xfs
[Top] [All Lists]

RE: 2.4.18 XFS 1.1 : Gave up on XFS - too many Oops

To: linux-xfs@xxxxxxxxxxx
Subject: RE: 2.4.18 XFS 1.1 : Gave up on XFS - too many Oops
From: Poul Petersen <petersp@xxxxxxxxxxxxx>
Date: Fri, 19 Jul 2002 15:54:37 -0700
Sender: owner-linux-xfs@xxxxxxxxxxx
> When you see a oops on any linux box you should always save 
> it, decode 
> it with ksymoops and send it to one of the development lists (in this
> case to linux-xfs). If you don't do that you should not be 
> surprised that 
> bugs don't get fixed and you can only blame yourself.

        Been there, done that. Spent many an anxious evening saving oops
outputs, running them through ksymoops and adding trying to make sense of it
all - I had plenty to choose from, and I shared them. A quick search finds:

http://marc.theaimsgroup.com/?l=linux-xfs&m=99558768600852&w=2
http://marc.theaimsgroup.com/?l=linux-xfs&m=99721971620728&w=2
http://marc.theaimsgroup.com/?l=linux-xfs&m=100402789630290&w=2
http://marc.theaimsgroup.com/?l=linux-xfs&m=102264819826742&w=2

        That last post evoked no responses and was pretty much the point
when we decided to bail. To be fair, one of the earlier ones was an oops
which resulted from a conflict between NFS and xfsdump (which I think was
resolved later - we just switched amanda to use tar). 

        There is no need to be defensive - I am not attacking XFS or saying
it is garbage, etc.  Our experience with XFS was a bad one, others not.
Maybe we have some bizarre configuration, maybe we are doing it wrong. But
we arrived at a point where the server was going down every week and we
couldn't get a similar system in a controlled environment to fail, which was
derailing any attempts to debug the problem - it's hard to spent time
figuring out what is going wrong or testing patches with 100+ people waiting
for the NFS server to come back up.

        Anyways, if every post was about how wonderful XFS was, what fun
would that be?

-poul


<Prev in Thread] Current Thread [Next in Thread>