We have not been able to hit this internally yet.
If those who can hit it could try these things:
Use the latest CVS code, or apply this patch:
--- linux/fs/xfs/xfs_rw.c_1.385 2003-10-30 10:33:20.000000000 -0600
+++ linux/fs/xfs/xfs_rw.c 2003-10-30 10:33:20.000000000 -0600
@@ -153,6 +153,9 @@
xfs_cmn_err(XFS_PTAG_SHUTDOWN_CORRUPT, CE_ALERT, mp,
"Corruption of in-memory data detected. Shutting down filesystem: %s",
mp->m_fsname);
+ if (XFS_ERRLEVEL_HIGH <= xfs_error_level) {
+ xfs_stack_trace();
+ }
} else if (!(flags & XFS_FORCE_UMOUNT)) {
if (logerror) {
xfs_cmn_err(XFS_PTAG_SHUTDOWN_LOGERROR, CE_ALERT, mp,
and
echo 5 > /proc/sys/fs/xfs/error_level
This should cause a stack trace to be dumped when the filesystem shuts
down.
Alternately (or additionally)
echo 2 > /proc/sys/fs/xfs/panic_mask
and this should cause the system to panic when it hits this log
reservation problem, which should also dump the stack.
Also, I will open a bugzilla bug for this, if the people who have hit
this could do a brain-dump to the bug, that will keep all the info in
one place.
Thanks,
-Eric
--
Eric Sandeen [C]XFS for Linux http://oss.sgi.com/projects/xfs
sandeen@xxxxxxx SGI, Inc. 651-683-3102
|