Re: reiserfs oops [2.5.65]

[Dave Jones <davej@codemonkey.org.uk>]

On Thu, Mar 20, 2003 at 04:59:41PM -0800, Andrew Morton wrote:
 > > 
 > > Slab corruption: start=c70c7044, expend=c70c7213 problemat=c70c7044
 > > Last user: [<c0280dcb>](reiserfs_alloc_inode+0x1b/0x30)
 > > Data: (lots of hex)
 > 
 > Alas, the "(lots of hex)" is important - it lets us determine which member of
 > struct reiserfs_inode was actually altered.

You're in luck, as noted in the follow up, I captured it all..
http://www.codemonkey.org.uk/cruft/oops.txt
 
 > It would be nice if we had a more robust way of capturing all this info,
 > especially the oops-while-running-X lossage.  Dump-to-floppy or something.

The tricky thing about capturing this one was that something in the VFS
was wedged hard, so I'm lucky the logs made it to disk.
Using sysrq, I synced and the disks made really nasty chugging noises.
umount read only made it write a bunch of stuff out, more chugging
noises for 15 minutes, kill all tasks then popped up a getty for me to
log in on. It took about 3 minutes to get a shell (disk IO was seriously
slow). I took a peek at vmstat 1 (no logs of this sorry), and nothing
out of the ordinary. Not much in swap, plenty of swap free.
short: the machine was horked. I synced, and rebooted, and thankfully
the logs were still there after fsck.ext2 recovered /var

		Dave