I thought someone more knowledgeable than I would have replied to you by
now?! I guess I'll take a wack at it!
Chris Frederick wrote:
> I have a server running Mandrake 9.0, 2 Western Digital 60Gig drives,
> and 1 Maxtor 20Gig drive. Lately I've noticed it making high pitched
You mention the makes and sizes of the drives but you don't mention
which one is making noise.
> noises (like a disk would make right before it crashes), and the machine
> itself has crashed a few times now. There's an e2fs error writing to
> certain blocks. They seem to be random, and in large numbers. I've ran
> e2fsck, checked for bad blocks in read only mode, and so far it's found
Is the noisy disk the same one that shows write errors?
> nothing. Changing /etc/fstab to mount all non essential disks read
> only, and removing a good number of cron jobs, will let it run without
> any errors, but that kinda defeats most of the functions of the server
> (backups).
>
> Is it possible that a disk can fail on a write, but not a read? Or am I
I would think that this is possible but I don't know.
> getting a different problem (I hope)? And if it is a disk crash, how
> long would you guess I have till it's no longer readable? I'm about
> three weeks away from affording a replacement disk.
After you have determined which disk is noisy/has errors my advice would
depend on what kind of data is on the drive:
If the data is:
valuable changes daily I suggest that you:
--------- ------------- ---------------------------------------
yes yes backup the data ASAP and stop using the drive
no yes replace the drive after it fails
yes no backup the data ASAP and replace the
drive when you can
no no replace the drive after it fails
>
> Thanks all...
--
Eric (Rick) Meyerhoff
_______________________________________________
TCLUG Mailing List - Minneapolis/St. Paul, Minnesota
http://www.mn-linux.org tclug-list at mn-linux.org
https://mailman.real-time.com/mailman/listinfo/tclug-list