WIP RDMA credits

.gitea/workflows/buildenv.Dockerfile

@@ -1,29 +1,28 @@
-FROM node:16-bookworm
+FROM node:16-bullseye
 
 WORKDIR /root
 
-ADD ./docker/etc/apt/trusted.gpg.d /etc/apt/trusted.gpg.d
+ADD ./docker/vitastor.gpg /etc/apt/trusted.gpg.d
 
-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; \
+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; \
     echo >> /etc/apt/preferences; \
     echo 'Package: *' >> /etc/apt/preferences; \
-    echo 'Pin: release n=bookworm-backports' >> /etc/apt/preferences; \
+    echo 'Pin: release a=bullseye-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 libasan8 \
+RUN apt-get -y install etcd qemu-system-x86 qemu-block-extra qemu-utils fio libasan5 \
     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 libc-ares-dev udev
+RUN apt-get update && apt-get -y install jq lp-solve sudo nfs-common fdisk parted
 RUN apt-get --download-only source fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`
 
 RUN set -ex; \

.gitea/workflows/test.yml

@@ -234,60 +234,6 @@ jobs:
           echo ""
         done
 
-  test_etcd_fail_https:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 10
-      run: ETCD_SCHEME=https /root/vitastor/tests/test_etcd_fail.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_etcd_fail_https_antietcd:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 10
-      run: ETCD_SCHEME=https ANTIETCD=1 /root/vitastor/tests/test_etcd_fail.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_snapshot_https:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: ETCD_SCHEME=https /root/vitastor/tests/test_snapshot.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_interrupted_rebalance:
     runs-on: ubuntu-latest
     needs: build
@@ -468,24 +414,6 @@ jobs:
           echo ""
         done
 
-  test_level_placement:
-    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_level_placement.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_snapshot:
     runs-on: ubuntu-latest
     needs: build
@@ -702,24 +630,6 @@ jobs:
           echo ""
         done
 
-  test_snapshot_chain_encrypted:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: ENCRYPTED=1 /root/vitastor/tests/test_snapshot_chain.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_snapshot_chain:
     runs-on: ubuntu-latest
     needs: build
@@ -1278,96 +1188,6 @@ 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_checksum_xxhash:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: TEST_NAME=xxhash OSD_ARGS="--data_csum_type xxh3_32" /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
@@ -1980,24 +1800,6 @@ jobs:
           echo ""
         done
 
-  test_old_partwr_csum:
-    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_partwr_csum.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_heal_old_csum_32k_dmj:
     runs-on: ubuntu-latest
     needs: build
@@ -2124,57 +1926,3 @@ jobs:
           echo ""
         done
 
-  test_nfs_unaligned_append:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: /root/vitastor/tests/test_nfs_unaligned_append.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_write_encrypted:
-    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_write_encrypted.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_write_encrypted_ec:
-    runs-on: ubuntu-latest
-    needs: build
-    container: ${{env.TEST_IMAGE}}:${{github.sha}}
-    steps:
-    - name: Run test
-      id: test
-      timeout-minutes: 3
-      run: SCHEME=ec /root/vitastor/tests/test_write_encrypted.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
-

.gitea/workflows/tests-to-yaml.pl

@@ -38,14 +38,6 @@ for my $line (<>)
             {
                 $test_name .= '_antietcd';
             }
-            elsif ($1 eq 'ETCD_SCHEME' && $2 eq 'https')
-            {
-                $test_name .= '_https';
-            }
-            elsif ($1 eq 'ENCRYPTED')
-            {
-                $test_name .= '_encrypted';
-            }
             elsif ($1 eq 'OLD')
             {
                 $test_name =~ s/^test_/test_old_/s;

.gitignore

@@ -3,4 +3,3 @@
 package-lock.json
 fio
 qemu
-node_modules

CMakeLists.txt

@@ -2,7 +2,7 @@ cmake_minimum_required(VERSION 2.8.12)
 
 project(vitastor)
 
-set(VITASTOR_VERSION "3.0.5")
+set(VITASTOR_VERSION "3.0.0")
 
 include(CTest)
 

README-ru.md

@@ -62,7 +62,6 @@ Vitastor поддерживает QEMU-драйвер, протоколы UBLK,
     - [Дисковые параметры OSD](docs/config/layout-osd.ru.md)
     - [Прочие параметры OSD](docs/config/osd.ru.md)
     - [Параметры мониторов](docs/config/monitor.ru.md)
-    - [Безопасность](docs/config/security.ru.md)
   - [Настройки пулов](docs/config/pool.ru.md)
   - [Метаданные образов в etcd](docs/config/inode.ru.md)
 - Использование

README.md

@@ -62,7 +62,6 @@ Read more details in the documentation. You can start from here: [Quick Start](d
     - [OSD Disk Layout](docs/config/layout-osd.en.md)
     - [OSD Runtime Parameters](docs/config/osd.en.md)
     - [Monitor](docs/config/monitor.en.md)
-    - [Security](docs/config/security.en.md)
   - [Pool configuration](docs/config/pool.en.md)
   - [Image metadata in etcd](docs/config/inode.en.md)
 - Usage

csi/Dockerfile

@@ -1,5 +1,5 @@
 # Compile stage
-FROM golang:trixie AS build
+FROM golang:bookworm 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:trixie
+FROM debian:bookworm
 
 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
-        libaio1t64 libc6 libfuse3-4 libglib2.0-0t64 libgmp10 libgnutls30t64 \
-        libhogweed6t64 libnettle8t64 libnuma1 libselinux1 liburing2 libzstd1 zlib1g && \
+        libnuma1 liburing2 libglib2.0-0 libfuse3-3 libaio1 libzstd1 libnettle8 \
+        libgmp10 libhogweed6 libp11-kit0 libidn2-0 libunistring2 libtasn1-6 libpcre2-8-0 libffi8 && \
     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 trixie main > /etc/apt/sources.list.d/vitastor.list) && \
+RUN (echo deb http://vitastor.io/debian bookworm 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-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 && \
+    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 && \
     dpkg -x qemu-utils*.deb tmp1 && \
     dpkg -x qemu-block-extra*.deb tmp1 && \
     cp -a tmp1/usr/bin/qemu-storage-daemon /usr/bin/ && \

csi/Dockerfile.local

@@ -1,5 +1,5 @@
 # Compile stage
-FROM golang:trixie AS build
+FROM golang:bookworm 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:trixie
+FROM debian:bookworm
 
 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-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 && \
+    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 && \
     dpkg -x qemu-utils*.deb tmp1 && \
     dpkg -x qemu-block-extra*.deb tmp1 && \
     cp -a tmp1/usr/bin/qemu-storage-daemon /usr/bin/ && \

csi/Makefile

@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.5
+VITASTOR_VERSION ?= v3.0.0
 
 all: build push
 

csi/deploy/004-csi-nodeplugin.yaml

@@ -49,7 +49,7 @@ spec:
             capabilities:
               add: ["SYS_ADMIN"]
             allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v3.0.5
+          image: vitalif/vitastor-csi:v3.0.0
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/deploy/007-csi-provisioner.yaml

@@ -121,7 +121,7 @@ spec:
             privileged: true
             capabilities:
               add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v3.0.5
+          image: vitalif/vitastor-csi:v3.0.0
           args:
             - "--node=$(NODE_ID)"
             - "--endpoint=$(CSI_ENDPOINT)"

csi/src/config.go

@@ -5,7 +5,7 @@ package vitastor
 
 const (
     vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "3.0.5"
+    vitastorCSIDriverVersion = "3.0.0"
 )
 
 // Config struct fills the parameters of request or user input

debian/build-vitastor-ubuntu-resolute.sh

@@ -1,5 +0,0 @@
-#!/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

debian/changelog

@@ -1,4 +1,4 @@
-vitastor (3.0.5-1) unstable; urgency=medium
+vitastor (3.0.0-1) unstable; urgency=medium
 
   * Bugfixes
 

debian/control

@@ -3,7 +3,7 @@ Section: admin
 Priority: optional
 Maintainer: Vitaliy Filippov <vitalif@yourcmc.ru>
 Build-Depends: debhelper, g++ (>= 8), libstdc++6 (>= 8),
-  linux-libc-dev, libgoogle-perftools-dev, libjerasure-dev, libgf-complete-dev, libc-ares-dev,
+  linux-libc-dev, libgoogle-perftools-dev, libjerasure-dev, libgf-complete-dev,
   libibverbs-dev, librdmacm-dev, libisal-dev, cmake, pkg-config, libnl-3-dev, libnl-genl-3-dev,
   node-bindings <!nocheck>, node-gyp, node-nan
 Standards-Version: 4.5.0

debian/vitastor-buildenv.Dockerfile

@@ -25,7 +25,7 @@ RUN set -e -x; \
     echo 'APT::Install-Suggests false;' >> /etc/apt/apt.conf
 
 RUN apt-get update && \
-    apt-get -y install fio libgoogle-perftools-dev devscripts libjerasure-dev cmake libc-ares-dev \
+    apt-get -y install fio libgoogle-perftools-dev devscripts libjerasure-dev cmake \
         libibverbs-dev librdmacm-dev libisal-dev libnl-3-dev libnl-genl-3-dev curl nodejs npm node-nan node-bindings && \
     apt-get -y build-dep fio && \
     apt-get --download-only source fio

docker/Dockerfile

@@ -1,6 +1,6 @@
 # Build Docker image with Vitastor packages
 
-FROM debian:trixie
+FROM debian:bookworm
 
 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

docker/Makefile

@@ -1,4 +1,4 @@
-VITASTOR_VERSION ?= v3.0.5
+VITASTOR_VERSION ?= v3.0.0
 
 all: build push
 

docker/etc/apt/preferences

@@ -1,3 +1,3 @@
 Package: *
-Pin: release n=trixie-backports
+Pin: release n=bookworm-backports
 Pin-Priority: 500

docker/etc/apt/sources.list.d/vitastor.list

@@ -1,2 +1,2 @@
-deb http://vitastor.io/debian trixie main
-#deb http://http.debian.net/debian/ trixie-backports main
+deb http://vitastor.io/debian bookworm main
+deb http://http.debian.net/debian/ bookworm-backports main

docker/etc/systemd/system/vitastor-host.service

@@ -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 -e SYSTEMD_IN_CHROOT=0 \
+ExecStart=bash -c 'docker run --rm -i -v /etc/vitastor:/etc/vitastor -v /dev:/dev -v /run:/run \
     --security-opt seccomp=unconfined --privileged --pid=host --log-driver none --network host --name vitastor vitastor:$VITASTOR_VERSION \
     sleep.sh'
 ExecStartPost=udevadm trigger

docker/etc/vitastor/docker.conf

@@ -4,7 +4,7 @@
 #
 
 # Desired Vitastor version
-VITASTOR_VERSION=v3.0.5
+VITASTOR_VERSION=v3.0.0
 
 # Additional arguments for all containers
 # For example, you may want to specify a custom logging driver here

docker/install.sh

@@ -2,7 +2,8 @@
 
 set -e
 
-cp -urv /etc/systemd/system/vitastor* /host-etc/systemd/system/
-cp -urv /etc/udev/rules.d /host-etc/udev/
+cp -urv /etc/default /host-etc/
+cp -urv /etc/systemd /host-etc/
+cp -urv /etc/udev /host-etc/
 cp -urnv /etc/vitastor /host-etc/
 cp -urnv /opt/scripts/* /host-bin/

docs/config.en.md

@@ -38,4 +38,3 @@ In the future, additional configuration methods may be added:
 - [OSD Disk Layout](config/layout-osd.en.md)
 - [OSD Runtime Parameters](config/osd.en.md)
 - [Monitor](config/monitor.en.md)
-- [Security Parameters](config/security.en.md)

docs/config.ru.md

@@ -41,4 +41,3 @@
 - [Дисковые параметры OSD](config/layout-osd.ru.md)
 - [Прочие параметры OSD](config/osd.ru.md)
 - [Параметры мониторов](config/monitor.ru.md)
-- [Параметры безопасности](config/security.ru.md)

docs/config/layout-osd.en.md

@@ -198,14 +198,8 @@ put a modified value into etcd key /vitastor/config/global.
 - Type: string
 - Default: none
 
-Data and metadata checksum type to use. May be "crc32c", "xxh3_32" or "none".
-Select crc32c or xxh3_32 and set csum_block_size to enable data checksums.
-
-Both crc32c and xxh3_32 are almost equally fast, xxh3_32 is safer. xxh3_32 is
-the xxhash3 algorithm truncated from 64 to 32 bits (which is still a good hash).
-
-Note that enabled data checksums either increase memory usage or reduce
-performance. Check details in [csum_block_size](#csum_block_size) description.
+Data checksum type to use. May be "crc32c" or "none". Set to "crc32c" to
+enable data checksums.
 
 ## csum_block_size
 

docs/config/layout-osd.ru.md

@@ -209,12 +209,8 @@ journal_block_size и meta_block_size. Однако на данный момен
 - Тип: строка
 - Значение по умолчанию: none
 
-Тип используемых OSD контрольных сумм данных и метаданных. Может быть "crc32c",
-"xxh3_32" или "none". Выберите crc32c или xxh3_32 и установите csum_block_size,
-чтобы включить контрольные суммы данных.
-
-И crc32c, и xxh3_32 примерно одинаково быстры, xxh3_32 надёжней. xxh3_32 - это
-алгоритм xxhash3, обрезанный с 64 до 32 бит (это всё равно хороший хеш).
+Тип используемых OSD контрольных сумм данных. Может быть "crc32c" или "none".
+Установите в "crc32c", чтобы включить расчёт и проверку контрольных сумм данных.
 
 Следует понимать, что контрольные суммы в зависимости от размера блока их
 расчёта либо увеличивают потребление памяти, либо снижают производительность.

docs/config/network.en.md

@@ -22,6 +22,7 @@ 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)
@@ -101,6 +102,11 @@ 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.
 
@@ -110,23 +116,6 @@ 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
@@ -198,6 +187,28 @@ 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

docs/config/network.ru.md

@@ -22,6 +22,7 @@
 - [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)
@@ -100,6 +101,12 @@ 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.
 
@@ -110,24 +117,6 @@ libibverbs < v32, например в Debian 10 Buster или CentOS 7.
 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
@@ -203,6 +192,29 @@ 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

docs/config/osd.en.md

@@ -38,7 +38,6 @@ 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)
@@ -68,8 +67,6 @@ with an OSD restart or, for some of them, even without restarting by updating co
 - [pg_lock_retry_interval_ms](#pg_lock_retry_interval_ms)
 - [atomic_write_size](#atomic_write_size)
 - [use_atomic_flag](#use_atomic_flag)
-- [pg_reshard_chunk_size](#pg_reshard_chunk_size)
-- [pg_reshard_chunk_pause_ms](#pg_reshard_chunk_pause_ms)
 
 ## bind_address
 
@@ -280,19 +277,13 @@ Maximum number of journal flushers (see above min_flusher_count).
 - Type: boolean
 - Default: true
 
-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.
+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.
 
 ## inmemory_journal
 
@@ -371,8 +362,6 @@ 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
@@ -386,17 +375,6 @@ 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
@@ -704,10 +682,7 @@ with replicated pools and reach the best possible write performance.
 
 Default value is auto-detected during OSD initialization from
 `/sys/block/xx/queue/atomic_write_max_bytes` or assumed to be 4096 bytes
-because all known disks support 4 KB atomic writes. Auto-detection is only used for
-NVMe disks because SAS disks require the explicit WRITE ATOMIC command which requires
-RWF_ATOMIC (see below [#use_atomic_flag]) but that flag works incorrectly in current
-Linux versions.
+because all known disks support 4 KB atomic writes.
 
 You can also check if your NVMe drives support atomic writes by running
 the command `nvme id-ctrl /dev/nvme0n1 | grep awupf`. If the reported value,
@@ -722,34 +697,12 @@ reducing Write Amplification and improving write performance up to 2 times.
 
 - Type: boolean
 
-This option controls whether Vitastor OSDs use RWF_ATOMIC write flag with atomic writes.
-This flag is supported since Linux 6.11 and adds some safety to atomic writes - the kernel
-guarantees to not fragment write requests with it and also to check them against the actual
-device atomic write capabilities.
+This option controls whether the Vitastor OSD uses RWF_ATOMIC write flag with atomic
+writes. This flag is only supported on Linux kernel since 6.11. Atomic writes are
+generally only safe to use with this flag because it tells the kernel to never fragment
+write requests and also to check the write against the actual atomic write capabilities
+of the device.
 
-However, the option is disabled by default because the flag is currently UNUSABLE - Linux
-incorrectly requires writes with that flag to be of power-of-2 length and length-aligned.
-I.e., for example, 12 KB writes and not-8-KB aligned 8 KB writes are forbidden by the kernel,
-even though the NVMe specification allows them.
-
-For NVMe disks with `scheduler=none` writes aren't fragmented anyway so it's not a big deal.
-However, you can rebuild your kernel with [this patch](../../patches/linux-fix-atomic-write-checks.diff)
-and turn this option on. It will make your atomic writes a bit safer.
-
-## pg_reshard_chunk_size
-
-- Type: integer
-- Default: 100000
-
-Pool PG count change is a CPU-intensive operation because OSDs store the full object database
-in memory and have to move all entries between old and new PGs. Thus it's performed in chunks,
-with pauses between chunks to prevent blocking OSD's event loop and other clients' operations.
-This option sets the maximum number of object is a chunk. Moving 100k objects usually takes
-50-100ms. Chunk size equal to 0 means unlimited.
-
-## pg_reshard_chunk_pause_ms
-
-- Type: milliseconds
-- Default: 100
-
-This option sets the interval between handling two PG count change chunks.
+This option is enabled by default when atomic_write_size is set to a value larger than 4 KB.
+You can disable it if you're sure that your disks support atomic writes and you want to
+bypass the Linux atomic write checks.

docs/config/osd.ru.md

@@ -39,7 +39,6 @@
 - [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)
@@ -69,8 +68,6 @@
 - [pg_lock_retry_interval_ms](#pg_lock_retry_interval_ms)
 - [atomic_write_size](#atomic_write_size)
 - [use_atomic_flag](#use_atomic_flag)
-- [pg_reshard_chunk_size](#pg_reshard_chunk_size)
-- [pg_reshard_chunk_pause_ms](#pg_reshard_chunk_pause_ms)
 
 ## bind_address
 
@@ -288,19 +285,13 @@ Flusher - это микро-поток (корутина), которая коп
 - Тип: булево (да/нет)
 - Значение по умолчанию: true
 
-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.
+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.
 
 ## inmemory_journal
 
@@ -383,8 +374,6 @@ 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)
@@ -400,18 +389,6 @@ 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)
@@ -740,9 +717,6 @@ pg_minsize OSD во время переключений, что может по
 Значение по умолчанию авто-определяется во время инициализации OSD из
 `/sys/block/xx/queue/atomic_write_max_bytes` либо принимается равным 4096,
 так как все известные диски поддерживают атомарную запись 4 КБ блоков.
-Автоопределение применяется только для NVMe-дисков, так как SAS диски требуют
-использования отдельной команды WRITE ATOMIC, а для неё нужен флаг RWF_ATOMIC
-(см. ниже [#use_atomic_flag]), а он в текущих версиях Linux работает некорректно.
 
 Вы также можете проверить, поддерживают ли ваши NVMe-диски атомарную запись,
 с помощью команды `nvme id-ctrl /dev/nvme0n1 | grep awupf`. Если значение awupf
@@ -761,35 +735,11 @@ pg_minsize OSD во время переключений, что может по
 - Тип: булево (да/нет)
 
 Данная опция контролирует использование Vitastor OSD флага RWF_ATOMIC при атомарной записи
-блоков. Этот флаг поддерживается, начиная с версии ядра Linux 6.11 и добавляет немного корректности
-атомарным записям - ядро гарантирует отсутствие фрагментации запросов записи с этим флагом и
-проверяет их на соответствие реальным возможностям устройства.
+блоков. Этот флаг поддерживается только в ядрах Linux начиная с 6.11. Атомарная запись
+является безопасной только при использовании этого флага, так как он сообщает ядру о том,
+что запрос записи нельзя фрагментировать и о том, что запрос нужно проверить на соответствие
+реальным возможностям атомарной записи устройства.
 
-Однако, данная опция по умолчанию отключена, так как флаг в текущих версиях Linux работает
-абсолютно НЕКОРРЕКТНО - при нём Linux требует, чтобы запросы записи имели длину, равную
-степени двойки и были выровнены на эту длину. То есть, например, 12 КБ запросы записи, а также
-8 КБ запросы записи по не-кратному 8 КБ смещению запрещаются ядром, хотя спецификация NVMe их
-разрешает.
-
-Для NVMe-дисков с `scheduler=none` запросы записи и так не фрагментируются, так что это не так
-уж и важно, однако вы можете пересобрать своё ядро с [этим патчем](../../patches/linux-fix-atomic-write-checks.diff)
-и включить данную опцию. Это сделает вашу атомарную запись капельку безопаснее.
-
-## pg_reshard_chunk_size
-
-- Тип: целое число
-- Значение по умолчанию: 100000
-
-Изменение числа PG в пуле заметно загружает процессор, так как OSD хранят полную базу данных
-объектов в памяти и им приходится перемещать все записи объектов между старыми и новыми PG.
-Поэтому изменение применяется порциями, с паузами между порциями, чтобы не блокировать обработку
-событий OSD и операции остальных клиентов. Данная опция задаёт максимальное число объектов
-в порции. Перемещение 100 тысяч объектов (значение по умолчанию) обычно занимает порядка
-50-100 миллисекунд. Значение опции 0 отключает лимит размера порции.
-
-## pg_reshard_chunk_pause_ms
-
-- Тип: миллисекунды
-- Значение по умолчанию: 100
-
-Данная опция задаёт интервал между обработкой двух порций изменения числа PG пулов.
+Опция включается по умолчанию, когда atomic_write_size устанавливается в значение больше 4 КБ.
+Вы можете явно отключить её, если уверены, что ваши диски поддерживают атомарную запись и
+хотите обойти проверки уровня ядра.

docs/config/security.en.md

@@ -1,150 +0,0 @@
-[Documentation](../../README.md#documentation) → [Configuration](../config.en.md) → Security Parameters
-
------
-
-[Читать на русском](security.ru.md)
-
-# Security Parameters
-
-These parameters affect your Vitastor installation security and apply to OSDs, monitors and clients.
-
-Most of them can be set in /etc/vitastor/vitastor.conf and in etcd, but don't support online modification.
-
-- [etcd_client_cert](#etcd_client_cert)
-- [etcd_client_key](#etcd_client_key)
-- [etcd_ca](#etcd_ca)
-- [osd_etcd_client_cert](#osd_etcd_client_cert)
-- [osd_etcd_client_key](#osd_etcd_client_key)
-- [mon_etcd_client_cert](#mon_etcd_client_cert)
-- [mon_etcd_client_key](#mon_etcd_client_key)
-- [vault_url](#vault_url)
-- [vault_secret_api_path](#vault_secret_api_path)
-- [vault_client_cert](#vault_client_cert)
-- [vault_client_key](#vault_client_key)
-- [vault_ca](#vault_ca)
-- [vault_timeout_ms](#vault_timeout_ms)
-- [vault_error_timeout_sec](#vault_error_timeout_sec)
-- [vault_refresh_leeway_sec](#vault_refresh_leeway_sec)
-- [max_aes_xts_pool_size](#max_aes_xts_pool_size)
-
-## etcd_client_cert
-
-- Type: string
-
-Client TLS certificate to use for Vitastor client (not OSD and not monitor)
-etcd https connections. May be path to a file or just a PEM string with certificate.
-In the latter case, string must begin with "-----BEGIN CERTIFICATE-----".
-
-## etcd_client_key
-
-- Type: string
-
-Private key for etcd_client_cert (also a file or a PEM string).
-
-## etcd_ca
-
-- Type: string
-
-Trusted TLS CA to verify etcd server certificate. May be path to a file,
-directory or just a PEM string with certificate.
-
-## osd_etcd_client_cert
-
-- Type: string
-
-Same as [etcd_client_cert](#etcd_client_cert), but only for OSDs.
-OSDs, clients and monitors should have different permissions, so they should
-use different certificates.
-
-## osd_etcd_client_key
-
-- Type: string
-
-Same as [etcd_client_key](#etcd_client_key), but only for OSDs.
-
-## mon_etcd_client_cert
-
-- Type: string
-
-Same as [etcd_client_cert](#etcd_client_cert), but only for Vitastor monitors.
-
-## mon_etcd_client_key
-
-- Type: string
-
-Same as [etcd_client_key](#etcd_client_key), but only for Vitastor monitors.
-
-## vault_url
-
-- Type: string
-
-Vault base URL.
-
-Vitastor clients support AES-256-XTS image data encryption with different per-image keys.
-Encryption is performed by the client, OSDs don't have access to decrypted data.
-
-Encryption keys may be stored in etcd or, for the increased security level, in an external
-[HashiCorp Vault](https://developer.hashicorp.com/vault/) or [OpenBao](https://openbao.org/)
-instance.
-
-Vitastor clients use [v1 k/v secrets engine](https://openbao.org/api-docs/secret/kv/kv-v1/)
-and [TLS authentication engine](https://openbao.org/api-docs/auth/cert/) in Vault.
-
-In that case, only key IDs are stored in etcd.
-
-## vault_secret_api_path
-
-- Type: string
-- Default: /v1/secret/
-
-Vault v1 secret API mount path to use.
-
-## vault_client_cert
-
-- Type: string
-
-Client TLS certificate to use for Vault connections. Just like [etcd_client_cert](#etcd_client_cert),
-may be path to a file or just a certificate in PEM string.
-
-## vault_client_key
-
-- Type: string
-
-Private key for vault_client_cert (also a file or a PEM string).
-
-## vault_ca
-
-- Type: string
-
-Trusted TLS CA to verify Vault server certificate. May be path to a file,
-directory or just a PEM string with certificate.
-
-## vault_timeout_ms
-
-- Type: integer
-- Default: 5000
-
-Timeout for Vault requests in milliseconds.
-
-## vault_error_timeout_sec
-
-- Type: integer
-- Default: 60
-
-Time (in seconds) to wait before retrying after receiving an error from Vault.
-
-## vault_refresh_leeway_sec
-
-- Type: integer
-- Default: 60
-
-Extra time (in seconds) before real Vault token lease_timeout to refresh it, just
-in case of system clock drift.
-
-## max_aes_xts_pool_size
-
-- Type: integer
-- Default: 256
-
-Maximum number of OpenSSL encryption contexts cached in OSD memory. Probably
-doesn't require modification.

docs/config/security.ru.md

@@ -1,154 +0,0 @@
-[Документация](../../README-ru.md#документация) → [Конфигурация](../config.ru.md) → Параметры безопасности
-
------
-
-[Read in English](security.en.md)
-
-# Параметры безопасности
-
-Данные параметры затрагивают безопасность инсталляций Vitastor и используются
-OSD, мониторами и клиентами.
-
-Большая их часть может задаваться в /etc/vitastor/vitastor.conf и в etcd, но не
-поддерживает онлайн-изменение.
-
-- [etcd_client_cert](#etcd_client_cert)
-- [etcd_client_key](#etcd_client_key)
-- [etcd_ca](#etcd_ca)
-- [osd_etcd_client_cert](#osd_etcd_client_cert)
-- [osd_etcd_client_key](#osd_etcd_client_key)
-- [mon_etcd_client_cert](#mon_etcd_client_cert)
-- [mon_etcd_client_key](#mon_etcd_client_key)
-- [vault_url](#vault_url)
-- [vault_secret_api_path](#vault_secret_api_path)
-- [vault_client_cert](#vault_client_cert)
-- [vault_client_key](#vault_client_key)
-- [vault_ca](#vault_ca)
-- [vault_timeout_ms](#vault_timeout_ms)
-- [vault_error_timeout_sec](#vault_error_timeout_sec)
-- [vault_refresh_leeway_sec](#vault_refresh_leeway_sec)
-- [max_aes_xts_pool_size](#max_aes_xts_pool_size)
-
-## etcd_client_cert
-
-- Тип: строка
-
-Клиентский TLS сертификат для https-подключений к etcd для клиентов Vitastor
-(не OSD и не мониторов). Может быть путём к файлу или просто строкой с
-сертификатом в формате PEM. В последнем случае строка должна начинаться с
-"-----BEGIN CERTIFICATE-----".
-
-## etcd_client_key
-
-- Тип: строка
-
-Закрытый ключ для сертификата etcd_client_cert (также путь к файлу или PEM строка).
-
-## etcd_ca
-
-- Тип: строка
-
-Доверенный корневой TLS-сертификат для проверки сертификата сервера etcd.
-Может быть путём к файлу, директории или просто строкой с сертификатом в
-формате PEM.
-
-## osd_etcd_client_cert
-
-- Тип: строка
-
-Аналогично [etcd_client_cert](#etcd_client_cert), но только для OSD.
-OSD, клиенты и мониторы должны иметь разные привилегии, поэтому они должны
-использовать разные сертификаты.
-
-## osd_etcd_client_key
-
-- Тип: строка
-
-Аналогично [etcd_client_key](#etcd_client_key), но только для OSD.
-
-## mon_etcd_client_cert
-
-- Тип: строка
-
-Аналогично [etcd_client_cert](#etcd_client_cert), но только для мониторов Vitastor.
-
-## mon_etcd_client_key
-
-- Тип: строка
-
-Аналогично [etcd_client_key](#etcd_client_key), но только для мониторов Vitastor.
-
-## vault_url
-
-- Тип: строка
-
-Базовый адрес Vault.
-
-Клиенты Vitastor поддерживают AES-256-XTS шифрование данных образов с отдельными ключами на
-каждый образ. Данные шифруются клиентами, OSD не имеют доступа к незашифрованным данным.
-
-Ключи шифрования могут храниться в etcd или, для повышенного уровня безопасности, во внешнем
-[HashiCorp Vault](https://developer.hashicorp.com/vault/) или [OpenBao](https://openbao.org/).
-
-Клиенты Vitastor используют [движок секретов v1](https://openbao.org/api-docs/secret/kv/kv-v1/)
-и [TLS-аутентификацию](https://openbao.org/api-docs/auth/cert/) в Vault.
-
-В этом случае, только ID ключей хранятся в etcd.
-
-## vault_secret_api_path
-
-- Тип: строка
-- Значение по умолчанию: /v1/secret/
-
-Путь к API секретов v1 для использования клиентами.
-
-## vault_client_cert
-
-- Тип: строка
-
-Клиентский TLS сертификат для подключений к Vault. Как и [etcd_client_cert](#etcd_client_cert),
-может быть путём к файлу или просто PEM-строкой с сертификатом.
-
-## vault_client_key
-
-- Тип: строка
-
-Закрытый ключ для сертификата vault_client_cert (также путь к файлу или PEM строка).
-
-## vault_ca
-
-- Тип: строка
-
-Доверенный корневой TLS-сертификат для проверки сертификата сервера Vault.
-Может быть путём к файлу, директории или просто строкой с сертификатом в
-формате PEM.
-
-## vault_timeout_ms
-
-- Тип: целое число
-- Значение по умолчанию: 5000
-
-Максимально время выполнения Vault-запросов в миллисекундах.
-
-## vault_error_timeout_sec
-
-- Тип: целое число
-- Значение по умолчанию: 60
-
-Время (в секундах) для ожидания перед повторной попыткой при получении ошибки от Vault.
-
-## vault_refresh_leeway_sec
-
-- Тип: целое число
-- Значение по умолчанию: 60
-
-Зазор времени (в секундах), чтобы обновлять токены Vault чуть раньше их реального
-lease_timeout, на случай "ухода" системных часов.
-
-## max_aes_xts_pool_size
-
-- Тип: целое число
-- Значение по умолчанию: 256
-
-Максимальное количество кэшируемых в памяти OSD контекстов шифрования OpenSSL.
-Вряд ли требует изменения.

docs/config/src/included.en.md

@@ -44,8 +44,6 @@
 
 {{../../config/monitor.en.md|indent=2}}
 
-{{../../config/security.en.md|indent=2}}
-
 {{../../config/pool.en.md|indent=2}}
 
 {{../../config/inode.en.md|indent=2}}

docs/config/src/included.ru.md

@@ -44,8 +44,6 @@
 
 {{../../config/monitor.ru.md|indent=2}}
 
-{{../../config/security.ru.md|indent=2}}
-
 {{../../config/pool.ru.md|indent=2}}
 
 {{../../config/inode.ru.md|indent=2}}

docs/config/src/layout-osd.yml

@@ -233,21 +233,11 @@
   type: string
   default: none
   info: |
-    Data and metadata checksum type to use. May be "crc32c", "xxh3_32" or "none".
-    Select crc32c or xxh3_32 and set csum_block_size to enable data checksums.
-
-    Both crc32c and xxh3_32 are almost equally fast, xxh3_32 is safer. xxh3_32 is
-    the xxhash3 algorithm truncated from 64 to 32 bits (which is still a good hash).
-
-    Note that enabled data checksums either increase memory usage or reduce
-    performance. Check details in [csum_block_size](#csum_block_size) description.
+    Data checksum type to use. May be "crc32c" or "none". Set to "crc32c" to
+    enable data checksums.
   info_ru: |
-    Тип используемых OSD контрольных сумм данных и метаданных. Может быть "crc32c",
-    "xxh3_32" или "none". Выберите crc32c или xxh3_32 и установите csum_block_size,
-    чтобы включить контрольные суммы данных.
-
-    И crc32c, и xxh3_32 примерно одинаково быстры, xxh3_32 надёжней. xxh3_32 - это
-    алгоритм xxhash3, обрезанный с 64 до 32 бит (это всё равно хороший хеш).
+    Тип используемых OSD контрольных сумм данных. Может быть "crc32c" или "none".
+    Установите в "crc32c", чтобы включить расчёт и проверку контрольных сумм данных.
 
     Следует понимать, что контрольные суммы в зависимости от размера блока их
     расчёта либо увеличивают потребление памяти, либо снижают производительность.

docs/config/src/network.yml

@@ -84,6 +84,11 @@
     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.
 
@@ -92,23 +97,6 @@
     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
@@ -117,6 +105,12 @@
     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.
 
@@ -126,24 +120,6 @@
     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: |
@@ -242,6 +218,45 @@
     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

docs/config/src/osd.yml

@@ -253,33 +253,21 @@
   type: bool
   default: true
   info: |
-    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.
+    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.
   info_ru: |
-    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.
+    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.
 - name: inmemory_journal
   type: bool
   default: true
@@ -398,15 +386,11 @@
     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
@@ -418,8 +402,6 @@
     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
@@ -430,20 +412,6 @@
     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
@@ -845,10 +813,7 @@
 
     Default value is auto-detected during OSD initialization from
     `/sys/block/xx/queue/atomic_write_max_bytes` or assumed to be 4096 bytes
-    because all known disks support 4 KB atomic writes. Auto-detection is only used for
-    NVMe disks because SAS disks require the explicit WRITE ATOMIC command which requires
-    RWF_ATOMIC (see below [#use_atomic_flag]) but that flag works incorrectly in current
-    Linux versions.
+    because all known disks support 4 KB atomic writes.
 
     You can also check if your NVMe drives support atomic writes by running
     the command `nvme id-ctrl /dev/nvme0n1 | grep awupf`. If the reported value,
@@ -869,9 +834,6 @@
     Значение по умолчанию авто-определяется во время инициализации OSD из
     `/sys/block/xx/queue/atomic_write_max_bytes` либо принимается равным 4096,
     так как все известные диски поддерживают атомарную запись 4 КБ блоков.
-    Автоопределение применяется только для NVMe-дисков, так как SAS диски требуют
-    использования отдельной команды WRITE ATOMIC, а для неё нужен флаг RWF_ATOMIC
-    (см. ниже [#use_atomic_flag]), а он в текущих версиях Linux работает некорректно.
 
     Вы также можете проверить, поддерживают ли ваши NVMe-диски атомарную запись,
     с помощью команды `nvme id-ctrl /dev/nvme0n1 | grep awupf`. Если значение awupf
@@ -887,54 +849,22 @@
 - name: use_atomic_flag
   type: bool
   info: |
-    This option controls whether Vitastor OSDs use RWF_ATOMIC write flag with atomic writes.
-    This flag is supported since Linux 6.11 and adds some safety to atomic writes - the kernel
-    guarantees to not fragment write requests with it and also to check them against the actual
-    device atomic write capabilities.
+    This option controls whether the Vitastor OSD uses RWF_ATOMIC write flag with atomic
+    writes. This flag is only supported on Linux kernel since 6.11. Atomic writes are
+    generally only safe to use with this flag because it tells the kernel to never fragment
+    write requests and also to check the write against the actual atomic write capabilities
+    of the device.
 
-    However, the option is disabled by default because the flag is currently UNUSABLE - Linux
-    incorrectly requires writes with that flag to be of power-of-2 length and length-aligned.
-    I.e., for example, 12 KB writes and not-8-KB aligned 8 KB writes are forbidden by the kernel,
-    even though the NVMe specification allows them.
-
-    For NVMe disks with `scheduler=none` writes aren't fragmented anyway so it's not a big deal.
-    However, you can rebuild your kernel with [this patch](../../patches/linux-fix-atomic-write-checks.diff)
-    and turn this option on. It will make your atomic writes a bit safer.
+    This option is enabled by default when atomic_write_size is set to a value larger than 4 KB.
+    You can disable it if you're sure that your disks support atomic writes and you want to
+    bypass the Linux atomic write checks.
   info_ru: |
     Данная опция контролирует использование Vitastor OSD флага RWF_ATOMIC при атомарной записи
-    блоков. Этот флаг поддерживается, начиная с версии ядра Linux 6.11 и добавляет немного корректности
-    атомарным записям - ядро гарантирует отсутствие фрагментации запросов записи с этим флагом и
-    проверяет их на соответствие реальным возможностям устройства.
+    блоков. Этот флаг поддерживается только в ядрах Linux начиная с 6.11. Атомарная запись
+    является безопасной только при использовании этого флага, так как он сообщает ядру о том,
+    что запрос записи нельзя фрагментировать и о том, что запрос нужно проверить на соответствие
+    реальным возможностям атомарной записи устройства.
 
-    Однако, данная опция по умолчанию отключена, так как флаг в текущих версиях Linux работает
-    абсолютно НЕКОРРЕКТНО - при нём Linux требует, чтобы запросы записи имели длину, равную
-    степени двойки и были выровнены на эту длину. То есть, например, 12 КБ запросы записи, а также
-    8 КБ запросы записи по не-кратному 8 КБ смещению запрещаются ядром, хотя спецификация NVMe их
-    разрешает.
-
-    Для NVMe-дисков с `scheduler=none` запросы записи и так не фрагментируются, так что это не так
-    уж и важно, однако вы можете пересобрать своё ядро с [этим патчем](../../patches/linux-fix-atomic-write-checks.diff)
-    и включить данную опцию. Это сделает вашу атомарную запись капельку безопаснее.
-- name: pg_reshard_chunk_size
-  type: int
-  default: 100000
-  info: |
-    Pool PG count change is a CPU-intensive operation because OSDs store the full object database
-    in memory and have to move all entries between old and new PGs. Thus it's performed in chunks,
-    with pauses between chunks to prevent blocking OSD's event loop and other clients' operations.
-    This option sets the maximum number of object is a chunk. Moving 100k objects usually takes
-    50-100ms. Chunk size equal to 0 means unlimited.
-  info_ru: |
-    Изменение числа PG в пуле заметно загружает процессор, так как OSD хранят полную базу данных
-    объектов в памяти и им приходится перемещать все записи объектов между старыми и новыми PG.
-    Поэтому изменение применяется порциями, с паузами между порциями, чтобы не блокировать обработку
-    событий OSD и операции остальных клиентов. Данная опция задаёт максимальное число объектов
-    в порции. Перемещение 100 тысяч объектов (значение по умолчанию) обычно занимает порядка
-    50-100 миллисекунд. Значение опции 0 отключает лимит размера порции.
-- name: pg_reshard_chunk_pause_ms
-  type: ms
-  default: 100
-  info: |
-    This option sets the interval between handling two PG count change chunks.
-  info_ru: |
-    Данная опция задаёт интервал между обработкой двух порций изменения числа PG пулов.
+    Опция включается по умолчанию, когда atomic_write_size устанавливается в значение больше 4 КБ.
+    Вы можете явно отключить её, если уверены, что ваши диски поддерживают атомарную запись и
+    хотите обойти проверки уровня ядра.

docs/config/src/package.json

@@ -1,5 +0,0 @@
-{
-  "dependencies": {
-    "yaml": "^2.8.2"
-  }
-}

docs/config/src/security.en.md

@@ -1,5 +0,0 @@
-# Security Parameters
-
-These parameters affect your Vitastor installation security and apply to OSDs, monitors and clients.
-
-Most of them can be set in /etc/vitastor/vitastor.conf and in etcd, but don't support online modification.

docs/config/src/security.ru.md

@@ -1,7 +0,0 @@
-# Параметры безопасности
-
-Данные параметры затрагивают безопасность инсталляций Vitastor и используются
-OSD, мониторами и клиентами.
-
-Большая их часть может задаваться в /etc/vitastor/vitastor.conf и в etcd, но не
-поддерживает онлайн-изменение.

docs/config/src/security.yml

@@ -1,131 +0,0 @@
-- name: etcd_client_cert
-  type: string
-  info: |
-    Client TLS certificate to use for Vitastor client (not OSD and not monitor)
-    etcd https connections. May be path to a file or just a PEM string with certificate.
-    In the latter case, string must begin with "-----BEGIN CERTIFICATE-----".
-  info_ru: |
-    Клиентский TLS сертификат для https-подключений к etcd для клиентов Vitastor
-    (не OSD и не мониторов). Может быть путём к файлу или просто строкой с
-    сертификатом в формате PEM. В последнем случае строка должна начинаться с
-    "-----BEGIN CERTIFICATE-----".
-- name: etcd_client_key
-  type: string
-  info: Private key for etcd_client_cert (also a file or a PEM string).
-  info_ru: Закрытый ключ для сертификата etcd_client_cert (также путь к файлу или PEM строка).
-- name: etcd_ca
-  type: string
-  info: |
-    Trusted TLS CA to verify etcd server certificate. May be path to a file,
-    directory or just a PEM string with certificate.
-  info_ru: |
-    Доверенный корневой TLS-сертификат для проверки сертификата сервера etcd.
-    Может быть путём к файлу, директории или просто строкой с сертификатом в
-    формате PEM.
-- name: osd_etcd_client_cert
-  type: string
-  info: |
-    Same as [etcd_client_cert](#etcd_client_cert), but only for OSDs.
-    OSDs, clients and monitors should have different permissions, so they should
-    use different certificates.
-  info_ru: |
-    Аналогично [etcd_client_cert](#etcd_client_cert), но только для OSD.
-    OSD, клиенты и мониторы должны иметь разные привилегии, поэтому они должны
-    использовать разные сертификаты.
-- name: osd_etcd_client_key
-  type: string
-  info: Same as [etcd_client_key](#etcd_client_key), but only for OSDs.
-  info_ru: Аналогично [etcd_client_key](#etcd_client_key), но только для OSD.
-- name: mon_etcd_client_cert
-  type: string
-  info: Same as [etcd_client_cert](#etcd_client_cert), but only for Vitastor monitors.
-  info_ru: Аналогично [etcd_client_cert](#etcd_client_cert), но только для мониторов Vitastor.
-- name: mon_etcd_client_key
-  type: string
-  info: Same as [etcd_client_key](#etcd_client_key), but only for Vitastor monitors.
-  info_ru: Аналогично [etcd_client_key](#etcd_client_key), но только для мониторов Vitastor.
-- name: vault_url
-  type: string
-  info: |
-    Vault base URL.
-
-    Vitastor clients support AES-256-XTS image data encryption with different per-image keys.
-    Encryption is performed by the client, OSDs don't have access to decrypted data.
-
-    Encryption keys may be stored in etcd or, for the increased security level, in an external
-    [HashiCorp Vault](https://developer.hashicorp.com/vault/) or [OpenBao](https://openbao.org/)
-    instance.
-
-    Vitastor clients use [v1 k/v secrets engine](https://openbao.org/api-docs/secret/kv/kv-v1/)
-    and [TLS authentication engine](https://openbao.org/api-docs/auth/cert/) in Vault.
-
-    In that case, only key IDs are stored in etcd.
-  info_ru: |
-    Базовый адрес Vault.
-
-    Клиенты Vitastor поддерживают AES-256-XTS шифрование данных образов с отдельными ключами на
-    каждый образ. Данные шифруются клиентами, OSD не имеют доступа к незашифрованным данным.
-
-    Ключи шифрования могут храниться в etcd или, для повышенного уровня безопасности, во внешнем
-    [HashiCorp Vault](https://developer.hashicorp.com/vault/) или [OpenBao](https://openbao.org/).
-
-    Клиенты Vitastor используют [движок секретов v1](https://openbao.org/api-docs/secret/kv/kv-v1/)
-    и [TLS-аутентификацию](https://openbao.org/api-docs/auth/cert/) в Vault.
-
-    В этом случае, только ID ключей хранятся в etcd.
-- name: vault_secret_api_path
-  type: string
-  default: /v1/secret/
-  info: Vault v1 secret API mount path to use.
-  info_ru: Путь к API секретов v1 для использования клиентами.
-- name: vault_client_cert
-  type: string
-  info: |
-    Client TLS certificate to use for Vault connections. Just like [etcd_client_cert](#etcd_client_cert),
-    may be path to a file or just a certificate in PEM string.
-  info_ru: |
-    Клиентский TLS сертификат для подключений к Vault. Как и [etcd_client_cert](#etcd_client_cert),
-    может быть путём к файлу или просто PEM-строкой с сертификатом.
-- name: vault_client_key
-  type: string
-  info: Private key for vault_client_cert (also a file or a PEM string).
-  info_ru: Закрытый ключ для сертификата vault_client_cert (также путь к файлу или PEM строка).
-- name: vault_ca
-  type: string
-  info: |
-    Trusted TLS CA to verify Vault server certificate. May be path to a file,
-    directory or just a PEM string with certificate.
-  info_ru: |
-    Доверенный корневой TLS-сертификат для проверки сертификата сервера Vault.
-    Может быть путём к файлу, директории или просто строкой с сертификатом в
-    формате PEM.
-- name: vault_timeout_ms
-  type: int
-  default: 5000
-  info: Timeout for Vault requests in milliseconds.
-  info_ru: Максимально время выполнения Vault-запросов в миллисекундах.
-- name: vault_error_timeout_sec
-  type: int
-  default: 60
-  info: |
-    Time (in seconds) to wait before retrying after receiving an error from Vault.
-  info_ru: |
-    Время (в секундах) для ожидания перед повторной попыткой при получении ошибки от Vault.
-- name: vault_refresh_leeway_sec
-  type: int
-  default: 60
-  info: |
-    Extra time (in seconds) before real Vault token lease_timeout to refresh it, just
-    in case of system clock drift.
-  info_ru: |
-    Зазор времени (в секундах), чтобы обновлять токены Vault чуть раньше их реального
-    lease_timeout, на случай "ухода" системных часов.
-- name: max_aes_xts_pool_size
-  type: int
-  default: 256
-  info: |
-    Maximum number of OpenSSL encryption contexts cached in OSD memory. Probably
-    doesn't require modification.
-  info_ru: |
-    Максимальное количество кэшируемых в памяти OSD контекстов шифрования OpenSSL.
-    Вряд ли требует изменения.

docs/installation/docker.en.md

@@ -26,37 +26,13 @@ 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:v3.0.5`
+   `docker pull vitalif/vitastor:v3.0.0`
 2. Install scripts to the host system: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.5 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.0 install.sh`
 3. Reload udev rules: \
    `udevadm control --reload-rules`
-4. Enable the vitastor-host service: \
-   `systemctl enable --now vitastor-host`
 
-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.
+And you can return to [Quick Start](../intro/quickstart.en.md).
 
 ## Upgrading Containers
 

docs/installation/docker.ru.md

@@ -25,39 +25,14 @@ Vitastor можно установить в Docker/Podman. При этом etcd,
 Инструкция по установке максимально простая.
 
 1. Скачайте Docker-образ желаемой версии: \
-   `docker pull vitalif/vitastor:v3.0.5`
+   `docker pull vitalif/vitastor:v3.0.0`
 2. Установите скрипты в хост-систему командой: \
-   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.5 install.sh`
+   `docker run --rm -it -v /etc:/host-etc -v /usr/bin:/host-bin vitalif/vitastor:v3.0.0 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),

docs/installation/packages.en.md

@@ -33,17 +33,15 @@
   - 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 9/10 clones: not required
+  - RHEL 9 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

docs/installation/packages.ru.md

@@ -33,17 +33,15 @@
   - 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 9/10: не нужно
+  - Клоны RHEL 9: 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

docs/installation/proxmox.en.md

@@ -6,7 +6,7 @@
 
 # Proxmox VE
 
-To enable Vitastor support in Proxmox Virtual Environment (6.4-9.x are supported):
+To enable Vitastor support in Proxmox Virtual Environment (6.4-8.x are supported):
 
 - Add the corresponding Vitastor Debian repository into sources.list on Proxmox hosts:
   trixie for 9.0+, bookworm for 8.1+, pve8.0 for 8.0, bullseye for 7.4, pve7.3 for 7.3, pve7.2 for 7.2, pve7.1 for 7.1, buster for 6.4

docs/installation/proxmox.ru.md

@@ -6,7 +6,7 @@
 
 # Proxmox VE
 
-Чтобы подключить Vitastor к Proxmox Virtual Environment (поддерживаются версии 6.4-9.x):
+Чтобы подключить Vitastor к Proxmox Virtual Environment (поддерживаются версии 6.4-8.x):
 
 - Добавьте соответствующий Debian-репозиторий Vitastor в sources.list на хостах Proxmox:
   trixie для 9.0+, bookworm для 8.1+, pve8.0 для 8.0, bullseye для 7.4, pve7.3 для 7.3, pve7.2 для 7.2, pve7.1 для 7.1, buster для 6.4

docs/installation/source.en.md

@@ -15,8 +15,8 @@
 - gcc and g++ 8 or newer, clang 10 or newer, or other compiler with C++11 plus
   designated initializers support from C++20
 - CMake
-- jerasure, c-ares headers and libraries
-- ISA-L, libibverbs, librdmacm, libnl3 headers and libraries (optional)
+- jerasure headers and libraries
+- ISA-L, libibverbs and librdmacm headers and libraries (optional)
 - tcmalloc (google-perftools-dev)
 
 ## Basic instructions

docs/installation/source.ru.md

@@ -15,8 +15,8 @@
 - gcc и g++ >= 8, либо clang >= 10, либо другой компилятор с поддержкой C++11 плюс
   назначенных инициализаторов (designated initializers) из C++20
 - CMake
-- Заголовки и библиотеки jerasure, c-ares
-- Опционально - заголовки и библиотеки ISA-L, libibverbs, librdmacm, libnl3
+- Заголовки и библиотеки jerasure
+- Опционально - заголовки и библиотеки ISA-L, libibverbs, librdmacm
 - tcmalloc (google-perftools-dev)
 
 ## Базовая инструкция

docs/intro/features.en.md

@@ -41,8 +41,6 @@
 - [Built-in Prometheus metric exporter](../config/monitor.en.md#enable_prometheus)
 - [NFS RDMA support](../usage/nfs.en.md#rdma) (probably also usable for GPUDirect)
 - [S3](../installation/s3.en.md)
-- [TLS support for etcd connections](../config/security.en.md)
-- [AES-256-XTS image encryption](../usage/cli.en.md#create) and [Vault support](../config/security.en.md#vault_url) for key storage
 
 ## Plugins and tools
 

docs/intro/features.ru.md

@@ -43,8 +43,6 @@
 - [Встроенный Prometheus-экспортер метрик](../config/monitor.ru.md#enable_prometheus)
 - [Поддержка NFS RDMA](../usage/nfs.ru.md#rdma) (вероятно, также подходящая для GPUDirect)
 - [S3](../installation/s3.ru.md)
-- [Поддержка TLS-соединений с etcd](../config/security.ru.md)
-- [AES-256-XTS шифрование данных](../usage/cli.ru.md#create) и [поддержка Vault](../config/security.ru.md#vault_url) для хранения ключей
 
 ## Драйверы и инструменты
 

docs/usage/cli.en.md

@@ -125,31 +125,18 @@ bench-kaveri         kaveri    10 G  10 G    0 B/s  0     0      0 us  0 B/s  0
 
 ## create
 
-`vitastor-cli create -s|--size SIZE [OPTIONS] <name>`
+`vitastor-cli create -s|--size <size> [-p|--pool <id|name>] [--parent <parent_name>[@<snapshot>]] <name>`
 
-Create an image. Options:
-
-* `-s|--size SIZE` - New image size in bytes or with a K/M/G/T unit suffix.
-* `-p|--pool POOL` - Specify pool for the new image (may be omitted if there is only 1 pool).
-* `--parent PARENT` - Create a copy-on-write image clone based on PARENT (or PARENT@SNAPSHOT).
-  If parent is not a snapshot, it must be a read-only image.
-* `--enc-key random` - Generate a new random AES-256-XTS encryption key for the new image.
-* `--enc-key HEX` - Set a specified AES-256-XTS key (64 bytes in hex) for the new image.
-* `--enc-key vault:ID` - Use an encryption key from an external Vault secret with specified ID.
+Create an image. You may use K/M/G/T suffixes for `<size>`. If `--parent` is specified,
+a copy-on-write image clone is created. Parent must be a snapshot (readonly image).
+Pool must be specified if there is more than one pool.
 
 ```
-vitastor-cli create --snapshot <snapshot> [OPTIONS] <image>
-vitastor-cli snap-create [OPTIONS] <image>@<snapshot>
+vitastor-cli create --snapshot <snapshot> [-p|--pool <id|name>] <image>
+vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 ```
 
-Create a snapshot of image `<image>`. May be used live if only a single writer is active.
-
-Options:
-
-* `-p|--pool POOL` - Move image to pool POOL, leaving the snapshot in the old pool.
-* `--enc-key random` - Change image encryption key to a new random AES-256-XTS key.
-* `--enc-key KEY` - Change image encryption key to a specified key, Vault key or to an empty key.
-  By default, the image retains its old encryption key when taking a snapshot.
+Create a snapshot of image `<name>` (either form can be used). May be used live if only a single writer is active.
 
 See also about [how to export snapshots](qemu.en.md#exporting-snapshots).
 
@@ -164,7 +151,6 @@ You should resize file system in the image, if present, before shrinking it.
 * `--deleted 1|0` - Set/clear 'deleted image' flag (set automatically during unfinished deletes).
 * `-f|--force` - Proceed with shrinking or setting readwrite flag even if the image has children.
 * `--down-ok` - Proceed with shrinking even if some data will be left on unavailable OSDs.
-* `--enc-key HEX` - Change image encryption key (allowed only with `--force`).
 
 ## dd
 

docs/usage/cli.ru.md

@@ -127,32 +127,19 @@ bench-kaveri         kaveri    10 G  10 G    0 B/s  0     0      0 us  0 B/s  0
 
 ## create
 
-`vitastor-cli create -s|--size SIZE [ОПЦИИ] <name>`
+`vitastor-cli create -s|--size <size> [-p|--pool <id|name>] [--parent <parent_name>[@<snapshot>]] <name>`
 
-Создать образ. Опции:
-
-* `-s|--size SIZE` - Размер нового образа в байтах или с суффиксом K/M/G/T (кило/мега/гига/терабайт).
-* `-p|--pool POOL` - Создать образ в заданном пуле (можно не указывать, если пул всего один).
-* `--parent PARENT` - Создать легковесный клон на основе образа `PARENT` или снимка `PARENT@SNAP`.
-  Если `PARENT` - не снимок, он должен быть помечен как образ только для чтения.
-* `--enc-key random` - Сгенерировать случайный ключ шифрования AES-256-XTS для нового образа.
-* `--enc-key HEX` - Установить заданный ключ AES-256-XTS (64 байта в hex) для нового образа.
-* `--enc-key vault:ID` - Использовать ключ из внешнего секрета с заданным ID из Vault.
+Создать образ. Для размера `<size>` можно использовать суффиксы K/M/G/T (килобайт-мегабайт-гигабайт-терабайт).
+Если указана опция `--parent`, создаётся клон образа. Родитель `<parent_name>[@<snapshot>]` должен быть
+снимком (или просто немодифицируемым образом). Пул обязательно указывать, если в кластере больше одного пула.
 
 ```
-vitastor-cli create --snapshot <snapshot> [ОПЦИИ] <image>
-vitastor-cli snap-create [ОПЦИИ] <image>@<snapshot>
+vitastor-cli create --snapshot <snapshot> [-p|--pool <id|name>] <image>
+vitastor-cli snap-create [-p|--pool <id|name>] <image>@<snapshot>
 ```
 
-Создать снимок образа `<image>` (можно использовать любую форму команды).
-Снимок можно создавать без остановки клиентов, если пишущих клиентов не больше одного.
-
-Опции:
-
-* `-p|--pool POOL` - Переместить образ в пул POOL, оставив снимок в старом пуле.
-* `--enc-key random` - Изменить ключ шифрования образа на новый случайный ключ AES-256-XTS.
-* `--enc-key KEY` - Изменить ключ шифрования образа на заданный ключ, ключ из Vault или пустой ключ.
-  По умолчанию шифрованные образы сохраняют старый ключ при снятии снимка.
+Создать снимок образа `<name>` (можно использовать любую форму команды). Снимок можно создавать без остановки
+клиентов, если пишущий клиент максимум 1.
 
 Смотрите также информацию о том, [как экспортировать снимки](qemu.ru.md#экспорт-снимков).
 
@@ -169,7 +156,6 @@ vitastor-cli snap-create [ОПЦИИ] <image>@<snapshot>
 * `--deleted 1|0` - Установить/снять флаг "образ удалён" (устанавливается при незавершённом удалении).
 * `-f|--force` - Разрешить уменьшение или перевод в чтение-запись образа, у которого есть клоны.
 * `--down-ok` - Разрешить уменьшение, даже если часть данных останется неудалённой на недоступных OSD.
-* `--enc-key HEX` - Изменить ключ шифрования образа (разрешено только с `--force`).
 
 ## dd
 

docs/usage/disk.en.md

@@ -95,8 +95,6 @@ Options (single-device mode):
 Options (both modes):
 
 ```
---tags tag1,tag2           Set new OSD tag(s)
---weight <number>          Set new OSD weight (between 0 to 1)
 --journal_size 1G/32M      Set journal size (area or partition size)
 --block_size 1M/128k       Set blockstore object size
 --bitmap_granularity 4k    Set bitmap granularity

docs/usage/disk.ru.md

@@ -96,8 +96,6 @@ vitastor-disk - инструмент командной строки для уп
 Опции для обоих режимов:
 
 ```
---tags tag1,tag2           Задать теги для новых OSD
---weight <number>          Задать вес для новых OSD (от 0 до 1)
 --journal_size 1G/32M      Задать размер журнала (области или раздела журнала)
 --block_size 1M/128k       Задать размер объекта хранилища
 --bitmap_granularity 4k    Задать гранулярность битовых карт

mon/antietcd_adapter.js

@@ -18,7 +18,7 @@ class AntiEtcdAdapter
                 cluster = cluster ? (''+(cluster||'')).split(/,+/) : [];
             cluster = Object.keys(cluster.reduce((a, url) =>
             {
-                a[url.toLowerCase().replace(/^(https?:\/\/)?(.*?)(\/.*)?$/, (m, m1, m2) => (m1||'http://')+m2)] = true;
+                a[url.toLowerCase().replace(/^(https?:\/\/)/, '').replace(/\/.*$/, '')] = true;
                 return a;
             }, {}));
             const cfg_port = config.antietcd_port;
@@ -26,8 +26,7 @@ class AntiEtcdAdapter
             is_local['0.0.0.0'] = true;
             is_local['::'] = true;
             is_local[''] = true;
-            // split :, 3 -> <schema>:<//ip>:<port>
-            const selected = cluster.map(s => s.split(':', 3)).filter(ip => is_local[ip[1].substr(2)] && (!cfg_port || ip[2] == cfg_port));
+            const selected = cluster.map(s => s.split(':', 2)).filter(ip => is_local[ip[0]] && (!cfg_port || ip[1] == cfg_port));
             if (selected.length > 1)
             {
                 console.error('More than 1 etcd_address matches local IPs, please specify port');
@@ -36,15 +35,12 @@ class AntiEtcdAdapter
             else if (selected.length == 1)
             {
                 const antietcd_config = {
-                    ip: selected[0][1].substr(2),
-                    port: selected[0][2],
-                    cert: config.antietcd_cert,
-                    key: config.antietcd_key,
-                    ca: config.etcd_ca,
-                    data: config.antietcd_data_file || ((config.antietcd_data_dir || '/var/lib/vitastor') + '/mon_'+selected[0][2]+'.json.gz'),
+                    ip: selected[0][0],
+                    port: selected[0][1],
+                    data: config.antietcd_data_file || ((config.antietcd_data_dir || '/var/lib/vitastor') + '/mon_'+selected[0][1]+'.json.gz'),
                     persist_filter: vitastor_persist_filter({ vitastor_prefix: config.etcd_prefix || '/vitastor' }),
-                    node_id: selected[0][1].substr(2)+':'+selected[0][2], // node_id = ip:port
-                    cluster: (cluster.length == 1 ? null : cluster.reduce((a, c) => { a[c.replace(/^(https?:\/\/)/, '')] = c; return a; }, {})),
+                    node_id: selected[0][0]+':'+selected[0][1], // node_id = ip:port
+                    cluster: (cluster.length == 1 ? null : cluster.reduce((a, c) => { a[c] = "http://"+c; return a; }, {})),
                     cluster_key: (config.etcd_prefix || '/vitastor'),
                     stale_read: 1,
                     log_level: 1,

mon/etcd_adapter.js

@@ -1,9 +1,7 @@
 // Copyright (c) Vitaliy Filippov, 2019+
 // License: VNPL-1.1 (see README.md for details)
 
-const fs = require('fs');
 const http = require('http');
-const https = require('https');
 const WebSocket = require('ws');
 const { b64, local_ips } = require('./utils.js');
 
@@ -17,30 +15,11 @@ class EtcdAdapter
         this.ws = null;
         this.ws_alive = false;
         this.ws_keepalive_timer = null;
-        this.opts = {};
     }
 
     parse_config(config)
     {
         this.parse_etcd_addresses(config.etcd_address||config.etcd_url);
-        if (config.mon_etcd_client_cert || config.etcd_client_cert)
-        {
-            this.opts.cert = config.mon_etcd_client_cert || config.etcd_client_cert;
-            if (this.opts.cert.substr(0, 5) != '-----')
-                this.opts.cert = fs.readFileSync(this.opts.cert, { encoding: 'utf-8' });
-        }
-        if (config.mon_etcd_client_key || config.etcd_client_key)
-        {
-            this.opts.key = config.mon_etcd_client_key || config.etcd_client_key;
-            if (this.opts.key.substr(0, 5) != '-----')
-                this.opts.key = fs.readFileSync(this.opts.key, { encoding: 'utf-8' });
-        }
-        if (config.etcd_ca)
-        {
-            this.opts.ca = config.etcd_ca;
-            if (this.opts.ca.substr(0, 5) != '-----')
-                this.opts.ca = fs.readFileSync(this.opts.ca, { encoding: 'utf-8' });
-        }
     }
 
     parse_etcd_addresses(addrs)
@@ -60,7 +39,7 @@ class EtcdAdapter
         for (let url of addrs)
         {
             let scheme = 'http';
-            url = url.trim().replace(/^(https?):\/\//i, (m, m1) => { scheme = m1.toLowerCase(); return ''; });
+            url = url.trim().replace(/^(https?):\/\//, (m, m1) => { scheme = m1; return ''; });
             const slash = url.indexOf('/');
             const colon = url.indexOf(':');
             const is_local = is_local_ip[colon >= 0 ? url.substr(0, colon) : (slash >= 0 ? url.substr(0, slash) : url)];
@@ -151,7 +130,7 @@ class EtcdAdapter
                     }
                     ok(false);
                 }, this.mon.config.etcd_mon_timeout);
-                this.ws = new WebSocket(base+'/watch', this.opts);
+                this.ws = new WebSocket(base+'/watch');
                 this.ws_used_url = cur_addr;
                 const fail = () =>
                 {
@@ -293,7 +272,7 @@ class EtcdAdapter
             {
                 throw new Error(MON_STOPPED);
             }
-            const res = await POST(base+path, body, timeout, this.opts);
+            const res = await POST(base+path, body, timeout);
             if (this.mon.stopped)
             {
                 throw new Error(MON_STOPPED);
@@ -319,7 +298,7 @@ class EtcdAdapter
     }
 }
 
-function POST(url, body, timeout, opts)
+function POST(url, body, timeout)
 {
     return new Promise(ok =>
     {
@@ -331,10 +310,10 @@ function POST(url, body, timeout, opts)
             req = null;
             ok({ error: 'timeout' });
         }, timeout) : null;
-        let req = (url.substr(0, 5) == 'https' ? https : http).request(url, { method: 'POST', headers: {
+        let req = http.request(url, { method: 'POST', headers: {
             'Content-Type': 'application/json',
             'Content-Length': body_text.length,
-        }, ...(opts||{}) }, (res) =>
+        } }, (res) =>
         {
             if (!req)
             {

mon/etcd_schema.js

@@ -45,14 +45,7 @@ const etcd_tree = {
             config_path: "/etc/vitastor/vitastor.conf",
             etcd_prefix: "/vitastor",
             // etcd connection - configurable online
-            etcd_address: "http://10.0.115.10:2379/v3",
-            etcd_client_cert: "",
-            etcd_client_key: "",
-            osd_etcd_client_cert: "",
-            osd_etcd_client_key: "",
-            mon_etcd_client_cert: "",
-            mon_etcd_client_key: "",
-            etcd_ca: "",
+            etcd_address: "10.0.115.10:2379/v3",
             // mon
             etcd_mon_ttl: 5, // min: 1
             etcd_mon_timeout: 1000, // ms. min: 0
@@ -224,8 +217,6 @@ const etcd_tree = {
                     parent_id?: <inode_t>,
                     readonly?: boolean,
                     deleted?: boolean,
-                    enc_key?: string,
-                    meta?: any,
                 }
             }
         }, */
@@ -392,7 +383,6 @@ const etcd_tree = {
             /* <name>: {
                 id: uint64_t,
                 pool_id: uint64_t,
-                // ...plus a copy of everything from config/inode/x/y
             }, */
         },
         maxid: {

mon/http_server.js

@@ -16,7 +16,7 @@ async function create_http_server(cfg, handler)
         };
         if (cfg.mon_https_ca)
         {
-            tls.ca = await fsp.readFile(cfg.mon_https_ca);
+            tls.mon_https_ca = await fsp.readFile(cfg.mon_https_ca);
         }
         if (cfg.mon_https_client_auth)
         {

mon/lp_optimizer/lp_optimizer.js

@@ -10,19 +10,16 @@ const NO_OSD = 'Z';
 async function lp_solve(text)
 {
     const cp = child_process.spawn('lp_solve');
-    let stdout = '', stderr = '', finish_cb, finished = 0;
+    let stdout = '', stderr = '', finish_cb;
     cp.stdout.on('data', buf => stdout += buf.toString());
     cp.stderr.on('data', buf => stderr += buf.toString());
-    cp.stdout.on('end', () => finish_cb());
-    cp.stderr.on('end', () => finish_cb());
+    cp.on('exit', () => finish_cb && finish_cb());
     cp.stdin.write(text);
     cp.stdin.end();
-    await new Promise(ok => (finish_cb = () =>
+    if (cp.exitCode == null)
     {
-        finished++;
-        if (finished == 2)
-            ok();
-    }));
+        await new Promise(ok => finish_cb = ok);
+    }
     if (!stdout.trim())
     {
         return null;

mon/osd_tree.js

@@ -87,7 +87,7 @@ function make_hier_tree(global_config, tree)
     tree[''] = { children: [] };
     for (const node_id in tree)
     {
-        if (node_id === '')
+        if (node_id === '' || !(tree[node_id].children||[]).length && (tree[node_id].size||0) <= 0)
         {
             continue;
         }

mon/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor-mon",
-  "version": "3.0.5",
+  "version": "3.0.0",
   "description": "Vitastor SDS monitor service",
   "main": "mon-main.js",
   "scripts": {
@@ -9,7 +9,7 @@
   "author": "Vitaliy Filippov",
   "license": "UNLICENSED",
   "dependencies": {
-    "antietcd": "^1.2.4",
+    "antietcd": "^1.1.3",
     "sprintf-js": "^1.1.2",
     "ws": "^7.2.5"
   },

mon/pg_gen.js

@@ -52,7 +52,6 @@ function recheck_primary(state, global_config, up_osds, osd_tree)
             continue;
         }
         const aff_osds = get_affinity_osds(pool_cfg, up_osds, osd_tree);
-        let paused = false;
         for (let pg_num = 1; pg_num <= pool_cfg.pg_count; pg_num++)
         {
             if (!state.pg.config.items[pool_id])
@@ -75,19 +74,6 @@ function recheck_primary(state, global_config, up_osds, osd_tree)
                     );
                     new_pg_config.items[pool_id][pg_num].primary = new_primary;
                 }
-                paused = paused || !!pg_cfg.pause;
-            }
-        }
-        if (paused)
-        {
-            if (!new_pg_config)
-            {
-                new_pg_config = JSON.parse(JSON.stringify(state.pg.config));
-            }
-            console.log(`Resuming paused pool ${pool_id}`);
-            for (const pg in new_pg_config.items[pool_id])
-            {
-                delete new_pg_config.items[pool_id][pg].pause;
             }
         }
     }
@@ -192,10 +178,10 @@ async function generate_pool_pgs(state, global_config, pool_id, osd_tree, levels
     const rules = use_rules ? get_pg_rules(pool_id, pool_cfg, global_config.placement_levels) : null;
     const folded = fold_failure_domains(Object.values(pool_tree), use_rules ? rules : [ [ [ pool_cfg.failure_domain ] ] ]);
     // FIXME: Remove/merge make_hier_tree() step somewhere, however it's needed to remove empty nodes
-    const folded_tree = make_hier_tree(global_config, folded.nodes.reduce((a, c) => { a[c.id] = c; return a; }, {}));
+    const folded_tree = make_hier_tree(global_config, folded.nodes);
     const old_pg_count = prev_pgs.length;
     const optimize_cfg = {
-        osd_weights: folded.nodes.reduce((a, c) => { if (/^\d+$/.exec(c.id) && c.size != null) { a[c.id] = c.size||0; } return a; }, {}),
+        osd_weights: folded.nodes.reduce((a, c) => { if (Number(c.id)) { a[c.id] = c.size; } return a; }, {}),
         combinator: use_rules
             // new algorithm:
             ? new RuleCombinator(folded_tree, rules, pool_cfg.max_osd_combinations)

node-binding/package.json

@@ -1,6 +1,6 @@
 {
   "name": "vitastor",
-  "version": "3.0.5",
+  "version": "3.0.0",
   "description": "Low-level native bindings to Vitastor client library",
   "main": "index.js",
   "keywords": [

patches/cinder-vitastor.py

@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils
 
-VITASTOR_VERSION = '3.0.5'
+VITASTOR_VERSION = '3.0.0'
 
 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.
@@ -291,14 +291,6 @@ 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."""
 
@@ -310,7 +302,7 @@ class VitastorDriver(driver.CloneableImageVD,
 
         LOG.debug("creating volume '%s'", vol_name)
 
-        self._cli('create volume', 'create', vol_name, '--size', size)
+        self._create_image(vol_name, { 'size': size })
 
         if volume.encryption_key_id:
             self._create_encrypted_volume(volume, volume.obj_context)
@@ -354,7 +346,7 @@ 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._cli('create snapshot', 'snap-create', vol_name+'@'+snap_name)
+        self._create_snapshot(vol_name, vol_name+'@'+snap_name)
 
     def snapshot_revert_use_temp_snapshot(self):
         """Disable the use of a temporary snapshot on revert."""
@@ -367,8 +359,21 @@ class VitastorDriver(driver.CloneableImageVD,
         snap_name = utils.convert_str(snapshot.name)
 
         # Delete the image and recreate it from the snapshot
-        self._cli('delete image', 'rm', vol_name)
-        self._cli('recreate image', 'create', '--parent', vol_name+'@'+snap_name, vol_name)
+        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)
 
     def delete_snapshot(self, snapshot):
         """Deletes a snapshot."""
@@ -376,7 +381,15 @@ class VitastorDriver(driver.CloneableImageVD,
         vol_name = utils.convert_str(snapshot.volume_name)
         snap_name = utils.convert_str(snapshot.name)
 
-        self._cli('remove snapshot', 'rm', vol_name+'@'+snap_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)
 
     def _child_count(self, parents):
         children = 0
@@ -414,7 +427,13 @@ 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
-            self._cli('create clone', 'create', '--parent', src_name, '--size', size, dest_name)
+            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'],
+            })
             return {}
 
         clone_snap = "%s@%s.clone_snap" % (src_name, dest_name)
@@ -427,12 +446,15 @@ class VitastorDriver(driver.CloneableImageVD,
             clone_snap = dest_name
             make_img = False
 
-        LOG.debug("creating snapshot '%s'", clone_snap)
-        self._cli('create base snapshot', 'snap-create', '--allow-existing', '1', clone_snap)
-
+        LOG.debug("creating layer '%s' under '%s'", clone_snap, src_name)
+        new_cfg = self._create_snapshot(src_name, clone_snap, True)
         if make_img:
             # Then create a clone from it
-            self._cli('create clone', 'create', '--parent', clone_snap, '--size', size, dest_name)
+            new_cfg = self._create_image(dest_name, {
+                'size': size,
+                'parent_id': new_cfg['parent_id'],
+                'parent_pool_id': new_cfg['parent_pool_id'],
+            })
 
         return {}
 
@@ -442,8 +464,7 @@ class VitastorDriver(driver.CloneableImageVD,
         vol_name = utils.convert_str(volume.name)
         snap_name = utils.convert_str(snapshot.name)
 
-        src_snap = 'volume-'+snapshot.volume_id+'@'+snap_name
-        snap = self._get_image(src_snap)
+        snap = self._get_image('volume-'+snapshot.volume_id+'@'+snap_name)
         if not snap:
             raise exception.SnapshotNotFound(snapshot_id = snap_name)
         snap_inode_id = int(resp['responses'][0]['kvs'][0]['value']['id'])
@@ -452,8 +473,12 @@ 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):
@@ -480,7 +505,49 @@ class VitastorDriver(driver.CloneableImageVD,
         """Deletes a logical volume."""
 
         vol_name = utils.convert_str(volume.name)
-        self._cli('delete volume', 'rm', '--matching', vol_name, vol_name+'@*', '--progress', '0')
+
+        # 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 })
 
     def retype(self, context, volume, new_type, diff, host):
         """Change extra type specifications for a volume."""
@@ -500,6 +567,98 @@ 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',
@@ -538,9 +697,13 @@ class VitastorDriver(driver.CloneableImageVD,
                     size = int(volume.size) * units.Gi
                     dest_name = utils.convert_str(volume.name)
                     # Find or create the base snapshot
-                    self._cli('create base snapshot', 'create', '--allow-existing', '1', base_vol.name+'@.clone_snap')
+                    snap_cfg = self._create_snapshot(base_vol.name, base_vol.name+'@.clone_snap', True)
                     # Then create a clone from it
-                    self._cli('create clone', 'create', dest_name, '--size', size, '--parent', base_vol.name+'@.clone_snap')
+                    new_cfg = self._create_image(dest_name, {
+                        'size': size,
+                        'parent_id': snap_cfg['parent_id'],
+                        'parent_pool_id': snap_cfg['parent_pool_id'],
+                    })
                     return ({}, True)
         return ({}, False)
 
@@ -607,8 +770,26 @@ class VitastorDriver(driver.CloneableImageVD,
     def extend_volume(self, volume, new_size):
         """Extend an existing volume."""
         vol_name = utils.convert_str(volume.name)
-        size = int(new_size) * units.Gi
-        self._cli('extend volume', 'modify', vol_name, '--resize', new_size)
+        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
         LOG.debug(
             "Extend volume from %(old_size)s GB to %(new_size)s GB.",
             {'old_size': volume.size, 'new_size': new_size}
@@ -681,7 +862,28 @@ class VitastorDriver(driver.CloneableImageVD,
         """
         from_name = self._get_existing_name(existing_ref)
         to_name = utils.convert_str(volume.name)
-        self._cli('rename', 'modify', from_name, '--rename', to_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
 
     def unmanage(self, volume):
         pass
@@ -754,7 +956,7 @@ 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._cli('rename', 'modify', from_name, '--rename', to_name)
+        self._rename(from_name, to_name)
 
     def unmanage_snapshot(self, snapshot):
         """Removes the specified snapshot from Cinder management."""

patches/linux-fix-atomic-write-checks.diff

@@ -1,39 +0,0 @@
-From 98d3f68a40130c438854f61db6025f9e9b099cb6 Mon Sep 17 00:00:00 2001
-From: Vitaliy Filippov <vitalifster@gmail.com>
-Date: Sat, 20 Dec 2025 14:44:35 +0300
-Subject: [PATCH] Do not require atomic writes to be power of 2 sized and
- aligned on length boundary
-
-It contradicts NVMe specification where alignment is only required when atomic
-write boundary (NABSPF/NABO) is set and highly limits usage of NVMe atomic writes
-
-Signed-off-by: Vitaliy Filippov <vitalifster@gmail.com>
----
- fs/read_write.c | 8 --------
- 1 file changed, 8 deletions(-)
-
-diff --git a/fs/read_write.c b/fs/read_write.c
-index 833bae068770..5467d710108d 100644
---- a/fs/read_write.c
-+++ b/fs/read_write.c
-@@ -1802,17 +1802,9 @@ int generic_file_rw_checks(struct file *file_in, struct file *file_out)
- 
- int generic_atomic_write_valid(struct kiocb *iocb, struct iov_iter *iter)
- {
--	size_t len = iov_iter_count(iter);
--
- 	if (!iter_is_ubuf(iter))
- 		return -EINVAL;
- 
--	if (!is_power_of_2(len))
--		return -EINVAL;
--
--	if (!IS_ALIGNED(iocb->ki_pos, len))
--		return -EINVAL;
--
- 	if (!(iocb->ki_flags & IOCB_DIRECT))
- 		return -EOPNOTSUPP;
- 
--- 
-2.51.0
-

rpm/build-tarball.sh

@@ -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 $(ls -d ~/rpmbuild/BUILD/fio*/ | grep -v SPECPARTS) fio
+ln -s ~/rpmbuild/BUILD/fio*/ fio
 sh copy-fio-includes.sh
 rm fio
 mv fio-copy fio

rpm/vitastor-el10.Dockerfile

@@ -1,17 +0,0 @@
-# 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 c-ares-devel
-RUN dnf download --source fio
-RUN rpm --nomd5 -i fio*.src.rpm
-RUN cd ~/rpmbuild/SPECS && dnf builddep -y --spec fio.spec

rpm/vitastor-el10.spec

@@ -1,199 +0,0 @@
-Name:           vitastor
-Version:        3.0.5
-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-3.0.5.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
-BuildRequires:  c-ares-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

rpm/vitastor-el7.Dockerfile

@@ -15,7 +15,7 @@ RUN yum -y --enablerepo=extras install centos-release-scl epel-release yum-utils
 RUN perl -i -pe 's!mirrorlist=!#mirrorlist=!s; s!#\s*baseurl=http://mirror.centos.org!baseurl=http://vault.centos.org!' /etc/yum.repos.d/CentOS-SCLo-scl*.repo
 RUN yum -y install https://vitastor.io/rpms/centos/7/vitastor-release-1.0-1.el7.noarch.rpm
 RUN yum -y install devtoolset-9-gcc-c++ devtoolset-9-libatomic-devel gcc make cmake gperftools-devel \
-    fio rh-nodejs12 jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libnl3-devel c-ares-devel
+    fio rh-nodejs12 jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libnl3-devel
 RUN yumdownloader --disablerepo=centos-sclo-rh --source fio
 RUN rpm --nomd5 -i fio*.src.rpm
 RUN rm -f /etc/yum.repos.d/CentOS-Media.repo

rpm/vitastor-el7.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.5
+Version:        3.0.0
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-3.0.5.el7.tar.gz
+Source0:        vitastor-3.0.0.el7.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  devtoolset-9-gcc-c++
@@ -17,7 +17,6 @@ BuildRequires:  gf-complete-devel
 BuildRequires:  rdma-core-devel
 BuildRequires:  cmake3
 BuildRequires:  libnl3-devel
-BuildRequires:  c-ares-devel
 Requires:       vitastor-osd = %{version}-%{release}
 Requires:       vitastor-mon = %{version}-%{release}
 Requires:       vitastor-client = %{version}-%{release}

rpm/vitastor-el8.Dockerfile

@@ -13,7 +13,7 @@ RUN dnf -y install centos-release-advanced-virtualization epel-release dnf-plugi
 RUN sed -i 's/^mirrorlist=/#mirrorlist=/; s!#baseurl=.*!baseurl=http://vault.centos.org/centos/8.4.2105/virt/$basearch/$avdir/!; s!^baseurl=.*Source/.*!baseurl=http://vault.centos.org/centos/8.4.2105/virt/Source/advanced-virtualization/!' /etc/yum.repos.d/CentOS-Advanced-Virtualization.repo
 RUN yum -y install https://vitastor.io/rpms/centos/8/vitastor-release-1.0-1.el8.noarch.rpm
 RUN dnf -y install gcc-toolset-9 gcc-toolset-9-gcc-c++ gperftools-devel \
-    fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel libibverbs-devel libarchive cmake libnl3-devel c-ares-devel
+    fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel libibverbs-devel libarchive cmake libnl3-devel
 RUN dnf download --source fio
 RUN rpm --nomd5 -i fio*.src.rpm
 RUN cd ~/rpmbuild/SPECS && dnf builddep -y --enablerepo=powertools --spec fio.spec

rpm/vitastor-el8.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.5
+Version:        3.0.0
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-3.0.5.el8.tar.gz
+Source0:        vitastor-3.0.0.el8.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-toolset-9-gcc-c++
@@ -16,7 +16,6 @@ BuildRequires:  gf-complete-devel
 BuildRequires:  rdma-core-devel
 BuildRequires:  cmake
 BuildRequires:  libnl3-devel
-BuildRequires:  c-ares-devel
 Requires:       vitastor-osd = %{version}-%{release}
 Requires:       vitastor-mon = %{version}-%{release}
 Requires:       vitastor-client = %{version}-%{release}

rpm/vitastor-el9.Dockerfile

@@ -10,7 +10,7 @@ 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/9/vitastor-release-1.0-1.el9.noarch.rpm
-RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libarchive cmake libnl3-devel c-ares-devel
+RUN dnf -y install gcc-c++ gperftools-devel fio nodejs rpm-build jerasure-devel libisa-l-devel gf-complete-devel rdma-core-devel libarchive 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

rpm/vitastor-el9.spec

@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        3.0.5
+Version:        3.0.0
 Release:        1%{?dist}
 Summary:        Vitastor, a fast software-defined clustered block storage
 
 License:        Vitastor Network Public License 1.1
 URL:            https://vitastor.io/
-Source0:        vitastor-3.0.5.el9.tar.gz
+Source0:        vitastor-3.0.0.el9.tar.gz
 
 BuildRequires:  gperftools-devel
 BuildRequires:  gcc-c++
@@ -16,7 +16,6 @@ BuildRequires:  gf-complete-devel
 BuildRequires:  rdma-core-devel
 BuildRequires:  cmake
 BuildRequires:  libnl3-devel
-BuildRequires:  c-ares-devel
 Requires:       vitastor-osd = %{version}-%{release}
 Requires:       vitastor-mon = %{version}-%{release}
 Requires:       vitastor-client = %{version}-%{release}

src/CMakeLists.txt

@@ -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="3.0.5")
+add_definitions(-DVITASTOR_VERSION="3.0.0")
 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})
@@ -75,14 +75,6 @@ if (RDMACM_LIBRARIES)
 	add_definitions(-DWITH_RDMACM)
 endif (RDMACM_LIBRARIES)
 
-find_package(OpenSSL REQUIRED)
-if (OPENSSL_FOUND)
-	add_definitions(-DWITH_OPENSSL)
-endif (OPENSSL_FOUND)
-
-pkg_check_modules(CARES REQUIRED libcares)
-include_directories(${CARES_INCLUDE_DIRS})
-
 if (${WITH_SYSTEM_LIBURING})
 	pkg_check_modules(LIBURING REQUIRED liburing>=2.10)
 	include_directories(${LIBURING_INCLUDE_DIRS})

src/blockstore/CMakeLists.txt

@@ -4,7 +4,7 @@ project(vitastor)
 
 # libvitastor_blk.a
 add_library(vitastor_blk STATIC
-	../util/allocator.cpp ../util/crc32c.c ../util/xxhash.c ../util/ringloop.cpp
+	../util/allocator.cpp ../util/crc32c.c ../util/ringloop.cpp
 	multilist.cpp blockstore_heap.cpp blockstore_disk.cpp
 	blockstore.cpp blockstore_impl.cpp blockstore_init.cpp blockstore_open.cpp
 	blockstore_flush.cpp blockstore_read.cpp blockstore_stable.cpp blockstore_sync.cpp blockstore_write.cpp

src/blockstore/blockstore.h

@@ -183,11 +183,6 @@ public:
     // Update configuration
     virtual void parse_config(blockstore_config_t & config) = 0;
 
-    // Reshard database for a pool in chunks
-    // 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;
-
     // Event loop
     virtual void loop() = 0;
 

src/blockstore/blockstore_disk.cpp

@@ -83,17 +83,13 @@ void blockstore_disk_t::parse_config(std::map<std::string, std::string> & config
     {
         data_csum_type = BLOCKSTORE_CSUM_CRC32C;
     }
-    else if (config["data_csum_type"] == "xxh3_32")
-    {
-        data_csum_type = BLOCKSTORE_CSUM_XXH3_32;
-    }
     else if (config["data_csum_type"] == "" || config["data_csum_type"] == "none")
     {
         data_csum_type = BLOCKSTORE_CSUM_NONE;
     }
     else
     {
-        throw std::runtime_error("data_csum_type="+config["data_csum_type"]+" is unsupported, only \"crc32c\", \"xxh3_32\" and \"none\" are supported");
+        throw std::runtime_error("data_csum_type="+config["data_csum_type"]+" is unsupported, only \"crc32c\" and \"none\" are supported");
     }
     csum_block_size = parse_size(config["csum_block_size"]);
     discard_on_start = config.find("discard_on_start") != config.end() &&
@@ -175,10 +171,6 @@ 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;
@@ -267,7 +259,7 @@ 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 + 7) / 8;
+    clean_entry_bitmap_size = data_block_size / bitmap_granularity / 8;
     clean_dyn_size = clean_entry_bitmap_size*2 + (csum_block_size
         ? data_block_size/csum_block_size*(data_csum_type & 0xFF) : 0);
 recalc:

src/blockstore/blockstore_disk.h

@@ -16,7 +16,6 @@
 #define BLOCKSTORE_CSUM_NONE 0
 // Lower byte of checksum type is its length
 #define BLOCKSTORE_CSUM_CRC32C 0x104
-#define BLOCKSTORE_CSUM_XXH3_32 0x204
 
 #define MOCK_DATA_FD 1000
 #define MOCK_META_FD 1001
@@ -27,14 +26,14 @@ class allocator_t;
 struct blockstore_disk_t
 {
     std::string data_device, meta_device, journal_device;
-    uint64_t data_block_size;
+    uint32_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
-    uint64_t journal_block_size = 4096;
+    uint32_t journal_block_size = 4096;
     // Metadata block size - minimum_io_size of the metadata device is the best choice
-    uint64_t meta_block_size = 4096;
+    uint32_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

src/blockstore/blockstore_flush.h

@@ -58,6 +58,7 @@ class journal_flusher_co
     int i, res;
     bool read_to_fill_incomplete;
     int copy_count;
+    bool do_repeat = false;
 
     friend class journal_flusher_t;
 

src/blockstore/blockstore_heap.cpp

@@ -12,7 +12,6 @@
 #include "blockstore_heap.h"
 #include "../util/allocator.h"
 #include "../util/crc32c.h"
-#include "../util/xxhash.h"
 #include "../util/malloc_or_die.h"
 
 #define BS_HEAP_FREE_MVCC 1
@@ -30,15 +29,6 @@
 #define IMAP_MALLOC_LOW_BITS ((size_t)0x0F)
 #define IMAP_MAX_LOW 16
 
-void inode_map_put(void* & inode_idx, heap_list_item_t* li);
-void inode_map_get(void *inode_idx, heap_inode_map_t::iterator & li_it, heap_list_item_t* & li, uint64_t stripe);
-void inode_map_free(void* inode_idx);
-bool inode_map_is_big(void* & inode_idx);
-void inode_map_iterate(void* & inode_idx, std::function<void(heap_list_item_t*)> cb);
-void inode_map_replace(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* new_li);
-void inode_map_erase(robin_hood::unordered_flat_map<inode_t, void*, i64hash_t> & pg_idx, void* & inode_idx,
-    const heap_inode_map_t::iterator & li_it, heap_list_item_t* li);
-
 static inline heap_list_item_t *list_item(heap_entry_t *wr)
 {
     return (heap_list_item_t*)((uint8_t*)wr - offsetof(struct heap_list_item_t, entry));
@@ -63,19 +53,19 @@ uint32_t blockstore_heap_t::get_simple_entry_size()
 uint32_t blockstore_heap_t::get_big_entry_size()
 {
     return sizeof(heap_big_write_t) + dsk->clean_entry_bitmap_size*2 +
-        (!dsk->csum_block_size ? 0 : dsk->data_block_size/dsk->csum_block_size * (dsk->data_csum_type & 0xFF));
+        (!dsk->data_csum_type ? 0 : dsk->data_block_size/dsk->csum_block_size * (dsk->data_csum_type & 0xFF));
 }
 
 uint32_t blockstore_heap_t::get_big_intent_entry_size()
 {
     return sizeof(heap_big_intent_t) + dsk->clean_entry_bitmap_size*2 +
-        (!dsk->csum_block_size ? 4 : dsk->data_block_size/dsk->csum_block_size * (dsk->data_csum_type & 0xFF));
+        (!dsk->data_csum_type ? 4 : dsk->data_block_size/dsk->csum_block_size * (dsk->data_csum_type & 0xFF));
 }
 
 uint32_t blockstore_heap_t::get_small_entry_size(uint32_t offset, uint32_t len)
 {
     return sizeof(heap_small_write_t) + dsk->clean_entry_bitmap_size +
-        (!dsk->csum_block_size ? 4 : (dsk->data_csum_type & 0xFF) *
+        (!dsk->data_csum_type ? 4 : (dsk->data_csum_type & 0xFF) *
             ((offset+len+dsk->csum_block_size-1)/dsk->csum_block_size - offset/dsk->csum_block_size));
 }
 
@@ -90,7 +80,7 @@ uint32_t blockstore_heap_t::get_csum_size(heap_entry_t *wr)
 
 uint32_t blockstore_heap_t::get_csum_size(uint32_t entry_type, uint32_t offset, uint32_t len)
 {
-    if (!dsk->csum_block_size)
+    if (!dsk->data_csum_type)
     {
         return 0;
     }
@@ -213,24 +203,15 @@ void heap_entry_t::set_big_location(blockstore_heap_t *heap, uint64_t location)
     big().block_num = location / heap->dsk->data_block_size;
 }
 
-uint32_t heap_entry_t::calc_checksum(blockstore_disk_t *dsk)
+uint32_t heap_entry_t::calc_crc32c()
 {
-    auto old_checksum = checksum;
-    checksum = 0;
-    uint32_t res = 0;
-    if (dsk->data_csum_type == BLOCKSTORE_CSUM_XXH3_32)
-        res = (uint32_t)XXH3_64bits(this, size);
-    else
-        res = ::crc32c(0, (uint8_t*)this, size);
-    checksum = old_checksum;
+    auto old_crc32c = crc32c;
+    crc32c = 0;
+    uint32_t res = ::crc32c(0, (uint8_t*)this, size);
+    crc32c = old_crc32c;
     return res;
 }
 
-uint32_t heap_entry_t::calc_checksum(blockstore_heap_t *heap)
-{
-    return calc_checksum(heap->dsk);
-}
-
 uint64_t blockstore_heap_t::get_pg_id(inode_t inode, uint64_t stripe)
 {
     uint64_t pg_num = 0;
@@ -311,13 +292,12 @@ 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 %ju bytes). ",
+                fprintf(stderr, "Error: entry is too large in metadata block %u at %u (%u > max %u 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
@@ -342,19 +322,7 @@ corrupted_block:
                     block_num, block_offset, wr->size, sizeof(heap_entry_t));
                 goto corrupted_block;
             }
-            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;
-            }
+            wr->entry_type &= ~BS_HEAP_GARBAGE;
             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)) ||
@@ -368,7 +336,6 @@ corrupted_object:
                 if (allow_corrupted)
                 {
                     fprintf(stderr, "Entry is corrupted, skipping\n");
-                    recheck_modified_blocks.insert(block_num);
                     block_offset += wr->size;
                     continue;
                 }
@@ -384,7 +351,7 @@ 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)\n",
+                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",
                     wr->inode, wr->stripe, wr->version, block_num, block_offset, wr->size, sizeof(heap_small_write_t));
                 goto corrupted_object;
             }
@@ -395,12 +362,12 @@ corrupted_object:
                 goto corrupted_object;
             }
             // Verify crc
-            uint32_t expected_checksum = wr->calc_checksum(this);
-            if (wr->checksum != expected_checksum)
+            uint32_t expected_crc32c = wr->calc_crc32c();
+            if (wr->crc32c != expected_crc32c)
             {
-                fprintf(stderr, "Error: entry %jx:%jx v%ju l%ju in metadata block %u at %u is corrupt (checksum mismatch: expected %08x, got %08x). ",
-                    wr->inode, wr->stripe, wr->version, wr->lsn,
-                    block_num, block_offset, expected_checksum, wr->checksum);
+                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,
+                    block_num, block_offset, expected_crc32c, wr->crc32c);
                 goto corrupted_object;
             }
             // Verify offset & len
@@ -409,7 +376,7 @@ 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. ",
+                fprintf(stderr, "Error: %s entry %jx:%jx v%ju has invalid offset/length: %u/%u. Metadata is incompatible with current parameters, aborting\n",
                     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;
@@ -419,7 +386,7 @@ 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. ",
+                fprintf(stderr, "Error: big_intent entry %jx:%jx v%ju has invalid offset/length: %u/%u. Metadata is incompatible with current parameters, aborting\n",
                     wr->inode, wr->stripe, wr->version, wr->big_intent().offset, wr->big_intent().len);
                 goto corrupted_object;
             }
@@ -446,7 +413,7 @@ int blockstore_heap_t::load_blocks(uint64_t disk_offset, uint64_t size, uint8_t
             next_lsn = wr->lsn;
         }
         entries_loaded++;
-        loaded_list_items.push_back(li);
+        insert_list_item(li);
         modify_alloc(block_num, [&](heap_block_info_t & inf)
         {
             if (!inf.entries.size())
@@ -482,22 +449,19 @@ 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)
         {
-            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;
-            }
+            commit_wr = wr;
             continue;
         }
         if (wr->entry_type & BS_HEAP_STABLE)
@@ -550,23 +514,6 @@ 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)
@@ -711,8 +658,7 @@ void blockstore_heap_t::recheck_buffer(heap_entry_t *cwr, uint8_t *buf)
     else if (!calc_checksums(cwr, buf, false))
     {
         // write entry is invalid, erase it and mark newer entries with garbage bit
-        auto & pg_idx = block_index[get_pg_id(cwr->inode, cwr->stripe)];
-        auto & inode_idx = pg_idx[cwr->inode];
+        auto & inode_idx = block_index[get_pg_id(cwr->inode, cwr->stripe)][cwr->inode];
         heap_inode_map_t::iterator li_it;
         heap_list_item_t *li = NULL;
         inode_map_get(inode_idx, li_it, li, cwr->stripe);
@@ -738,7 +684,7 @@ void blockstore_heap_t::recheck_buffer(heap_entry_t *cwr, uint8_t *buf)
         {
             fprintf(stderr, "Notice: the whole object %jx:%jx only has unfinished writes, rolling back\n",
                 cwr->inode, cwr->stripe);
-            inode_map_erase(pg_idx, inode_idx, li_it, li);
+            inode_map_erase(inode_idx, li_it, li);
         }
         free_entry(li);
     }
@@ -753,7 +699,6 @@ 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;
     }
@@ -845,7 +790,7 @@ std::vector<uint32_t> blockstore_heap_t::get_recheck_modified_blocks()
     return modified;
 }
 
-int blockstore_heap_t::finish_recheck()
+int blockstore_heap_t::finish_load(bool allow_corrupted)
 {
     if (!marked_used_blocks)
     {
@@ -882,17 +827,14 @@ 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)
             len = wr->big_intent().len;
         else
             assert(0);
-        uint32_t real_csum = 0;
-        if (dsk->data_csum_type == BLOCKSTORE_CSUM_XXH3_32)
-            real_csum = (uint32_t)XXH3_64bits(data, len);
-        else
-            real_csum = crc32c(0, data, len);
+        uint32_t real_csum = crc32c(0, data, len);
         if (set)
         {
             *wr_csum = real_csum;
@@ -902,14 +844,13 @@ 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) + bi.offset/dsk->csum_block_size * (dsk->data_csum_type & 0xFF)),
+        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), bi.offset, bi.offset+bi.len, set, NULL);
     }
     assert(wr->type() == BS_HEAP_SMALL_WRITE || wr->type() == BS_HEAP_INTENT_WRITE);
@@ -942,26 +883,11 @@ static uint32_t crc32c_iter(uint32_t prev_crc, const std::function<uint8_t*(uint
     return prev_crc;
 }
 
-static void xxh3_iter(XXH3_state_t* xxh3_state, const std::function<uint8_t*(uint32_t start, uint32_t & len)> & next, uint32_t pos, uint32_t size)
-{
-    uint32_t cur_len = 0;
-    while (size > 0)
-    {
-        uint8_t *data = next(pos, cur_len);
-        assert(data);
-        cur_len = (cur_len < size ? cur_len : size);
-        XXH3_64bits_update(xxh3_state, data, cur_len);
-        pos += cur_len;
-        size -= cur_len;
-    }
-}
-
 bool blockstore_heap_t::calc_block_checksums(uint32_t *block_csums, uint8_t *bitmap,
     uint32_t start, uint32_t end, std::function<uint8_t*(uint32_t start, uint32_t & len)> next,
     bool set, std::function<void(uint32_t, uint32_t, uint32_t)> bad_block_cb)
 {
     bool res = true;
-    XXH3_state_t* xxh3_state = NULL;
     uint32_t pos = start;
     uint32_t block_end = (start/dsk->csum_block_size + 1)*dsk->csum_block_size;
     uint32_t block_crc = 0;
@@ -978,214 +904,83 @@ bool blockstore_heap_t::calc_block_checksums(uint32_t *block_csums, uint8_t *bit
                     pos += dsk->bitmap_granularity;
                 // zero padding at the beginning or at the end of the block is not counted
                 if (pos > prev && prev > 0 && pos < block_end)
-                {
-                    if (dsk->data_csum_type == BLOCKSTORE_CSUM_XXH3_32)
-                    {
-                        if (!xxh3_state)
-                        {
-                            xxh3_state = XXH3_createState();
-                            XXH3_64bits_reset(xxh3_state);
-                        }
-                        uint32_t zeropad = pos-prev;
-                        while (zeropad > 0)
-                        {
-                            uint32_t zerolen = zeropad > 4096 ? 4096 : zeropad;
-                            XXH3_64bits_update(xxh3_state, zero_page, zerolen);
-                            zeropad -= zerolen;
-                        }
-                    }
-                    else
-                        block_crc = crc32c_pad(block_crc, NULL, 0, pos-prev, 0);
-                }
+                    block_crc = crc32c_pad(block_crc, NULL, 0, pos-prev, 0);
                 prev = pos;
                 while (pos < end && pos < block_end && (bitmap[pos/dsk->bitmap_granularity/8] & (1 << ((pos/dsk->bitmap_granularity) % 8))))
                     pos += dsk->bitmap_granularity;
                 if (pos > prev)
                 {
                     isset = true;
-                    if (dsk->data_csum_type == BLOCKSTORE_CSUM_XXH3_32)
-                    {
-                        if (!xxh3_state)
-                        {
-                            xxh3_state = XXH3_createState();
-                            XXH3_64bits_reset(xxh3_state);
-                        }
-                        xxh3_iter(xxh3_state, next, prev, pos-prev);
-                    }
-                    else
-                        block_crc = crc32c_iter(block_crc, next, prev, pos-prev);
+                    block_crc = crc32c_iter(block_crc, next, prev, pos-prev);
                 }
                 prev = pos;
             }
         }
         else
         {
-            if (dsk->data_csum_type == BLOCKSTORE_CSUM_XXH3_32)
-            {
-                if (!xxh3_state)
-                {
-                    xxh3_state = XXH3_createState();
-                    XXH3_64bits_reset(xxh3_state);
-                }
-                xxh3_iter(xxh3_state, next, pos, (end > block_end ? block_end : end)-pos);
-            }
-            else
-                block_crc = crc32c_iter(block_crc, next, pos, (end > block_end ? block_end : end)-pos);
+            block_crc = crc32c_iter(block_crc, next, pos, (end > block_end ? block_end : end)-pos);
             pos = (end > block_end ? block_end : end);
             isset = true;
         }
-        if (dsk->data_csum_type == BLOCKSTORE_CSUM_XXH3_32 && xxh3_state)
-        {
-            block_crc = (uint32_t)XXH3_64bits_digest(xxh3_state);
-            XXH3_64bits_reset(xxh3_state);
-        }
         if (set)
         {
             *block_csums = block_crc;
         }
         else if (isset && block_crc != *block_csums)
         {
-            res = false;
             if (bad_block_cb)
+            {
                 bad_block_cb(blk_start, *block_csums, block_crc);
+                res = false;
+            }
             else
-                break;
+                return false;
         }
         block_end += dsk->csum_block_size;
         block_crc = 0;
         block_csums++;
     }
-    if (dsk->data_csum_type == BLOCKSTORE_CSUM_XXH3_32 && xxh3_state)
-    {
-        block_crc = (uint32_t)XXH3_64bits_digest(xxh3_state);
-        XXH3_freeState(xxh3_state);
-        xxh3_state = NULL;
-    }
     return res;
 }
 
-struct heap_reshard_state_t
-{
-    int state = 0;
-    uint64_t pool_id = 0;
-    uint32_t old_pg_count = 0;
-    uint32_t pg_count = 0;
-    uint32_t pg_stripe_size = 0;
-    uint64_t chunk_size = 0;
-    heap_block_index_t new_shards;
-    heap_block_index_t old_shards;
-    heap_block_index_t::iterator sh_it;
-    robin_hood::unordered_flat_map<inode_t, void*, i64hash_t>::iterator inode_it;
-    heap_inode_map_t *stripe_map = NULL;
-    heap_inode_map_t::iterator stripe_it;
-
-    void add(heap_list_item_t *li);
-    bool run(uint64_t chunk_limit);
-};
-
-void heap_reshard_state_t::add(heap_list_item_t *li)
-{
-    // like map_to_pg()
-    uint64_t pg_num = (li->entry.stripe / pg_stripe_size) % pg_count + 1;
-    uint64_t shard_id = (pool_id << (64-POOL_ID_BITS)) | pg_num;
-    inode_map_put(new_shards[shard_id][li->entry.inode], li);
-    chunk_size++;
-}
-
-bool heap_reshard_state_t::run(uint64_t chunk_limit)
-{
-    chunk_size = 0;
-    if (state == 1)
-        goto resume_1;
-    else if (state == 2)
-        goto resume_2;
-    sh_it = old_shards.begin();
-    for (; sh_it != old_shards.end(); sh_it++)
-    {
-        inode_it = sh_it->second.begin();
-        for (; inode_it != sh_it->second.end(); inode_it++)
-        {
-            if (!inode_map_is_big(inode_it->second))
-            {
-                if (chunk_limit > 0 && chunk_size >= chunk_limit)
-                {
-                    state = 1;
-                    return false;
-                }
-resume_1:
-                inode_map_iterate(inode_it->second, [&](heap_list_item_t *li) { add(li); });
-            }
-            else
-            {
-                stripe_map = (heap_inode_map_t*)inode_it->second;
-                stripe_it = stripe_map->begin();
-                for (; stripe_it != stripe_map->end(); stripe_it++)
-                {
-                    if (chunk_limit > 0 && chunk_size >= chunk_limit)
-                    {
-                        state = 2;
-                        return false;
-                    }
-resume_2:
-                    add(*stripe_it);
-                }
-            }
-            inode_map_free(inode_it->second);
-        }
-    }
-    return true;
-}
-
-void* blockstore_heap_t::reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit)
+void blockstore_heap_t::reshard(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size)
 {
     auto & pool_settings = pool_shard_settings[pool];
     if (pool_settings.pg_count == pg_count && pool_settings.pg_stripe_size == pg_stripe_size)
     {
-        return NULL;
+        return;
     }
-    heap_reshard_state_t *st = new heap_reshard_state_t;
-    st->pool_id = (uint64_t)pool;
-    st->pg_count = pg_count;
-    st->pg_stripe_size = pg_stripe_size;
-    st->old_pg_count = !pool_settings.pg_count ? 1 : pool_settings.pg_count;
-    for (uint32_t pg_num = 0; pg_num <= st->old_pg_count; pg_num++)
+    uint32_t old_pg_count = !pool_settings.pg_count ? 1 : pool_settings.pg_count;
+    uint64_t pool_id = (uint64_t)pool;
+    heap_block_index_t new_shards;
+    for (uint32_t pg_num = 0; pg_num <= old_pg_count; pg_num++)
     {
-        auto sh_it = block_index.find((st->pool_id << (64-POOL_ID_BITS)) | pg_num);
-        if (sh_it != block_index.end())
+        auto sh_it = block_index.find((pool_id << (64-POOL_ID_BITS)) | pg_num);
+        if (sh_it == block_index.end())
         {
-            st->old_shards[pg_num] = std::move(sh_it->second);
-            block_index.erase(sh_it);
+            continue;
         }
+        for (auto & inode_pair: sh_it->second)
+        {
+            inode_map_iterate(inode_pair.second, [&](heap_list_item_t *li)
+            {
+                // like map_to_pg()
+                uint64_t pg_num = (li->entry.stripe / pg_stripe_size) % pg_count + 1;
+                uint64_t shard_id = (pool_id << (64-POOL_ID_BITS)) | pg_num;
+                inode_map_put(new_shards[shard_id][li->entry.inode], li);
+            });
+            inode_map_free(inode_pair.second);
+        }
+        block_index.erase(sh_it);
     }
-    bool finished = reshard_continue(st, chunk_limit);
-    return finished ? NULL : st;
-}
-
-bool blockstore_heap_t::reshard_continue(void *reshard_state, uint64_t chunk_limit)
-{
-    heap_reshard_state_t *st = (heap_reshard_state_t*)reshard_state;
-    if (!st->run(chunk_limit))
-    {
-        return false;
-    }
-    for (auto sh_it = st->new_shards.begin(); sh_it != st->new_shards.end(); sh_it++)
+    for (auto sh_it = new_shards.begin(); sh_it != new_shards.end(); sh_it++)
     {
         block_index[sh_it->first] = std::move(sh_it->second);
     }
-    pool_shard_settings[st->pool_id] = (pool_shard_settings_t){
-        .pg_count = st->pg_count,
-        .pg_stripe_size = st->pg_stripe_size,
+    pool_settings = (pool_shard_settings_t){
+        .pg_count = pg_count,
+        .pg_stripe_size = pg_stripe_size,
     };
-    delete st;
-    return true;
-}
-
-bool blockstore_heap_t::reshard_check(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size)
-{
-    auto set_it = pool_shard_settings.find(pool);
-    return (set_it != pool_shard_settings.end() &&
-        set_it->second.pg_count == pg_count &&
-        set_it->second.pg_stripe_size == pg_stripe_size);
 }
 
 heap_entry_t *blockstore_heap_t::lock_and_read_entry(object_id oid)
@@ -1200,6 +995,27 @@ 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);
@@ -1236,35 +1052,6 @@ 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)
@@ -1331,7 +1118,31 @@ 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);
-    gc_block(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;
+    }
     *block_num = last_allocated_block;
     modify_alloc(last_allocated_block, [&](heap_block_info_t & inf)
     {
@@ -1412,7 +1223,7 @@ int blockstore_heap_t::add_entry(uint32_t wr_size, uint32_t *modified_block,
     insert_list_item(li);
     li->block_num = block_num;
     new_wr->size = wr_size;
-    new_wr->checksum = new_wr->calc_checksum(this);
+    new_wr->crc32c = new_wr->calc_crc32c();
     return 0;
 }
 
@@ -1472,7 +1283,7 @@ int blockstore_heap_t::add_big_write(object_id oid, heap_entry_t *old_head, bool
             memset(wr->get_ext_bitmap(this), 0, dsk->clean_entry_bitmap_size);
         memset(wr->get_int_bitmap(this), 0, dsk->clean_entry_bitmap_size);
         bitmap_set(wr->get_int_bitmap(this), offset, len, dsk->bitmap_granularity);
-        if (dsk->csum_block_size)
+        if (dsk->data_csum_type)
         {
             memset(wr->get_checksums(this), 0, get_csum_size(wr));
             calc_checksums(wr, (uint8_t*)data, true, offset, len);
@@ -1501,9 +1312,9 @@ int blockstore_heap_t::add_redirect_intent(object_id oid, heap_entry_t **obj_ptr
             memset(wr->get_ext_bitmap(this), 0, dsk->clean_entry_bitmap_size);
         memset(wr->get_int_bitmap(this), 0, dsk->clean_entry_bitmap_size);
         bitmap_set(wr->get_int_bitmap(this), offset, len, dsk->bitmap_granularity);
-        if (dsk->csum_block_size)
+        if (dsk->data_csum_type)
             memset(wr->get_checksums(this), 0, get_csum_size(wr));
-        calc_checksums(wr, (uint8_t*)data, true);
+        calc_checksums(wr, (uint8_t*)data, true, offset, len);
         *obj_ptr = wr;
     });
 }
@@ -1539,14 +1350,14 @@ int blockstore_heap_t::add_big_intent(object_id oid, heap_entry_t **obj_ptr, uin
             memcpy(wr->get_ext_bitmap(this), obj->get_ext_bitmap(this), dsk->clean_entry_bitmap_size);
         memcpy(wr->get_int_bitmap(this), obj->get_int_bitmap(this), dsk->clean_entry_bitmap_size);
         bitmap_set(wr->get_int_bitmap(this), offset, len, dsk->bitmap_granularity);
-        if (dsk->csum_block_size)
+        if (dsk->data_csum_type)
         {
             if (checksums)
-                memcpy(wr->get_checksums(this), checksums, get_csum_size(wr));
+                memcpy(wr->get_checksums(this), checksums, dsk->clean_entry_bitmap_size);
             else
             {
-                memcpy(wr->get_checksums(this), obj->get_checksums(this), get_csum_size(wr));
-                calc_checksums(wr, (uint8_t*)data, true);
+                memcpy(wr->get_checksums(this), obj->get_checksums(this), dsk->clean_entry_bitmap_size);
+                calc_checksums(wr, (uint8_t*)data, true, offset, len);
             }
         }
         else
@@ -1586,7 +1397,7 @@ int blockstore_heap_t::add_compact(heap_entry_t *obj, uint64_t compact_version,
         new_wr->set_big_location(this, compact_location);
         memcpy(new_wr->get_int_bitmap(this), new_int_bitmap, dsk->clean_entry_bitmap_size);
         memcpy(new_wr->get_ext_bitmap(this), new_ext_bitmap, dsk->clean_entry_bitmap_size);
-        if (dsk->csum_block_size && new_csums)
+        if (dsk->data_csum_type && new_csums)
             memcpy(new_wr->get_checksums(this), new_csums, dsk->data_block_size/dsk->csum_block_size*(dsk->data_csum_type & 0xFF));
     });
 }
@@ -1667,7 +1478,7 @@ int blockstore_heap_t::add_commit(heap_entry_t *obj, uint64_t version, uint32_t
     }
     if (!uncommitted)
     {
-        return 0;
+        return EBUSY;
     }
     return add_simple(obj, version, modified_block, BS_HEAP_COMMIT);
 }
@@ -1677,32 +1488,23 @@ 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;
-    uint64_t rollback_version = UINT64_MAX;
-    while (wr)
+    while (wr && !wr->is_overwrite())
     {
         if (wr->type() == BS_HEAP_ROLLBACK)
         {
-            if (wr->version <= version)
-            {
-                // All previous writes are already rolled back, stop
-                break;
-            }
-            rollback_version = wr->version;
+            auto rollback_version = wr->version;
             wr = prev(wr);
+            while (wr->version > rollback_version)
+            {
+                assert(!(wr->entry_type & BS_HEAP_STABLE));
+                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;
         }
@@ -1713,10 +1515,14 @@ int blockstore_heap_t::add_rollback(heap_entry_t *obj, uint64_t version, uint32_
             {
                 return EBUSY;
             }
-            else
+            else if (wr->version == version)
             {
                 break;
             }
+            else if (wr->version < version)
+            {
+                return ENOENT;
+            }
         }
         else if (wr->version > version)
         {
@@ -1897,9 +1703,9 @@ void blockstore_heap_t::iterate_with_stable(heap_entry_t *obj, uint64_t max_lsn,
         }
         else
         {
-            // 1) 1 2 3 ROLLBACK(2) COMMIT(3) -> 3 is unstable
+            // 1) 1 2 3 ROLLBACK(2) COMMIT(3) -> impossible
             // 2) 1 2 3 4 ROLLBACK(3) COMMIT(2) -> OK
-            // 3) 1 2 3 ROLLBACK(2) 3 COMMIT(3) -> first 3 is unstable
+            // 3) 1 2 3 ROLLBACK(2) 3 COMMIT(3) -> first 3 shouldn't be treated as stable
             // 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)
@@ -2056,21 +1862,24 @@ int blockstore_heap_t::list_objects(uint32_t pg_num, object_id min_oid, object_i
                 return;
             }
             uint64_t stable_version = 0;
-            iterate_with_stable(obj, UINT64_MAX, [&](heap_entry_t* wr, bool stable)
+            auto first_wr = obj;
+            for (auto wr = first_wr; wr; wr = prev(wr))
             {
-                if (stable)
+                if ((wr->entry_type & BS_HEAP_STABLE) || wr->type() == BS_HEAP_COMMIT || wr->type() == BS_HEAP_ROLLBACK)
                 {
                     stable_version = wr->version;
-                    return false;
+                    break;
                 }
-                if (unstable_size >= unstable_alloc)
+                else
                 {
-                    unstable_alloc = (!unstable_alloc ? 128 : unstable_alloc*2);
-                    unstable = (obj_ver_id*)realloc_or_die(unstable, sizeof(obj_ver_id) * 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[unstable_size++] = (obj_ver_id){ .oid = oid, .version = wr->version };
-                return true;
-            });
+            }
             if (stable_version)
             {
                 if (res_size >= res_alloc)
@@ -2132,13 +1941,7 @@ 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);
-    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);
-    }
+    inode_space_stats[inode] -= dsk->data_block_size;
     data_used_space -= dsk->data_block_size;
 }
 
@@ -2367,15 +2170,12 @@ 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];
+            auto & inode_idx = block_index[get_pg_id(wr->inode, wr->stripe)][wr->inode];
             heap_inode_map_t::iterator li_it;
             heap_list_item_t *old_li = NULL;
             inode_map_get(inode_idx, li_it, old_li, wr->stripe);
-            inode_map_erase(pg_idx, inode_idx, li_it, old_li);
+            inode_map_erase(inode_idx, li_it, old_li);
         }
         else
         {
@@ -2390,15 +2190,6 @@ void blockstore_heap_t::apply_inflight(heap_inflight_lsn_t & inflight)
     }
 }
 
-bool blockstore_heap_t::is_lsn_completed(uint64_t lsn)
-{
-    if (lsn <= completed_lsn)
-        return true;
-    assert(lsn-first_inflight_lsn < inflight_lsn.size());
-    auto it = inflight_lsn.begin() + (lsn-first_inflight_lsn);
-    return (it->flags & HEAP_INFLIGHT_DONE);
-}
-
 uint64_t blockstore_heap_t::get_completed_lsn()
 {
     return completed_lsn;
@@ -2423,7 +2214,7 @@ 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 == 2);
+            recalc_inode_space_stats(ps.first, ps.second.no_inode_stats == 1);
         ps.second.no_inode_stats &= 1;
     }
 }
@@ -2434,8 +2225,8 @@ 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;
-    for (uint32_t pg_num = pg_count ? 1 : 0; pg_num <= pg_count; pg_num++)
+    uint32_t pg_count = ps.pg_count ? ps.pg_count : 1;
+    for (uint32_t pg_num = 1; pg_num <= pg_count; pg_num++)
     {
         auto & pg_idx = block_index[(pool_id << (64-POOL_ID_BITS)) | pg_num];
         for (auto & ip: pg_idx)
@@ -2468,15 +2259,12 @@ void blockstore_heap_t::recalc_inode_space_stats(uint64_t pool_id, bool per_inod
 // This is some really crazy shit but it seems to work well :)
 // At the same time it has almost zero overhead and works just as fast for fat inodes.
 
-void inode_map_get(void *inode_idx, heap_inode_map_t::iterator & li_it, heap_list_item_t* & li, uint64_t stripe)
+void blockstore_heap_t::inode_map_get(void *inode_idx, heap_inode_map_t::iterator & li_it, heap_list_item_t* & li, uint64_t stripe)
 {
     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)
@@ -2498,7 +2286,7 @@ void inode_map_get(void *inode_idx, heap_inode_map_t::iterator & li_it, heap_lis
     }
 }
 
-void inode_map_free(void* inode_idx)
+void blockstore_heap_t::inode_map_free(void* inode_idx)
 {
     size_t n = ((size_t)inode_idx & IMAP_MALLOC_LOW_BITS);
     if (!n)
@@ -2511,12 +2299,7 @@ void inode_map_free(void* inode_idx)
     }
 }
 
-bool inode_map_is_big(void* & inode_idx)
-{
-    return !((size_t)inode_idx & IMAP_MALLOC_LOW_BITS);
-}
-
-void inode_map_iterate(void* & inode_idx, std::function<void(heap_list_item_t*)> cb)
+void blockstore_heap_t::inode_map_iterate(void* & inode_idx, std::function<void(heap_list_item_t*)> cb)
 {
     size_t n = ((size_t)inode_idx & IMAP_MALLOC_LOW_BITS);
     if (!n)
@@ -2543,7 +2326,7 @@ void inode_map_iterate(void* & inode_idx, std::function<void(heap_list_item_t*)>
     }
 }
 
-void inode_map_put(void* & inode_idx, heap_list_item_t* li)
+void blockstore_heap_t::inode_map_put(void* & inode_idx, heap_list_item_t* li)
 {
     if (!inode_idx)
     {
@@ -2612,7 +2395,7 @@ void inode_map_put(void* & inode_idx, heap_list_item_t* li)
     }
 }
 
-void inode_map_replace(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* new_li)
+void blockstore_heap_t::inode_map_replace(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* new_li)
 {
     size_t map_n = ((size_t)inode_idx & IMAP_MALLOC_LOW_BITS);
     if (!map_n)
@@ -2638,8 +2421,7 @@ void inode_map_replace(void* & inode_idx, const heap_inode_map_t::iterator & li_
     }
 }
 
-void inode_map_erase(robin_hood::unordered_flat_map<inode_t, void*, i64hash_t> & pg_idx, void* & inode_idx,
-    const heap_inode_map_t::iterator & li_it, heap_list_item_t* li)
+void blockstore_heap_t::inode_map_erase(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* li)
 {
     size_t map_n = ((size_t)inode_idx & IMAP_MALLOC_LOW_BITS);
     if (!map_n)
@@ -2664,7 +2446,7 @@ void inode_map_erase(robin_hood::unordered_flat_map<inode_t, void*, i64hash_t> &
     else if (map_n == 1)
     {
         // Erase
-        pg_idx.erase(li->entry.inode);
+        block_index[get_pg_id(li->entry.inode, li->entry.stripe)].erase(li->entry.inode);
     }
     else
     {

src/blockstore/blockstore_heap.h

@@ -43,7 +43,7 @@ struct __attribute__((__packed__)) heap_entry_t
 {
     uint16_t size;
     uint16_t entry_type;
-    uint32_t checksum;
+    uint32_t crc32c;
     uint64_t lsn;
     uint64_t inode;
     uint64_t stripe;
@@ -69,8 +69,7 @@ struct __attribute__((__packed__)) heap_entry_t
     uint32_t *get_checksum(blockstore_heap_t *heap);
     uint64_t big_location(blockstore_heap_t *heap);
     void set_big_location(blockstore_heap_t *heap, uint64_t location);
-    uint32_t calc_checksum(blockstore_heap_t *heap);
-    uint32_t calc_checksum(blockstore_disk_t *dsk);
+    uint32_t calc_crc32c();
 };
 
 struct __attribute__((__packed__)) heap_small_write_t
@@ -81,7 +80,7 @@ struct __attribute__((__packed__)) heap_small_write_t
     uint32_t offset;
     uint32_t len;
 
-    // Also includes 1 bitmap and 1 checksum after the bitmap if block checksums are disabled
+    // Also includes 1 bitmap and 1 crc32c after the bitmap if checksums are disabled
 };
 
 struct __attribute__((__packed__)) heap_big_write_t
@@ -99,7 +98,7 @@ struct __attribute__((__packed__)) heap_big_intent_t
     uint32_t offset;
     uint32_t len;
 
-    // Also includes 2 bitmaps and 1 checksums if block checksums are disabled
+    // Also includes 2 bitmaps and 1 crc32c if checksums are disabled
 };
 
 struct __attribute__((__packed__)) heap_list_item_t
@@ -138,8 +137,6 @@ struct heap_compact_t
     bool do_delete;
 };
 
-struct heap_reshard_state_t;
-
 struct heap_li_hash
 {
     size_t operator()(const heap_list_item_t* li) const noexcept
@@ -164,7 +161,7 @@ using heap_mvcc_map_t = robin_hood::unordered_flat_map<object_id, heap_object_mv
 
 class blockstore_heap_t
 {
-    friend struct heap_entry_t;
+    friend class heap_entry_t;
 
     blockstore_disk_t *dsk = NULL;
     uint8_t* buffer_area = NULL;
@@ -201,7 +198,6 @@ 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;
@@ -209,15 +205,20 @@ class blockstore_heap_t
     std::function<void(bool is_data, uint64_t offset, uint64_t len, uint8_t* buf, std::function<void()>)> recheck_cb;
     int recheck_queue_depth = 0;
 
+    void inode_map_put(void* & inode_idx, heap_list_item_t* li);
+    void inode_map_get(void *inode_idx, heap_inode_map_t::iterator & li_it, heap_list_item_t* & li, uint64_t stripe);
+    void inode_map_free(void* inode_idx);
+    void inode_map_iterate(void* & inode_idx, std::function<void(heap_list_item_t*)> cb);
+    void inode_map_replace(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* new_li);
+    void inode_map_erase(void* & inode_idx, const heap_inode_map_t::iterator & li_it, heap_list_item_t* li);
+
     uint64_t get_pg_id(inode_t inode, uint64_t stripe);
     bool validate_object(heap_entry_t *obj);
     void fill_recheck_queue();
     int mark_used_blocks();
     void recheck_buffer(heap_entry_t *cwr, uint8_t *buf);
     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,
@@ -240,29 +241,28 @@ 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 - should be called after load_blocks
-    void finish_load();
+    // finish loading
+    int finish_load(bool allow_corrupted = false);
     // 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(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size);
     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 = UINT32_MAX, uint32_t len = UINT32_MAX);
+    bool calc_checksums(heap_entry_t *wr, uint8_t *data, bool set, uint32_t offset = 0, uint32_t len = 0);
     // 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);
@@ -314,7 +314,6 @@ public:
     void start_block_write(uint32_t block_num);
     void complete_block_write(uint32_t block_num);
     void complete_lsn_write(uint64_t lsn);
-    bool is_lsn_completed(uint64_t lsn);
     uint64_t get_completed_lsn();
     uint64_t get_fsynced_lsn();
     void mark_lsn_fsynced(uint64_t lsn);

src/blockstore/blockstore_impl.cpp

@@ -23,7 +23,6 @@ blockstore_impl_t::blockstore_impl_t(blockstore_config_t & config, ring_loop_i *
         dsk.open_meta();
         dsk.open_journal();
         dsk.calc_lengths();
-        dsk.check_lengths();
     }
     catch (std::exception & e)
     {
@@ -32,13 +31,16 @@ blockstore_impl_t::blockstore_impl_t(blockstore_config_t & config, ring_loop_i *
     }
     meta_superblock = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, dsk.meta_block_size);
     memset(meta_superblock, 0, dsk.meta_block_size);
+}
+
+void blockstore_impl_t::init()
+{
     flusher = new journal_flusher_t(this);
     if (dsk.inmemory_journal)
     {
         buffer_area = (uint8_t*)memalign_or_die(MEM_ALIGNMENT, dsk.journal_len);
     }
     heap = new blockstore_heap_t(&dsk, buffer_area, log_level);
-    ringloop->wakeup();
 }
 
 blockstore_impl_t::~blockstore_impl_t()
@@ -193,12 +195,12 @@ 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;
@@ -323,13 +325,9 @@ void blockstore_impl_t::process_list(blockstore_op_t *op)
         FINISH_OP(op);
         return;
     }
-    // Check if the DB is sharded correctly
-    if (!heap->reshard_check(INODE_POOL(min_inode), pg_count, pg_stripe_size))
-    {
-        op->retval = -EAGAIN;
-        FINISH_OP(op);
-        return;
-    }
+    // Check if the DB needs resharding
+    // (we don't know about PGs from the beginning, we only create "shards" here)
+    heap->reshard(INODE_POOL(min_inode), pg_count, pg_stripe_size);
     obj_ver_id *result = NULL;
     size_t stable_count = 0, unstable_count = 0;
     int res = heap->list_objects(list_pg, op->min_oid, op->max_oid, &result, &stable_count, &unstable_count);
@@ -396,13 +394,3 @@ std::string blockstore_impl_t::get_op_diag(blockstore_op_t *op)
         snprintf(buf, sizeof(buf), "state=%d", priv->op_state);
     return std::string(buf);
 }
-
-void* blockstore_impl_t::reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit)
-{
-    return heap->reshard_start(pool, pg_count, pg_stripe_size, chunk_limit);
-}
-
-bool blockstore_impl_t::reshard_continue(void *reshard_state, uint64_t chunk_limit)
-{
-    return heap->reshard_continue(reshard_state, chunk_limit);
-}

src/blockstore/blockstore_impl.h

@@ -78,7 +78,6 @@ 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;
@@ -143,6 +142,7 @@ public:
     int metadata_buf_size;
     blockstore_init_meta* metadata_init_reader;
 
+    void init();
     void check_wait(blockstore_op_t *op);
     void init_op(blockstore_op_t *op);
 
@@ -190,9 +190,6 @@ public:
     void parse_config(blockstore_config_t & config);
     void parse_config(blockstore_config_t & config, bool init);
 
-    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);
-
     // Event loop
     void loop();
 

src/blockstore/blockstore_init.cpp

@@ -72,6 +72,7 @@ resume_1:
     }
     if (is_zero((uint64_t*)bs->meta_superblock, bs->dsk.meta_block_size))
     {
+        bs->dsk.check_lengths();
         {
             blockstore_meta_header_v3_t *hdr = (blockstore_meta_header_v3_t *)bs->meta_superblock;
             hdr->zero = 0;
@@ -140,12 +141,12 @@ resume_1:
             hdr->bitmap_granularity != bs->dsk.bitmap_granularity ||
             hdr->data_csum_type != bs->dsk.data_csum_type ||
             hdr->csum_block_size != bs->dsk.csum_block_size ||
-            hdr->meta_area_size != bs->dsk.meta_area_size)
+            hdr->meta_area_size > bs->dsk.meta_area_size)
         {
             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 (%ju/%ju/%u, %u/%u, %ju).\n",
+                " differs from OSD configuration (%u/%u/%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,
@@ -153,7 +154,15 @@ resume_1:
             );
             exit(1);
         }
+        bs->dsk.meta_area_size = hdr->meta_area_size;
+        if (bs->dsk.meta_format != hdr->version)
+        {
+            bs->dsk.meta_format = hdr->version;
+            bs->dsk.calc_lengths();
+        }
+        bs->dsk.check_lengths();
     }
+    bs->init();
     bs->heap->start_load(((blockstore_meta_header_v3_t *)bs->meta_superblock)->completed_lsn);
     if (bs->dsk.inmemory_journal)
     {
@@ -225,7 +234,7 @@ 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, bs->skip_corrupted_meta_entries, loaded);
+            int r = bs->heap->load_blocks(bufs[i].offset-bs->dsk.meta_block_size, bufs[i].size, bufs[i].buf, false, loaded);
             if (r != 0)
                 exit(1);
             entries_loaded += loaded;
@@ -239,7 +248,6 @@ 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)
     {
@@ -270,7 +278,7 @@ resume_6:
         }
         GET_SQE();
         data->iov = (iovec){ buf, len };
-        data->callback = [offset, cb](ring_data_t *data)
+        data->callback = [this, offset, cb](ring_data_t *data)
         {
             if (data->res < 0)
             {
@@ -285,7 +293,7 @@ resume_6:
     }, bs->meta_write_recheck_parallelism);
     return 1;
 resume_7:
-    if (bs->heap->finish_recheck() != 0)
+    if (bs->heap->finish_load() != 0)
     {
         exit(1);
     }

src/blockstore/blockstore_open.cpp

@@ -28,7 +28,6 @@ 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);

src/blockstore/blockstore_read.cpp

@@ -22,7 +22,7 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
     uint64_t result_version = 0;
     bool found = false;
     uint32_t skip_csum = 0;
-    uint32_t blk_start = op->offset, blk_end = op->offset+op->len;
+    uint32_t blk_start = 0, blk_end = 0;
     bool need_skip = dsk.csum_block_size > dsk.bitmap_granularity && !perfect_csum_update;
     if (need_skip)
     {
@@ -32,28 +32,12 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
         if (blk_end % dsk.csum_block_size)
             blk_end += dsk.csum_block_size - (blk_end % dsk.csum_block_size);
     }
-    bool need_wait = false;
     heap->iterate_with_stable(obj, obj->lsn, [&](heap_entry_t *wr, bool stable)
     {
         if (wr->type() == BS_HEAP_DELETE)
         {
             return false;
         }
-        if (!heap->is_lsn_completed(wr->lsn))
-        {
-            if (wr->type() == BS_HEAP_BIG_INTENT && wr->big_intent().offset < blk_end && wr->big_intent().offset+wr->big_intent().len > blk_start ||
-                wr->type() == BS_HEAP_INTENT_WRITE && wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start)
-            {
-                // Wait until intent write is completed
-                need_wait = true;
-                return false;
-            }
-            else if (wr->type() == BS_HEAP_SMALL_WRITE && wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start)
-            {
-                // Skip entry and read the previous one
-                return true;
-            }
-        }
         if (op->version >= wr->version && !found)
         {
             found = true;
@@ -63,6 +47,12 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
                 memcpy(op->bitmap, wr->get_ext_bitmap(heap), dsk.clean_entry_bitmap_size);
             }
         }
+        if (need_skip && wr->lsn < heap->get_completed_lsn() &&
+            (wr->type() == BS_HEAP_BIG_INTENT && wr->big_intent().offset < blk_end && wr->big_intent().offset+wr->big_intent().len > blk_start ||
+            wr->type() == BS_HEAP_INTENT_WRITE && wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start))
+        {
+            skip_csum = COPY_BUF_SKIP_CSUM;
+        }
         if (op->version >= wr->version)
         {
             fulfilled += prepare_read(PRIV(op)->read_vec, obj, wr, op->offset, op->offset+op->len,
@@ -75,23 +65,13 @@ int blockstore_impl_t::dequeue_read(blockstore_op_t *op)
                 return false;
             }
         }
-        if (need_skip && wr->type() == BS_HEAP_SMALL_WRITE &&
+        if (need_skip && (wr->type() == BS_HEAP_SMALL_WRITE || wr->type() == BS_HEAP_INTENT_WRITE) &&
             wr->small().offset < blk_end && wr->small().offset+wr->small().len > blk_start)
         {
-            // Small write may mutate big write checksums during flush
             skip_csum = COPY_BUF_SKIP_CSUM;
         }
         return true;
     });
-    if (need_wait)
-    {
-undo_wait:
-        // Need to wait. undo added requests, unlock lsn
-        heap->unlock_entry(op->oid);
-        free_read_buffers(rv);
-        rv.clear();
-        return 0;
-    }
     if (!found)
     {
         // May happen if there are entries but all of them are > requested version
@@ -104,7 +84,11 @@ undo_wait:
     assert(fulfilled == op->len);
     if (!fulfill_read(op))
     {
-        goto undo_wait;
+        // Need to wait. undo added requests, unlock lsn
+        heap->unlock_entry(op->oid);
+        free_read_buffers(rv);
+        rv.clear();
+        return 0;
     }
     op->version = result_version;
     if (!PRIV(op)->pending_ops)

src/blockstore/blockstore_stable.cpp

@@ -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:

src/blockstore/blockstore_write.cpp

@@ -22,23 +22,21 @@ 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](ring_data_t *data)
+    data->callback = [this, modified_block, buf](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);
         }
-        auto it = modified_blocks.find(modified_block);
-        assert(it != modified_blocks.end());
-        free(it->second.buf);
-        modified_blocks.erase(it);
+        modified_blocks.erase(modified_block);
         heap->complete_block_write(modified_block);
         ringloop->wakeup();
     };
     io_uring_prep_writev(
-        sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + ((uint64_t)modified_block+1)*dsk.meta_block_size
+        sqe, dsk.meta_fd, &data->iov, 1, dsk.meta_offset + (modified_block+1)*dsk.meta_block_size
     );
     unsynced_meta_write_count++;
     pending_modified_blocks.push_back(modified_block);
@@ -251,12 +249,13 @@ 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);
@@ -349,7 +348,6 @@ 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)
@@ -406,12 +404,11 @@ 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)->pending_ops--;
+                PRIV(op)->op_state = 8;
                 ringloop->wakeup();
             });
             return 1;

src/blockstore/v1/impl.cpp

@@ -407,77 +407,32 @@ blockstore_clean_db_t& blockstore_impl_t::clean_db_shard(object_id oid)
     return clean_db_shards[(pool_id << (64-POOL_ID_BITS)) | pg_num];
 }
 
-struct bs_reshard_state_t
+void blockstore_impl_t::reshard_clean_db(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size)
 {
-    int state = 0;
-    uint64_t pool_id = 0;
-    uint32_t pg_count = 0;
-    uint32_t pg_stripe_size = 0;
-    uint64_t chunk_size = 0;
-    std::map<pool_pg_id_t, blockstore_clean_db_t> old_shards;
+    uint64_t pool_id = (uint64_t)pool;
     std::map<pool_pg_id_t, blockstore_clean_db_t> new_shards;
-    std::map<pool_pg_id_t, blockstore_clean_db_t>::iterator sh_it;
-    blockstore_clean_db_t::iterator obj_it;
-};
-
-void* blockstore_impl_t::reshard_start(pool_id_t pool, uint32_t pg_count, uint32_t pg_stripe_size, uint64_t chunk_limit)
-{
-    auto & settings = clean_db_settings[pool];
-    if (settings.pg_count == pg_count && settings.pg_stripe_size == pg_stripe_size)
-    {
-        return NULL;
-    }
-    bs_reshard_state_t *st = new bs_reshard_state_t;
-    st->state = 0;
-    st->pool_id = pool;
-    st->pg_count = pg_count;
-    st->pg_stripe_size = pg_stripe_size;
-    auto sh_it = clean_db_shards.lower_bound((st->pool_id << (64-POOL_ID_BITS)));
+    auto sh_it = clean_db_shards.lower_bound((pool_id << (64-POOL_ID_BITS)));
     while (sh_it != clean_db_shards.end() &&
-        (sh_it->first >> (64-POOL_ID_BITS)) == st->pool_id)
+        (sh_it->first >> (64-POOL_ID_BITS)) == pool_id)
     {
-        st->old_shards[sh_it->first] = std::move(sh_it->second);
+        for (auto & pair: sh_it->second)
+        {
+            // like map_to_pg()
+            uint64_t pg_num = (pair.first.stripe / pg_stripe_size) % pg_count + 1;
+            uint64_t shard_id = (pool_id << (64-POOL_ID_BITS)) | pg_num;
+            new_shards[shard_id][pair.first] = pair.second;
+        }
         clean_db_shards.erase(sh_it++);
     }
-    bool finished = reshard_continue(st, chunk_limit);
-    return finished ? NULL : st;
-}
-
-bool blockstore_impl_t::reshard_continue(void *reshard_state, uint64_t chunk_limit)
-{
-    bs_reshard_state_t *st = (bs_reshard_state_t*)reshard_state;
-    uint64_t chunk_size = 0;
-    if (st->state == 1)
-        goto resume_1;
-    for (st->sh_it = st->old_shards.begin(); st->sh_it != st->old_shards.end(); )
-    {
-        for (st->obj_it = st->sh_it->second.begin(); st->obj_it != st->sh_it->second.end(); st->obj_it++)
-        {
-            if (chunk_limit > 0 && chunk_size >= chunk_limit)
-            {
-                st->state = 1;
-                return false;
-            }
-resume_1:
-            // like map_to_pg()
-            uint64_t pg_num = (st->obj_it->first.stripe / st->pg_stripe_size) % st->pg_count + 1;
-            uint64_t shard_id = (st->pool_id << (64-POOL_ID_BITS)) | pg_num;
-            st->new_shards[shard_id][st->obj_it->first] = st->obj_it->second;
-            chunk_size++;
-        }
-        st->old_shards.erase(st->sh_it++);
-    }
-    for (auto sh_it = st->new_shards.begin(); sh_it != st->new_shards.end(); sh_it++)
+    for (sh_it = new_shards.begin(); sh_it != new_shards.end(); sh_it++)
     {
         auto & to = clean_db_shards[sh_it->first];
         to.swap(sh_it->second);
     }
-    clean_db_settings[st->pool_id] = (pool_shard_settings_t){
-        .pg_count = st->pg_count,
-        .pg_stripe_size = st->pg_stripe_size,
+    clean_db_settings[pool_id] = (pool_shard_settings_t){
+        .pg_count = pg_count,
+        .pg_stripe_size = pg_stripe_size,
     };
-    delete st;
-    return true;
 }
 
 void blockstore_impl_t::process_list(blockstore_op_t *op)
@@ -510,10 +465,7 @@ void blockstore_impl_t::process_list(blockstore_op_t *op)
                 sh_it->second.pg_count != pg_count ||
                 sh_it->second.pg_stripe_size != pg_stripe_size)
             {
-                // Sharding mismatch
-                op->retval = -EAGAIN;
-                FINISH_OP(op);
-                return;
+                reshard_clean_db(pool_id, pg_count, pg_stripe_size);
             }
             first_shard = last_shard = ((uint64_t)pool_id << (64-POOL_ID_BITS)) | list_pg;
         }

src/blockstore/v1/impl.h

@@ -202,6 +202,7 @@ class blockstore_impl_t: public blockstore_i
     uint8_t* get_clean_entry_bitmap(uint64_t block_loc, int offset);
 
     blockstore_clean_db_t& clean_db_shard(object_id oid);
+    void reshard_clean_db(pool_id_t pool_id, uint32_t pg_count, uint32_t pg_stripe_size);
     void recalc_inode_space_stats(uint64_t pool_id, bool per_inode);
 
     // Journaling
@@ -287,10 +288,6 @@ public:
     void parse_config(blockstore_config_t & config);
     void parse_config(blockstore_config_t & config, bool init);
 
-    // 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);
-
     // Event loop
     void loop();
 

src/blockstore/v1/init.cpp

@@ -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 (%ju/%ju/%u, %u/%u).\n",
+                " differs from OSD configuration (%u/%u/%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,

src/blockstore/v1/read.cpp

@@ -620,7 +620,8 @@ 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);
+        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));
         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))
         {

src/blockstore/v1/write.cpp

@@ -183,7 +183,7 @@ bool blockstore_impl_t::enqueue_write(blockstore_op_t *op)
         uint32_t end = (op->offset+op->len-1) / dsk.csum_block_size;
         auto fn = state & BS_ST_BIG_WRITE ? crc32c_pad : crc32c_nopad;
         if (start == end)
-            data_csums[0] = fn(0, op->buf, op->len, op->offset - start*dsk.csum_block_size, (end+1)*dsk.csum_block_size - (op->offset+op->len));
+            data_csums[0] = fn(0, op->buf, op->len, op->offset - start*dsk.csum_block_size, end*dsk.csum_block_size - (op->offset+op->len));
         else
         {
             // First block

src/client/CMakeLists.txt

@@ -13,10 +13,10 @@ if (RDMACM_LIBRARIES)
 endif (RDMACM_LIBRARIES)
 add_library(vitastor_common STATIC
 	../util/epoll_manager.cpp etcd_state_client.cpp messenger.cpp ../util/addr_util.cpp
-	msgr_encrypt.cpp msgr_stop.cpp msgr_op.cpp msgr_send.cpp msgr_receive.cpp ../util/ringloop.cpp ../../json11/json11.cpp
+	msgr_stop.cpp msgr_op.cpp msgr_send.cpp msgr_receive.cpp ../util/ringloop.cpp ../../json11/json11.cpp
 	http_client.cpp osd_ops.cpp pg_states.cpp ../util/timerfd_manager.cpp ../util/str_util.cpp ../util/json_util.cpp ${MSGR_RDMA} ${MSGR_RDMACM}
 )
-target_link_libraries(vitastor_common pthread ${OPENSSL_LIBRARIES} ${CARES_LIBRARIES})
+target_link_libraries(vitastor_common pthread)
 target_compile_options(vitastor_common PUBLIC -fPIC)
 
 # libvitastor_client.so
@@ -24,7 +24,6 @@ add_library(vitastor_client SHARED
 	cluster_client.cpp
 	cluster_client_list.cpp
 	cluster_client_wb.cpp
-	cluster_client_icache.cpp
 	vitastor_c.cpp
 )
 set_target_properties(vitastor_client PROPERTIES PUBLIC_HEADER "client/vitastor_c.h")
@@ -34,7 +33,6 @@ target_link_libraries(vitastor_client
 	${LIBURING_LIBRARIES}
 	${IBVERBS_LIBRARIES}
 	${RDMACM_LIBRARIES}
-	${OPENSSL_LIBRARIES}
 )
 set_target_properties(vitastor_client PROPERTIES VERSION ${VITASTOR_VERSION} SOVERSION 0)
 configure_file(vitastor.pc.in vitastor.pc @ONLY)
@@ -100,10 +98,9 @@ endif (${WITH_QEMU})
 add_executable(test_cluster_client
 	EXCLUDE_FROM_ALL
 	../test/test_cluster_client.cpp
-	pg_states.cpp osd_ops.cpp cluster_client.cpp cluster_client_list.cpp cluster_client_wb.cpp cluster_client_icache.cpp msgr_op.cpp ../test/mock/messenger.cpp msgr_stop.cpp msgr_encrypt.cpp
+	pg_states.cpp osd_ops.cpp cluster_client.cpp cluster_client_list.cpp cluster_client_wb.cpp msgr_op.cpp ../test/mock/messenger.cpp msgr_stop.cpp
 	etcd_state_client.cpp ../util/timerfd_manager.cpp ../util/addr_util.cpp ../util/str_util.cpp ../util/json_util.cpp ../../json11/json11.cpp
 )
-target_link_libraries(test_cluster_client ${OPENSSL_LIBRARIES})
 target_compile_definitions(test_cluster_client PUBLIC -D__MOCK__)
 target_include_directories(test_cluster_client BEFORE PUBLIC ${CMAKE_SOURCE_DIR}/src/test/mock)
 add_dependencies(build_tests test_cluster_client)

src/client/cluster_client.cpp

@@ -62,7 +62,6 @@ cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd
     st_cli.on_change_node_placement_hook = [this]() { on_change_node_placement_hook(); };
     st_cli.on_load_pgs_hook = [this](bool success) { on_load_pgs_hook(success); };
     st_cli.on_reload_hook = [this]() { st_cli.load_global_config(); };
-    st_cli.on_inode_change_hook = [this](uint64_t inode, bool removed) { on_change_inode_hook(inode, removed); };
 
     st_cli.parse_config(config);
     st_cli.infinite_start = false;
@@ -78,7 +77,6 @@ cluster_client_t::cluster_client_t(ring_loop_t *ringloop, timerfd_manager_t *tfd
 
 cluster_client_t::~cluster_client_t()
 {
-    vault_destroy();
     if (retry_timeout_id >= 0)
     {
         tfd->clear_timer(retry_timeout_id);
@@ -483,8 +481,6 @@ void cluster_client_t::on_load_config_hook(json11::Json::object & etcd_global_co
         self_tree_metrics.clear();
         client_hostname = new_hostname;
     }
-    // vault
-    vault_parse_config();
     msgr.parse_config(config);
     st_cli.parse_config(config);
     st_cli.load_pgs();
@@ -611,9 +607,6 @@ void cluster_client_t::on_change_pool_config_hook()
             pg_counts[pool_item.first] = pool_item.second.real_pg_count;
         }
     }
-    inode_cache.clear();
-    inode_cache_children.clear();
-    vault_keys.clear();
     continue_ops();
 }
 
@@ -680,10 +673,6 @@ bool cluster_client_t::flush()
 {
     if (!ringloop)
     {
-        if (vault_loading)
-        {
-            return false;
-        }
         if (wb->writeback_queue.size())
         {
             wb->start_writebacks(this, 0);
@@ -706,7 +695,7 @@ bool cluster_client_t::flush()
         sync_done = true;
     };
     execute(sync);
-    while (!sync_done || vault_loading)
+    while (!sync_done)
     {
         ringloop->loop();
         if (!sync_done)
@@ -879,8 +868,7 @@ void cluster_client_t::execute_cas(cluster_op_t *op)
     {
         int expected = part->req.hdr.opcode == OSD_OP_DELETE ? 0 : part->req.rw.len;
         op->retval = part->reply.hdr.retval;
-        if (op->retval != expected && op->retval >= 0)
-            op->retval = -EIO;
+        op->retval = op->retval == expected ? 0 : (op->retval >= 0 ? -EIO : op->retval);
         op->retval = op->retval == -EPIPE ? -EINTR : op->retval;
         auto peer_it = msgr.osd_peer_fds.find(op->parts[0].osd_num);
         if (op->retval != 0 || (op->flags & OP_IMMEDIATE_COMMIT))
@@ -909,13 +897,10 @@ void cluster_client_t::execute_cas(cluster_op_t *op)
                         .opcode = OSD_OP_SYNC,
                     },
                 },
-                .callback = [op](osd_op_t *part)
+                .callback = [this, op](osd_op_t *part)
                 {
-                    if (part->reply.hdr.retval != 0)
-                    {
-                        op->retval = part->reply.hdr.retval;
-                        op->retval = op->retval == -EPIPE ? -EINTR : op->retval;
-                    }
+                    op->retval = part->reply.hdr.retval;
+                    op->retval = op->retval == -EPIPE ? -EINTR : op->retval;
                     auto cb = std::move(op->callback);
                     cb(op);
                 },
@@ -969,40 +954,10 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
     {
         op->flags |= OP_IMMEDIATE_COMMIT;
     }
-    bool searched = false;
-    std::shared_ptr<inode_cache_t> icache;
-    if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_WRITE)
-    {
-        if (!searched)
-        {
-            icache = inode_cache_get(op->inode);
-            searched = true;
-        }
-        if (icache && icache->has_parent_loop && op->opcode == OSD_OP_READ)
-        {
-            op->retval = -EINVAL;
-            auto cb = std::move(op->callback);
-            cb(op);
-            return false;
-        }
-        if (icache && icache->op_enc)
-        {
-            // Use shared_ptr aliasing to attach op_enc to the inode cache entry
-            op->enc = std::shared_ptr<osd_op_enc_t>(icache, icache->op_enc);
-        }
-        else
-            op->enc.reset();
-    }
-    else
-        op->enc.reset();
     if ((op->opcode == OSD_OP_WRITE || op->opcode == OSD_OP_DELETE) && !(op->flags & OSD_OP_IGNORE_READONLY))
     {
-        if (!searched)
-        {
-            icache = inode_cache_get(op->inode);
-            searched = true;
-        }
-        if (icache && icache->readonly)
+        auto ino_it = st_cli.inode_config.find(op->inode);
+        if (ino_it != st_cli.inode_config.end() && ino_it->second.readonly)
         {
             op->retval = -EROFS;
             auto cb = std::move(op->callback);
@@ -1013,39 +968,33 @@ bool cluster_client_t::check_rw(cluster_op_t *op)
     op->deoptimise_snapshot = false;
     if (enable_writeback && (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP))
     {
-        if (!searched)
+        auto ino_it = st_cli.inode_config.find(op->inode);
+        if (ino_it != st_cli.inode_config.end())
         {
-            icache = inode_cache_get(op->inode);
-            searched = true;
-        }
-        if (icache)
-        {
-            for (auto & parent: icache->chain)
+            int chain_size = 0;
+            while (ino_it != st_cli.inode_config.end() && ino_it->second.parent_id)
             {
-                if (INODE_POOL(parent) == INODE_POOL(op->inode) && wb->has_inode(parent))
+                // Check for loops - FIXME check it in etcd_state_client
+                if (ino_it->second.parent_id == op->inode ||
+                    chain_size > st_cli.inode_config.size())
+                {
+                    op->retval = -EINVAL;
+                    auto cb = std::move(op->callback);
+                    cb(op);
+                    return false;
+                }
+                if (INODE_POOL(ino_it->second.parent_id) == INODE_POOL(ino_it->first) &&
+                    wb->has_inode(ino_it->second.parent_id))
                 {
                     // Deoptimise reads - we have dirty data for one of the parent layer(s).
                     op->deoptimise_snapshot = true;
                     break;
                 }
+                chain_size++;
+                ino_it = st_cli.inode_config.find(ino_it->second.parent_id);
             }
         }
     }
-    if (icache && icache->err_code)
-    {
-        if (icache->err_code == EPERM)
-        {
-            op->retval = -EPERM;
-            auto cb = std::move(op->callback);
-            cb(op);
-            return false;
-        }
-        else if (icache->err_code == EAGAIN)
-        {
-            key_wait_ops.push_back(op);
-            return false;
-        }
-    }
     return true;
 }
 
@@ -1168,33 +1117,31 @@ resume_2:
         // because if some operations were invalid for the new PG count we'd get errors
         if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
         {
-            uint64_t next_inode = 0;
-            auto icache = inode_cache_get(op->cur_inode);
-            if (icache)
+            // Check parent inode
+            auto ino_it = st_cli.inode_config.find(op->cur_inode);
+            // Skip parents from the same pool
+            int skipped = 0;
+            while (!op->deoptimise_snapshot &&
+                ino_it != st_cli.inode_config.end() && ino_it->second.parent_id &&
+                INODE_POOL(ino_it->second.parent_id) == INODE_POOL(op->cur_inode))
             {
-                if (icache->has_parent_loop)
+                // Check for loops - FIXME check it in etcd_state_client
+                if (ino_it->second.parent_id == op->inode ||
+                    skipped > st_cli.inode_config.size())
                 {
                     op->retval = -EINVAL;
                     erase_op(op);
                     return 1;
                 }
-                if (op->deoptimise_snapshot)
-                {
-                    if (icache->chain.size() > 1)
-                        next_inode = icache->chain[1];
-                }
-                else
-                {
-                    if (icache->other_pool_parent_id)
-                        next_inode = icache->other_pool_parent_id;
-                }
+                skipped++;
+                ino_it = st_cli.inode_config.find(ino_it->second.parent_id);
             }
-            if (next_inode)
+            if (ino_it != st_cli.inode_config.end() &&
+                ino_it->second.parent_id &&
+                ino_it->second.parent_id != op->inode)
             {
                 // Continue reading from the parent inode
-                icache = inode_cache_get(next_inode);
-                op->cur_inode = next_inode;
-                op->enc = (icache && icache->op_enc ? std::shared_ptr<osd_op_enc_t>(icache, icache->op_enc) : nullptr);
+                op->cur_inode = ino_it->second.parent_id;
                 op->parts.clear();
                 op->done_count = 0;
                 goto resume_0;
@@ -1301,18 +1248,14 @@ void cluster_client_t::slice_rw(cluster_op_t *op)
         // Allocate memory for the bitmap
         unsigned object_bitmap_size = ((op->len / pool_cfg.bitmap_granularity + 7) / 8);
         object_bitmap_size = (object_bitmap_size < 8 ? 8 : object_bitmap_size);
-        unsigned bitmap_mem = object_bitmap_size +
-            op->parts.size() * pg_data_size *
-            (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8
-            // read chain info - 1 byte per block
-            + (op->enc ? op->len/pool_cfg.bitmap_granularity : 0));
+        unsigned bitmap_mem = object_bitmap_size + (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8 * pg_data_size) * op->parts.size();
         if (!op->bitmap_buf || op->bitmap_buf_size < bitmap_mem)
         {
             op->bitmap_buf = realloc_or_die(op->bitmap_buf, bitmap_mem);
             op->part_bitmaps = (uint8_t*)op->bitmap_buf + object_bitmap_size;
-            memset((uint8_t*)op->bitmap_buf+op->bitmap_buf_size, 0, bitmap_mem-op->bitmap_buf_size);
             op->bitmap_buf_size = bitmap_mem;
         }
+        memset(op->bitmap_buf, 0, bitmap_mem);
     }
     int iov_idx = 0;
     size_t iov_pos = 0;
@@ -1458,11 +1401,11 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
         if (peer_it != msgr.osd_peer_fds.end())
         {
             int peer_fd = peer_it->second;
-            part->flags |= PART_SENT|PART_VALID;
+            part->flags |= PART_SENT;
             op->inflight_count++;
-            uint32_t pg_data_size = (pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks);
-            uint64_t pg_bitmap_size = pg_data_size * (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8
-                + (op->opcode == OSD_OP_READ && op->enc ? pool_cfg.data_block_size/pool_cfg.bitmap_granularity : 0));
+            uint64_t pg_bitmap_size = (pool_cfg.data_block_size / pool_cfg.bitmap_granularity / 8) * (
+                pool_cfg.scheme == POOL_SCHEME_REPLICATED ? 1 : pool_cfg.pg_size-pool_cfg.parity_chunks
+            );
             uint64_t meta_rev = 0;
             if (op->opcode != OSD_OP_READ_BITMAP && op->opcode != OSD_OP_DELETE && !op->deoptimise_snapshot)
             {
@@ -1481,7 +1424,6 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
                     .inode = op->cur_inode,
                     .offset = part->offset,
                     .len = part->len,
-                    .flags = op->opcode == OSD_OP_READ && op->enc && !op->deoptimise_snapshot ? OSD_OP_RETURN_CHAIN : 0,
                     .meta_revision = meta_rev,
                     .version = op->opcode == OSD_OP_WRITE || op->opcode == OSD_OP_DELETE ? op->version : 0,
                 } },
@@ -1489,7 +1431,6 @@ int cluster_client_t::try_send(cluster_op_t *op, int i, std::function<void(osd_o
                     ? (uint8_t*)op->part_bitmaps + pg_bitmap_size*i : NULL),
                 .bitmap_len = (unsigned)(op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP
                     ? pg_bitmap_size : 0),
-                .enc = op->enc,
                 .callback = cb ? cb : [this, part](osd_op_t *op_part)
                 {
                     handle_op_part(part);
@@ -1673,11 +1614,14 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
             dirty_osds.insert(part->osd_num);
         part->flags |= PART_DONE;
         op->done_count++;
-        if ((op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
-            && op->inode == op->cur_inode)
+        if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
         {
-            // Read only returns the version of the uppermost layer
-            op->version = op->parts.size() == 1 ? part->op.reply.rw.version : 0;
+            copy_part_bitmap(op, part);
+            if (op->inode == op->cur_inode)
+            {
+                // Read only returns the version of the uppermost layer
+                op->version = op->parts.size() == 1 ? part->op.reply.rw.version : 0;
+            }
         }
         else if (op->opcode == OSD_OP_WRITE || op->opcode == OSD_OP_DELETE)
         {
@@ -1685,13 +1629,6 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
         }
         if (op->inflight_count == 0 && !op->retry_after)
         {
-            // Copy part bitmaps only after finishing all part reads
-            if (op->opcode == OSD_OP_READ || op->opcode == OSD_OP_READ_BITMAP || op->opcode == OSD_OP_READ_CHAIN_BITMAP)
-            {
-                for (auto & part: op->parts)
-                    if (part.flags == (PART_SENT|PART_VALID|PART_DONE))
-                        copy_part_bitmap(op, &part);
-            }
             if (op->opcode == OSD_OP_SYNC)
                 continue_sync(op);
             else

src/client/cluster_client.h

@@ -5,7 +5,6 @@
 
 #include "messenger.h"
 #include "etcd_state_client.h"
-#include "../util/robin_hood.h"
 
 #define DEFAULT_CLIENT_MAX_DIRTY_BYTES 32*1024*1024
 #define DEFAULT_CLIENT_MAX_DIRTY_OPS 1024
@@ -72,7 +71,6 @@ protected:
     cluster_op_t *prev = NULL, *next = NULL;
     int prev_wait = 0;
     uint64_t flush_id = 0;
-    std::shared_ptr<osd_op_enc_t> enc;
     friend class cluster_client_t;
     friend class writeback_cache_t;
 };
@@ -82,25 +80,6 @@ struct inode_list_osd_t;
 struct inode_list_pg_t;
 class writeback_cache_t;
 
-struct inode_cache_t
-{
-    std::vector<inode_t> chain;
-    uint8_t *key_data = NULL;
-    osd_op_enc_t *op_enc = NULL;
-    bool readonly = false;
-    bool has_parent_loop = false;
-    inode_t other_pool_parent_id = 0;
-    int err_code = 0;
-
-    ~inode_cache_t();
-};
-
-struct vault_load_key_t
-{
-    int key_state = 0;
-    std::string key;
-};
-
 // FIXME: Split into public and private interfaces
 class __attribute__((visibility("default"))) cluster_client_t
 {
@@ -110,8 +89,8 @@ public:
     timerfd_manager_t *tfd = NULL;
     ring_loop_t *ringloop = NULL;
 
-    // config:
-
+    std::map<pool_id_t, uint64_t> pg_counts;
+    std::map<pool_pg_num_t, osd_num_t> pg_primary;
     // client_max_dirty_* is actually "max unsynced", for the case when immediate_commit is off
     uint64_t client_max_dirty_bytes = 0;
     uint64_t client_max_dirty_ops = 0;
@@ -123,23 +102,12 @@ public:
     uint64_t client_max_writeback_iodepth = 0;
     std::string conf_hostname;
 
-    std::string vault_url;
-    std::string vault_client_cert;
-    std::string vault_client_key;
-    std::string vault_ca;
-    std::string vault_secret_api_path;
-    uint64_t vault_timeout_ms = 0;
-    uint64_t vault_error_timeout_sec = 0;
-    uint64_t vault_refresh_leeway_sec = 0;
-
     int log_level = 0;
     int client_retry_interval = 50; // ms
     int client_eio_retry_interval = 1000; // ms
     bool client_retry_enospc = true;
     int client_wait_up_timeout = 16; // sec (for listings)
 
-    // state:
-
     std::string client_hostname;
     std::map<std::string, int> self_tree_metrics;
     std::map<osd_num_t, int> osd_tree_metrics;
@@ -147,7 +115,6 @@ public:
     int retry_timeout_id = -1;
     int retry_timeout_duration = 0;
     std::vector<cluster_op_t*> offline_ops;
-    std::vector<cluster_op_t*> key_wait_ops;
     cluster_op_t *op_queue_head = NULL, *op_queue_tail = NULL;
     writeback_cache_t *wb = NULL;
     std::set<osd_num_t> dirty_osds;
@@ -156,22 +123,7 @@ public:
     void *scrap_buffer = NULL;
     unsigned scrap_buffer_size = 0;
 
-    // inodes require some extra state for read/write, it's stored here.
-    // moreover, robin_hood access is slightly faster than std::map :)
-    robin_hood::unordered_flat_map<inode_t, std::shared_ptr<inode_cache_t>> inode_cache;
-    std::set<std::pair<inode_t, inode_t>> inode_cache_children;
-
-    http_context_t *vault_http_ctx = NULL;
-    http_co_t *vault_http_cli = NULL;
-    bool vault_loading = false;
-    std::string vault_token;
-    bool vault_auth_error = false;
-    timespec vault_token_expire = {};
-    std::vector<std::string> vault_key_load_queue;
-    std::map<std::string, vault_load_key_t> vault_keys;
-
     bool pgs_loaded = false;
-    std::map<pool_id_t, uint64_t> pg_counts;
     ring_consumer_t consumer;
     std::vector<std::function<void(void)>> on_ready_hooks;
     int list_retry_timeout_id = -1;
@@ -211,13 +163,6 @@ protected:
 #endif
     void continue_ops(int time_passed = 0);
 
-    std::shared_ptr<inode_cache_t> inode_cache_get(inode_t ino);
-    void vault_parse_config();
-    bool vault_check_token();
-    void vault_load_keys();
-    void vault_destroy();
-    void vault_parse_secret(const std::string & key_id, const std::string & err, json11::Json data);
-
 protected:
     bool affects_osd(uint64_t inode, uint64_t offset, uint64_t len, osd_num_t osd);
     bool affects_pg(uint64_t inode, uint64_t offset, uint64_t len, pool_id_t pool_id, pg_num_t pg_num);
@@ -228,7 +173,6 @@ protected:
     void on_change_pg_state_hook(pool_id_t pool_id, pg_num_t pg_num, osd_num_t prev_primary);
     void on_change_osd_state_hook(uint64_t peer_osd);
     void on_change_node_placement_hook();
-    void on_change_inode_hook(uint64_t inode, bool removed);
 
     void execute_internal(cluster_op_t *op);
     void execute_cas(cluster_op_t *op);
@@ -245,7 +189,6 @@ protected:
     void erase_op(cluster_op_t *op);
     void calc_wait(cluster_op_t *op);
     void inc_wait(uint64_t opcode, uint64_t flags, cluster_op_t *next, int inc);
-
     void continue_lists();
     bool continue_listing(inode_list_t *lst);
     bool restart_listing(inode_list_t* lst);