Release 0.8.9

- The tests are now stable and run in a CI system based on Gitea CI - The release includes final bug fixes for EC: - Implement missing EC recovery of allocation bitmap when built with ISA-L - Fix broken snapshot export with EC (allocation bitmap reads were giving incorrect results previously) - Also fixed bugs manifesting under heavy load: - Fix monitor possibly applying incorrect PG history on retries - Fix monitor incorrectly changing PG count when last_clean_pgs contains less PGs than the new number - Allow writes to wait for free space again, but now correctly (previously dropped in 0.8.2) - Fix a rare segfault in client (handle client stop during incoming stream handling in 1 more place) - Make monitor correctly handle etcd connection errors - it could die instead of connecting to another etcd - Fix OSD rarely being unable to report PG states after a PG was taken over by another OSD - Fixed return code for incomplete EC objects (now EIO) and made cluster client retry this error - Made other small changes for tests: timeouts, nice/ionice for etcd, waiting conditions, NBD device checks and so on
Add tests to CI
2023-05-14 01:25:09 +03:00 · 2023-05-14 00:06:09 +03:00 · 2023-05-14 00:06:09 +03:00 · 2023-05-13 23:59:44 +03:00 · 2023-05-13 15:16:28 +03:00 · 2023-05-12 23:18:07 +03:00
52 changed files with 1061 additions and 158 deletions
--- a/.gitea/workflows/buildenv.Dockerfile
+++ b/.gitea/workflows/buildenv.Dockerfile
@@ -0,0 +1,36 @@
+FROM node:16-bullseye
+
+WORKDIR /root
+
+ADD ./docker/vitastor.gpg /etc/apt/trusted.gpg.d
+
+RUN echo 'deb http://deb.debian.org/debian bullseye-backports main' >> /etc/apt/sources.list; \
+    echo 'deb http://vitastor.io/debian bullseye main' >> /etc/apt/sources.list; \
+    echo >> /etc/apt/preferences; \
+    echo 'Package: *' >> /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; \
+    grep '^deb ' /etc/apt/sources.list | perl -pe 's/^deb/deb-src/' >> /etc/apt/sources.list; \
+    echo 'APT::Install-Recommends false;' >> /etc/apt/apt.conf; \
+    echo 'APT::Install-Suggests false;' >> /etc/apt/apt.conf
+
+RUN apt-get update
+RUN apt-get -y install etcd qemu-system-x86 qemu-block-extra qemu-utils fio libasan5 \
+    liburing1 liburing-dev libgoogle-perftools-dev devscripts libjerasure-dev cmake libibverbs-dev libisal-dev
+RUN apt-get -y build-dep fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`
+RUN apt-get -y install jq lp-solve sudo
+RUN apt-get --download-only source fio qemu=`dpkg -s qemu-system-x86|grep ^Version:|awk '{print $2}'`
+
+RUN set -ex; \
+    mkdir qemu-build; \
+    cd qemu-build; \
+    dpkg-source -x /root/qemu*.dsc; \
+    cd qemu*/; \
+    debian/rules configure-qemu || debian/rules b/configure-stamp; \
+    cd b/qemu; \
+    make -j8 config-poison.h || true; \
+    make -j8 qapi/qapi-builtin-types.h
--- a/.gitea/workflows/test.Dockerfile
+++ b/.gitea/workflows/test.Dockerfile
@@ -0,0 +1,16 @@
+FROM git.yourcmc.ru/vitalif/vitastor/buildenv
+
+ADD . /root/vitastor
+
+RUN set -e -x; \
+    mkdir -p /root/fio-build/; \
+    cd /root/fio-build/; \
+    dpkg-source -x /root/fio*.dsc; \
+    cd /root/vitastor; \
+    ln -s /root/fio-build/fio-*/ ./fio; \
+    ln -s /root/qemu-build/qemu-*/ ./qemu; \
+    ls /usr/include/linux/raw.h || cp ./debian/raw.h /usr/include/linux/raw.h; \
+    mkdir build; \
+    cd build; \
+    cmake .. -DWITH_ASAN=yes -DWITH_QEMU=yes; \
+    make -j16
--- a/.gitea/workflows/test.yml
+++ b/.gitea/workflows/test.yml
@@ -0,0 +1,552 @@
+name: Test
+
+on:
+  push:
+    branches:
+    - '*'
+    paths:
+    - '.gitea/**'
+    - 'src/**'
+    - 'mon/**'
+    - 'json11'
+    - 'cpp-btree'
+    - 'tests/**'
+
+env:
+  BUILDENV_IMAGE: git.yourcmc.ru/vitalif/vitastor/buildenv
+  TEST_IMAGE: git.yourcmc.ru/vitalif/vitastor/test
+  OSD_ARGS: '--etcd_quick_timeout 2000'
+
+concurrency:
+  group: ci-${{ github.ref }}
+  cancel-in-progress: true
+
+jobs:
+
+  buildenv:
+    runs-on: ubuntu-latest
+    container: git.yourcmc.ru/vitalif/gitea-ci-dind
+    steps:
+    - uses: actions/checkout@v3
+
+    - name: Build and push
+      run: |
+        set -ex
+        if ! docker manifest inspect $BUILDENV_IMAGE >/dev/null; then
+          docker build -t $BUILDENV_IMAGE -f .gitea/workflows/buildenv.Dockerfile .
+          docker login git.yourcmc.ru -u vitalif -p "${{secrets.TOKEN}}"
+          docker push $BUILDENV_IMAGE
+        fi
+
+  build:
+    runs-on: ubuntu-latest
+    needs: buildenv
+    container: git.yourcmc.ru/vitalif/gitea-ci-dind
+    steps:
+    - uses: actions/checkout@v3
+      with:
+        submodules: true
+
+    - name: Build and push
+      run: |
+        set -ex
+        if ! docker manifest inspect $TEST_IMAGE:$GITHUB_SHA >/dev/null; then
+          docker build -t $TEST_IMAGE:$GITHUB_SHA -f .gitea/workflows/test.Dockerfile .
+          docker login git.yourcmc.ru -u vitalif -p "${{secrets.TOKEN}}"
+          docker push $TEST_IMAGE:$GITHUB_SHA
+        fi
+
+  make_test:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    # leak sanitizer sometimes crashes
+    - run: cd /root/vitastor/build && ASAN_OPTIONS=detect_leaks=0 make -j16 test
+
+  test_add_osd:
+    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_add_osd.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_cas:
+    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_cas.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_change_pg_count:
+    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_change_pg_count.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_change_pg_count_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_change_pg_count.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_change_pg_size:
+    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_change_pg_size.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_create_nomaxid:
+    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_create_nomaxid.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:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 10
+      run: /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_failure_domain:
+    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_failure_domain.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
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 10
+      run: /root/vitastor/tests/test_interrupted_rebalance.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_imm:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 10
+      run: IMMEDIATE_COMMIT=1 /root/vitastor/tests/test_interrupted_rebalance.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_ec:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 10
+      run: SCHEME=ec /root/vitastor/tests/test_interrupted_rebalance.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_ec_imm:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 10
+      run: SCHEME=ec IMMEDIATE_COMMIT=1 /root/vitastor/tests/test_interrupted_rebalance.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_minsize_1:
+    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_minsize_1.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_move_reappear:
+    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_move_reappear.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_rebalance_verify:
+    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_rebalance_verify.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_rebalance_verify_imm:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: IMMEDIATE_COMMIT=1 /root/vitastor/tests/test_rebalance_verify.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_rebalance_verify_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_rebalance_verify.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_rebalance_verify_ec_imm:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: SCHEME=ec IMMEDIATE_COMMIT=1 /root/vitastor/tests/test_rebalance_verify.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_rm:
+    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_rm.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
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: /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_snapshot_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_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_splitbrain:
+    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_splitbrain.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:
+    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.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_xor:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 3
+      run: SCHEME=xor /root/vitastor/tests/test_write.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_no_same:
+    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_no_same.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_pg_size_2:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 10
+      run: PG_SIZE=2 /root/vitastor/tests/test_heal.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_ec:
+    runs-on: ubuntu-latest
+    needs: build
+    container: ${{env.TEST_IMAGE}}:${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: 10
+      run: SCHEME=ec /root/vitastor/tests/test_heal.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
+
--- a/.gitea/workflows/tests-to-yaml.pl
+++ b/.gitea/workflows/tests-to-yaml.pl
@@ -0,0 +1,68 @@
+#!/usr/bin/perl
+
+use strict;
+
+for my $line (<>)
+{
+    if ($line =~ /\.\/(test_[^\.]+)/s)
+    {
+        chomp $line;
+        my $test_name = $1;
+        my $timeout = 3;
+        if ($test_name eq 'test_etcd_fail' || $test_name eq 'test_heal' || $test_name eq 'test_interrupted_rebalance')
+        {
+            $timeout = 10;
+        }
+        while ($line =~ /([^\s=]+)=(\S+)/gs)
+        {
+            if ($1 eq 'SCHEME' && $2 eq 'ec')
+            {
+                $test_name .= '_ec';
+            }
+            elsif ($1 eq 'SCHEME' && $2 eq 'xor')
+            {
+                $test_name .= '_xor';
+            }
+            elsif ($1 eq 'IMMEDIATE_COMMIT')
+            {
+                $test_name .= '_imm';
+            }
+            else
+            {
+                $test_name .= '_'.lc($1).'_'.$2;
+            }
+        }
+        $line =~ s!\./test_!/root/vitastor/tests/test_!;
+        # Gitea CI doesn't support artifacts yet, lol
+        #- name: Upload results
+        #  uses: actions/upload-artifact\@v3
+        #  if: always()
+        #  with:
+        #    name: ${test_name}_result
+        #    path: |
+        #      /root/vitastor/testdata
+        #      !/root/vitastor/testdata/*.bin
+        #    retention-days: 5
+        print <<"EOF"
+  $test_name:
+    runs-on: ubuntu-latest
+    needs: build
+    container: \${{env.TEST_IMAGE}}:\${{github.sha}}
+    steps:
+    - name: Run test
+      id: test
+      timeout-minutes: $timeout
+      run: $line
+    - 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
+
+EOF
+;
+    }
+}
--- a/CMakeLists.txt
+++ b/CMakeLists.txt
@@ -2,6 +2,6 @@ cmake_minimum_required(VERSION 2.8.12)

 project(vitastor)

-set(VERSION "0.8.8")
+set(VERSION "0.8.9")

 add_subdirectory(src)
--- a/csi/Makefile
+++ b/csi/Makefile
@@ -1,4 +1,4 @@
-VERSION ?= v0.8.8
+VERSION ?= v0.8.9

 all: build push

--- a/csi/deploy/004-csi-nodeplugin.yaml
+++ b/csi/deploy/004-csi-nodeplugin.yaml
@@ -49,7 +49,7 @@ spec:
            capabilities:
              add: ["SYS_ADMIN"]
            allowPrivilegeEscalation: true
-          image: vitalif/vitastor-csi:v0.8.8
+          image: vitalif/vitastor-csi:v0.8.9
          args:
            - "--node=$(NODE_ID)"
            - "--endpoint=$(CSI_ENDPOINT)"
--- a/csi/deploy/007-csi-provisioner.yaml
+++ b/csi/deploy/007-csi-provisioner.yaml
@@ -116,7 +116,7 @@ spec:
            privileged: true
            capabilities:
              add: ["SYS_ADMIN"]
-          image: vitalif/vitastor-csi:v0.8.8
+          image: vitalif/vitastor-csi:v0.8.9
          args:
            - "--node=$(NODE_ID)"
            - "--endpoint=$(CSI_ENDPOINT)"
--- a/csi/src/config.go
+++ b/csi/src/config.go
@@ -5,7 +5,7 @@ package vitastor

 const (
    vitastorCSIDriverName    = "csi.vitastor.io"
-    vitastorCSIDriverVersion = "0.8.8"
+    vitastorCSIDriverVersion = "0.8.9"
 )

 // Config struct fills the parameters of request or user input
--- a/debian/changelog
+++ b/debian/changelog
@@ -1,10 +1,10 @@
-vitastor (0.8.8-1) unstable; urgency=medium
+vitastor (0.8.9-1) unstable; urgency=medium

  * Bugfixes

 -- Vitaliy Filippov <vitalif@yourcmc.ru>  Fri, 03 Jun 2022 02:09:44 +0300

-vitastor (0.8.8-1) unstable; urgency=medium
+vitastor (0.8.9-1) unstable; urgency=medium

  * Implement NFS proxy
  * Add documentation
--- a/debian/vitastor.Dockerfile
+++ b/debian/vitastor.Dockerfile
@@ -34,8 +34,8 @@ RUN set -e -x; \
    mkdir -p /root/packages/vitastor-$REL; \
    rm -rf /root/packages/vitastor-$REL/*; \
    cd /root/packages/vitastor-$REL; \
-    cp -r /root/vitastor vitastor-0.8.8; \
-    cd vitastor-0.8.8; \
+    cp -r /root/vitastor vitastor-0.8.9; \
+    cd vitastor-0.8.9; \
    ln -s /root/fio-build/fio-*/ ./fio; \
    FIO=$(head -n1 fio/debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
    ls /usr/include/linux/raw.h || cp ./debian/raw.h /usr/include/linux/raw.h; \
@@ -48,8 +48,8 @@ RUN set -e -x; \
    rm -rf a b; \
    echo "dep:fio=$FIO" > debian/fio_version; \
    cd /root/packages/vitastor-$REL; \
-    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.8.8.orig.tar.xz vitastor-0.8.8; \
-    cd vitastor-0.8.8; \
+    tar --sort=name --mtime='2020-01-01' --owner=0 --group=0 --exclude=debian -cJf vitastor_0.8.9.orig.tar.xz vitastor-0.8.9; \
+    cd vitastor-0.8.9; \
    V=$(head -n1 debian/changelog | perl -pe 's/^.*\((.*?)\).*$/$1/'); \
    DEBFULLNAME="Vitaliy Filippov <vitalif@yourcmc.ru>" dch -D $REL -v "$V""$REL" "Rebuild for $REL"; \
    DEB_BUILD_OPTIONS=nocheck dpkg-buildpackage --jobs=auto -sa; \
--- a/docker/vitastor.gpg
+++ b/docker/vitastor.gpg
--- a/mon/PGUtil.js
+++ b/mon/PGUtil.js
@@ -43,16 +43,16 @@ function finish_pg_history(merged_history)
    merged_history.all_peers = Object.values(merged_history.all_peers);
 }

-function scale_pg_count(prev_pgs, prev_pg_history, new_pg_history, new_pg_count)
+function scale_pg_count(prev_pgs, real_prev_pgs, prev_pg_history, new_pg_history, new_pg_count)
 {
-    const old_pg_count = prev_pgs.length;
+    const old_pg_count = real_prev_pgs.length;
    // Add all possibly intersecting PGs to the history of new PGs
    if (!(new_pg_count % old_pg_count))
    {
        // New PG count is a multiple of old PG count
        for (let i = 0; i < new_pg_count; i++)
        {
-            add_pg_history(new_pg_history, i, prev_pgs, prev_pg_history, i % old_pg_count);
+            add_pg_history(new_pg_history, i, real_prev_pgs, prev_pg_history, i % old_pg_count);
            finish_pg_history(new_pg_history[i]);
        }
    }
@@ -64,7 +64,7 @@ function scale_pg_count(prev_pgs, prev_pg_history, new_pg_history, new_pg_count)
        {
            for (let j = 0; j < mul; j++)
            {
-                add_pg_history(new_pg_history, i, prev_pgs, prev_pg_history, i+j*new_pg_count);
+                add_pg_history(new_pg_history, i, real_prev_pgs, prev_pg_history, i+j*new_pg_count);
            }
            finish_pg_history(new_pg_history[i]);
        }
@@ -76,7 +76,7 @@ function scale_pg_count(prev_pgs, prev_pg_history, new_pg_history, new_pg_count)
        let merged_history = {};
        for (let i = 0; i < old_pg_count; i++)
        {
-            add_pg_history(merged_history, 1, prev_pgs, prev_pg_history, i);
+            add_pg_history(merged_history, 1, real_prev_pgs, prev_pg_history, i);
        }
        finish_pg_history(merged_history[1]);
        for (let i = 0; i < new_pg_count; i++)
@@ -90,15 +90,15 @@ function scale_pg_count(prev_pgs, prev_pg_history, new_pg_history, new_pg_count)
        new_pg_history[i] = null;
    }
    // Just for the lp_solve optimizer - pick a "previous" PG for each "new" one
-    if (old_pg_count < new_pg_count)
+    if (prev_pgs.length < new_pg_count)
    {
-        for (let i = old_pg_count; i < new_pg_count; i++)
+        for (let i = prev_pgs.length; i < new_pg_count; i++)
        {
-            prev_pgs[i] = prev_pgs[i % old_pg_count];
+            prev_pgs[i] = prev_pgs[i % prev_pgs.length];
        }
    }
-    else if (old_pg_count > new_pg_count)
+    else if (prev_pgs.length > new_pg_count)
    {
-        prev_pgs.splice(new_pg_count, old_pg_count-new_pg_count);
+        prev_pgs.splice(new_pg_count, prev_pgs.length-new_pg_count);
    }
 }
--- a/mon/mon.js
+++ b/mon/mon.js
@@ -956,7 +956,7 @@ class Mon
        return alive_set[this.rng() % alive_set.length];
    }

-    save_new_pgs_txn(request, pool_id, up_osds, osd_tree, prev_pgs, new_pgs, pg_history)
+    save_new_pgs_txn(save_to, request, pool_id, up_osds, osd_tree, prev_pgs, new_pgs, pg_history)
    {
        const aff_osds = this.get_affinity_osds(this.state.config.pools[pool_id], up_osds, osd_tree);
        const pg_items = {};
@@ -1009,14 +1009,14 @@ class Mon
                });
            }
        }
-        this.state.config.pgs.items = this.state.config.pgs.items || {};
+        save_to.items = save_to.items || {};
        if (!new_pgs.length)
        {
-            delete this.state.config.pgs.items[pool_id];
+            delete save_to.items[pool_id];
        }
        else
        {
-            this.state.config.pgs.items[pool_id] = pg_items;
+            save_to.items[pool_id] = pg_items;
        }
    }

@@ -1160,6 +1160,7 @@ class Mon
        if (this.state.config.pgs.hash != tree_hash)
        {
            // Something has changed
+            const new_config_pgs = JSON.parse(JSON.stringify(this.state.config.pgs));
            const etcd_request = { compare: [], success: [] };
            for (const pool_id in (this.state.config.pgs||{}).items||{})
            {
@@ -1180,7 +1181,7 @@ class Mon
                    etcd_request.success.push({ requestDeleteRange: {
                        key: b64(this.etcd_prefix+'/pool/stats/'+pool_id),
                    } });
-                    this.save_new_pgs_txn(etcd_request, pool_id, up_osds, osd_tree, prev_pgs, [], []);
+                    this.save_new_pgs_txn(new_config_pgs, etcd_request, pool_id, up_osds, osd_tree, prev_pgs, [], []);
                }
            }
            for (const pool_id in this.state.config.pools)
@@ -1234,7 +1235,7 @@ class Mon
                            return;
                        }
                        const new_pg_history = [];
-                        PGUtil.scale_pg_count(prev_pgs, pg_history, new_pg_history, pool_cfg.pg_count);
+                        PGUtil.scale_pg_count(prev_pgs, real_prev_pgs, pg_history, new_pg_history, pool_cfg.pg_count);
                        pg_history = new_pg_history;
                    }
                    for (const pg of prev_pgs)
@@ -1287,14 +1288,15 @@ class Mon
                    key: b64(this.etcd_prefix+'/pool/stats/'+pool_id),
                    value: b64(JSON.stringify(this.state.pool.stats[pool_id])),
                } });
-                this.save_new_pgs_txn(etcd_request, pool_id, up_osds, osd_tree, real_prev_pgs, optimize_result.int_pgs, pg_history);
+                this.save_new_pgs_txn(new_config_pgs, etcd_request, pool_id, up_osds, osd_tree, real_prev_pgs, optimize_result.int_pgs, pg_history);
            }
-            this.state.config.pgs.hash = tree_hash;
-            await this.save_pg_config(etcd_request);
+            new_config_pgs.hash = tree_hash;
+            await this.save_pg_config(new_config_pgs, etcd_request);
        }
        else
        {
            // Nothing changed, but we still want to recheck the distribution of primaries
+            let new_config_pgs;
            let changed = false;
            for (const pool_id in this.state.config.pools)
            {
@@ -1314,31 +1316,35 @@ class Mon
                        const new_primary = this.pick_primary(pool_id, pg_cfg.osd_set, up_osds, aff_osds);
                        if (pg_cfg.primary != new_primary)
                        {
+                            if (!new_config_pgs)
+                            {
+                                new_config_pgs = JSON.parse(JSON.stringify(this.state.config.pgs));
+                            }
                            console.log(
                                `Moving pool ${pool_id} (${pool_cfg.name || 'unnamed'}) PG ${pg_num}`+
                                ` primary OSD from ${pg_cfg.primary} to ${new_primary}`
                            );
                            changed = true;
-                            pg_cfg.primary = new_primary;
+                            new_config_pgs.items[pool_id][pg_num].primary = new_primary;
                        }
                    }
                }
            }
            if (changed)
            {
-                await this.save_pg_config();
+                await this.save_pg_config(new_config_pgs);
            }
        }
    }

-    async save_pg_config(etcd_request = { compare: [], success: [] })
+    async save_pg_config(new_config_pgs, etcd_request = { compare: [], success: [] })
    {
        etcd_request.compare.push(
            { key: b64(this.etcd_prefix+'/mon/master'), target: 'LEASE', lease: ''+this.etcd_lease_id },
            { key: b64(this.etcd_prefix+'/config/pgs'), target: 'MOD', mod_revision: ''+this.etcd_watch_revision, result: 'LESS' },
        );
        etcd_request.success.push(
-            { requestPut: { key: b64(this.etcd_prefix+'/config/pgs'), value: b64(JSON.stringify(this.state.config.pgs)) } },
+            { requestPut: { key: b64(this.etcd_prefix+'/config/pgs'), value: b64(JSON.stringify(new_config_pgs)) } },
        );
        const res = await this.etcd_call('/kv/txn', etcd_request, this.config.etcd_mon_timeout, 0);
        if (!res.succeeded)
@@ -1817,6 +1823,7 @@ function POST(url, body, timeout)
            clearTimeout(timer_id);
            let res_body = '';
            res.setEncoding('utf8');
+            res.on('error', (error) => ok({ error }));
            res.on('data', chunk => { res_body += chunk; });
            res.on('end', () =>
            {
@@ -1836,6 +1843,8 @@ function POST(url, body, timeout)
                }
            });
        });
+        req.on('error', (error) => ok({ error }));
+        req.on('close', () => ok({ error: new Error('Connection closed prematurely') }));
        req.write(body_text);
        req.end();
    });
--- a/patches/cinder-vitastor.py
+++ b/patches/cinder-vitastor.py
@@ -50,7 +50,7 @@ from cinder.volume import configuration
 from cinder.volume import driver
 from cinder.volume import volume_utils

-VERSION = '0.8.8'
+VERSION = '0.8.9'

 LOG = logging.getLogger(__name__)

--- a/rpm/build-tarball.sh
+++ b/rpm/build-tarball.sh
@@ -24,4 +24,4 @@ rm fio
 mv fio-copy fio
 FIO=`rpm -qi fio | perl -e 'while(<>) { /^Epoch[\s:]+(\S+)/ && print "$1:"; /^Version[\s:]+(\S+)/ && print $1; /^Release[\s:]+(\S+)/ && print "-$1"; }'`
 perl -i -pe 's/(Requires:\s*fio)([^\n]+)?/$1 = '$FIO'/' $VITASTOR/rpm/vitastor-el$EL.spec
-tar --transform 's#^#vitastor-0.8.8/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.8.8$(rpm --eval '%dist').tar.gz *
+tar --transform 's#^#vitastor-0.8.9/#' --exclude 'rpm/*.rpm' -czf $VITASTOR/../vitastor-0.8.9$(rpm --eval '%dist').tar.gz *
--- a/rpm/vitastor-el7.Dockerfile
+++ b/rpm/vitastor-el7.Dockerfile
@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
    cd /root/vitastor/rpm; \
    sh build-tarball.sh; \
-    cp /root/vitastor-0.8.8.el7.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.8.9.el7.tar.gz ~/rpmbuild/SOURCES; \
    cp vitastor-el7.spec ~/rpmbuild/SPECS/vitastor.spec; \
    cd ~/rpmbuild/SPECS/; \
    rpmbuild -ba vitastor.spec; \
--- a/rpm/vitastor-el7.spec
+++ b/rpm/vitastor-el7.spec
@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.8.8
+Version:        0.8.9
 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-0.8.8.el7.tar.gz
+Source0:        vitastor-0.8.9.el7.tar.gz

 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
--- a/rpm/vitastor-el8.Dockerfile
+++ b/rpm/vitastor-el8.Dockerfile
@@ -35,7 +35,7 @@ ADD . /root/vitastor
 RUN set -e; \
    cd /root/vitastor/rpm; \
    sh build-tarball.sh; \
-    cp /root/vitastor-0.8.8.el8.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.8.9.el8.tar.gz ~/rpmbuild/SOURCES; \
    cp vitastor-el8.spec ~/rpmbuild/SPECS/vitastor.spec; \
    cd ~/rpmbuild/SPECS/; \
    rpmbuild -ba vitastor.spec; \
--- a/rpm/vitastor-el8.spec
+++ b/rpm/vitastor-el8.spec
@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.8.8
+Version:        0.8.9
 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-0.8.8.el8.tar.gz
+Source0:        vitastor-0.8.9.el8.tar.gz

 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
--- a/rpm/vitastor-el9.Dockerfile
+++ b/rpm/vitastor-el9.Dockerfile
@@ -18,7 +18,7 @@ ADD . /root/vitastor
 RUN set -e; \
    cd /root/vitastor/rpm; \
    sh build-tarball.sh; \
-    cp /root/vitastor-0.8.8.el9.tar.gz ~/rpmbuild/SOURCES; \
+    cp /root/vitastor-0.8.9.el9.tar.gz ~/rpmbuild/SOURCES; \
    cp vitastor-el9.spec ~/rpmbuild/SPECS/vitastor.spec; \
    cd ~/rpmbuild/SPECS/; \
    rpmbuild -ba vitastor.spec; \
--- a/rpm/vitastor-el9.spec
+++ b/rpm/vitastor-el9.spec
@@ -1,11 +1,11 @@
 Name:           vitastor
-Version:        0.8.8
+Version:        0.8.9
 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-0.8.8.el9.tar.gz
+Source0:        vitastor-0.8.9.el9.tar.gz

 BuildRequires:  liburing-devel >= 0.6
 BuildRequires:  gperftools-devel
--- a/src/CMakeLists.txt
+++ b/src/CMakeLists.txt
@@ -16,7 +16,7 @@ if("${CMAKE_INSTALL_PREFIX}" MATCHES "^/usr/local/?$")
 	set(CMAKE_INSTALL_RPATH "${CMAKE_INSTALL_PREFIX}/${CMAKE_INSTALL_LIBDIR}")
 endif()

-add_definitions(-DVERSION="0.8.8")
+add_definitions(-DVERSION="0.8.9")
 add_definitions(-Wall -Wno-sign-compare -Wno-comment -Wno-parentheses -Wno-pointer-arith -fdiagnostics-color=always -I ${CMAKE_SOURCE_DIR}/src)
 if (${WITH_ASAN})
 	add_definitions(-fsanitize=address -fno-omit-frame-pointer)
--- a/src/blockstore_impl.cpp
+++ b/src/blockstore_impl.cpp
@@ -307,6 +307,18 @@ void blockstore_impl_t::check_wait(blockstore_op_t *op)
        }
        PRIV(op)->wait_for = 0;
    }
+    else if (PRIV(op)->wait_for == WAIT_FREE)
+    {
+        if (!data_alloc->get_free_count() && big_to_flush > 0)
+        {
+#ifdef BLOCKSTORE_DEBUG
+            printf("Still waiting for free space on the data device\n");
+#endif
+            return;
+        }
+        flusher->release_trim();
+        PRIV(op)->wait_for = 0;
+    }
    else
    {
        throw std::runtime_error("BUG: op->wait_for value is unexpected");
--- a/src/blockstore_impl.h
+++ b/src/blockstore_impl.h
@@ -160,6 +160,8 @@ struct __attribute__((__packed__)) dirty_entry
 #define WAIT_JOURNAL 3
 // Suspend operation until the next journal sector buffer is free
 #define WAIT_JOURNAL_BUFFER 4
+// Suspend operation until there is some free space on the data device
+#define WAIT_FREE 5

 struct fulfill_read_t
 {
@@ -263,6 +265,7 @@ class blockstore_impl_t

    struct journal_t journal;
    journal_flusher_t *flusher;
+    int big_to_flush = 0;
    int write_iodepth = 0;

    bool live = false, queue_stall = false;
--- a/src/blockstore_rollback.cpp
+++ b/src/blockstore_rollback.cpp
@@ -201,6 +201,11 @@ void blockstore_impl_t::erase_dirty(blockstore_dirty_db_t::iterator dirty_start,
    }
    while (1)
    {
+        if ((IS_BIG_WRITE(dirty_it->second.state) || IS_DELETE(dirty_it->second.state)) &&
+            IS_STABLE(dirty_it->second.state))
+        {
+            big_to_flush--;
+        }
        if (IS_BIG_WRITE(dirty_it->second.state) && dirty_it->second.location != clean_loc &&
            dirty_it->second.location != UINT64_MAX)
        {
--- a/src/blockstore_stable.cpp
+++ b/src/blockstore_stable.cpp
@@ -446,6 +446,7 @@ void blockstore_impl_t::mark_stable(const obj_ver_id & v, bool forget_dirty)
                    {
                        inode_space_stats[dirty_it->first.oid.inode] += dsk.data_block_size;
                    }
+                    big_to_flush++;
                }
                else if (IS_DELETE(dirty_it->second.state))
                {
@@ -454,6 +455,7 @@ void blockstore_impl_t::mark_stable(const obj_ver_id & v, bool forget_dirty)
                        sp -= dsk.data_block_size;
                    else
                        inode_space_stats.erase(dirty_it->first.oid.inode);
+                    big_to_flush++;
                }
            }
            if (forget_dirty && (IS_BIG_WRITE(dirty_it->second.state) ||
--- a/src/blockstore_write.cpp
+++ b/src/blockstore_write.cpp
@@ -271,6 +271,13 @@ int blockstore_impl_t::dequeue_write(blockstore_op_t *op)
        if (loc == UINT64_MAX)
        {
            // no space
+            if (big_to_flush > 0)
+            {
+                // hope that some space will be available after flush
+                flusher->request_trim();
+                PRIV(op)->wait_for = WAIT_FREE;
+                return 0;
+            }
            cancel_all_writes(op, dirty_it, -ENOSPC);
            return 2;
        }
--- a/src/cluster_client.cpp
+++ b/src/cluster_client.cpp
@@ -743,15 +743,16 @@ resume_3:
        erase_op(op);
        return 1;
    }
-    else if (op->retval != 0 && op->retval != -EPIPE)
+    else if (op->retval != 0 && op->retval != -EPIPE && op->retval != -EIO && op->retval != -ENOSPC)
    {
-        // Fatal error (not -EPIPE)
+        // Fatal error (neither -EPIPE, -EIO nor -ENOSPC)
+        // FIXME: Add a parameter to allow to not wait for EIOs (incomplete or corrupted objects) to heal
        erase_op(op);
        return 1;
    }
    else
    {
-        // -EPIPE - clear the error and retry
+        // Non-fatal error - clear the error and retry
        op->retval = 0;
        if (op->needs_reslice)
        {
@@ -1048,7 +1049,7 @@ resume_1:
                uw_it->second.state = CACHE_DIRTY;
            }
        }
-        if (op->retval == -EPIPE)
+        if (op->retval == -EPIPE || op->retval == -EIO || op->retval == -ENOSPC)
        {
            // Retry later
            op->parts.clear();
@@ -1119,13 +1120,13 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
    {
        // Operation failed, retry
        part->flags |= PART_ERROR;
-        if (!op->retval || op->retval == -EPIPE)
+        if (!op->retval || op->retval == -EPIPE || part->op.reply.hdr.retval == -EIO)
        {
-            // Don't overwrite other errors with -EPIPE
+            // Error priority: EIO > ENOSPC > EPIPE
            op->retval = part->op.reply.hdr.retval;
        }
        int stop_fd = -1;
-        if (op->retval != -EINTR && op->retval != -EIO)
+        if (op->retval != -EINTR && op->retval != -EIO && op->retval != -ENOSPC)
        {
            stop_fd = part->op.peer_fd;
            fprintf(
@@ -1133,21 +1134,25 @@ void cluster_client_t::handle_op_part(cluster_op_part_t *part)
                osd_op_names[part->op.req.hdr.opcode], part->osd_num, part->op.reply.hdr.retval, expected
            );
        }
+        else
+        {
+            fprintf(
+                stderr, "%s operation failed on OSD %lu: retval=%ld (expected %d)\n",
+                osd_op_names[part->op.req.hdr.opcode], part->osd_num, part->op.reply.hdr.retval, expected
+            );
+        }
        // All next things like timer, continue_sync/rw and stop_client may affect the operation again
        // So do all these things after modifying operation state, otherwise we may hit reenterability bugs
        // FIXME postpone such things to set_immediate here to avoid bugs
-        if (part->op.reply.hdr.retval == -EPIPE)
+        // Mark op->up_wait = true to retry operation after a short pause (not immediately)
+        op->up_wait = true;
+        if (!retry_timeout_id)
        {
-            // Mark op->up_wait = true before stopping the client
-            op->up_wait = true;
-            if (!retry_timeout_id)
+            retry_timeout_id = tfd->set_timer(up_wait_retry_interval, false, [this](int)
            {
-                retry_timeout_id = tfd->set_timer(up_wait_retry_interval, false, [this](int)
-                {
-                    retry_timeout_id = 0;
-                    continue_ops(true);
-                });
-            }
+                retry_timeout_id = 0;
+                continue_ops(true);
+            });
        }
        if (op->inflight_count == 0)
        {
--- a/src/epoll_manager.cpp
+++ b/src/epoll_manager.cpp
@@ -54,6 +54,13 @@ void epoll_manager_t::set_fd_handler(int fd, bool wr, std::function<void(int, in
        ev.events = (wr ? EPOLLOUT : 0) | EPOLLIN | EPOLLRDHUP | EPOLLET;
        if (epoll_ctl(epoll_fd, exists ? EPOLL_CTL_MOD : EPOLL_CTL_ADD, fd, &ev) < 0)
        {
+            if (errno == ENOENT)
+            {
+                // The FD is probably already closed
+                epoll_ctl(epoll_fd, EPOLL_CTL_DEL, fd, NULL);
+                epoll_handlers.erase(fd);
+                return;
+            }
            throw std::runtime_error(std::string("epoll_ctl: ") + strerror(errno));
        }
        epoll_handlers[fd] = handler;
--- a/src/msgr_receive.cpp
+++ b/src/msgr_receive.cpp
@@ -103,7 +103,10 @@ bool osd_messenger_t::handle_read(int result, osd_client_t *cl)
            cl->recv_list.eat(result);
            if (cl->recv_list.done >= cl->recv_list.count)
            {
-                handle_finished_read(cl);
+                if (!handle_finished_read(cl))
+                {
+                    goto fin;
+                }
            }
        }
        if (result >= cl->read_iov.iov_len)
--- a/src/osd_cluster.cpp
+++ b/src/osd_cluster.cpp
@@ -812,11 +812,21 @@ void osd_t::report_pg_states()
                    pg_it->second.cur_state != 0)
                {
                    pg_state_exists = true;
+                    if (pg.state == PG_OFFLINE && pg_it->second.cur_primary != this->osd_num)
+                    {
+                        // Nothing to check or report, PG is already taken over by another OSD
+                        continue;
+                    }
                }
            }
        }
        if (!pg_state_exists)
        {
+            if (pg.state == PG_OFFLINE)
+            {
+                // Nothing to check or report, PG is already stopped
+                continue;
+            }
            // Check that the PG key does not exist
            // Failed check indicates an unsuccessful PG lock attempt in this case
            checks.push_back(json11::Json::object {
@@ -901,6 +911,15 @@ void osd_t::report_pg_states()
    {
        etcd_reporting_pg_state = false;
        if (!data["succeeded"].bool_value())
+        {
+            std::string rpgnames = "";
+            for (auto pp: reporting_pgs)
+            {
+                rpgnames += (rpgnames.size() ? ", " : "")+std::to_string(pp.pool_pg_num.pool_id)+"/"+std::to_string(pp.pool_pg_num.pg_num);
+            }
+            printf("Error reporting PG %s states, will repeat the attempt: %s\n", rpgnames.c_str(), err.c_str());
+        }
+        if (!data["succeeded"].bool_value())
        {
            // One of PG state updates failed, put dirty flags back
            for (auto pp: reporting_pgs)
--- a/src/osd_ops.h
+++ b/src/osd_ops.h
@@ -191,7 +191,7 @@ struct __attribute__((__packed__)) osd_op_rw_t
    uint64_t inode;
    // offset
    uint64_t offset;
-    // length
+    // length. 0 means to read all bitmaps of the specified range, but no data.
    uint32_t len;
    // flags (for future)
    uint32_t flags;
--- a/src/osd_primary.cpp
+++ b/src/osd_primary.cpp
@@ -186,10 +186,22 @@ void osd_t::continue_primary_read(osd_op_t *cur_op)
    cur_op->reply.rw.bitmap_len = 0;
    {
        auto & pg = pgs.at({ .pool_id = INODE_POOL(op_data->oid.inode), .pg_num = op_data->pg_num });
-        for (int role = 0; role < op_data->pg_data_size; role++)
+        if (cur_op->req.rw.len == 0)
        {
-            op_data->stripes[role].read_start = op_data->stripes[role].req_start;
-            op_data->stripes[role].read_end = op_data->stripes[role].req_end;
+            // len=0 => bitmap read
+            for (int role = 0; role < op_data->pg_data_size; role++)
+            {
+                op_data->stripes[role].read_start = 0;
+                op_data->stripes[role].read_end = UINT32_MAX;
+            }
+        }
+        else
+        {
+            for (int role = 0; role < op_data->pg_data_size; role++)
+            {
+                op_data->stripes[role].read_start = op_data->stripes[role].req_start;
+                op_data->stripes[role].read_end = op_data->stripes[role].req_end;
+            }
        }
        // Determine version
        auto vo_it = pg.ver_override.find(op_data->oid);
--- a/src/osd_primary_subops.cpp
+++ b/src/osd_primary_subops.cpp
@@ -151,6 +151,13 @@ int osd_t::submit_primary_subop_batch(int submit_type, inode_t inode, uint64_t o
        {
            int stripe_num = rep ? 0 : role;
            osd_op_t *subop = op_data->subops + i;
+            uint32_t subop_len = wr
+                ? stripes[stripe_num].write_end - stripes[stripe_num].write_start
+                : stripes[stripe_num].read_end - stripes[stripe_num].read_start;
+            if (!wr && stripes[stripe_num].read_end == UINT32_MAX)
+            {
+                subop_len = 0;
+            }
            if (role_osd_num == this->osd_num)
            {
                clock_gettime(CLOCK_REALTIME, &subop->tv_begin);
@@ -169,7 +176,7 @@ int osd_t::submit_primary_subop_batch(int submit_type, inode_t inode, uint64_t o
                    },
                    .version = op_version,
                    .offset = wr ? stripes[stripe_num].write_start : stripes[stripe_num].read_start,
-                    .len = wr ? stripes[stripe_num].write_end - stripes[stripe_num].write_start : stripes[stripe_num].read_end - stripes[stripe_num].read_start,
+                    .len = subop_len,
                    .buf = wr ? stripes[stripe_num].write_buf : stripes[stripe_num].read_buf,
                    .bitmap = stripes[stripe_num].bmp_buf,
                });
@@ -199,7 +206,7 @@ int osd_t::submit_primary_subop_batch(int submit_type, inode_t inode, uint64_t o
                    },
                    .version = op_version,
                    .offset = wr ? stripes[stripe_num].write_start : stripes[stripe_num].read_start,
-                    .len = wr ? stripes[stripe_num].write_end - stripes[stripe_num].write_start : stripes[stripe_num].read_end - stripes[stripe_num].read_start,
+                    .len = subop_len,
                    .attr_len = wr ? clean_entry_bitmap_size : 0,
                };
 #ifdef OSD_DEBUG
@@ -218,9 +225,9 @@ int osd_t::submit_primary_subop_batch(int submit_type, inode_t inode, uint64_t o
                }
                else
                {
-                    if (stripes[stripe_num].read_end > stripes[stripe_num].read_start)
+                    if (subop_len > 0)
                    {
-                        subop->iov.push_back(stripes[stripe_num].read_buf, stripes[stripe_num].read_end - stripes[stripe_num].read_start);
+                        subop->iov.push_back(stripes[stripe_num].read_buf, subop_len);
                    }
                }
                subop->callback = [cur_op, this](osd_op_t *subop)
--- a/src/osd_primary_write.cpp
+++ b/src/osd_primary_write.cpp
@@ -81,7 +81,7 @@ resume_1:
        if (!cur_op->rmw_buf)
        {
            // Refuse partial overwrite of an incomplete object
-            cur_op->reply.hdr.retval = -EINVAL;
+            cur_op->reply.hdr.retval = -EIO;
            goto continue_others;
        }
    }
@@ -197,6 +197,11 @@ resume_5:
    }
    if (op_data->errors > 0)
    {
+        // FIXME: Handle ENOSPC. If one of the subops fail with ENOSPC here,
+        // next writes to the same object will also fail because they'll try
+        // to overwrite the same version number which will result in EEXIST.
+        // To fix it, we should mark the object as degraded for replicas,
+        // and rollback successful part updates in case of EC.
        pg_cancel_write_queue(pg, cur_op, op_data->oid, op_data->errcode);
        return;
    }
--- a/src/osd_rmw.cpp
+++ b/src/osd_rmw.cpp
@@ -28,7 +28,9 @@ static inline void extend_read(uint32_t start, uint32_t end, osd_rmw_stripe_t &
    }
    else
    {
-        if (stripe.read_end < end)
+        if (stripe.read_end < end && end != UINT32_MAX ||
+            // UINT32_MAX means that stripe only needs bitmap, end != 0 => needs also data
+            stripe.read_end == UINT32_MAX && end != 0)
            stripe.read_end = end;
        if (stripe.read_start > start)
            stripe.read_start = start;
@@ -105,24 +107,30 @@ void reconstruct_stripes_xor(osd_rmw_stripe_t *stripes, int pg_size, uint32_t bi
                    }
                    else if (prev >= 0)
                    {
-                        assert(stripes[role].read_start >= stripes[prev].read_start &&
-                            stripes[role].read_start >= stripes[other].read_start);
-                        memxor(
-                            (uint8_t*)stripes[prev].read_buf + (stripes[role].read_start - stripes[prev].read_start),
-                            (uint8_t*)stripes[other].read_buf + (stripes[role].read_start - stripes[other].read_start),
-                            stripes[role].read_buf, stripes[role].read_end - stripes[role].read_start
-                        );
+                        if (stripes[role].read_end != UINT32_MAX)
+                        {
+                            assert(stripes[role].read_start >= stripes[prev].read_start &&
+                                stripes[role].read_start >= stripes[other].read_start);
+                            memxor(
+                                (uint8_t*)stripes[prev].read_buf + (stripes[role].read_start - stripes[prev].read_start),
+                                (uint8_t*)stripes[other].read_buf + (stripes[role].read_start - stripes[other].read_start),
+                                stripes[role].read_buf, stripes[role].read_end - stripes[role].read_start
+                            );
+                        }
                        memxor(stripes[prev].bmp_buf, stripes[other].bmp_buf, stripes[role].bmp_buf, bitmap_size);
                        prev = -1;
                    }
                    else
                    {
-                        assert(stripes[role].read_start >= stripes[other].read_start);
-                        memxor(
-                            stripes[role].read_buf,
-                            (uint8_t*)stripes[other].read_buf + (stripes[role].read_start - stripes[other].read_start),
-                            stripes[role].read_buf, stripes[role].read_end - stripes[role].read_start
-                        );
+                        if (stripes[role].read_end != UINT32_MAX)
+                        {
+                            assert(stripes[role].read_start >= stripes[other].read_start);
+                            memxor(
+                                stripes[role].read_buf,
+                                (uint8_t*)stripes[other].read_buf + (stripes[role].read_start - stripes[other].read_start),
+                                stripes[role].read_buf, stripes[role].read_end - stripes[role].read_start
+                            );
+                        }
                        memxor(stripes[role].bmp_buf, stripes[other].bmp_buf, stripes[role].bmp_buf, bitmap_size);
                    }
                }
@@ -356,20 +364,23 @@ void reconstruct_stripes_ec(osd_rmw_stripe_t *stripes, int pg_size, int pg_minsi
    uint64_t read_start = 0, read_end = 0;
    auto recover_seq = [&]()
    {
-        int orig = 0;
-        for (int other = 0; other < pg_size; other++)
+        if (read_end != UINT32_MAX)
        {
-            if (stripes[other].read_end != 0 && !stripes[other].missing)
+            int orig = 0;
+            for (int other = 0; other < pg_size && orig < pg_minsize; other++)
            {
-                assert(stripes[other].read_start <= read_start);
-                assert(stripes[other].read_end >= read_end);
-                data_ptrs[orig++] = (uint8_t*)stripes[other].read_buf + (read_start - stripes[other].read_start);
+                if (stripes[other].read_end != 0 && !stripes[other].missing)
+                {
+                    assert(stripes[other].read_start <= read_start);
+                    assert(stripes[other].read_end >= read_end);
+                    data_ptrs[orig++] = (uint8_t*)stripes[other].read_buf + (read_start - stripes[other].read_start);
+                }
            }
+            ec_encode_data(
+                read_end-read_start, pg_minsize, wanted, dectable + wanted_base*32*pg_minsize,
+                data_ptrs, data_ptrs + pg_minsize
+            );
        }
-        ec_encode_data(
-            read_end-read_start, pg_minsize, wanted, dectable + wanted_base*32*pg_minsize,
-            data_ptrs, data_ptrs + pg_minsize
-        );
        wanted_base += wanted;
        wanted = 0;
    };
@@ -391,6 +402,32 @@ void reconstruct_stripes_ec(osd_rmw_stripe_t *stripes, int pg_size, int pg_minsi
    {
        recover_seq();
    }
+    // Recover bitmaps
+    if (bitmap_size > 0)
+    {
+        for (int role = 0; role < pg_minsize; role++)
+        {
+            if (stripes[role].read_end != 0 && stripes[role].missing)
+            {
+                data_ptrs[pg_minsize + (wanted++)] = (uint8_t*)stripes[role].bmp_buf;
+            }
+        }
+        if (wanted > 0)
+        {
+            int orig = 0;
+            for (int other = 0; other < pg_size && orig < pg_minsize; other++)
+            {
+                if (stripes[other].read_end != 0 && !stripes[other].missing)
+                {
+                    data_ptrs[orig++] = (uint8_t*)stripes[other].bmp_buf;
+                }
+            }
+            ec_encode_data(
+                bitmap_size, pg_minsize, wanted, dectable,
+                data_ptrs, data_ptrs + pg_minsize
+            );
+        }
+    }
 }
 #else
 void reconstruct_stripes_ec(osd_rmw_stripe_t *stripes, int pg_size, int pg_minsize, uint32_t bitmap_size)
@@ -412,7 +449,8 @@ void reconstruct_stripes_ec(osd_rmw_stripe_t *stripes, int pg_size, int pg_minsi
        if (stripes[role].read_end != 0 && stripes[role].missing)
        {
            recovered = true;
-            if (stripes[role].read_end > stripes[role].read_start)
+            if (stripes[role].read_end > stripes[role].read_start &&
+                stripes[role].read_end != UINT32_MAX)
            {
                for (int other = 0; other < pg_size; other++)
                {
@@ -531,7 +569,8 @@ void* alloc_read_buffer(osd_rmw_stripe_t *stripes, int read_pg_size, uint64_t ad
    uint64_t buf_size = add_size;
    for (int role = 0; role < read_pg_size; role++)
    {
-        if (stripes[role].read_end != 0)
+        if (stripes[role].read_end != 0 &&
+            stripes[role].read_end != UINT32_MAX)
        {
            buf_size += stripes[role].read_end - stripes[role].read_start;
        }
@@ -541,7 +580,8 @@ void* alloc_read_buffer(osd_rmw_stripe_t *stripes, int read_pg_size, uint64_t ad
    uint64_t buf_pos = add_size;
    for (int role = 0; role < read_pg_size; role++)
    {
-        if (stripes[role].read_end != 0)
+        if (stripes[role].read_end != 0 &&
+            stripes[role].read_end != UINT32_MAX)
        {
            stripes[role].read_buf = (uint8_t*)buf + buf_pos;
            buf_pos += stripes[role].read_end - stripes[role].read_start;
--- a/src/osd_rmw.h
+++ b/src/osd_rmw.h
@@ -23,6 +23,7 @@ struct osd_rmw_stripe_t
    void *read_buf, *write_buf;
    void *bmp_buf;
    uint32_t req_start, req_end;
+    // read_end=UINT32_MAX means to only read bitmap, but not data
    uint32_t read_start, read_end;
    uint32_t write_start, write_end;
    bool missing;
--- a/src/osd_rmw_test.cpp
+++ b/src/osd_rmw_test.cpp
@@ -27,6 +27,7 @@ void test13();
 void test14();
 void test15(bool second);
 void test16();
+void test_recover_22_d2();

 int main(int narg, char *args[])
 {
@@ -61,6 +62,8 @@ int main(int narg, char *args[])
    test15(true);
    // Test 16
    test16();
+    // Test 17
+    test_recover_22_d2();
    // End
    printf("all ok\n");
    return 0;
@@ -1045,7 +1048,12 @@ void test16()
    assert(stripes[3].read_buf == (uint8_t*)read_buf+2*128*1024);
    set_pattern(stripes[1].read_buf, 128*1024, PATTERN2);
    memcpy(stripes[3].read_buf, rmw_buf, 128*1024);
+    memset(stripes[0].bmp_buf, 0xa8, bmp);
+    memset(stripes[2].bmp_buf, 0xb7, bmp);
+    assert(bitmaps[1] == 0xFFFFFFFF);
+    assert(bitmaps[3] == 0xF1F1F1F1);
    reconstruct_stripes_ec(stripes, 4, 2, bmp);
+    assert(*(uint32_t*)stripes[3].bmp_buf == 0xF1F1F1F1);
    assert(bitmaps[0] == 0xFFFFFFFF);
    check_pattern(stripes[0].read_buf, 128*1024, PATTERN1);
    free(read_buf);
@@ -1054,3 +1062,47 @@ void test16()
    free(write_buf);
    use_ec(4, 2, false);
 }
+
+/***
+
+17. EC 2+2 recover second data block
+
+***/
+
+void test_recover_22_d2()
+{
+    const int bmp = 128*1024 / 4096 / 8;
+    use_ec(4, 2, true);
+    osd_num_t osd_set[4] = { 1, 0, 3, 4 };
+    osd_rmw_stripe_t stripes[4] = {};
+    unsigned bitmaps[4] = { 0 };
+    // Read 0-256K
+    split_stripes(2, 128*1024, 0, 256*1024, stripes);
+    assert(stripes[0].req_start == 0 && stripes[0].req_end == 128*1024);
+    assert(stripes[1].req_start == 0 && stripes[1].req_end == 128*1024);
+    assert(stripes[2].req_start == 0 && stripes[2].req_end == 0);
+    assert(stripes[3].req_start == 0 && stripes[3].req_end == 0);
+    uint8_t *data_buf = (uint8_t*)malloc_or_die(128*1024*4);
+    for (int i = 0; i < 4; i++)
+    {
+        stripes[i].read_start = stripes[i].req_start;
+        stripes[i].read_end = stripes[i].req_end;
+        stripes[i].read_buf = data_buf + i*128*1024;
+        stripes[i].bmp_buf = bitmaps + i;
+    }
+    // Read using parity
+    assert(extend_missing_stripes(stripes, osd_set, 2, 4) == 0);
+    assert(stripes[2].read_start == 0 && stripes[2].read_end == 128*1024);
+    assert(stripes[3].read_start == 0 && stripes[3].read_end == 0);
+    bitmaps[0] = 0xffffffff;
+    bitmaps[2] = 0;
+    set_pattern(stripes[0].read_buf, 128*1024, PATTERN1);
+    set_pattern(stripes[2].read_buf, 128*1024, PATTERN1^PATTERN2);
+    // Reconstruct
+    reconstruct_stripes_ec(stripes, 4, 2, bmp);
+    check_pattern(stripes[1].read_buf, 128*1024, PATTERN2);
+    assert(bitmaps[1] == 0xFFFFFFFF);
+    free(data_buf);
+    // Done
+    use_ec(4, 2, false);
+}
--- a/src/qemu_driver.c
+++ b/src/qemu_driver.c
@@ -218,6 +218,7 @@ static void coroutine_fn vitastor_co_get_metadata(VitastorRPC *task)
    }
 }

+// FIXME: Fix thread safety of the driver - now it segfaults when iothread is enabled in QEMU
 static void vitastor_aio_set_fd_handler(void *ctx, int fd, int unused1, IOHandler *fd_read, IOHandler *fd_write, void *unused2, void *opaque)
 {
    aio_set_fd_handler(ctx, fd,
--- a/src/vitastor.pc.in
+++ b/src/vitastor.pc.in
@@ -6,7 +6,7 @@ includedir=${prefix}/@CMAKE_INSTALL_INCLUDEDIR@

 Name: Vitastor
 Description: Vitastor client library
-Version: 0.8.8
+Version: 0.8.9
 Libs: -L${libdir} -lvitastor_client
 Cflags: -I${includedir}

--- a/tests/common.sh
+++ b/tests/common.sh
@@ -36,12 +36,12 @@ for i in $(seq 2 $ETCD_COUNT); do
    ETCD_URL="$ETCD_URL,http://$ETCD_IP:$((ETCD_PORT+2*i-2))"
    ETCD_CLUSTER="$ETCD_CLUSTER,etcd$i=http://$ETCD_IP:$((ETCD_PORT+2*i-1))"
 done
-ETCDCTL="${ETCD}ctl --endpoints=$ETCD_URL"
+ETCDCTL="${ETCD}ctl --endpoints=$ETCD_URL --dial-timeout=5s --command-timeout=10s"

 start_etcd()
 {
    local i=$1
-    $ETCD -name etcd$i --data-dir ./testdata/etcd$i \
+    ionice -c2 -n0 $ETCD -name etcd$i --data-dir ./testdata/etcd$i \
        --advertise-client-urls http://$ETCD_IP:$((ETCD_PORT+2*i-2)) --listen-client-urls http://$ETCD_IP:$((ETCD_PORT+2*i-2)) \
        --initial-advertise-peer-urls http://$ETCD_IP:$((ETCD_PORT+2*i-1)) --listen-peer-urls http://$ETCD_IP:$((ETCD_PORT+2*i-1)) \
        --initial-cluster-token vitastor-tests-etcd --initial-cluster-state new \
@@ -53,9 +53,12 @@ start_etcd()
 for i in $(seq 1 $ETCD_COUNT); do
    start_etcd $i
 done
-if [ $ETCD_COUNT -gt 1 ]; then
-    sleep 1
-fi
+for i in {1..30}; do
+    ${ETCD}ctl --endpoints=$ETCD_URL --dial-timeout=1s --command-timeout=1s member list >/dev/null && break
+    if [[ $i = 30 ]]; then
+        format_error "Failed to start etcd"
+    fi
+done

 echo leak:fio >> testdata/lsan-suppress.txt
 echo leak:tcmalloc >> testdata/lsan-suppress.txt
--- a/tests/run_3osds.sh
+++ b/tests/run_3osds.sh
@@ -39,19 +39,19 @@ done
 cd mon
 npm install
 cd ..
-node mon/mon-main.js --etcd_url $ETCD_URL --etcd_prefix "/vitastor" --verbose 1 &>./testdata/mon.log &
+(while true; do node mon/mon-main.js --etcd_url $ETCD_URL --etcd_prefix "/vitastor" --verbose 1 || true; done) &>./testdata/mon.log &
 MON_PID=$!

 if [ "$SCHEME" = "ec" ]; then
    PG_SIZE=${PG_SIZE:-5}
-    PG_MINSIZE=${PG_MINSIZE:-3}
-    PG_DATA_SIZE=$PG_MINSIZE
-    POOLCFG='"scheme":"ec","parity_chunks":'$((PG_SIZE-PG_MINSIZE))
+    PG_MINSIZE=${PG_MINSIZE:-4}
+    PG_DATA_SIZE=${PG_DATA_SIZE:-3}
+    POOLCFG='"scheme":"ec","parity_chunks":'$((PG_SIZE-PG_DATA_SIZE))
 elif [ "$SCHEME" = "xor" ]; then
    PG_SIZE=${PG_SIZE:-3}
-    PG_MINSIZE=${PG_MINSIZE:-2}
-    PG_DATA_SIZE=$PG_MINSIZE
-    POOLCFG='"scheme":"xor","parity_chunks":'$((PG_SIZE-PG_MINSIZE))
+    PG_MINSIZE=${PG_MINSIZE:-3}
+    PG_DATA_SIZE=$((PG_SIZE-1))
+    POOLCFG='"scheme":"xor","parity_chunks":1'
 else
    PG_SIZE=${PG_SIZE:-2}
    PG_MINSIZE=${PG_MINSIZE:-2}
@@ -100,13 +100,13 @@ wait_finish_rebalance()
    sec=$1
    i=0
    while [[ $i -lt $sec ]]; do
-        ($ETCDCTL get --prefix /vitastor/pg/state/ --print-value-only | jq -s -e '([ .[] | select(.state == ["active"]) ] | length) == 32') && \
+        ($ETCDCTL get --prefix /vitastor/pg/state/ --print-value-only | jq -s -e '([ .[] | select(.state == ["active"] or .state == ["active", "left_on_dead"]) ] | length) == '$PG_COUNT) && \
            break
-        if [ $i -eq 60 ]; then
-            format_error "Rebalance couldn't finish in $sec seconds"
-        fi
        sleep 1
        i=$((i+1))
+        if [ $i -eq $sec ]; then
+            format_error "Rebalance couldn't finish in $sec seconds"
+        fi
    done
 }

@@ -117,3 +117,14 @@ check_qemu()
        sudo ln -s "$(realpath .)/build/src/block-vitastor.so" /usr/lib/x86_64-linux-gnu/qemu/block-vitastor.so
    fi
 }
+
+check_nbd()
+{
+    if [[ -d /sys/module/nbd && ! -e /dev/nbd0 ]]; then
+        max_part=$(cat /sys/module/nbd/parameters/max_part)
+        nbds_max=$(cat /sys/module/nbd/parameters/nbds_max)
+        for i in $(seq 1 $nbds_max); do
+            mknod /dev/nbd$((i-1)) b 43 $(((i-1)*(max_part+1)))
+        done
+    fi
+}
--- a/tests/run_tests.sh
+++ b/tests/run_tests.sh
@@ -44,5 +44,5 @@ SCHEME=xor ./test_write.sh

 ./test_write_no_same.sh

-./test_heal.sh
-SCHEME=ec PG_MINSIZE=2 ./test_heal.sh
+PG_SIZE=2 ./test_heal.sh
+SCHEME=ec ./test_heal.sh
--- a/tests/test_add_osd.sh
+++ b/tests/test_add_osd.sh
@@ -15,10 +15,10 @@ done

 sleep 2

-for i in {1..10}; do
+for i in {1..30}; do
    ($ETCDCTL get /vitastor/config/pgs --print-value-only |\
        jq -s -e '([ .[0].items["1"] | map(.osd_set)[][] ] | sort | unique == ["1","2","3","4"])') && \
-        ($ETCDCTL get --prefix /vitastor/pg/state/ --print-value-only | jq -s -e '([ .[] | select(.state == ["active"]) ] | length) == '$PG_COUNT'') && \
+        ($ETCDCTL get --prefix /vitastor/pg/state/ --print-value-only | jq -s -e '([ .[] | select(.state == ["active"]) ] | length) == '$PG_COUNT) && \
        break
    sleep 1
 done
@@ -28,7 +28,7 @@ if ! ($ETCDCTL get /vitastor/config/pgs --print-value-only |\
    format_error "FAILED: OSD NOT ADDED INTO DISTRIBUTION"
 fi

-wait_finish_rebalance 10
+wait_finish_rebalance 60

 sleep 1
 kill -9 $OSD4_PID
@@ -37,7 +37,7 @@ build/src/vitastor-cli --etcd_address $ETCD_URL rm-osd --force 4

 sleep 2

-for i in {1..10}; do
+for i in {1..30}; do
    ($ETCDCTL get /vitastor/config/pgs --print-value-only |\
        jq -s -e '([ .[0].items["1"] | map(.osd_set)[][] ] | sort | unique == ["1","2","3"])') && \
        ($ETCDCTL get --prefix /vitastor/pg/state/ --print-value-only | jq -s -e '([ .[] | select(.state == ["active"] or .state == ["active", "left_on_dead"]) ] | length) == '$PG_COUNT'') && \
@@ -50,6 +50,6 @@ if ! ($ETCDCTL get /vitastor/config/pgs --print-value-only |\
    format_error "FAILED: OSD NOT REMOVED FROM DISTRIBUTION"
 fi

-wait_finish_rebalance 10
+wait_finish_rebalance 60

 format_green OK
--- a/tests/test_change_pg_count.sh
+++ b/tests/test_change_pg_count.sh
@@ -21,7 +21,7 @@ try_change()

    $ETCDCTL put /vitastor/config/pools '{"1":{'$POOLCFG',"pg_size":'$PG_SIZE',"pg_minsize":'$PG_MINSIZE',"pg_count":'$n'}}'

-    for i in {1..10}; do
+    for i in {1..60}; do
        ($ETCDCTL get /vitastor/config/pgs --print-value-only | jq -s -e '(.[0].items["1"] | map((.osd_set | select(. > 0)) | length == 2) | length) == '$n) && \
            ($ETCDCTL get --prefix /vitastor/pg/state/ --print-value-only | jq -s -e '([ .[] | select(.state == ["active"] or .state == ["active", "has_misplaced"]) ] | length) == '$n'') && \
            break
--- a/tests/test_heal.sh
+++ b/tests/test_heal.sh
@@ -2,7 +2,11 @@

 # Kill OSDs while writing

-PG_SIZE=3
+PG_SIZE=${PG_SIZE:-3}
+if [[ "$SCHEME" = "ec" ]]; then
+    PG_DATA_SIZE=${PG_DATA_SIZE:-2}
+    PG_MINSIZE=${PG_MINSIZE:-3}
+fi
 OSD_COUNT=7
 PG_COUNT=32
 . `dirname $0`/run_3osds.sh
--- a/tests/test_interrupted_rebalance.sh
+++ b/tests/test_interrupted_rebalance.sh
@@ -30,8 +30,11 @@ try_reweight 4 1

 try_reweight 5 1

+# Allow rebalance to start
+sleep 5
+
 # Wait for the rebalance to finish
-wait_finish_rebalance 60
+wait_finish_rebalance 300

 # Check that PGs never had degraded objects !
 # FIXME: In fact, the test doesn't guarantee it because PGs aren't always peered only with full prior OSD sets :-(
--- a/tests/test_move_reappear.sh
+++ b/tests/test_move_reappear.sh
@@ -15,11 +15,14 @@ $ETCDCTL put /vitastor/config/pools '{"1":{"name":"testpool","scheme":"replicate

 $ETCDCTL put /vitastor/config/pgs '{"items":{"1":{"1":{"osd_set":[1,0],"primary":1}}}}'

-sleep 2
-
-if ! ($ETCDCTL get /vitastor/pg/state/1/1 --print-value-only | jq -s -e '(. | length) != 0 and .[0].state == ["active","degraded"]'); then
-    format_error "Failed to start the PG active+degraded"
-fi
+for i in {1..30}; do
+    sleep 1
+    if ($ETCDCTL get /vitastor/pg/state/1/1 --print-value-only | jq -s -e '(. | length) != 0 and .[0].state == ["active","degraded"]'); then
+        break
+    elif [[ $i = 30 ]]; then
+        format_error "Failed to start the PG active+degraded"
+    fi
+done

 LD_PRELOAD="build/src/libfio_vitastor.so" \
 fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bs=4M -direct=1 -iodepth=1 -fsync=1 -rw=write \
@@ -27,45 +30,60 @@ fio -thread -name=test -ioengine=build/src/libfio_vitastor.so -bs=4M -direct=1 -

 $ETCDCTL put /vitastor/config/pgs '{"items":{"1":{"1":{"osd_set":[1,0],"primary":0}}}}'

-sleep 2
-
-if [ "`$ETCDCTL get /vitastor/pg/state/1/1 --print-value-only`" != "" ]; then
-    format_error "Failed to stop the PG"
-fi
+for i in {1..30}; do
+    sleep 1
+    if [[ "`$ETCDCTL get /vitastor/pg/state/1/1 --print-value-only`" = "" ]]; then
+        break
+    elif [[ $i = 30 ]]; then
+        format_error "Failed to stop the PG"
+    fi
+done

 $ETCDCTL put /vitastor/pg/history/1/1 '{"all_peers":[1,2,3]}'

 $ETCDCTL put /vitastor/config/pgs '{"items":{"1":{"1":{"osd_set":[4,5],"primary":4}}}}'

 sleep 5
-
-if ! ($ETCDCTL get /vitastor/pg/state/1/1 --print-value-only | jq -s -e '(. | length) != 0 and .[0].state == ["active"]'); then
-    format_error "Failed to move degraded objects to the clean OSD set"
-fi
+for i in {1..30}; do
+    sleep 1
+    if ($ETCDCTL get /vitastor/pg/state/1/1 --print-value-only | jq -s -e '(. | length) != 0 and .[0].state == ["active"]'); then
+        break
+    elif [[ $i = 30 ]]; then
+        format_error "Failed to move degraded objects to the clean OSD set"
+    fi
+done

 $ETCDCTL put /vitastor/config/pgs '{"items":{"1":{"1":{"osd_set":[4,5],"primary":0}}}}'

 $ETCDCTL put /vitastor/pg/history/1/1 '{"all_peers":[1,2,3]}'

-sleep 2
-
-if [ "`$ETCDCTL get /vitastor/pg/state/1/1 --print-value-only`" != "" ]; then
-    format_error "Failed to stop the PG after degraded recovery"
-fi
+for i in {1..30}; do
+    sleep 1
+    if [ "`$ETCDCTL get /vitastor/pg/state/1/1 --print-value-only`" = "" ]; then
+        break
+    elif [[ $i = 30 ]]; then
+        format_error "Failed to stop the PG after degraded recovery"
+    fi
+done

 cp testdata/osd4.log testdata/osd4_pre.log
 >testdata/osd4.log

 $ETCDCTL put /vitastor/config/pgs '{"items":{"1":{"1":{"osd_set":[4,5],"primary":4}}}}'

-sleep 2
-
-if grep -q 'PG 1/1.*is.*has_' testdata/osd4.log; then
-    format_error "PG has degraded or misplaced objects after a full re-peer following a degraded recovery"
-fi
-
-if ! ($ETCDCTL get /vitastor/pg/state/1/1 --print-value-only | jq -s -e '(. | length) != 0 and .[0].state == ["active"]'); then
-    format_error "PG not active+clean after a full re-peer following a degraded recovery"
-fi
+for i in {1..30}; do
+    sleep 1
+    if grep -q 'PG 1/1.*is.*has_' testdata/osd4.log; then
+        if [[ $i = 30 ]]; then
+            format_error "PG has degraded or misplaced objects after a full re-peer following a degraded recovery"
+        fi
+    elif ! ($ETCDCTL get /vitastor/pg/state/1/1 --print-value-only | jq -s -e '(. | length) != 0 and .[0].state == ["active"]'); then
+        if [[ $i = 30 ]]; then
+            format_error "PG not active+clean after a full re-peer following a degraded recovery"
+        fi
+    else
+        break
+    fi
+done

 format_green OK
--- a/tests/test_rebalance_verify.sh
+++ b/tests/test_rebalance_verify.sh
@@ -4,11 +4,13 @@ OSD_COUNT=7
 PG_COUNT=32
 . `dirname $0`/run_3osds.sh

+check_nbd
+
 IMG_SIZE=256

 $ETCDCTL put /vitastor/config/inode/1/1 '{"name":"testimg","size":'$((IMG_SIZE*1024*1024))'}'

-NBD_DEV=$(sudo build/src/vitastor-nbd map --etcd_address $ETCD_URL --image testimg --logfile ./testdata/nbd.log &)
+NBD_DEV=$(sudo build/src/vitastor-nbd map --nbd_timeout 180 --etcd_address $ETCD_URL --image testimg --logfile ./testdata/nbd.log &)

 trap "sudo build/src/vitastor-nbd unmap $NBD_DEV"'; kill -9 $(jobs -p)' EXIT

--- a/tests/test_splitbrain.sh
+++ b/tests/test_splitbrain.sh
@@ -21,7 +21,7 @@ LD_PRELOAD="build/src/libfio_vitastor.so" \
 # Kill OSD 2, start OSD 1

 kill $OSD2_PID
-build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 $OSD_ARGS --etcd_address $ETCD_URL \
+build/src/vitastor-osd --osd_num 1 --bind_address 127.0.0.1 $NO_SAME $OSD_ARGS --etcd_address $ETCD_URL \
    $(build/src/vitastor-disk simple-offsets --format options --device ./testdata/test_osd2.bin 2>/dev/null) >>./testdata/osd2.log 2>&1 &
 sleep 2

--- a/tests/test_vm_cont.sh
+++ b/tests/test_vm_cont.sh
@@ -14,7 +14,7 @@ for i in $(seq 1 $OSD_COUNT); do
    eval OSD${i}_PID=$!
 done

-node mon/mon-main.js --etcd_url $ETCD_URL --etcd_prefix "/vitastor" &>./testdata/mon.log &
+(while true; do node mon/mon-main.js --etcd_url $ETCD_URL --etcd_prefix "/vitastor" --verbose 1 || true; done) &>./testdata/mon.log &
 MON_PID=$!

 sleep 3
Author	SHA1	Message	Date
Vitaliy Filippov	5a9e1ede52	Release 0.8.9 - The tests are now stable and run in a CI system based on Gitea CI - The release includes final bug fixes for EC: - Implement missing EC recovery of allocation bitmap when built with ISA-L - Fix broken snapshot export with EC (allocation bitmap reads were giving incorrect results previously) - Also fixed bugs manifesting under heavy load: - Fix monitor possibly applying incorrect PG history on retries - Fix monitor incorrectly changing PG count when last_clean_pgs contains less PGs than the new number - Allow writes to wait for free space again, but now correctly (previously dropped in 0.8.2) - Fix a rare segfault in client (handle client stop during incoming stream handling in 1 more place) - Make monitor correctly handle etcd connection errors - it could die instead of connecting to another etcd - Fix OSD rarely being unable to report PG states after a PG was taken over by another OSD - Fixed return code for incomplete EC objects (now EIO) and made cluster client retry this error - Made other small changes for tests: timeouts, nice/ionice for etcd, waiting conditions, NBD device checks and so on	2023-05-14 01:25:09 +03:00
Vitaliy Filippov	1c9a188600	Add tests to CI	2023-05-14 00:06:09 +03:00
Vitaliy Filippov	de3e609166	Add a FIXME about QEMU driver thread safety	2023-05-14 00:06:09 +03:00
Vitaliy Filippov	11481170f5	Add a FIXME about ENOSPC	2023-05-13 23:59:44 +03:00
Vitaliy Filippov	e69d459d43	Allow rebalance to start in test_interrupted_rebalance, raise etcd start timeout	2023-05-13 15:16:28 +03:00
Vitaliy Filippov	da82754baa	Wait for conditions in test_move_reappear instead of waiting a fixed amount of time	2023-05-12 23:18:07 +03:00
Vitaliy Filippov	d356aca030	Add missing $NO_SAME OSD argument to test_splitbrain	2023-05-12 23:18:07 +03:00
Vitaliy Filippov	04a273d213	Raise NBD timeout in tests	2023-05-12 23:18:07 +03:00
Vitaliy Filippov	6442010f93	Skip offline PGs during state reporting when the state is already deleted or taken over by another OSD This fixes OSDs being unable to report PG states in rare conditions	2023-05-12 23:17:45 +03:00
Vitaliy Filippov	6f4dc16c59	Handle etcd connection errors correctly in mon (unhandled error events)	2023-05-11 11:02:44 +03:00
Vitaliy Filippov	ce4a8067b5	Handle client stop during incoming stream handling in 1 more place	2023-05-11 01:53:41 +03:00
Vitaliy Filippov	e431ecb715	Make tests more stable in CI	2023-05-11 01:53:41 +03:00
Vitaliy Filippov	8cac795445	Return EIO instead of EINVAL for incomplete EC objects	2023-05-11 01:15:23 +03:00
Vitaliy Filippov	a409598b16	Wait for free space again, but count on big_write flushes instead of just flusher activity	2023-05-10 01:51:02 +03:00
Vitaliy Filippov	f4c6765522	Ignore ENOENT in epoll_ctl	2023-05-08 20:39:20 +03:00
Vitaliy Filippov	ad2916068a	Fix test_add_osd rebalance timeout check	2023-05-08 20:39:20 +03:00
Vitaliy Filippov	321cb435a6	Fix monitor incorrectly changing PG count when last_clean_pgs contains less PGs than the new number	2023-05-08 20:39:20 +03:00
Vitaliy Filippov	cfcf4f4355	Support checking /dev/nbdX nodes in Docker	2023-05-08 20:39:20 +03:00
Vitaliy Filippov	e0fb17bfee	Make etcd more stable in tests (add ionice and raise timeout)	2023-05-08 20:36:00 +03:00
Vitaliy Filippov	5b9031fecc	Fix monitor possibly applying incorrect PG history under heavy load Monitor could deceive itself by immediately saving PG configuration changes which weren't applied to etcd yet in memory, and apply incorrect PG history changes next time if the first update fails. This usually only happened under heavy load and was caught in CI. :-)	2023-05-07 23:23:00 +03:00
Vitaliy Filippov	5da1d8e1b5	Fix EC just-bitmap reads (len=0) (fixes SCHEME=ec test_snapshot.sh)	2023-05-07 14:00:08 +03:00
Vitaliy Filippov	44f86f1999	Add a basic EC 2+2 recovery test (not really required, but let it be there)	2023-05-07 11:26:27 +03:00
Vitaliy Filippov	2d9a80c6f6	Implement missing bitmap recovery with ISA-L \(°□°)/	2023-05-07 11:25:51 +03:00