From: Cowley, David E <de.cowley@pn...> - 2004-10-01 17:12:05
First of all, thank you Bruce for providing smartmontools! I wish I'd
heard about them much sooner.
I have many many storage servers stuffed full of Western Digital drives
(WD2000JBs and WD2500JBs) on 3ware adapters (configured JBOD),
configured into linux kernel RAID5 volumes with ext3 filesystems on top.
After having some problems with drives unexpectedly getting kicked out
of the RAID volumes, I've been checking the drives' attributes w/
I have a number of drives that have values of Offline_Uncorrectable
and/or Current_Pending_Sector greater than zero.
Questions I have are:
If I run a long selftest, should it take care of Current_Pending_Sectors
that aren't associated with unreadable sectors?
If the long selftest clears up Current_Pending_Sectors, is my
If I have Offline_Uncorrectable errors, and the manufacturer's
diagnostic utilities can fix them up (I would pull the drive and run
diags on another machine), is my filesystem/RAID OK?
I would really love to be able to identify the files containing the bad
sectors I expect I have, but do not know how to trace through the linux
kernel RAID 5 layer to map bad sectors to filenames. Does anybody know
of resources I should check out? I have thought of the linux-raid
mailing list, but haven't asked there yet.
- Dave Cowley
David Cowley, Unix System Administrator Pacific Northwest National
PO Box 999, Richland, WA 99352 Internet E-Mail:
Get latest updates about Open Source Projects, Conferences and News.