Drive Failure or User Error?

[Lowell Gilbert]

Jason Morgan <jwm-freebsd-questions at sentinelchicken.net> writes:

> I was setting up a new server (6.1 i386 STABLE) - more specifically, I
> was mirroring the functioning server drive - when I suddenly got this:
>
> ad0 FAILURE - READ_DMA48 status=51<READY,DSC,ERROR>
> error=40<UNCORRECTABLE> LBA=611703808
> GEOM_MIRROR: Request failed (error=5). ad0[READ(offset=313192349696, 
> length=131072)]
>
> Along with several more errors, which were very similar. At this point,
> the server pretty much froze and would repeat the error at reboot, and
> as gmirror began resyncing the drive, the server would crash. I've tried
> disabling the mirror, fscking (multiple times), removing disks, and I
> just got done reinstalling (which went just fine) and resyncing. I still
> get the error and the system becomes unusable.
>
> So, my question is - and I suspect this is the case - is this a drive 
> failure or some issue with the mirroring process?

It *is* a drive failure, but I don't understand all of what's
happening there.  It is possible that this is not a FATAL drive
failure, but it's hard to be certain from this information.  If you
can figure out which file contains the bad sector, you can rewrite
that file and the drive may be able to recover.