This issue touches a number of different places, but I got a working
environment from the Dual Primary DRBD + OCFS2 HOWTO, so I hope there's
enough expertise here to answer my question. Background on these
systems is Debian Squeeze, nothing from backports, configuration mostly
from the above listed HOWTO.
I've been able to (mostly) successfully build a Pacemaker-based
LVM/DRBD/OCFS2 filesystem shared between two hosts. Among the sticking
points so far is how I can successfully extend the filesystem when it
starts getting close to its limit.
In the past when we were active/passive, the process was:
- lvextend the underlying DRBD partitions
- drbdadmin resize the drbd device
- extend the ext3 filesystem
With OCFS2 (or active/active?) this appears to be more complicated. The
DRBD documentation says that you have to be active/passive in order for
online extensions to work.
This then implies the process is:
- lvextend DRBD
- shut down one side (stop corosync and start just drbd and make it
secondary)?
- drbdadmin resize the drbd device on the active side
- tune.ocfs2 -S on the active side
- stop drbd and restart corosync on the passive side
- profit?
Since this is backing a TWiki install, taking one server offline
temporarily to do this isn't a problem, but I'm not able to get this to
work, and often requires reboots of both systems and eventually the new
extended size is seen by ocfs2.
-Mark
--
Mark Komarinski mark_komarin...@hms.harvard.edu
Manager http://ritg.med.harvard.edu
Research IT Group
Harvard Medical School
_______________________________________________
Pacemaker mailing list: Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
Project Home: http://www.clusterlabs.org
Getting started: http://www.clusterlabs.org/doc/Cluster_from_Scratch.pdf
Bugs: http://developerbugs.linux-foundation.org/enter_bug.cgi?product=Pacemaker
