Merge pull request #12250 from simondeziel/ceph-pool

Try to improve ceph pool creation reliability

.github/workflows/tests.yml

@@ -134,7 +134,7 @@ jobs:
 
       - name: Install Go (tip)
         run: |
-          curl -sL https://storage.googleapis.com/go-build-snap/go/linux-amd64/$(git ls-remote https://github.com/golang/go.git HEAD | awk '{print $1;}').tar.gz -o gotip.tar.gz
+          curl -sSfL https://storage.googleapis.com/go-build-snap/go/linux-amd64/$(git ls-remote https://github.com/golang/go.git HEAD | awk '{print $1;}').tar.gz -o gotip.tar.gz
           ls -lah gotip.tar.gz
           mkdir -p ~/sdk/gotip
           tar -C ~/sdk/gotip -xzf gotip.tar.gz
@@ -150,6 +150,7 @@ jobs:
           sudo apt-get update
 
           sudo snap remove lxd --purge
+          sudo snap remove core20 --purge || true
           sudo apt-get autopurge moby-containerd docker uidmap -y
           sudo ip link delete docker0
           sudo nft flush ruleset
@@ -218,7 +219,7 @@ jobs:
         if: ${{ matrix.backend == 'ceph' }}
         run: |
           set -x
-          sudo apt-get install --no-install-recommends -y snapd ceph-common
+          sudo apt-get install --no-install-recommends -y ceph-common
           sudo snap install microceph --edge
           sleep 5
           sudo microceph cluster bootstrap
@@ -230,17 +231,11 @@ jobs:
           for flag in nosnaptrim noscrub nobackfill norebalance norecover noscrub nodeep-scrub; do
               sudo microceph.ceph osd set $flag
           done
-          # Use ephemeral disk mounted on /mnt for ceph OSD.
-          # The block-devices plug doesn't allow accessing /dev/loopX devices so we make those same devices
-          # available under alternate names (/dev/sdiY) that are not used inside GitHub Action runners.
+          # Repurpose the ephemeral disk for ceph OSD.
           sudo swapoff /mnt/swapfile
-          sudo rm -f /mnt/swapfile
-          loop_file="/mnt/ceph-osd.img"
-          sudo fallocate -l 10G "${loop_file}"
-          loop_dev="$(sudo losetup --show --direct-io=on --nooverlap -f "${loop_file}")"
-          devInfo=($(sudo stat -c '%t %T' "${loop_dev}"))
-          sudo mknod -m 0660 /dev/sdia b 0x"${devInfo[0]}" 0x"${devInfo[1]}"
-          sudo microceph disk add --wipe /dev/sdia
+          ephemeral_disk="$(findmnt --noheadings --output SOURCE --target /mnt | sed 's/[0-9]\+$//')"
+          sudo umount /mnt
+          sudo microceph disk add --wipe "${ephemeral_disk}"
           sudo rm -rf /etc/ceph
           sudo ln -s /var/snap/microceph/current/conf/ /etc/ceph
           sudo microceph enable rgw

test/backends/ceph.sh

@@ -15,7 +15,7 @@ ceph_configure() {
 
   echo "==> Configuring CEPH backend in ${LXD_DIR}"
 
-  lxc storage create "lxdtest-$(basename "${LXD_DIR}")" ceph volume.size=25MiB ceph.osd.pg_num=16
+  lxc storage create "lxdtest-$(basename "${LXD_DIR}")" ceph volume.size=25MiB ceph.osd.pg_num=8
   lxc profile device add default root disk path="/" pool="lxdtest-$(basename "${LXD_DIR}")"
 }
 

test/main.sh

@@ -83,6 +83,10 @@ cleanup() {
     read -r _
   fi
 
+  echo ""
+  echo "df -h output:"
+  df -h
+
   if [ -n "${GITHUB_ACTIONS:-}" ]; then
     echo "==> Skipping cleanup (GitHub Action runner detected)"
   else