http://bugzilla.kernel.org/show_bug.cgi?id=10421
------- Comment #4 from dgc@xxxxxxx 2008-04-08 17:45 -------
> xfs_force_shutdown(sdd1,0x1) called from line 420 of file
> fs/xfs/xfs_rw.c. Return address = 0xf8b79dd2
(sdd1,0x1) means that the shutdown flags were 0x1:
#define SHUTDOWN_META_IO_ERROR 0x0001 /* write attempt to metadata failed */
There are a total of 5 places this can come from. From:
- xfs_trans_read_buf() we'd see I/O error alerts in the log prior
to the shutdown
- xfs_inactive() we'd see an error saying 'xfs_ifree returned an error'
- xfs_bwrite() won't log an extra error
- xfs_buf_error_relse() won't log an extra error
- xfs_bmap_finish() won't log an extra error
The three that don't log extra errors are a result of I/O errors or
object validation errors. Object validation errors typicaly log lots
more errors and stack dumps, though, so I'm tending towards a new
set of I/O errors coming from below XFS at this point....
What I really need to know is what function is at the return address
0xf8b79dd2. 'ksymoops -A 0xf8b79dd2' or something similar should give us
the symbol....
Also, can you run xfs_check on the broken filesystem before you run
xfs_repair so we can see all the errors on disk as a result of the shutdown.
--
Configure bugmail: http://bugzilla.kernel.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
|