cluster HA LVM service fails beacuse of a bug

[nanonettr]

Hello,

I was playing with proxmox cluster for active/passive SAN configuration using HA LVM and ext4.

I configured cluster.conf and lvm.conf but my service was not starting with error;

Sep 02 00:04:34 rgmanager Starting disabled service service:GlusterHA
Sep 02 00:04:35 rgmanager [lvm] HA LVM:  Improper setup detected
Sep 02 00:04:35 rgmanager [lvm] * initrd image needs to be newer than lvm.conf
Sep 02 00:04:35 rgmanager start on lvm "lvmSAS" returned 1 (generic error)
Sep 02 00:04:35 rgmanager #68: Failed to start service:GlusterHA; return value: 1
Sep 02 00:04:35 rgmanager Stopping service service:GlusterHA
Sep 02 00:04:35 rgmanager [ip] 10.0.0.81/24 is not configured
Sep 02 00:04:35 rgmanager [fs] unmounting /SANSAS
Sep 02 00:04:35 rgmanager [lvm] HA LVM:  Improper setup detected
Sep 02 00:04:35 rgmanager [lvm] * initrd image needs to be newer than lvm.conf
Sep 02 00:04:35 rgmanager [lvm] WARNING: An improper setup can cause data corruption!
Sep 02 00:04:35 rgmanager [lvm] Deactivating sansas/sansas01
Sep 02 00:04:35 rgmanager [lvm] Making resilient : lvchange -an sansas/sansas01
Sep 02 00:04:35 rgmanager [lvm] Resilient command: lvchange -an sansas/sansas01 --config devices{filter=["a|/dev/mapper/sasvol-part1|","a|/dev/mapper/saSep 02 00:04:36 rgmanager [lvm] Removing ownership tag (node01) from sansas/sansas01
Sep 02 00:04:36 rgmanager Service service:GlusterHA is recovering
Sep 02 00:04:36 rgmanager #71: Relocating failed service service:GlusterHA
Sep 02 00:04:36 rgmanager #70: Failed to relocate service:GlusterHA; restarting locally
Sep 02 00:04:36 rgmanager Recovering failed service service:GlusterHA
Sep 02 00:04:36 rgmanager [lvm] HA LVM:  Improper setup detected
Sep 02 00:04:36 rgmanager [lvm] * initrd image needs to be newer than lvm.conf
Sep 02 00:04:36 rgmanager start on lvm "lvmSAS" returned 1 (generic error)
Sep 02 00:04:36 rgmanager #68: Failed to start service:GlusterHA; return value: 1
Sep 02 00:04:36 rgmanager Stopping service service:GlusterHA
Sep 02 00:04:36 rgmanager [ip] 10.0.0.81/24 is not configured
Sep 02 00:04:36 rgmanager [fs] stop: Could not match /dev/mapper/sansas-sansas01 with a real device
Sep 02 00:04:36 rgmanager stop on fs "fsSAS" returned 2 (invalid argument(s))
Sep 02 00:04:36 rgmanager [lvm] HA LVM:  Improper setup detected
Sep 02 00:04:36 rgmanager [lvm] * initrd image needs to be newer than lvm.conf
Sep 02 00:04:36 rgmanager [lvm] WARNING: An improper setup can cause data corruption!
Sep 02 00:04:36 rgmanager [lvm] Deactivating sansas/sansas01
Sep 02 00:04:37 rgmanager [lvm] Making resilient : lvchange -an sansas/sansas01
Sep 02 00:04:37 rgmanager [lvm] Resilient command: lvchange -an sansas/sansas01 --config devices{filter=["a|/dev/mapper/sasvol-part1|","a|/dev/mapper/saSep 02 00:04:37 rgmanager [lvm] Removing ownership tag (node01) from sansas/sansas01
Sep 02 00:04:37 rgmanager #12: RG service:GlusterHA failed to stop; intervention required
Sep 02 00:04:37 rgmanager Service service:GlusterHA is failed
Sep 02 00:04:37 rgmanager #2: Service service:GlusterHA returned failure code.  Last Owner: node01
Sep 02 00:04:37 rgmanager #4: Administrator intervention required.

I updated initrd images again and rebooted but nothing changed.

After a bit debugging i found the failing command in "/usr/share/cluster/lvm.sh"

...
if [ "$(find /boot -name *.img -newer /etc/lvm/lvm.conf)" == "" ]; then
...

this command did not find my initrd images but some other files.
In default proxmox install the file naming "initrd.img-<VERSION>-pve".

So the correct command needs to be "find /boot -maxdepth 1 -name initrd.img*"

I wanted to report this, but where ?
is this a bug of Proxmox, RedHat or Debian ?

 

[dietmar]

This file is part of the resource-agents project:

https://github.com/ClusterLabs/resource-agents

 

[nanonettr]

This file is part of the resource-agents project:

https://github.com/ClusterLabs/resource-agents

thanks, problem reported.
https://github.com/ClusterLabs/resource-agents/issues/671