[Pacemaker] "ERROR: Wrong stack o2cb" when trying to start o2cb service in Pacemaker cluster

[David Guyot]

Hello, everybody.

I'm trying to configure Pacemaker for using DRBD + OCFS2 storage, but
I'm stuck with DRBD and controld up and o2cb doggedly displaying "not
installed" errors. To do this, I followed the DRBD guide (
http://www.drbd.org/users-guide-8.3/ch-ocfs2.html), with the difference
that I was forced to disable DRBD fencing because it was interfering
with Pacemaker fencing and stopping each nodes as often as it could.

Now, as I said, I'm stuck with these errors (Malastare and Vindemiatrix
being the 2 nodes of my cluster) :
Failed actions:
    p_o2cb_start_0 (node=Vindemiatrix, call=11, rc=5, status=complete):
not installed
    p_o2cb_monitor_0 (node=Malastare, call=4, rc=5, status=complete):
not installed

Looking into logs, I find these messages :
o2cb[19904]:    2012/06/20_09:04:35 ERROR: Wrong stack o2cb
o2cb[19904]:    2012/06/20_09:04:35 ERROR: Wrong stack o2cb

I tried to manually test ocf:pacemaker:o2cb, but I got this result :
root at Malastare:/home/david# export OCF_ROOT="/usr/lib/ocf"
root at Malastare:/home/david# /usr/lib/ocf/resource.d/pacemaker/o2cb monitor
o2cb[22387]: ERROR: Wrong stack o2cb
root at Malastare:/home/david# echo $?
5

I tried the solution described on this message (
http://oss.clusterlabs.org/pipermail/pacemaker/2009-December/004112.html),
but tunefs.ocfs2 failed :
root at Malastare:/home/david# cat /etc/ocfs2/cluster.conf
node:
    name = Malastare
    cluster = ocfs2
    number = 0
    ip_address = 10.88.0.1
    ip_port = 7777
node:
    name = Vindemiatrix
    cluster = ocfs2
    number = 1
    ip_address = 10.88.0.2
    ip_port = 7777
cluster:
    name = ocfs2
    node_count = 2
root at Malastare:/home/david# /etc/init.d/ocfs2 start
root at Malastare:/home/david# tunefs.ocfs2 --update-cluster-stack /dev/drbd1
Updating on-disk cluster information to match the running cluster.
DANGER: YOU MUST BE ABSOLUTELY SURE THAT NO OTHER NODE IS USING THIS
FILESYSTEM BEFORE MODIFYING ITS CLUSTER CONFIGURATION.
Update the on-disk cluster information? yes
tunefs.ocfs2: Unable to access cluster service - unable to update the
cluster stack information on device "/dev/drbd1"
root at Malastare:/home/david# /etc/init.d/o2cb start
root at Malastare:/home/david# tunefs.ocfs2 --update-cluster-stack /dev/drbd1
Updating on-disk cluster information to match the running cluster.
DANGER: YOU MUST BE ABSOLUTELY SURE THAT NO OTHER NODE IS USING THIS
FILESYSTEM BEFORE MODIFYING ITS CLUSTER CONFIGURATION.
Update the on-disk cluster information? yes
tunefs.ocfs2: Unable to access cluster service - unable to update the
cluster stack information on device "/dev/drbd1"

I also tried to manually set the cluster stack using "echo pcmk >
/sys/fs/ocfs2/cluster_stack" on both nodes, then restarting Corosync and
Pacemaker, but error messages stayed. Same thing with resetting OCFS2
FSs while cluster is on-line. Now I'm stuck, desperately waiting for
help  Seriously, if some do-gooder would like to help me, I would
greatly appreciate his or her help.

Thank you in advance for your answers.

Kind regards.


-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 554 bytes
Desc: OpenPGP digital signature
URL: <http://oss.clusterlabs.org/pipermail/pacemaker/attachments/20120620/ab4317d6/attachment.sig>