[Linux-disciples] Lost Inodes and Hard Drive Failure Questions

[Dylan Thurston]

I can't help you with all of this, but...

On Sat, Jun 19, 2004 at 02:50:55PM -0400, Adam Kessel wrote:
> Once I get all the data off, I'm wondering what the best way is to
> determine whether the drive just needs to be junked.  It's a 200.0G
> drive, and less than one year old, so if it's still useable I'd like to
> keep it.  I know e2fsck -c will mark unuseable bad blocks.  Is there some
> "stress test" that anyone can recommend to see if the drive is likely to
> fail again?

Have you looked at the SMART pre-failure attributes?

> Interestingly, there is absolutely nothing in the logs indicating what
> actually went wrong.  I woke up one day to find the drive remounted
> read-only, and then lots of errors on e2fsck.  Also, cfdisk is unable to
> read the disk's partition table.  The drive appears to be functioning
> perfectly well now as I recover the data onto a known good drive.

Probably there was a message sent to syslog, but it couldn't write the
log file on the bad drive.

Peace,
	Dylan
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
Url : http://lists.bostoncoop.net/pipermail/linux-disciples/attachments/20040619/7ba6e494/attachment.pgp