On Tue, Apr 23, 2002 at 06:35:46PM +0200, Willi Langenberger wrote:
> Hi!
Hi there,
> The problems began with disappearing directories. We tried a
> xfs_repair (1.2.8), but without luck. Unfortunatly, my colleague
> hasn't saved the output from that run.
>
> Then, we upgraded the xfsprogs to 2.0.3, an now we get a core dump:
>
> # /sbin/xfs_repair /dev/sdb
> Phase 1 - find and verify superblock...
> Phase 2 - using internal log
> - zero log...
> xfs_repair: xfs_log_recover.c:159: xlog_find_verify_log_record: Assertion
> `start_blk != 0 || *last_blk != start_blk' failed.
> Aborted (core dumped)
>
> Should this happen? The Kernel Upgrade from 2.4.3-SGI_XFS_1.0.1 to
> 2.4.9-31SGI_XFS_1.1 made no difference.
Ah, I see the problem. You have an unclean log and corruption in
the log which is causing the code in xfs_repair to get confused when
searching for the log head/tail.
First try mount, then unmount, and then run xfs_repair. The initial
mount/unmount will cause the log to be replayed, if it can be.
Failing that, try "xfs_repair -L /dev/sdb" which skips the check for
an empty log, zeroes it, then goes ahead with repairing.
Failing that (!) we can get libxlog built non-debug so this assertion
doesn't get tripped, and see how we go from there.
cheers.
--
Nathan
|