http://oss.sgi.com/bugzilla/show_bug.cgi?id=206
hmh@xxxxxxxxxx changed:
What |Removed |Added
----------------------------------------------------------------------------
Version|Current |1.1.x
------- Additional Comments From hmh@xxxxxxxxxx 2003-02-07 16:31 -------
I've duplicated the problem as well. Linux 2.4.18, xfs 1.1 patches from SGI,
SMP.
userspace is Debian 3.0r1.
Upon remount ro of /dev/md0 (root filesystem, RAID1 array) followed by a reboot,
the xfs fs would be corrupted. xfs_repair fixes the problem, which among other
things causes the primary xfs superblock to be trashed.
I am not 100% sure the corruption happens at umount (well, I suppose a mount
ro,remount since it happened on the root partition at reboot time), raid1
shutdown (maybe xfs doesn't flush state right at raid1 read-only remount for
root fs shutdown?), or raid1 resync.
The problem was very easy to duplicate: wait for the raid to resync, reboot,
and there would it go. xfs_repair time, again.
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
|