[Ocfs2-users] fsck in startup scripts

[Henrik Carlqvist]

Thanks for a quick reply!

> Hmm... meaning the corrupted area was not touched by the running fs.
> Because if it had, the fs would have been remounted ro with a message
> asking the user to run fsck.

If so, the corrupted fs was probably not the cause of an initial problem
which made me run fsck. I had all nfsd daemons in D state waiting for disk
on both servers, those daemons were not even killable with "kill -9". At
reboot one of the file systems refused to unmount so I had to do a hard
reset with the power switch. Some hour later the problem came back and I
then started up only one node to run fsck. Since that fsck the problem
hasn't come back and I thought or maybe rather hoped that the problem
could be explained by a corrupted fs.

Now it seems more likely that my initial problem was caused by broken
hardware or maybe related to the READDIRPLUS problem described at
http://www.mail-archive.com/ocfs2-announce@oss.oracle.com/msg00016.html
I don't have any FreeBSD clients, only Linux and Solaris Clients, but from
the description of the READDIRPLUS problem it seems as it can occur also
with other clients than FreeBSD.

The 2-node cluster is running Slackware 12.0 with a slightly patched
2.6.21.5 kernel, the ocfs2 version is the one included in the kernel, but
if I remember right one of the patches applied is related to ocfs2.

The NFS clients are running Slackware 8, Slackware 9.1, Slackware 12.0,
Slamd64 12.0, Solaris 2.6 (sparc) and Solaris 8 (sparc). There are
probably also some more Solaris clients and maybe also some RedHat
clients.

> Remember, ocfs2 is a journaled fs. So there is no need to run fsck

Thanks, yes I was a little surprised to find those corrupted file systems.
Maybe my raid disk is behaving oddly when losing power.

regards Henrik
-- 
NOTE: Dear Outlook users: Please remove me from your address books.
      Read this article and you know why:
      http://newsforge.com/article.pl?sid=03/08/21/143258