[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: ext3_get_inode_loc: bad inode number:

On Wed, Dec 03, 2003 at 04:26:07PM -0800, Poul Petersen wrote:
> 	I took this disc offline and ran fsck twice. It found a few errors
> the first time and none the second. I then rebooted the server, but the
> errors returned. The obvious next step would seem to be to build a new
> file-system and tar copy the data. Does it seem reasonable that the shrink
> could have corrupted the inode table? There is a lot going on with this
> setup (could be LVM, ext3online_resize, etc), so mostly I'm just curious if
> it is reasonable that the shrink could have been responsible or if there is
> some other component I should examine. Thoughts?

If fsck found filesystem corruptions, and the errors returned after
running fsck, I would be very suspicious about hardware problems;
perhaps a memory problem, or a loose SCSI/IDE cable, or a flakey disk
about to go bad.  So the *first* thing I would do is a full backup,
before doing any more investigation, just in case this is a warning
sign before a full fledge hard disk head crash.

It doesn't seem reasonable that the shrink has anything to do with
this; even if the shrink had corrupted the filesystem, fsck would have
fixed it.  That's why the fact that filesystem errors occured right
after an fsck makes me very suspicious about your hardware.  It's
possible it might also be caused by a kernel bug, but if you haven't
changed your kernel in a while, and it's been working fine, the
component to be most suspicious of is your hardware....

						- Ted

[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]