RE: Fwd: xfs_repair -n report

["Vincent Roy" <vincentr@geoglobal.com>]

[-- Attachment #1: Type: text/plain, Size: 1596 bytes --]

That isn't good news. What do you suggest I do before totally trashing it,
what xfs_repair arguments should I at least try? The data volume size is as
it was before the RAID failure so the files are likely there and the RAID is
stable. I will try an xfsdump prior.
Regards,
Vincent

-----Original Message-----
From: David Chinner [mailto:dgc@sgi.com] 
Sent: Monday, April 21, 2008 4:31 PM
To: support@geoglobal.com
Cc: xfs@oss.sgi.com; vincent.roy@geoglobal.com
Subject: Re: Fwd: xfs_repair -n report

On Mon, Apr 21, 2008 at 02:52:01PM -0600, support@geoglobal.com wrote:
> Sorry, I truncated the original logfile of 128MB to a short version for 
> review. Again, all comments welcome.

Firstly, please post plain text attachments - that will cut down the
size of them as well...

> 
> Subject: xfs_repair -n report
> Date: Saturday 19 April 2008 23:22
> From: support@geoglobal.com
> To: xfs@oss.sgi.com, vincent.roy@geoglobal.com
> 
> RAID 5 - 4 disks on 9550 AMCC production machine / backup disk failed also
> verified without error - rebuilt after abrupt power shut down
> cannot mount - inconsistencies indicated - could not dump fs

Secondly, your raid has completely screwed the filesystem. I would
strongly suggest that you confirm that the rebuild of the RAID put
stuff back together properly (check things like the order of disks
in the stripe, etc). If it claims nothing is wrong, then it has been
lying to you.

I hope you have backups, because that's the only way you're going
to recover this mess.

Cheers,

Dave.
-- 
Dave Chinner
Principal Engineer
SGI Australian Software Group

[-- Attachment #2: testfile_raid.rtf --]
[-- Type: application/msword, Size: 190853 bytes --]