Howdy, folks. Today, shortly after boot, I got the following in my
logs:
Dec 30 12:51:32 sto-kerrig kernel: xfs_force_shutdown(lvm(58,2),0x8) called
from line 4079 of file xfs_bmap.c. Return address = 0xc016db34
Dec 30 12:51:32 sto-kerrig kernel: Corruption of in-memory data detected.
Shutting down filesystem: lvm(58,2)
Dec 30 12:51:32 sto-kerrig kernel: Please umount the filesystem, and
rectify the problem(s)
followed by some oopses (attached).
The kernel is 2.4.17-xfs, built on the 24th and in use since then.
The tree I built from was updated shortly before the build.
I rebooted into single-user and successfully activated the volume
group (once I figured out that vgscan fails very ungracefully if / is
mounted ro and that [*phew*] my volume groups hadn't evaporated). I
then attempted to run xfs_repair on /usr, which told me there was
unreplayed log data. I mounted /usr, saw messages indicating a
successful recovery and then umounted it. xfs_repair gave me the same
"unreplayed log" message, so I mounted and umounted /usr a few more
times (seeing a recovery each time) before deciding to chance it and
run xfs_repair with the -L option. The repair found only a couple of
errors in the filesystem, and returned successfully. Subsequent
mounts of /usr have been clean and error-free.
I've gone back to my most recent 2.4.16 while I build a fresh 2.4.17
from today's CVS. I will also run a memtest86 at some point today.
xfs.txt
Description: oopsen from syslog
--
///////////////// | | The spark of a pin
<sneakums@xxxxxxxx> | (require 'gnu) | dropping, falling feather-like.
\\\\\\\\\\\\\\\\\ | | There is too much noise.
|