[Pacemaker] cannot mount gfs2 filesystem

Andrew Beekhof andrew at beekhof.net
Wed Oct 31 21:01:04 EDT 2012


On Wed, Oct 31, 2012 at 5:04 PM, Soni Maula Harriz
<soni.harriz at sangkuriang.co.id> wrote:
> and this is what happen in cluster1


> [root at cluster1 ~]# grep -i error /var/log/messages
> Oct 31 11:10:44 cluster1 pengine[1921]:  warning: unpack_rsc_op: Processing
> failed op WebFS:0_last_failure_0 on cluster1: unknown exec error (-2)
> Oct 31 11:10:49 cluster1 dlm_controld[1467]: open "/sys/kernel/dlm/web/id"
> error -1 2
> Oct 31 11:10:49 cluster1 dlm_controld[1467]: open
> "/sys/kernel/dlm/web/control" error -1 2
> Oct 31 11:10:49 cluster1 dlm_controld[1467]: open
> "/sys/kernel/dlm/web/event_done" error -1 2

The DLM is clearly not very happy.
I'd recommend taking Pacemaker out of the picture for now, until you
can get the filesystem mounted manually.
Once that works, only then get Pacemaker to try.

As to why the dlm is unhappy, I'm afraid I'm not much of an expert.
The developers hang out on the linux-cluster mailing list though, you
may have more luck there.
They will probably want to see more than just the errors lines though.




More information about the Pacemaker mailing list