----- Original Message -----
From: "Raoul Bhatia [IPAX]" <r.bha...@ipax.at>
To: <pacemaker@oss.clusterlabs.org>; "Daniel Qian" <dan...@bestningning.com>
Sent: Thursday, December 10, 2009 11:10 AM
Subject: Re: [Pacemaker] Manual fsck required on passive node upon failover
On 12/10/2009 05:01 PM, Daniel Qian wrote:
I installed a two-node cluster following this link on clusterlabs.org
http://www.howtoforge.com/installation-and-setup-guide-for-drbd-openais-pacemaker-xen-on-opensuse-11.1
The guide is for OpenSuse but I did it on Centos 5 as all the packages
are available there.
Basically what I did so far is a Xen guest instance sitting on the
DRBD-backed file system. Everything seems to be working fine except for
one thing - When I shutdown openais (service openais stop) on the active
node for failover testing, the passive node was trying to take over all
the resources but got stuck at the point of file system resource. Logs
show it required file system check thus manual intervention is required
to bring it online. However if I do 'crm resource move xen_rsc
passive_node' the failover transits fine every time.
can you reproduce the problem or has this been a one time issue?
This is not a one time issue. it happened more than half of the times I did
the test. My hunch is the Xen instance did not get shutdown cleanly in time
or the file system did not get umounted on the active node. I ended up with
primary/primary for drbd roles of the two nodes when that happened.
did you turn off the "periodic fsck" on e.g. ext3? see:
# tune2fs -l /dev/md0
...
Filesystem state: clean
Errors behavior: Continue
..
Mount count: 2
Maximum mount count: 23
Last checked: Sat Aug 8 16:17:11 2009
Check interval: 15552000 (6 months)
Next check after: Thu Feb 4 15:17:11 2010
..
This doesnt seem to be the issue
dev/drbd0 691435652 1945052 654367656 1% /xen
[r...@ilo142 ~]# tune2fs -l /dev/drbd0
tune2fs 1.39 (29-May-2006)
Filesystem volume name: <none>
Last mounted on: <not available>
Filesystem UUID: 94351cb7-d51d-4d63-8c31-e261e12f8647
Filesystem magic number: 0xEF53
Filesystem revision #: 1 (dynamic)
Filesystem features: has_journal ext_attr resize_inode dir_index
filetype needs_recovery sparse_super large_file
Default mount options: (none)
Filesystem state: clean
Errors behavior: Continue
Filesystem OS type: Linux
Inode count: 87818240
Block count: 175614727
Reserved block count: 8780736
Free blocks: 172388127
Free inodes: 87818227
First block: 0
Block size: 4096
Fragment size: 4096
Reserved GDT blocks: 982
Blocks per group: 32768
Fragments per group: 32768
Inodes per group: 16384
Inode blocks per group: 512
Thanks,
Daniel
_______________________________________________
Pacemaker mailing list
Pacemaker@oss.clusterlabs.org
http://oss.clusterlabs.org/mailman/listinfo/pacemaker
