[CentOS] OT - Offline uncorrectable sectors

Fri Aug 22 16:29:27 UTC 2008
Akemi Yagi <amyagi at gmail.com>

On Fri, Aug 22, 2008 at 9:26 AM, nate <centos at linuxpowered.net> wrote:
> Lorenzo Quatrini wrote:

> Not sure myself but the manufacturer's testing tools have
> non destructive ways of detecting and re-mapping bad sectors.
> Of course a downside to the manufacturer's tools is they often
> only support a limited number of disk controllers.
> It's probably been since the IBM Deathstar 75GXP that I last recall
> having drives with bad sectors on them but typically at least at that
> time, when the OS encountered a bad sector it didn't handle it too
> gracefully, often times had to reboot the system. Perhaps the linux
> kernel is more robust for those things these days (I had roughly 75%
> of my 75GXP drives fail - more than 30).
> Interesting that the man page for e2fsck in RHEL 4 doesn't describe
> the -c option, but the man page for it in RHEL 3 does, not sure if
> that is significant(RHEL4 man page mentions the option, but no
> clear description of what it does). Haven't checked RHEL/CentOS 5.
> from RHEL 3 manpage:
>       -c     This option causes e2fsck to run  the  badblocks(8)
>              program  to find  any blocks which are bad on the
>              filesystem, and then marks them as bad by adding
>              them to the  bad  block  inode.   If  this option
>              is specified twice, then the bad block scan will
>              be done using a non-destructive read-write test.
> So if you haven't heard of it already, try e2fsck -c <device> ?
> I recall using this off and on about 10 years ago but found the
> manufacturer's tools to be more accurate.
>> And yes, I know that a disk with bad blocks isn't reliable, but you
>> remember?
>> I'm too lazy to send my home disks back to the manufacturer ;)
> Ahh ok, I see...just keep in mind that it's quite possible the
> bad sector count will continue to mount as time goes on..

There is a thread on this topic in the CentOS forum: