inode_space_stats.erase(sp_it);
+    }
    data_used_space -= dsk->data_block_size;
 }

@@ -2274,7 +2291,9 @@ void blockstore_heap_t::apply_inflight(heap_inflight_lsn_t & inflight)
        }
        if (!next)
        {
+            // The last freed entry must be a deletion
            assert(!prev);
+            assert(wr->entry_type == BS_HEAP_DELETE|BS_HEAP_STABLE);
            auto & pg_idx = block_index[get_pg_id(wr->inode, wr->stripe)];
            auto & inode_idx = pg_idx[wr->inode];
            heap_inode_map_t::iterator li_it;
@@ -2328,7 +2347,7 @@ void blockstore_heap_t::set_no_inode_stats(const std::vector<uint64_t> & pool_id
    {
        // Recalculate if changed
        if (ps.second.no_inode_stats == 2 || ps.second.no_inode_stats == 1)
-            recalc_inode_space_stats(ps.first, ps.second.no_inode_stats == 1);
+            recalc_inode_space_stats(ps.first, ps.second.no_inode_stats == 2);
        ps.second.no_inode_stats &= 1;
    }
 }
@@ -2339,8 +2358,8 @@ void blockstore_heap_t::recalc_inode_space_stats(uint64_t pool_id, bool per_inod
    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);
-    uint32_t pg_count = ps.pg_count ? ps.pg_count : 1;
-    for (uint32_t pg_num = 1; pg_num <= pg_count; pg_num++)
+    uint32_t pg_count = ps.pg_count;
+    for (uint32_t pg_num = pg_count ? 1 : 0; pg_num <= pg_count; pg_num++)
    {
        auto & pg_idx = block_index[(pool_id << (64-POOL_ID_BITS)) | pg_num];
        for (auto & ip: pg_idx)
@@ -2378,7 +2397,10 @@ void inode_map_get(void *inode_idx, heap_inode_map_t::iterator & li_it, heap_lis
    size_t map_n = ((size_t)inode_idx & IMAP_MALLOC_LOW_BITS);
    if (!map_n)
    {
+#pragma GCC diagnostic push
+#pragma GCC diagnostic ignored "-Warray-bounds"
        li_it = ((heap_inode_map_t*)inode_idx)->find(list_item_key(&stripe));
+#pragma GCC diagnostic pop
        li = li_it != ((heap_inode_map_t*)inode_idx)->end() ? *li_it : NULL;
    }
    else if (map_n == 1)
@@ -163,7 +163,7 @@ using heap_mvcc_map_t = robin_hood::unordered_flat_map<object_id, heap_object_mv

 class blockstore_heap_t
 {
-    friend class heap_entry_t;
+    friend struct heap_entry_t;

    blockstore_disk_t *dsk = NULL;
    uint8_t* buffer_area = NULL;
@@ -200,6 +200,7 @@ class blockstore_heap_t

    bool marked_used_blocks = false;
    bool recheck_queue_filled = false;
+    std::vector<heap_list_item_t*> loaded_list_items;
    std::set<uint32_t> recheck_modified_blocks;
    std::deque<heap_entry_t*> recheck_queue;
    int recheck_in_progress = 0;
@@ -215,6 +216,7 @@ class blockstore_heap_t
    void defragment_block(uint32_t block_num);
    void reshard_add(heap_reshard_state_t *st, heap_list_item_t *li);

+    void gc_block(heap_block_info_t & inf);
    int allocate_entry(uint32_t entry_size, uint32_t *block_num, bool allow_last_free);
    void insert_list_item(heap_list_item_t *li);
    int add_entry(uint32_t wr_size, uint32_t *modified_block, bool allow_last_free,
@@ -237,31 +239,29 @@ public:
        std::function<void(uint32_t, uint32_t, uint8_t*)> handle_block);
    int load_blocks(uint64_t disk_offset, uint64_t size, uint8_t *buf,
        bool allow_corrupted, uint64_t &entries_loaded);
-    // finish loading
-    int finish_load(bool allow_corrupted = false);
+    // finish loading - should be called after load_blocks
+    void finish_load();
    // get blocks which are modified during loading and should be written to the disk
    // before finishing initialization if not R/O
    std::vector<uint32_t> get_recheck_modified_blocks();
    // recheck small write data after reading the database from disk
    bool recheck_small_writes(std::function<void(bool is_data, uint64_t offset, uint64_t len, uint8_t* buf, std::function<void()>)> read_buffer, int queue_depth);
+    int finish_recheck();
    // reshard database according to the pool's PG count
    void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit);
    bool reshard_continue(void* reshard_state, uint64_t chunk_limit);
    bool reshard_check(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size);
-    void reshard_abort(void* reshard_state);
    void set_no_inode_stats(const std::vector<uint64_t> & pool_ids);
    void recalc_inode_space_stats(uint64_t pool_id, bool per_inode);
    // read an object entry and lock it against removal
    // in the future, may become asynchronous
    heap_entry_t *lock_and_read_entry(object_id oid);
-    // re-read a locked object entry with the given lsn (pointer may be invalidated)
-    heap_entry_t *read_locked_entry(object_id oid, uint64_t lsn);
    // read an object entry without locking it
    heap_entry_t *read_entry(object_id oid);
    // unlock an entry
    bool unlock_entry(object_id oid);
    // set or verify checksums in a write request
-    bool calc_checksums(heap_entry_t *wr, uint8_t *data, bool set, uint32_t offset = 0, uint32_t len = 0);
+    bool calc_checksums(heap_entry_t *wr, uint8_t *data, bool set, uint32_t offset = UINT32_MAX, uint32_t len = UINT32_MAX);
    // set or verify raw block checksums
    bool calc_block_checksums(uint32_t *block_csums, uint8_t *data, uint8_t *bitmap, uint32_t start, uint32_t end,
        bool set, std::function<void(uint32_t, uint32_t, uint32_t)> bad_block_cb);
@@ -193,12 +193,12 @@ void blockstore_impl_t::loop()
            heap->start_block_write(block_num);
            mb.sent = true;
        }
+        pending_modified_blocks.clear();
        int ret = ringloop->submit();
        if (ret < 0)
        {
            throw std::runtime_error(std::string("io_uring_submit: ") + strerror(-ret));
        }
-        pending_modified_blocks.clear();
        if ((initial_ring_space - ringloop->space_left()) > 0)
        {
            live = true;
@@ -406,8 +406,3 @@ bool blockstore_impl_t::reshard_continue(void *reshard_state, uint64_t chunk_lim
 {
    return heap->reshard_continue(reshard_state, chunk_limit);
 }
-
-void blockstore_impl_t::reshard_abort(void *reshard_state)
-{
-    return heap->reshard_abort(reshard_state);
-}
@@ -78,6 +78,7 @@ public:
    // Suitable only for server SSDs with capacitors, requires disabled data and journal fsyncs
    int immediate_commit = IMMEDIATE_NONE;
    bool inmemory_meta = false;
+    bool skip_corrupted_meta_entries = false;
    uint32_t meta_write_recheck_parallelism = 0;
    // Maximum and minimum flusher count
    unsigned max_flusher_count = 0, min_flusher_count = 0;
@@ -191,7 +192,6 @@ public:

    void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit);
    bool reshard_continue(void *reshard_state, uint64_t chunk_limit);
-    void reshard_abort(void *reshard_state);

    // Event loop
    void loop();
@@ -145,7 +145,7 @@ 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, meta_area_size=%ju)"
-                " differs from OSD configuration (%u/%u/%u, %u/%u, %ju).\n",
+                " differs from OSD configuration (%ju/%ju/%u, %u/%u, %ju).\n",
                hdr->meta_block_size, hdr->data_block_size, hdr->bitmap_granularity,
                hdr->data_csum_type, hdr->csum_block_size, hdr->meta_area_size,
                bs->dsk.meta_block_size, bs->dsk.data_block_size, bs->dsk.bitmap_granularity,
@@ -225,7 +225,7 @@ resume_4:
        {
            // Handle result
            uint64_t loaded = 0;
-            int r = bs->heap->load_blocks(bufs[i].offset-bs->dsk.meta_block_size, bufs[i].size, bufs[i].buf, false, loaded);
+            int r = bs->heap->load_blocks(bufs[i].offset-bs->dsk.meta_block_size, bufs[i].size, bufs[i].buf, bs->skip_corrupted_meta_entries, loaded);
            if (r != 0)
                exit(1);
            entries_loaded += loaded;
@@ -239,6 +239,7 @@ resume_4:
        return 1;
    }
    // metadata read finished
+    bs->heap->finish_load();
    printf("Metadata entries loaded: %ju, used blocks: %ju / %ju\n", entries_loaded, bs->heap->get_data_used_space() / bs->dsk.data_block_size, bs->dsk.block_count);
    if (zero_on_init && !bs->dsk.disable_meta_fsync)
    {
@@ -269,7 +270,7 @@ resume_6:
        }
        GET_SQE();
        data->iov = (iovec){ buf, len };
-        data->callback = [this, offset, cb](ring_data_t *data)
+        data->callback = [offset, cb](ring_data_t *data)
        {
            if (data->res < 0)
            {
@@ -284,7 +285,7 @@ resume_6:
    }, bs->meta_write_recheck_parallelism);
    return 1;
 resume_7:
-    if (bs->heap->finish_load() != 0)
+    if (bs->heap->finish_recheck() != 0)
    {
        exit(1);
    }
@@ -28,6 +28,7 @@ void blockstore_impl_t::parse_config(blockstore_config_t & config, bool init)
    throttle_target_parallelism = strtoull(config["throttle_target_parallelism"].c_str(), NULL, 10);
    throttle_threshold_us = strtoull(config["throttle_threshold_us"].c_str(), NULL, 10);
    perfect_csum_update = config["perfect_csum_update"] == "true" || config["perfect_csum_update"] == "1" || config["perfect_csum_update"] == "yes";
+    skip_corrupted_meta_entries = config["skip_corrupted_meta_entries"] == "true" || config["skip_corrupted_meta_entries"] == "1" || config["skip_corrupted_meta_entries"] == "yes";
    if (config["autosync_writes"] != "")
    {
        autosync_writes = strtoull(config["autosync_writes"].c_str(), NULL, 10);
@@ -57,9 +57,9 @@ int blockstore_impl_t::dequeue_stable(blockstore_op_t *op)
            }
            assert(res == 0);
        }
+resume_1:
        if (priv->modified_block != UINT32_MAX && priv->modified_block2 != priv->modified_block)
        {
-resume_1:
            BS_SUBMIT_CHECK_SQES(1);
            prepare_meta_block_write(priv->modified_block);
 resume_2:
@@ -22,21 +22,23 @@ void blockstore_impl_t::prepare_meta_block_write(uint32_t modified_block)
    ring_data_t *data = ((ring_data_t*)sqe->user_data);
    uint8_t *buf = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, dsk.meta_block_size);
    data->iov = (struct iovec){ buf, (size_t)dsk.meta_block_size };
-    data->callback = [this, modified_block, buf](ring_data_t *data)
+    data->callback = [this, modified_block](ring_data_t *data)
    {
-        free(buf);
        live = true;
        if (data->res != data->iov.iov_len)
        {
            // FIXME: our state becomes corrupted after a write error. maybe do something better than just die
            disk_error_abort("data write", data->res, data->iov.iov_len);
        }
-        modified_blocks.erase(modified_block);
+        auto it = modified_blocks.find(modified_block);
+        assert(it != modified_blocks.end());
+        free(it->second.buf);
+        modified_blocks.erase(it);
        heap->complete_block_write(modified_block);
        ringloop->wakeup();
    };
    io_uring_prep_writev(
-        sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + (modified_block+1)*dsk.meta_block_size
+        sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + ((uint64_t)modified_block+1)*dsk.meta_block_size
    );
    unsynced_meta_write_count++;
    pending_modified_blocks.push_back(modified_block);
@@ -249,13 +251,12 @@ enospc:
            goto enospc;
        assert(res == 0);
        PRIV(op)->lsn = obj->lsn;
-        if (op->len)
-            heap->use_buffer_area(op->oid.inode, loc, op->len);
        prepare_meta_block_write(PRIV(op)->modified_block);
        PRIV(op)->pending_ops++;
        if (op->len > 0)
        {
            // Prepare buffered data write
+            heap->use_buffer_area(op->oid.inode, loc, op->len);
            if (dsk.inmemory_journal)
            {
                memcpy((uint8_t*)buffer_area + loc, op->buf, op->len);
@@ -348,6 +349,7 @@ resume_12:
    }
 resume_4:
    {
+        BS_SUBMIT_CHECK_SQES(1);
        auto obj = heap->read_entry(op->oid);
        int res = 0;
        if (PRIV(op)->write_type == _REDIRECT_INTENT)
@@ -404,11 +406,12 @@ resume_6:
        if (ref_us > exec_us + throttle_threshold_us)
        {
            // Pause reply
+            PRIV(op)->pending_ops++;
            PRIV(op)->op_state = 7;
            // Remember that the timer can in theory be called right here
            tfd->set_timer_us(ref_us-exec_us, false, [this, op](int timer_id)
            {
-                PRIV(op)->op_state = 8;
+                PRIV(op)->pending_ops--;
                ringloop->wakeup();
            });
            return 1;
@@ -480,17 +480,6 @@ resume_1:
    return true;
 }

-void blockstore_impl_t::reshard_abort(void *reshard_state)
-{
-    bs_reshard_state_t *st = (bs_reshard_state_t*)reshard_state;
-    for (auto sh_it = st->old_shards.begin(); sh_it != st->old_shards.end(); sh_it++)
-    {
-        auto & to = clean_db_shards[sh_it->first];
-        to.swap(sh_it->second);
-    }
-    delete st;
-}
-
 void blockstore_impl_t::process_list(blockstore_op_t *op)
 {
    uint32_t list_pg = op->pg_number+1;
@@ -290,7 +290,6 @@ public:
    // Reshard database for a pool
    void* reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit);
    bool reshard_continue(void *reshard_state, uint64_t chunk_limit);
-    void reshard_abort(void *reshard_state);

    // Event loop
    void loop();
@@ -189,7 +189,7 @@ 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 (%u/%u/%u, %u/%u).\n",
+                " differs from OSD configuration (%ju/%ju/%u, %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,
@@ -620,8 +620,7 @@ bool blockstore_impl_t::fulfill_clean_read(blockstore_op_t *read_op, uint64_t &
    else if (from_journal)
    {
        // Don't scan bitmap - journal writes don't have holes (internal bitmap)!
-        uint8_t *csum = !dsk.csum_block_size ? 0 : (clean_entry_bitmap + dsk.clean_entry_bitmap_size +
-            item_start/dsk.csum_block_size*(dsk.data_csum_type & 0xFF));
+        uint8_t *csum = !dsk.csum_block_size ? 0 : (clean_entry_bitmap + dsk.clean_entry_bitmap_size);
        if (!fulfill_read(read_op, fulfilled, item_start, item_end,
            (BS_ST_BIG_WRITE | BS_ST_STABLE), 0, clean_loc + item_start, 0, csum, dyn_data))
        {
@@ -898,7 +898,7 @@ void cluster_client_t::execute_cas(cluster_op_t *op)
                        .opcode = OSD_OP_SYNC,
                    },
                },
-                .callback = [this, op](osd_op_t *part)
+                .callback = [op](osd_op_t *part)
                {
                    if (part->reply.hdr.retval != 0)
                    {
@@ -1257,7 +1257,7 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
        {
            op->bitmap_buf = realloc_or_die(op->bitmap_buf, bitmap_mem);
            op->part_bitmaps = (uint8_t*)op->bitmap_buf + object_bitmap_size;
-            memset(op->bitmap_buf+op->bitmap_buf_size, 0, bitmap_mem-op->bitmap_buf_size);
+            memset((uint8_t*)op->bitmap_buf+op->bitmap_buf_size, 0, bitmap_mem-op->bitmap_buf_size);
            op->bitmap_buf_size = bitmap_mem;
        }
    }
@@ -414,7 +414,10 @@ void etcd_state_client_t::start_etcd_watcher()
                }
                // Save revision only if it's present in the message - because sometimes etcd sends something without a header, like:
                // {"error": {"grpc_code": 14, "http_code": 503, "http_status": "Service Unavailable", "message": "error reading from server: EOF"}}
-                if (etcd_watches_initialised == ETCD_TOTAL_WATCHES && !data["result"]["header"]["revision"].is_null())
+                // Also don't save revision from the initial created: true messages because they always contain the latest revision
+                if (etcd_watches_initialised == ETCD_TOTAL_WATCHES &&
+                    !data["result"]["header"]["revision"].is_null() &&
+                    !data["result"]["created"].bool_value())
                {
                    // Restart watchers from the same revision number as in the last received message,
                    // not from the next one to protect against revision being split into multiple messages,
@@ -6,7 +6,7 @@
 #include <set>

 #include "json11/json11.hpp"
-#include "osd_id.h"
+#include "object_id.h"
 #include "timerfd_manager.h"

 #define ETCD_CONFIG_WATCH_ID 1
@@ -145,7 +145,7 @@ void osd_messenger_t::init()
        rdma_contexts = msgr_rdma_context_t::create_all(
            osd_num && osd_cluster_network_masks.size() ? osd_cluster_network_masks : osd_network_masks,
            rdma_device != "" ? rdma_device.c_str() : NULL,
-            rdma_port_num, rdma_gid_index, rdma_mtu, rdma_odp, log_level
+            rdma_port_num, rdma_gid_index, rdma_mtu, log_level
        );
        if (!rdma_contexts.size())
        {
@@ -322,7 +322,6 @@ void osd_messenger_t::parse_config(const json11::Json & config)
    this->rdma_max_msg = config["rdma_max_msg"].uint64_value();
    if (!this->rdma_max_msg || this->rdma_max_msg > 128*1024*1024)
        this->rdma_max_msg = 129*1024;
-    this->rdma_odp = config["rdma_odp"].bool_value();
 #endif
    if (!osd_num)
        this->iothread_count = (uint32_t)config["client_iothread_count"].uint64_value();
@@ -740,14 +739,6 @@ void osd_messenger_t::check_peer_config(osd_client_t *cl)
                    fprintf(stderr, "Connected to OSD %ju using RDMA\n", cl->osd_num);
                }
                cl->peer_state = PEER_RDMA;
-                tfd->set_fd_handler(cl->peer_fd, false, [this](int peer_fd, int epoll_events)
-                {
-                    // Do not miss the disconnection!
-                    if (epoll_events & EPOLLRDHUP)
-                    {
-                        handle_peer_epoll(peer_fd, epoll_events);
-                    }
-                });
                // Add the initial receive request
                init_recv_rdma(cl);
            }
@@ -200,7 +200,6 @@ protected:
    std::vector<msgr_rdma_context_t *> rdma_contexts;
    uint64_t rdma_max_sge = 0, rdma_max_send = 0, rdma_max_recv = 0;
    uint64_t rdma_max_msg = 0;
-    bool rdma_odp = false;
    rdma_event_channel *rdmacm_evch = NULL;
    std::map<rdma_cm_id*, osd_client_t*> rdmacm_connections;
    std::map<rdma_cm_id*, rdmacm_connecting_t*> rdmacm_connecting;
@@ -210,7 +209,7 @@ protected:
    std::vector<int> read_ready_clients;
    std::vector<int> write_ready_clients;
    // We don't use ringloop->set_immediate here because we may have no ringloop in client :)
-    std::vector<osd_op_t*> set_immediate_ops;
+    std::deque<osd_op_t*> set_immediate_ops;

 public:
    timerfd_manager_t *tfd = NULL;
@@ -287,8 +286,6 @@ protected:

 #ifdef WITH_RDMA
    void try_send_rdma(osd_client_t *cl);
-    void try_send_rdma_odp(osd_client_t *cl);
-    void try_send_rdma_nodp(osd_client_t *cl);
    bool init_recv_rdma(osd_client_t *cl);
    void handle_rdma_events(msgr_rdma_context_t *rdma_context);
    msgr_rdma_context_t* choose_rdma_context(osd_client_t *cl);
@@ -59,8 +59,6 @@ msgr_rdma_context_t::~msgr_rdma_context_t()
        ibv_destroy_cq(cq);
    if (channel)
        ibv_destroy_comp_channel(channel);
-    if (mr)
-        ibv_dereg_mr(mr);
    if (pd)
        ibv_dealloc_pd(pd);
    if (context && !is_cm)
@@ -182,7 +180,7 @@ static int match_port_gid(const std::vector<addr_mask_t> & osd_network_masks, ib
 #endif

 std::vector<msgr_rdma_context_t*> msgr_rdma_context_t::create_all(const std::vector<addr_mask_t> & osd_network_masks,
-    const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, bool odp, int log_level)
+    const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, int log_level)
 {
    int res;
    std::vector<msgr_rdma_context_t*> ret;
@@ -271,7 +269,7 @@ std::vector<msgr_rdma_context_t*> msgr_rdma_context_t::create_all(const std::vec
                {
                    if (log_level > 0)
                        log_rdma_dev_port_gid(dev, port_num, best_gid_idx, port_mtu, best_gidx);
-                    auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, odp, log_level);
+                    auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, log_level);
                    if (ctx)
                    {
                        ctx->net_mask = osd_network_masks[net_num];
@@ -291,7 +289,7 @@ std::vector<msgr_rdma_context_t*> msgr_rdma_context_t::create_all(const std::vec
                    log_rdma_dev_port_gid(dev, port_num, best_gid_idx, port_mtu, gidx);
                }
 #endif
-                auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, odp, log_level);
+                auto ctx = msgr_rdma_context_t::create(dev, portinfo, port_num, best_gid_idx, port_mtu, log_level);
                if (ctx)
                    ret.push_back(ctx);
            }
@@ -306,7 +304,7 @@ cleanup:
    return ret;
 }

-msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr & portinfo, int ib_port, int gid_index, uint32_t mtu, bool odp, int log_level)
+msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr & portinfo, int ib_port, int gid_index, uint32_t mtu, int log_level)
 {
    msgr_rdma_context_t *ctx = new msgr_rdma_context_t();
    ibv_context *context = ibv_open_device(dev);
@@ -346,30 +344,6 @@ msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr
        goto cleanup;
    }

-    ctx->odp = odp;
-    if (ctx->odp)
-    {
-        if (!(ctx->attrx.odp_caps.general_caps & IBV_ODP_SUPPORT) ||
-            !(ctx->attrx.odp_caps.general_caps & IBV_ODP_SUPPORT_IMPLICIT) ||
-            !(ctx->attrx.odp_caps.per_transport_caps.rc_odp_caps & IBV_ODP_SUPPORT_SEND) ||
-            !(ctx->attrx.odp_caps.per_transport_caps.rc_odp_caps & IBV_ODP_SUPPORT_RECV))
-        {
-            ctx->odp = false;
-            if (log_level > 0)
-                fprintf(stderr, "The RDMA device isn't implicit ODP (On-Demand Paging) capable, disabling it\n");
-        }
-    }
-
-    if (ctx->odp)
-    {
-        ctx->mr = ibv_reg_mr(ctx->pd, NULL, SIZE_MAX, IBV_ACCESS_LOCAL_WRITE | IBV_ACCESS_ON_DEMAND);
-        if (!ctx->mr)
-        {
-            fprintf(stderr, "Couldn't register RDMA memory region\n");
-            goto cleanup;
-        }
-    }
-
    ctx->channel = ibv_create_comp_channel(ctx->context);
    if (!ctx->channel)
    {
@@ -388,8 +362,6 @@ msgr_rdma_context_t *msgr_rdma_context_t::create(ibv_device *dev, ibv_port_attr
    return ctx;

 cleanup:
-    if (context)
-        ibv_close_device(context);
    delete ctx;
    return NULL;
 }
@@ -605,52 +577,7 @@ static int try_send_rdma_copy(osd_client_t *cl, uint8_t *dst, int dst_len)
    return total_dst_len-dst_len;
 }

-void osd_messenger_t::try_send_rdma_odp(osd_client_t *cl)
-{
-    auto rc = cl->rdma_conn;
-    if (!cl->send_list.size() || rc->cur_send >= rc->max_send)
-    {
-        return;
-    }
-    uint64_t op_size = 0, op_sge = 0;
-    ibv_sge sge[rc->max_sge];
-    while (rc->send_pos < cl->send_list.size())
-    {
-        iovec & iov = cl->send_list[rc->send_pos];
-        if (op_size >= rc->max_msg || op_sge >= rc->max_sge)
-        {
-            rc->send_sizes.push_back(op_size);
-            try_send_rdma_wr(cl, sge, op_sge);
-            op_sge = 0;
-            op_size = 0;
-            if (rc->cur_send >= rc->max_send)
-            {
-                break;
-            }
-        }
-        uint32_t len = (uint32_t)(op_size+iov.iov_len-rc->send_buf_pos < rc->max_msg
-            ? iov.iov_len-rc->send_buf_pos : rc->max_msg-op_size);
-        sge[op_sge++] = {
-            .addr = (uintptr_t)((uint8_t*)iov.iov_base+rc->send_buf_pos),
-            .length = len,
-            .lkey = rc->ctx->mr->lkey,
-        };
-        op_size += len;
-        rc->send_buf_pos += len;
-        if (rc->send_buf_pos >= iov.iov_len)
-        {
-            rc->send_pos++;
-            rc->send_buf_pos = 0;
-        }
-    }
-    if (op_sge > 0)
-    {
-        rc->send_sizes.push_back(op_size);
-        try_send_rdma_wr(cl, sge, op_sge);
-    }
-}
-
-void osd_messenger_t::try_send_rdma_nodp(osd_client_t *cl)
+void osd_messenger_t::try_send_rdma(osd_client_t *cl)
 {
    auto rc = cl->rdma_conn;
    if (!rc->send_out_size)
@@ -658,14 +585,11 @@ void osd_messenger_t::try_send_rdma_nodp(osd_client_t *cl)
        // Allocate send ring buffer, if not yet
        rc->send_out_size = rc->max_msg*rdma_max_send;
        rc->send_out.buf = (uint8_t*)malloc_or_die(rc->send_out_size);
-        if (!rc->ctx->odp)
+        rc->send_out.mr = ibv_reg_mr(rc->ctx->pd, rc->send_out.buf, rc->send_out_size, 0);
+        if (!rc->send_out.mr)
        {
-            rc->send_out.mr = ibv_reg_mr(rc->ctx->pd, rc->send_out.buf, rc->send_out_size, 0);
-            if (!rc->send_out.mr)
-            {
-                fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
-                exit(1);
-            }
+            fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
+            exit(1);
        }
    }
    // Copy data into the buffer and send it
@@ -690,7 +614,7 @@ void osd_messenger_t::try_send_rdma_nodp(osd_client_t *cl)
            ibv_sge sge = {
                .addr = (uintptr_t)dst,
                .length = (uint32_t)copied,
-                .lkey = rc->ctx->odp ? rc->ctx->mr->lkey : rc->send_out.mr->lkey,
+                .lkey = rc->send_out.mr->lkey,
            };
            try_send_rdma_wr(cl, &sge, 1);
            rc->send_sizes.push_back(copied);
@@ -698,20 +622,12 @@ void osd_messenger_t::try_send_rdma_nodp(osd_client_t *cl)
    }
 }

-void osd_messenger_t::try_send_rdma(osd_client_t *cl)
-{
-    if (cl->rdma_conn->ctx->odp)
-        try_send_rdma_odp(cl);
-    else
-        try_send_rdma_nodp(cl);
-}
-
 static void try_recv_rdma_wr(osd_client_t *cl, void *buf)
 {
    ibv_sge sge = {
        .addr = (uintptr_t)buf,
        .length = (uint32_t)cl->rdma_conn->max_msg,
-        .lkey = cl->rdma_conn->ctx->odp ? cl->rdma_conn->ctx->mr->lkey : cl->rdma_conn->recv_buf.mr->lkey,
+        .lkey = cl->rdma_conn->recv_buf.mr->lkey,
    };
    ibv_recv_wr *bad_wr = NULL;
    ibv_recv_wr wr = {
@@ -733,14 +649,11 @@ bool osd_messenger_t::init_recv_rdma(osd_client_t *cl)
    auto rc = cl->rdma_conn;
    assert(!rc->recv_buf.buf);
    rc->recv_buf.buf = (uint8_t*)malloc_or_die(rc->max_msg * rc->max_recv);
-    if (!rc->ctx->odp)
+    rc->recv_buf.mr = ibv_reg_mr(rc->ctx->pd, rc->recv_buf.buf, rc->max_msg * rc->max_recv, IBV_ACCESS_LOCAL_WRITE);
+    if (!rc->recv_buf.mr)
    {
-        rc->recv_buf.mr = ibv_reg_mr(rc->ctx->pd, rc->recv_buf.buf, rc->max_msg * rc->max_recv, IBV_ACCESS_LOCAL_WRITE);
-        if (!rc->recv_buf.mr)
-        {
-            fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
-            exit(1);
-        }
+        fprintf(stderr, "Failed to register RDMA memory region: %s\n", strerror(errno));
+        exit(1);
    }
    for (uint32_t i = 0; i < rc->max_recv; i++)
    {
@@ -783,6 +696,10 @@ void osd_messenger_t::handle_rdma_events(msgr_rdma_context_t *rdma_context)
                continue;
            }
            osd_client_t *cl = cl_it->second;
+            if (cl->peer_state == PEER_STOPPED)
+            {
+                continue;
+            }
            auto rc = cl->rdma_conn;
            if (wc[i].status != IBV_WC_SUCCESS)
            {
@@ -816,14 +733,11 @@ void osd_messenger_t::handle_rdma_events(msgr_rdma_context_t *rdma_context)
                rc->cur_send--;
                uint64_t sent_size = rc->send_sizes.at(0);
                rc->send_sizes.erase(rc->send_sizes.begin(), rc->send_sizes.begin()+1);
-                if (!rdma_context->odp)
-                {
-                    rc->send_done_pos += sent_size;
-                    rc->send_out_full = false;
-                    if (rc->send_done_pos == rc->send_out_size)
-                        rc->send_done_pos = 0;
-                    assert(rc->send_done_pos < rc->send_out_size);
-                }
+                rc->send_done_pos += sent_size;
+                rc->send_out_full = false;
+                if (rc->send_done_pos == rc->send_out_size)
+                    rc->send_done_pos = 0;
+                assert(rc->send_done_pos < rc->send_out_size);
                int send_pos = 0, send_buf_pos = 0;
                while (sent_size > 0)
                {
@@ -26,8 +26,6 @@ struct msgr_rdma_context_t
    ibv_context *context = NULL;
    ibv_device_attr_ex attrx;
    ibv_pd *pd = NULL;
-    bool odp = false;
-    ibv_mr *mr = NULL;
    ibv_comp_channel *channel = NULL;
    ibv_cq *cq = NULL;
    ibv_port_attr portinfo;
@@ -43,9 +41,9 @@ struct msgr_rdma_context_t
    int cm_refs = 0;

    static std::vector<msgr_rdma_context_t*> create_all(const std::vector<addr_mask_t> & osd_network_masks,
-        const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, bool odp, int log_level);
+        const char *sel_dev_name, int sel_port_num, int sel_gid_index, uint32_t sel_mtu, int log_level);
    static msgr_rdma_context_t *create(ibv_device *dev, ibv_port_attr & portinfo,
-        int ib_port, int gid_index, uint32_t mtu, bool odp, int log_level);
+        int ib_port, int gid_index, uint32_t mtu, int log_level);
    static msgr_rdma_context_t* create_cm(ibv_context *ctx);
    bool reserve_cqe(int n);

@@ -178,7 +178,6 @@ msgr_rdma_context_t* msgr_rdma_context_t::create_cm(ibv_context *ctx)
        delete rdma_context;
        return NULL;
    }
-    rdma_context->odp = false;
    rdma_context->channel = ibv_create_comp_channel(rdma_context->context);
    if (!rdma_context->channel)
    {
@@ -9,7 +9,8 @@ void osd_messenger_t::read_requests()
    {
        int peer_fd = read_ready_clients[i];
        auto cl_it = clients.find(peer_fd);
-        if (cl_it == clients.end() || !cl_it->second || cl_it->second->read_msg.msg_iovlen)
+        if (cl_it == clients.end() || !cl_it->second || cl_it->second->read_msg.msg_iovlen ||
+            cl_it->second->peer_state == PEER_RDMA || cl_it->second->peer_state == PEER_RDMA_CONNECTING)
        {
            continue;
        }
@@ -42,6 +43,7 @@ void osd_messenger_t::read_requests()
            }
            if (!sqe)
            {
+                cl->refs--;
                cl->read_msg.msg_iovlen = 0;
                read_ready_clients.erase(read_ready_clients.begin(), read_ready_clients.begin() + i);
                return;
@@ -74,6 +76,10 @@ bool osd_messenger_t::handle_read(int result, osd_client_t *cl)
    int peer_fd = cl->peer_fd;
    cl->read_msg.msg_iovlen = 0;
    cl->refs--;
+    if (cl->peer_state == PEER_RDMA)
+    {
+        return true;
+    }
    if (cl->peer_state == PEER_STOPPED)
    {
        if (cl->refs <= 0)
@@ -159,8 +165,10 @@ void osd_messenger_t::clear_immediate_ops(int peer_fd)

 void osd_messenger_t::handle_immediate_ops()
 {
-    for (auto op: set_immediate_ops)
+    while (set_immediate_ops.size())
    {
+        auto op = set_immediate_ops.front();
+        set_immediate_ops.pop_front();
        if (op->op_type == OSD_OP_IN)
        {
            exec_op(op);
@@ -171,7 +179,6 @@ void osd_messenger_t::handle_immediate_ops()
            std::function<void(osd_op_t*)>(op->callback)(op);
        }
    }
-    set_immediate_ops.clear();
 }

 bool osd_messenger_t::handle_read_buffer(osd_client_t *cl, void *curbuf, int remain)
@@ -251,7 +251,7 @@ void osd_messenger_t::send_replies()
    {
        int peer_fd = write_ready_clients[i];
        auto cl_it = clients.find(peer_fd);
-        if (cl_it != clients.end() && !try_send(cl_it->second))
+        if (cl_it != clients.end() && cl_it->second->peer_state != PEER_RDMA && !try_send(cl_it->second))
        {
            write_ready_clients.erase(write_ready_clients.begin(), write_ready_clients.begin() + i);
            return;
@@ -349,21 +349,12 @@ void osd_messenger_t::handle_send(int result, bool prev, bool more, osd_client_t
 #ifdef WITH_RDMA
        if (cl->rdma_conn && !cl->outbox.size() && cl->peer_state == PEER_RDMA_CONNECTING)
        {
-            // FIXME: Do something better than just forgetting the FD
            // FIXME: Ignore pings during RDMA state transition
            if (log_level > 0)
            {
                fprintf(stderr, "Successfully connected with client %d using RDMA\n", cl->peer_fd);
            }
            cl->peer_state = PEER_RDMA;
-            tfd->set_fd_handler(cl->peer_fd, false, [this](int peer_fd, int epoll_events)
-            {
-                // Do not miss the disconnection!
-                if (epoll_events & EPOLLRDHUP)
-                {
-                    handle_peer_epoll(peer_fd, epoll_events);
-                }
-            });
            // Add the initial receive request
            init_recv_rdma(cl);
        }
@@ -57,6 +57,7 @@ void osd_messenger_t::stop_client(int peer_fd, bool force, bool force_delete)
    {
        return;
    }
+    clear_immediate_ops(peer_fd);
    if (log_level > 0)
    {
        if (cl->osd_num)
@@ -85,6 +86,16 @@ void osd_messenger_t::stop_client(int peer_fd, bool force, bool force_delete)
            osd_peer_fds.erase(osd_it);
        }
    }
+#ifdef WITH_RDMA
+    if (cl->rdma_conn && cl->rdma_conn->cmid)
+    {
+        auto rdma_it = rdmacm_connections.find(cl->rdma_conn->cmid);
+        if (rdma_it != rdmacm_connections.end() && rdma_it->second == cl)
+        {
+            rdmacm_connections.erase(rdma_it);
+        }
+    }
+#endif
 #ifndef __MOCK__
    // Then remove FD from the eventloop so we don't accidentally read something
    tfd->set_fd_handler(peer_fd, false, NULL);
@@ -20,6 +20,15 @@ typedef uint64_t inode_t;
 // Pool ID is 16 bits long
 typedef uint32_t pool_id_t;

+typedef uint64_t osd_num_t;
+typedef uint32_t pg_num_t;
+
+struct pool_pg_num_t
+{
+    pool_id_t pool_id;
+    pg_num_t pg_num;
+};
+
 // 16 bytes per object/stripe id
 // stripe = (start of the parity stripe + peer role)
 // i.e. for example (256KB + one of 0,1,2)
@@ -61,6 +70,21 @@ inline bool operator < (const obj_ver_id & a, const obj_ver_id & b)
    return a.oid < b.oid || a.oid == b.oid && a.version < b.version;
 }

+inline bool operator < (const pool_pg_num_t & a, const pool_pg_num_t & b)
+{
+    return a.pool_id < b.pool_id || a.pool_id == b.pool_id && a.pg_num < b.pg_num;
+}
+
+inline bool operator == (const pool_pg_num_t & a, const pool_pg_num_t & b)
+{
+    return a.pool_id == b.pool_id && a.pg_num == b.pg_num;
+}
+
+inline bool operator != (const pool_pg_num_t & a, const pool_pg_num_t & b)
+{
+    return a.pool_id != b.pool_id || a.pg_num != b.pg_num;
+}
+
 namespace std
 {
    template<> struct hash<object_id>
@@ -4,7 +4,6 @@
 #pragma once

 #include "object_id.h"
-#include "osd_id.h"

 // Magic numbers
 #define SECONDARY_OSD_OP_MAGIC      0x2bd7b10325434553l
@@ -6,7 +6,7 @@ includedir=${prefix}/@CMAKE_INSTALL_INCLUDEDIR@

 Name: Vitastor
 Description: Vitastor client library
-Version: 3.0.2
+Version: 3.0.6
 Libs: -L${libdir} -lvitastor_client
 Cflags: -I${includedir}

@@ -7,7 +7,6 @@

 #include "json11/json11.hpp"
 #include "object_id.h"
-#include "osd_id.h"
 #include "ringloop.h"
 #include <functional>

@@ -442,7 +442,7 @@ struct cli_dd_t
                }
                delete cur_read;
            }
-            else if (!is_zero(read_op->bitmap_buf, read_op->len/iinfo.in_granularity/8))
+            else if (!is_zero(read_op->bitmap_buf, (read_op->len/iinfo.in_granularity+7)/8))
            {
                vitastor_read(cur_read);
            }
@@ -890,12 +890,12 @@ resume_2:
        clock_gettime(CLOCK_REALTIME, &tv_begin);
        tv_progress = tv_begin;
 resume_3:
-        while ((ignore_errors || !copy_error) && (!in_eof || read_buffers.size() || in_waiting > 0 || out_waiting > 0))
+        while ((ignore_errors || !copy_error) && (!in_eof || read_buffers.size() || in_waiting > 0 || out_waiting > 0 || short_writes.size()))
        {
            print_progress(false);
            while ((ignore_errors || !copy_error) &&
                (!in_eof && in_waiting < in_iodepth && read_buffers.size() < out_iodepth ||
-                read_buffers.size() && out_waiting < out_iodepth))
+                (read_buffers.size() || short_writes.size()) && out_waiting < out_iodepth))
            {
                if (!in_eof && in_waiting < in_iodepth && read_buffers.size() < out_iodepth)
                {
@@ -904,7 +904,7 @@ resume_3:
                        break;
                    }
                }
-                if (read_buffers.size() && out_waiting < out_iodepth)
+                if ((read_buffers.size() || short_writes.size()) && out_waiting < out_iodepth)
                {
                    if (!add_write_op())
                    {
@@ -54,7 +54,7 @@ int disk_tool_t::trim_data(std::string device)
    fprintf(stderr, "Reading metadata\n");
    data_alloc = new allocator_t(dsk.block_count);
    r = process_meta(
-        [this](blockstore_meta_header_v3_t *hdr) {},
+        [](blockstore_meta_header_v3_t *hdr) {},
        [this](blockstore_heap_t *heap, heap_entry_t *obj, uint32_t meta_block_num)
        {
            for (auto wr = obj; wr; wr = heap->prev(wr))
@@ -3,7 +3,6 @@

 #include "disk_tool.h"
 #include "rw_blocking.h"
-#include "osd_id.h"
 #include "json_util.h"
 #include "malloc_or_die.h"

@@ -109,6 +108,7 @@ close_error:
            r = heap->load_blocks(meta_pos-dsk.meta_block_size, read_len, data, true, entries_loaded);
            meta_pos += read_len;
        }
+        heap->finish_load();
        heap->iterate_objects([&](heap_entry_t* obj, uint32_t meta_block_num)
        {
            obj_fn(heap, obj, meta_block_num);
@@ -331,7 +331,7 @@ void disk_tool_t::dump_meta_header(blockstore_meta_header_v3_t *hdr)
    }
    else
    {
-        printf("{\"version\":\"0.5\",\"meta_block_size\":%u,\"entries\":[\n", dsk.meta_block_size);
+        printf("{\"version\":\"0.5\",\"meta_block_size\":%ju,\"entries\":[\n", dsk.meta_block_size);
    }
    first_entry = true;
 }
@@ -4,7 +4,6 @@
 #include "disk_tool.h"
 #include "str_util.h"
 #include "json_util.h"
-#include "osd_id.h"

 void disk_tool_t::parse_meta_reserve()
 {
@@ -150,9 +149,12 @@ int disk_tool_t::prepare_one(std::map<std::string, std::string> options, int is_
    {
        if (options["block_size"] == "")
            options["block_size"] = "1M";
+        if (is_hybrid && options["atomic_write_size"] == "")
+            options["atomic_write_size"] = "0";
        if (is_hybrid && options["throttle_small_writes"] == "")
            options["throttle_small_writes"] = "1";
-        if (!is_hybrid && options.find("data_csum_type") != options.end() && options.at("data_csum_type") != "")
+        if (!is_hybrid && options.find("data_csum_type") != options.end() && options.at("data_csum_type") != "" &&
+            options["csum_block_size"] == "")
            options["csum_block_size"] = "32k";
    }
    else if (!json_is_true(options["disable_data_fsync"]))
@@ -212,11 +212,11 @@ void disk_tool_t::resize_init(blockstore_meta_header_v3_t *hdr)
    dsk.calc_lengths();
    if (((new_data_offset-dsk.data_offset) % dsk.data_block_size))
    {
-        fprintf(stderr, "Data alignment mismatch: old data offset is 0x%jx, new is 0x%jx, but alignment on %x should be equal\n",
+        fprintf(stderr, "Data alignment mismatch: old data offset is 0x%jx, new is 0x%jx, but alignment on %jx should be equal\n",
            dsk.data_offset, new_data_offset, dsk.data_block_size);
        exit(1);
    }
-    data_idx_diff = ((int64_t)(dsk.data_offset-new_data_offset)) / dsk.data_block_size;
+    data_idx_diff = ((int64_t)(dsk.data_offset-new_data_offset))/((int64_t)dsk.data_block_size);
    free_first = new_data_offset > dsk.data_offset ? (new_data_offset-dsk.data_offset) / dsk.data_block_size : 0;
    free_last = (new_data_offset+new_data_len < dsk.data_offset+dsk.data_len)
        ? (dsk.data_offset+dsk.data_len-new_data_offset-new_data_len) / dsk.data_block_size
@@ -351,7 +351,7 @@ int disk_tool_t::resize_copy_data()
                        if (data->res != dsk.data_block_size)
                        {
                            fprintf(
-                                stderr, "Failed to read %u bytes at %ju from %s: %s\n", dsk.data_block_size,
+                                stderr, "Failed to read %ju 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"
                            );
@@ -376,7 +376,7 @@ int disk_tool_t::resize_copy_data()
                        if (data->res != dsk.data_block_size)
                        {
                            fprintf(
-                                stderr, "Failed to write %u bytes at %ju to %s: %s\n", dsk.data_block_size,
+                                stderr, "Failed to write %ju 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"
                            );
@@ -140,7 +140,10 @@ uint32_t disk_tool_t::write_osd_superblock(std::string device, json11::Json para
    }
    close(fd);
    free(buf);
-    shell_exec({ "udevadm", "trigger", "--settle", device }, "", NULL, NULL);
+    if (!test_mode)
+    {
+        shell_exec({ "udevadm", "trigger", "--settle", device }, "", NULL, NULL);
+    }
    return sb_size;
 }

@@ -61,7 +61,7 @@ resume_1:
    }
    if (st->ientry["type"].string_value() != "file" &&
        st->ientry["type"].string_value() != "" &&
-        !st->set_attrs["size"].is_null())
+        st->set_attrs.find("size") != st->set_attrs.end())
    {
        auto cb = std::move(st->cb);
        cb(-EINVAL);
@@ -96,8 +96,13 @@ resume_1:
        nfs_kv_continue_setattr(st, 2);
    }, [st](int res, const std::string & cas_value)
    {
+        if ((res == 0 || res == -ENOENT && st->ino == KV_ROOT_INODE) && cas_value == st->ientry_text)
+        {
+            st->cas_res = 0;
+            return true;
+        }
        st->cas_res = res;
-        return (res == 0 || res == -ENOENT && st->ino == KV_ROOT_INODE) && cas_value == st->ientry_text;
+        return false;
    });
    return;
 resume_2:
@@ -809,21 +809,17 @@ again:
        goto again;
    }
    auto & pool_cfg = pool_it->second;
-    bool done = false;
-    if (pool_cfg.real_pg_count != pool_cfg.applied_pg_count ||
-        pool_cfg.pg_stripe_size != pool_cfg.applied_pg_stripe_size)
+    bool done = bs->reshard_continue(pool_cfg.reshard_state, pg_reshard_chunk_size);
+    if (done &&
+        (pool_cfg.real_pg_count != pool_cfg.applied_pg_count ||
+        pool_cfg.pg_stripe_size != pool_cfg.applied_pg_stripe_size))
    {
        // PG count changed again, reshard again
-        bs->reshard_abort(pool_cfg.reshard_state);
        pool_cfg.applied_pg_count = pool_cfg.real_pg_count;
        pool_cfg.applied_pg_stripe_size = pool_cfg.pg_stripe_size;
        pool_cfg.reshard_state = bs->reshard_start(pool_id, pool_cfg.real_pg_count, pool_cfg.pg_stripe_size, pg_reshard_chunk_size);
        done = !pool_cfg.reshard_state;
    }
-    else
-    {
-        done = bs->reshard_continue(pool_cfg.reshard_state, pg_reshard_chunk_size);
-    }
    if (done)
    {
        // Pool is resharded
@@ -1,30 +0,0 @@
-// Copyright (c) Vitaliy Filippov, 2019+
-// License: VNPL-1.1 or GNU GPL-2.0+ (see README.md for details)
-
-#pragma once
-
-#include "object_id.h"
-
-typedef uint64_t osd_num_t;
-typedef uint32_t pg_num_t;
-
-struct pool_pg_num_t
-{
-    pool_id_t pool_id;
-    pg_num_t pg_num;
-};
-
-inline bool operator < (const pool_pg_num_t & a, const pool_pg_num_t & b)
-{
-    return a.pool_id < b.pool_id || a.pool_id == b.pool_id && a.pg_num < b.pg_num;
-}
-
-inline bool operator == (const pool_pg_num_t & a, const pool_pg_num_t & b)
-{
-    return a.pool_id == b.pool_id && a.pg_num == b.pg_num;
-}
-
-inline bool operator != (const pool_pg_num_t & a, const pool_pg_num_t & b)
-{
-    return a.pool_id != b.pool_id || a.pg_num != b.pg_num;
-}
@@ -584,7 +584,7 @@ void osd_t::submit_list_subop(osd_num_t role_osd, pg_peering_state_t *ps)
                    .opcode = OSD_OP_SEC_LIST,
                },
                .list_pg = ps->pg_num,
-                .pg_count = pool_cfg.applied_pg_count,
+                .pg_count = (uint32_t)pool_cfg.applied_pg_count,
                .pg_stripe_size = pool_cfg.applied_pg_stripe_size,
                .min_inode = ((uint64_t)(ps->pool_id) << (64 - POOL_ID_BITS)),
                .max_inode = ((uint64_t)(ps->pool_id+1) << (64 - POOL_ID_BITS)) - 1,
@@ -334,36 +334,44 @@ pg_osd_set_state_t* pg_t::add_object_to_state(const object_id oid, const uint64_
    if (it == state_dict.end())
    {
        std::vector<osd_num_t> read_target;
+        bool found = false;
+        uint32_t bad_mask = (LOC_OUTDATED | LOC_CORRUPTED);
+retry:
        if (scheme == POOL_SCHEME_REPLICATED)
        {
            for (auto & o: osd_set)
            {
-                if (!(o.loc_bad & (LOC_OUTDATED | LOC_CORRUPTED)))
+                if (!(o.loc_bad & bad_mask))
                {
                    read_target.push_back(o.osd_num);
+                    found = true;
                }
            }
-            while (read_target.size() < pg_size)
+            if (read_target.size() < pg_size)
            {
                // FIXME: This is because we then use .data() and assume it's at least <pg_size> long
-                read_target.push_back(0);
+                read_target.resize(pg_size);
            }
        }
        else
        {
            read_target.resize(pg_size);
-            for (int i = 0; i < pg_size; i++)
-            {
-                read_target[i] = 0;
-            }
            for (auto & o: osd_set)
            {
-                if (!(o.loc_bad & (LOC_OUTDATED | LOC_CORRUPTED)))
+                if (!(o.loc_bad & bad_mask))
                {
                    read_target[o.role] = o.osd_num;
+                    found = true;
                }
            }
        }
+        if (!found && (bad_mask & LOC_CORRUPTED))
+        {
+            // Allow to try reading corrupted copies in rare cases when the object is corrupted on all OSDs
+            bad_mask = LOC_OUTDATED;
+            read_target.clear();
+            goto retry;
+        }
        state_dict[osd_set] = {
            .read_target = read_target,
            .osd_set = osd_set,
@@ -271,7 +271,6 @@ resume_0:
            }
            cur_op->buf = alloc_read_buffer(op_data->stripes, pg ? pg->pg_data_size : 1, 0);
            submit_primary_subops(SUBMIT_RMW_READ, op_data->target_ver, op_data->prev_set, cur_op);
-            op_data->st = 1;
        }
        else
        {
@@ -284,11 +283,14 @@ resume_0:
            op_data->degraded = 1;
            cur_op->buf = alloc_read_buffer(op_data->stripes, pg->pg_size, 0);
            submit_primary_subops(SUBMIT_RMW_READ, op_data->target_ver, op_data->prev_set, cur_op);
-            op_data->st = 1;
        }
    }
 resume_1:
-    return;
+    if (op_data->n_subops > 0)
+    {
+        op_data->st = 1;
+        return;
+    }
 resume_2:
    if (op_data->errors > 0)
    {
@@ -296,8 +298,9 @@ resume_2:
        {
            // I/O or checksum error
            // FIXME: ref = true ideally... because new_state != state is not necessarily true if it's freed and recreated
-            op_data->object_state = mark_object_corrupted(*pg, op_data->oid, op_data->object_state, op_data->stripes, false);
-            goto resume_0;
+            auto new_object_state = mark_object_corrupted(*pg, op_data->oid, op_data->object_state, op_data->stripes, false);
+            if (new_object_state != op_data->object_state)
+                goto resume_0;
        }
        finish_op(cur_op, op_data->errcode);
        return;
@@ -736,8 +739,11 @@ resume_1:
    submit_primary_subops(SUBMIT_RMW_READ, UINT64_MAX, op_data->prev_set, cur_op);
    op_data->prev_set = NULL;
 resume_2:
-    op_data->st = 2;
-    return;
+    if (op_data->n_subops > 0)
+    {
+        op_data->st = 2;
+        return;
+    }
 resume_3:
    if (op_data->errors > 0)
    {
@@ -134,11 +134,17 @@ void osd_t::submit_primary_subops(int submit_type, uint64_t op_version, const ui
        n_subops = 1;
    else
        zero_read = -1;
-    osd_op_t *subops = new osd_op_t[n_subops];
    op_data->fact_ver = 0;
    op_data->done = op_data->errors = op_data->drops = op_data->errcode = 0;
    op_data->n_subops = n_subops;
-    op_data->subops = subops;
+    if (!n_subops)
+    {
+        op_data->errcode = -EIO;
+        op_data->subops = NULL;
+        op_data->errors = 1;
+        return;
+    }
+    op_data->subops = new osd_op_t[n_subops];
    int sent = submit_primary_subop_batch(submit_type, op_data->oid.inode, op_version, op_data->stripes, osd_set, cur_op, 0, zero_read);
    assert(sent == n_subops);
 }
@@ -404,8 +410,9 @@ void osd_t::handle_primary_subop(osd_op_t *subop, osd_op_t *cur_op)
            if (op_data->fact_ver != 0 && op_data->fact_ver != version)
            {
                fprintf(
-                    stderr, "different fact_versions returned from %s %jx:%jx subops: %ju vs %ju\n",
-                    osd_op_names[opcode], subop->req.sec_rw.oid.inode, subop->req.sec_rw.oid.stripe, version, op_data->fact_ver
+                    stderr, "different fact_versions returned from %s %jx:%jx subops for a %s op: %ju vs %ju\n",
+                    osd_op_names[opcode], subop->req.sec_rw.oid.inode, subop->req.sec_rw.oid.stripe,
+                    osd_op_names[cur_op->req.hdr.opcode], version, op_data->fact_ver
                );
                retval = -ERANGE;
            }
@@ -108,57 +108,31 @@ retry_1:
        }
    }
    // Read required blocks
-    {
-        if (op_data->object_state && (op_data->object_state->state & OBJ_INCOMPLETE))
-        {
-            // Allow to read version number (just version number!) from corrupted chunks
-            // to allow full overwrite of a corrupted object
-            bool found = false;
-            for (int role = 0; role < pg.pg_size; role++)
-            {
-                if (op_data->prev_set[role] != 0 || op_data->stripes[role].read_end > op_data->stripes[role].read_start)
-                {
-                    found = true;
-                    break;
-                }
-            }
-            if (!found)
-            {
-                osd_num_t corrupted_target[pg.pg_size];
-                for (int role = 0; role < pg.pg_size; role++)
-                {
-                    corrupted_target[role] = 0;
-                }
-                for (auto & loc: op_data->object_state->osd_set)
-                {
-                    if (!(loc.loc_bad & LOC_OUTDATED) && !corrupted_target[loc.role])
-                    {
-                        corrupted_target[loc.role] = loc.osd_num;
-                    }
-                }
-                submit_primary_subops(SUBMIT_RMW_READ, UINT64_MAX, corrupted_target, cur_op);
-                goto resume_2;
-            }
-        }
-        submit_primary_subops(SUBMIT_RMW_READ, UINT64_MAX, op_data->prev_set, cur_op);
-    }
+    submit_primary_subops(SUBMIT_RMW_READ, UINT64_MAX, op_data->prev_set, cur_op);
 resume_2:
-    op_data->st = 2;
-    return;
+    if (op_data->n_subops > 0)
+    {
+        op_data->st = 2;
+        return;
+    }
 resume_3:
    if (op_data->errors > 0)
    {
        if (op_data->errcode == -EIO || op_data->errcode == -EDOM)
        {
            // Mark object corrupted and retry
-            op_data->object_state = mark_object_corrupted(pg, op_data->oid, op_data->object_state, op_data->stripes, true);
-            op_data->prev_set = op_data->object_state ? op_data->object_state->read_target.data() : pg.cur_set.data();
-            if (cur_op->rmw_buf)
+            pg_osd_set_state_t *new_object_state = mark_object_corrupted(pg, op_data->oid, op_data->object_state, op_data->stripes, true);
+            if (new_object_state != op_data->object_state)
            {
-                free(cur_op->rmw_buf);
-                cur_op->rmw_buf = NULL;
+                op_data->object_state = new_object_state;
+                op_data->prev_set = op_data->object_state ? op_data->object_state->read_target.data() : pg.cur_set.data();
+                if (cur_op->rmw_buf)
+                {
+                    free(cur_op->rmw_buf);
+                    cur_op->rmw_buf = NULL;
+                }
+                goto retry_1;
            }
-            goto retry_1;
        }
        deref_object_state(pg, &op_data->object_state, true);
        pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->errcode);
@@ -6,7 +6,6 @@
 #include <stdint.h>
 #include <vector>
 #include "object_id.h"
-#include "osd_id.h"

 struct buf_len_t
 {
@@ -57,3 +57,7 @@ json11::Json::object osd_messenger_t::merge_configs(const json11::Json::object &
 {
    return cli_config;
 }
+
+void osd_messenger_t::clear_immediate_ops(int peer_fd)
+{
+}
@@ -199,7 +199,7 @@ void disk_mock_t::erase_buffers(uint64_t begin, uint64_t end)
        {
            // Cut beginning & end & stop
            uint8_t *ce = (uint8_t*)malloc_or_die(be-end);
-            memcpy(ce, it->second.iov_base + (end-bs), be-end);
+            memcpy(ce, (uint8_t*)it->second.iov_base + (end-bs), be-end);
            uint8_t *cs = (uint8_t*)realloc(it->second.iov_base, begin-bs);
            if (!cs)
                throw std::bad_alloc();
@@ -221,7 +221,7 @@ void disk_mock_t::erase_buffers(uint64_t begin, uint64_t end)
            // Cut end & stop
            assert(be > end);
            uint8_t *ce = (uint8_t*)malloc_or_die(be-end);
-            memcpy(ce, it->second.iov_base + (end-bs), be-end);
+            memcpy(ce, (uint8_t*)it->second.iov_base + (end-bs), be-end);
            buffers[be] = (iovec){ .iov_base = ce, .iov_len = be-end };
            buffers.erase(it);
            break;
@@ -308,7 +308,7 @@ void disk_mock_t::read_item(uint8_t *to, uint64_t offset, uint64_t len)
            last = offset;
        }
        uint64_t cur_end = be < offset+len ? be : offset+len;
-        memcpy(to+last-offset, it->second.iov_base+last-bs, cur_end-last);
+        memcpy(to+last-offset, (uint8_t*)it->second.iov_base+last-bs, cur_end-last);
        last = be;
    }
    if (last < offset+len)
@@ -96,7 +96,7 @@ std::string str_replace(const std::string & in, const std::string & needle, cons
    {
        res += in.substr(pos, p2-pos);
        res += replacement;
-        pos = p2 + replacement.size();
+        pos = p2 + needle.size();
    }
    if (!pos)
    {
@@ -96,6 +96,11 @@ TEST_NAME=local_read POOLCFG='"local_reads":"random",' ./test_heal.sh
 SCHEME=ec ./test_heal.sh
 ANTIETCD=1 ./test_heal.sh

+./test_checksum.sh
+OLD=1 ./test_checksum.sh
+./test_corrupt_all.sh
+OLD=1 ./test_corrupt_all.sh
+
 ./test_reweight_half.sh
 ./test_snapshot_pool2.sh
 ./test_snapshot_read_bitmap.sh
@@ -0,0 +1,46 @@
+#!/bin/bash -ex
+
+SCHEME=replicated
+PG_SIZE=2
+OSD_COUNT=2
+IMG_SIZE=128
+OSD_ARGS="--data_csum_type crc32c --csum_block_size 4k --inmemory_journal false --journal_trim_interval $((IMG_SIZE*8)) $OSD_ARGS"
+OFFSET_ARGS="--data_csum_type crc32c --csum_block_size 4k $OFFSET_ARGS"
+GLOBAL_CONFIG=',"client_eio_retry_interval":0'
+. `dirname $0`/run_3osds.sh
+check_qemu
+
+$ETCDCTL put /vitastor/config/inode/1/1 '{"name":"testimg","size":'$((IMG_SIZE*1024*1024))'}'
+
+# Write
+$VITASTOR_FIO -bs=1M -direct=1 -iodepth=4 -end_fsync=1 -rw=write -image=testimg -runtime=10
+#$VITASTOR_FIO -bs=4k -direct=1 -iodepth=16 -end_fsync=1 -rw=randwrite -image=testimg -number_ios=10000
+
+# Intentionally corrupt OSD data and restart both of them
+kill $OSD1_PID $OSD2_PID
+data_offset=$(build/src/disk_tool/vitastor-disk simple-offsets ./testdata/bin/test_osd1.bin $OFFSET_ARGS | grep data_offset | awk '{print $2}')
+truncate -s $data_offset ./testdata/bin/test_osd1.bin
+dd if=/dev/zero of=./testdata/bin/test_osd1.bin bs=1024 count=1 seek=$((OSD_SIZE*1024-1))
+truncate -s $data_offset ./testdata/bin/test_osd2.bin
+dd if=/dev/zero of=./testdata/bin/test_osd2.bin bs=1024 count=1 seek=$((OSD_SIZE*1024-1))
+start_osd 1
+start_osd 2
+
+# Wait until start
+wait_up 10
+
+# Trigger scrub
+$ETCDCTL put /vitastor/pg/history/1/1 `$ETCDCTL get --print-value-only /vitastor/pg/history/1/1 | jq -s -c '(.[0] // {}) + {"next_scrub":1}'`
+
+# Wait for scrub to finish
+wait_condition 300 "$ETCDCTL get --prefix /vitastor/pg/history/ --print-value-only | jq -s -e '([ .[] | select(.next_scrub == 0 or .next_scrub == null) ] | length) == $PG_COUNT'" Scrubbing
+
+# Verify that ALL objects are now corrupted+incomplete
+$VITASTOR_CLI describe --json &>./testdata/describe.json
+$VITASTOR_CLI describe --json | jq -e '[ .[] | select(.corrupted) ] | length == '$((IMG_SIZE * 8 * PG_SIZE))
+
+# Check that we can remove or overwrite them
+$VITASTOR_FIO -bs=4M -direct=1 -iodepth=1 -end_fsync=1 -rw=write -offset=$((IMG_SIZE/2))M -image=testimg -runtime=10
+$VITASTOR_CLI rm-data --pool 1 --inode 1
+
+format_green OK
@@ -15,6 +15,8 @@ sudo mount localhost:/ ./testdata/nfs -o port=2050,mountport=2050,nfsvers=3,soft
 MNT=$(pwd)/testdata/nfs
 trap "sudo umount -f $MNT"' || true; kill -9 $(jobs -p)' EXIT

+chown 1000:1000 ./testdata/nfs
+
 touch ./testdata/nfs/f1
 chown 1000:1000 ./testdata/nfs/f1
 chmod 600 ./testdata/nfs/f1
@@ -15,7 +15,7 @@ trap "kill -9 $(jobs -p) || true; sudo losetup -d $LOOP1 $LOOP2"' || true' EXIT
 # also test prepare --hybrid :)
 # non-vitastor random type UUID to prevent udev activation
 mount | grep '/dev type devtmpfs' || sudo mount udev /dev/ -t devtmpfs
-sudo build/src/disk_tool/vitastor-disk-test prepare $OFFSET_ARGS --no_init 1 --meta_reserve 1x,1M \
+sudo -E build/src/disk_tool/vitastor-disk-test prepare $OFFSET_ARGS --no_init 1 --meta_reserve 1x,1M \
    --block_size 131072 --osd_num 987654 --part_type_uuid 0df42ae0-3695-4395-a957-7d5ff3645c56 \
    --hybrid --fast-devices $LOOP2 $LOOP1

@@ -27,8 +27,8 @@ console.log(JSON.stringify([
    {"type":"big_write_instant","inode":"0x1000000000001","stripe":"0xc60000","ver":"10","offset":0,"len":131072,"loc":"0x18ffdc0000","bitmap":"ffffffff"}
 ]));
 EOF
-    sudo build/src/disk_tool/vitastor-disk write-journal ${LOOP1}p1 < ./testdata/journal.json
-    sudo build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
+    sudo -E build/src/disk_tool/vitastor-disk write-journal ${LOOP1}p1 < ./testdata/journal.json
+    sudo -E build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
    jq -S '[ .[] + {"valid":true} ]' < ./testdata/journal.json > ./testdata/j1.json
    diff ./testdata/j1.json ./testdata/j2.json
 fi
@@ -84,15 +84,15 @@ EOF
 fi

 # also test write & dump
-sudo build/src/disk_tool/vitastor-disk write-meta ${LOOP1}p1 < ./testdata/meta.json
-sudo build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 > ./testdata/compare.json
+sudo -E build/src/disk_tool/vitastor-disk write-meta ${LOOP1}p1 < ./testdata/meta.json
+sudo -E build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 > ./testdata/compare.json
 jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' < ./testdata/meta.json > ./testdata/1.json
 jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' < ./testdata/compare.json > ./testdata/2.json
 diff ./testdata/1.json ./testdata/2.json

 # move journal & meta back, data will become smaller; end indexes should be shifted by -1251
-sudo build/src/disk_tool/vitastor-disk-test resize --move-journal '' --move-meta '' ${LOOP1}p1
-sudo build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
+sudo -E build/src/disk_tool/vitastor-disk-test resize --move-journal '' --move-meta '' ${LOOP1}p1
+sudo -E build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
 if [[ -n "$OLD" ]]; then
    jq -S '. + {"entries": ([ .entries[] | (. + { "block": (.block-1251) }) ] | sort_by(.stripe))}' < ./testdata/meta.json > ./testdata/1.json
 else
@@ -100,24 +100,24 @@ else
 fi
 diff ./testdata/1.json ./testdata/2.json
 if [[ -n "$OLD" ]]; then
-    sudo build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
+    sudo -E build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
    jq -S '[ (.[] + {"valid":true}) | (if .type == "big_write_instant" then . + {"loc":"0x18f6160000"} else . end) ]' < ./testdata/journal.json > ./testdata/j1.json
    diff ./testdata/j1.json ./testdata/j2.json
 fi

 # move journal & meta out, data will become larger; end indexes should be shifted back by +1251
-sudo build/src/disk_tool/vitastor-disk-test resize --move-journal ${LOOP2}p1 --move-meta ${LOOP2}p2 ${LOOP1}p1
-sudo build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
+sudo -E build/src/disk_tool/vitastor-disk-test resize --move-journal ${LOOP2}p1 --move-meta ${LOOP2}p2 ${LOOP1}p1
+sudo -E build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
 jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' < ./testdata/meta.json > ./testdata/1.json
 diff ./testdata/1.json ./testdata/2.json
 if [[ -n "$OLD" ]]; then
    jq -S '[ .[] + {"valid":true} ]' < ./testdata/journal.json > ./testdata/j1.json
-    sudo build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
+    sudo -E build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
 fi

 # reduce data device size by exactly 128k * 99 (occupied blocks); exactly 1 should be left in place :)
-sudo build/src/disk_tool/vitastor-disk-test resize --data-size $((DATA_DEV_SIZE-128*1024*99)) ${LOOP1}p1
-sudo build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
+sudo -E build/src/disk_tool/vitastor-disk-test resize --data-size $((DATA_DEV_SIZE-128*1024*99)) ${LOOP1}p1
+sudo -E build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
 if [[ -n "$OLD" ]]; then
    jq -S '. + {"entries": ([ .entries[] | (. + { "block": (.block | if . > '$BLOCK_COUNT'-100 then .-('$BLOCK_COUNT'-100+1) else '$BLOCK_COUNT'-100 end) }) ]
        | .[1:] + [ .[0] ]) | sort_by(.stripe)}' < ./testdata/meta.json > ./testdata/1.json
@@ -128,12 +128,12 @@ fi
 diff ./testdata/1.json ./testdata/2.json
 if [[ -n "$OLD" ]]; then
    jq -S '[ .[] + {"valid":true} ]' < ./testdata/journal.json > ./testdata/j1.json
-    sudo build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
+    sudo -E build/src/disk_tool/vitastor-disk dump-journal --json --format data ${LOOP1}p1 | jq -S '[ .[] | del(.crc32, .crc32_prev) ]' > ./testdata/j2.json
 fi

 # extend data device size to maximum
-sudo build/src/disk_tool/vitastor-disk-test resize --data-size max ${LOOP1}p1
-sudo build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
+sudo -E build/src/disk_tool/vitastor-disk-test resize --data-size max ${LOOP1}p1
+sudo -E build/src/disk_tool/vitastor-disk dump-meta ${LOOP1}p1 | jq -S '. + {"entries": (.entries | sort_by(.stripe)) }' > ./testdata/2.json
 diff ./testdata/1.json ./testdata/2.json

 format_green OK
@@ -0,0 +1,47 @@
+#!/bin/bash -ex
+
+# Test for Issue #112: Integer division bug causes crash with block_size < 32KB
+# This test verifies that small block sizes (4KB, 8KB, 16KB) work correctly
+# with 4KB bitmap_granularity, both for aligned and unaligned I/O
+
+# Arrange: Set up test environment with small block sizes
+export SCHEME=replicated
+export OSD_COUNT=3
+export PG_COUNT=1
+export PG_SIZE=2
+export OSD_SIZE=256
+
+# Test with 16KB block_size and 4KB bitmap_granularity
+# This should trigger the bug: 16384 / 4096 / 8 = 4 / 8 = 0 bytes
+export OFFSET_ARGS="--data_block_size 16384 --bitmap_granularity 4096"
+
+. `dirname $0`/run_3osds.sh
+
+# Act: Run I/O tests that exercise the bitmap code paths
+
+echo "Test 1: 128KB aligned I/O (may work even with bug)"
+LD_PRELOAD="build/src/client/libfio_vitastor.so" \
+    fio -thread -name=test -ioengine=build/src/client/libfio_vitastor.so \
+        -bs=128k -direct=1 -iodepth=4 -rw=randwrite \
+        -etcd=$ETCD_URL -pool=1 -inode=1 -size=64M -runtime=5
+
+echo "Test 2: 4KB random I/O (will crash with bug)"
+LD_PRELOAD="build/src/client/libfio_vitastor.so" \
+    fio -thread -name=test -ioengine=build/src/client/libfio_vitastor.so \
+        -bs=4k -direct=1 -iodepth=16 -rw=randwrite \
+        -etcd=$ETCD_URL -pool=1 -inode=1 -size=64M -runtime=5
+
+echo "Test 3: Mixed read/write with 4KB I/O"
+LD_PRELOAD="build/src/client/libfio_vitastor.so" \
+    fio -thread -name=test -ioengine=build/src/client/libfio_vitastor.so \
+        -bs=4k -direct=1 -iodepth=16 -rw=randrw -rwmixread=50 \
+        -etcd=$ETCD_URL -pool=1 -inode=1 -size=64M -runtime=5
+
+echo "Test 4: Sequential 4KB writes"
+LD_PRELOAD="build/src/client/libfio_vitastor.so" \
+    fio -thread -name=test -ioengine=build/src/client/libfio_vitastor.so \
+        -bs=4k -direct=1 -iodepth=1 -rw=write \
+        -etcd=$ETCD_URL -pool=1 -inode=1 -size=64M
+
+# Assert: If we reach here without crash, test passed
+format_green "OK: 16KB block_size with 4KB bitmap_granularity works correctly"
Author	SHA1	Message	Date
Vitaliy Filippov	538620b400	Release 3.0.6 - Fix one more bug in old store checksums O_o - checksum verification wasn't working correctly on reads quickly after small initial offsetted writes (when a BIG_WRITE was read from the journal). - Fix clients possibly not destroyed after being switched to RDMA. Not 100% verified, but in theory it could rarely lead to PGs hanging in peering/stopping states with RDMA enabled. - Use vitastor-cli in the Cinder driver for modifications instead of direct etcd interaction. - Fix mon_https_ca parsing. - Fix Docker install.sh, add instructions for Podman.	2026-03-22 22:49:51 +03:00
Vitaliy Filippov	4c34a47179	Fix str_replace	2026-03-19 01:19:29 +03:00
Vitaliy Filippov	df16ab627a	Fix another bug in old store checksums O_o Checksum verification wasn't working correctly on reads of small initial offsetted writes from the journal (until the write was flushed to "clean" metadata).	2026-03-17 22:26:38 +00:00
Vitaliy Filippov	44c895dc30	Fix clients possibly not destroyed after being switched to RDMA It could in theory lead to hung operations on RDMA client disconnections. Workflow (probable): * OSD switches an outbound connection to PEER_RDMA and removes the FD from epoll * But a receive request in io_uring is still active * Receive response is never handled, refs remains > 0 * The connection is dropped, but osd_client_t is not destroyed because refs > 0 * Outbound ops are not canceled, suboperations hang, primary operations hang too	2026-03-17 01:04:17 +03:00
Vitaliy Filippov	fdea595913	Use vitastor-cli in the Cinder driver for modifications instead of direct etcd interaction	2026-03-15 01:44:30 +03:00
Vitaliy Filippov	ef4c91ecc8	Fix mon_https_ca parsing	2026-03-15 01:01:41 +03:00
Vitaliy Filippov	5192c6cf50	Fix Docker install.sh, add instructions for Podman	2026-03-13 20:08:18 +03:00
Vitaliy Filippov	cb639a130d	Release 3.0.5 A single urgent fix for the new store: metadata was written incorrectly when filled metadata area exceeded 4 GB due to a simple & stupid 32-bit integer overflow. This was leading to OSDs corrupting metadata and being unable to start on the next run. The bug affected only the new store (meta_format=3) and OSDs with metadata area size larger than 4 GB. You can check your OSD metadata area size by checking `data_offset` in `vitastor-disk read-sb /dev/vitastor/osdXX-data` for single-disk OSDs or `blockdev --getsize64 /dev/vitastor/osdXX-meta` for hybrid OSDs. If you're affected, it's highly recommended to: - Install updated packages without restarting OSDs - Restart OSDs one by one - If a restart of an OSD triggers "has_degraded" rebalance - wait for it to finish before restarting another OSD - If an OSD fails to start with "double-claimed" errors in the log - purge and recreate it, and again wait for "has_degraded" rebalance to finish before proceeding to the next OSD	2026-03-07 13:54:44 +03:00
Vitaliy Filippov	892e3a8b6d	Add Ubuntu 26.04 build	2026-03-07 13:54:44 +03:00
Vitaliy Filippov	0bd9c26620	Make block_sizes uint64_t to prevent other integer overflow errors in the future	2026-03-07 12:02:57 +03:00
Vitaliy Filippov	ae2b1f7802	Some useless code changes for beauty	2026-03-07 00:15:16 +03:00
Vitaliy Filippov	883b2e45da	Fix uint32 overflow in the new store :-X	2026-03-06 17:21:22 +03:00
Vitaliy Filippov	09df74cfac	Release 3.0.4 New store bug fixes: - Fix very slow OSD startup (hours) possible in unlucky cases - Fix incorrect garbage entry validation on start in some cases, preventing OSD startup - Fix small_write_throttling not functioning and either hanging or crashing OSDs - Fix "different fact_versions returned from subops" error possible with EC in some cases - Add a recovery option for the new store: skip_corrupted_meta_entries - Disable intent writes (even 4k) for hybrid OSDs by default for better write buffering Other changes: - Fix chmod/chown of a new empty VitastorFS root directory not working - Fix vitastor-dd treating all < 32k reads as always empty - Fix vitastor-disk always setting 32k csum_block_size for HDDs when checksums are enabled - Fix OSD missing PG state updates and PGs hanging in "starting" in rare cases - Fix a possible OSD crash with "map::at" message in rare cases when stopping a client - Fix a possible OSD crash on start when RDMA is available but fails to initialize - Fix removal and overwriting of totaled objects (objects with all corrupted copies) - Do not allow data_block_size < 8*bitmap_granularity - Remove RDMA ODP (On-Demand Paging) support - Update antietcd to 1.2.4 - Update Docker build to Debian Trixie - Add packages for AlmaLinux 10	2026-03-03 20:12:31 +03:00
Vitaliy Filippov	2ef3f012c9	Fix Docker builds	2026-03-03 20:09:23 +03:00
Vitaliy Filippov	7175d99c64	Add packages for Alma 10 to docs	2026-03-03 11:17:20 +03:00
Vitaliy Filippov	230a26772e	Do not allow data_block_size < 8*bitmap_granularity	2026-03-03 02:09:16 +03:00
Vitaliy Filippov	637684c579	Fix chmod/chown of empty NFS root directory	2026-03-03 02:05:37 +03:00
Vitaliy Filippov	c541dd422f	Add skip_corrupted_meta_entries docs	2026-03-03 01:52:22 +03:00
Vitaliy Filippov	22c39561cf	Add a recovery option for the new store: skip_corrupted_meta_entries	2026-03-02 02:51:09 +03:00
Vitaliy Filippov	d4c67b4879	Ignore array-bounds warning in bs_heap	2026-03-02 02:49:56 +03:00
Vitaliy Filippov	d1b7167861	Fix vitastor-disk always overwriting csum_block_size with 32k, add default atomic_write_size 0 for hybrid OSDs	2026-03-01 23:47:35 +00:00
Vitaliy Filippov	33b561b73a	Fix throttling with the new store	2026-03-01 23:46:30 +00:00
Vitaliy Filippov	a95a60c600	Add sudo -E to preserve asan suppressions	2026-03-02 01:48:50 +03:00
Vitaliy Filippov	b3bc815354	Sorry, udevadm trigger --settle sometimes just fails in docker :-(	2026-03-02 01:48:50 +03:00
Vitaliy Filippov	42fc45d6da	Use iterate_with_stable in list_objects too Fixes "different fact_versions returned from subops" with EC in some cases	2026-03-02 01:48:33 +03:00
Vitaliy Filippov	3fbe2e7f8a	Add more details for "different fact_versions"	2026-03-02 01:48:33 +03:00
Vitaliy Filippov	310c512b43	Correctly delete handle_immediate_ops when stopping a client Fixes map::at crash in some cases	2026-03-02 01:48:19 +03:00
Vitaliy Filippov	fd3e3b4ef0	Update buildenv to bookworm	2026-03-02 01:47:05 +03:00
Vitaliy Filippov	93cf69f89f	Merge osd_id.h into object_id.h	2026-03-02 01:46:17 +03:00
Vitaliy Filippov	a8dd8bf06c	Update antietcd to 1.2.4	2026-03-02 01:45:56 +03:00
Vitaliy Filippov	7e23b57014	Do not save watch revision from initial created: true messages	2026-02-27 01:11:18 +03:00
Vitaliy Filippov	b2427836a1	Fix dd treating all < 32k reads as always empty :-X	2026-02-22 14:19:13 +03:00
Vitaliy Filippov	2ed6760447	Remove ODP support	2026-02-17 02:35:46 +03:00
Vitaliy Filippov	a1d215ea2f	Fix garbage validation, skip garbage entries on load	2026-02-17 01:12:26 +03:00
Vitaliy Filippov	4fa442a5de	Sort items on load to fix O(N^2) startup	2026-02-17 01:12:26 +03:00
Vitaliy Filippov	3a057ed5af	Update docker to trixie	2026-02-13 01:09:34 +03:00
Vitaliy Filippov	0ad042b28b	Add packages for Alma 10	2026-02-12 02:11:06 +03:00
Vitaliy Filippov	4d75c6c8f9	Fix removal and overwriting of totaled objects (objects with all corrupted copies)	2026-02-11 01:24:58 +03:00
Vitaliy Filippov	9ec18f0aa1	Fix double ibv_close_device on RDMA init failure	2026-02-11 00:32:12 +03:00
Vitaliy Filippov	facaa2cc6a	Release 3.0.3 - Fix csum_block_size > 0 unusable with atomic writes in the new store (almost all atomic write requests generated invalid checksums with csum_block_size > 0) - Fix monitor sometimes randomly failing to optimize PGs with the "problem is infeasible or unbounded" message due to not waiting to read full stdout of lp_solve - Remove reshard_abort optimisation to fix chunked resharding introduced in 3.0.2 possibly corrupting in-memory OSD state when handling multiple rapid PG count change requests - Fix removed inodes not disappearing from OSD statistics in the new store, leading to bloating of the statistics with old inodes - Increase test coverage for the new store and fix several minor bugs: - Enabling/disabling of used_for_app was recalculating inode space statistics incorrectly - Fix object metadata validation on OSD startup rejecting some correct sequences of entries, possibly leading to OSD being unable to start - Fix PG activation with EC failing in rare cases with EBUSY when requesting to commit an already committed write - Fix a theoretically possible metadata writeback issue on ENOSPC during commit - Fix monitor failing to optimize PGs in presence of a host with name convertible to js Number (like 04e278988710) :D - Fix vitastor-cli dd sometimes (rarely) truncating the image when writing to stdout - Fix a theoretically possible client connection object leak when io_uring is full - Fix a leak of RDMA-CM connection objects - Fix crashes with data_block_size < 32KB (useless setup, but anyway) (#113)	2026-02-08 01:39:57 +03:00
Vitaliy Filippov	d2ac8e3827	Fix block writeback on enospc during commit	2026-02-07 16:27:26 +03:00
Vitaliy Filippov	d6dacc67db	Do not return EBUSY when already stable in the new store	2026-02-07 16:10:39 +03:00
Vitaliy Filippov	cbe51595cb	Remove reshard_abort	2026-02-07 16:05:40 +03:00
Vitaliy Filippov	6bd0830ab8	Add a test for chunked resharding	2026-02-07 16:02:17 +03:00
Vitaliy Filippov	796e82f34d	Check get_compacted_count and get_compact_queue_size	2026-02-07 15:44:54 +03:00
Vitaliy Filippov	db39283970	Add a test for complete_lsn_write	2026-02-07 15:36:46 +03:00
Vitaliy Filippov	9f8c686321	Add a test for redirect_intent	2026-02-07 15:09:58 +03:00
Vitaliy Filippov	2618e559d1	Add rollback tests, simplify rollback logic	2026-02-07 15:09:58 +03:00
Vitaliy Filippov	037d2bc162	Add a test for recalc_inode_stats in the new store and fix a bug in it	2026-02-07 14:00:10 +03:00
Vitaliy Filippov	f0b64adb32	Fix removed inodes not disappearing from statistics in the new store	2026-02-07 02:48:49 +03:00
Vitaliy Filippov	f06d64d879	Fix validate_object commit validation bug	2026-02-06 18:52:37 +03:00
Vitaliy Filippov	d164499a1c	Add more iterate_compaction test cases	2026-02-06 18:52:19 +03:00
Vitaliy Filippov	725e9aa8ae	Remove unused read_locked_entry	2026-02-06 17:38:59 +03:00
Vitaliy Filippov	0715feffa1	Wait for finished read, not for process exit in lp_optimizer	2026-02-04 02:33:08 +03:00
Vitaliy Filippov	e9f37e8dc3	Fix NaN PG optimisation error when a hostname has a name convertible to js Number... some shit like 04e278988710	2026-02-04 15:08:03 +03:00
Vitaliy Filippov	4fbe4b5654	Fix BIG_INTENT rechecks with enabled checksums	2026-02-04 02:33:50 +03:00
Vitaliy Filippov	9fb645693b	Fix checksums with BIG_INTENT writes in the new store	2026-02-04 02:33:50 +03:00
Vitaliy Filippov	9e47828383	Add a test for big_intent with checksums	2026-02-04 02:33:50 +03:00
Vitaliy Filippov	ac2ce48cb2	Fix missing wait for short writes in vitastor-cli dd	2026-02-03 01:28:13 +03:00
Vitaliy Filippov	9cc2beed95	Fix new compile warnings from merged PR	2026-01-30 00:47:58 +03:00
MoonBoi9001andVitaliy Filippov	fb1c870f5c	Fix: integer division bug with block_size < 32KB (#113 ) Test scenarios: - 128KB aligned I/O (may work despite bug) - 4KB random I/O (triggers assertion failure) - 4KB mixed read/write (triggers crash) - 4KB sequential write (triggers crash) I accept Vitastor CLA agreement: https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/CLA-en.md	2026-01-29 02:04:30 +03:00
ston3luandVitaliy Filippov	2d616d8058	Fix VLA init in test_heap I accept Vitastor CLA agreement: https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/CLA-en.md	2026-01-29 02:01:21 +03:00
Vitaliy Filippov	3f7f6f442b	Fix some compilation warnings	2026-01-29 02:00:25 +03:00
ston3luandVitaliy Filippov	7e7b95eeb4	Fix void* arithmetic in memset/memcpy I accept Vitastor CLA agreement: https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/CLA-en.md	2026-01-29 01:59:35 +03:00
ston3luandVitaliy Filippov	dd588a0783	Fix ref leak on read path when SQE allocation fails I accept Vitastor CLA agreement: https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/CLA-en.md	2026-01-29 01:52:07 +03:00
ston3luandVitaliy Filippov	028a6cab68	Clean RDMA entry I accept Vitastor CLA agreement: https://git.yourcmc.ru/vitalif/vitastor/src/branch/master/CLA-en.md	2026-01-29 01:51:31 +03:00