hey,
last night on of our filesystems filled up (which is another problem ;), and
when that happened, i noticed the following error twice in the syslog:
XFS: Internal error XFS_WANT_CORRUPTED_GOTO at line 1320 of file
/build/buildd-linux_3.2.35-1-i386-9hA0hl/linux-3.2.35/fs/xfs/xfs_alloc.c.
Caller 0xffffffffa03e9abd
more complete log excerpt + call trace attached.
i reacted by mounting that filesystem read-only, running an 'xfs_repair -n'
on it. because the apparent lack of trouble there, i continued by unmounting
the filesystem in all, then running a normal 'xfs_repair' (both output of
dry-run and real thing attached).
can somebody explain what happened here?
the filesystem rides on top of an LVM logical volume here if that might be
of any relevance. there were (some) free physical extents in the volume
group it was part of (used to extend the LV, and grow the FS -- none of
which happened until the xfs_repair was completed)
let me know if you want some other specific information.
kind regards,
--
/ Koen Breugelmans (koen.br@xxxxxxxxxxxx)
/ systeembeheer GRID-VFX
xfs-rushes-repair-info.txt
Description: Text document
xfs-rushes-repair-n.txt
Description: Text document
xfs-rushes-repair.txt
Description: Text document
|