http://oss.sgi.com/bugzilla/show_bug.cgi?id=411
--- Comment #16 from Andras Korn <korn-sgi.com@xxxxxxxxxxxxxxxxxxxxxx>
2009-02-15 14:17:48 CST ---
> Was this filesystem written ontop of luks? In that case I suspect it's doing
> something very weird with the write I/Os which causes this log corruption.
It was on top of LVM on top of LUKS on top of softraid.
The hardware write cache on the underlying SATA disks was enabled when the
kernel crashed for the first time and hosed the filesystem (it was from then on
that it crashed reliably when trying to replay the log). I don't know which
combination of these factors was what led to those data structures being
zeroed. I also don't know what caused the first crash, because nothing was
logged.
If the patch prevents the subsequent crashes, that's great; it would be even
better if it weren't necessary to run xfs_repair on the filesystem (i.e. if the
corrupt log data could just be discarded or something), but I guess that's not
of critical importance.
Thanks for the patch, I hope it'll be in the official kernel soon.
--
Configure bugmail: http://oss.sgi.com/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.
|