[geeks] HD/IDE question

[velociraptor]

On 9/20/06, Michael C. Vergallen <mvergall at pandora.be> wrote:
> Jochen Kunz wrote:
>
> >On Tue, 19 Sep 2006 10:45:19 -0400
> >velociraptor <velociraptor at gmail.com> wrote:
> >>I'm wondering if the problems could be an artifact of a flaky IDE
> >>controller (on-board IDE and the RAID chipset is also one of the
> >>Promise "Lite" RAID set ups) rather than the disk.
> >>
> >Once I had a strange data corruption problem wirth a IDE disk. Finaly I
> >found that is was a broken IDE cable... (The machine in question was
> >new.)
> >
> I'll second that... This is the most likely couse of the problem however
> when replacing the cable with a
>  known good one and the problem persists  I  would replace the disk next
> because some combinations
> of seamingly good  disks  can  lead  to disasters ...I've  had the same
> problem with one the systems here and it
>  turned out to be a disk that was flaky ....I've only found that out
> after a while ... anyway try to check the parts in
> order of expense ....That is how I've learded to work because the cost
> most of the time gives an indication of the
> testing that has taken plase at the production site.... also check the
> flash bios updates and make shure you have the
> latest rev of the bios or eprom.

I am going to do the surface analysis on the Mac as my first step, as
that will isolate if the disk itself is the problem (I currently have
another Seagate disk of the same model in the FW cannister that I will
be using, so everything will be "known good" other than the disk).

I am ignorant of Windows' low-level stuff so I don't know if assuming
the Promise RAID Lite works like I'd expect a *NIX HW RAID is a
reasonable thing to do.  E.g., while I might see errors in the SMART
report, I *should not* see complete failure errors in Windows event
manager because the HW RAID would mask them by getting the data off
the good disk.  At most, I should see timeout errors.

I do have other IDE cables I can try, so after a surface analysis,
I'll try that to isolate cable vs. mobo issues.  I suppose should also
test the continuity on the cables, and the power out of the PSU, just
to cover all the bases.

My biggest hurdle at this point converting the other RAID file system
on my 3Ware card from NTFS to something Linux/FreeBSD can handle since
I'm not going to roll the dice on an in-place conversion of the file
system without a backup even though I have had good luck with things
like Partition Magic and the like.  I suppose if the disks from the
boot mirror are OK and I can get the little bit of data copied off of
those, I can then stripe the two boot disks and use that for a backup
of the big file system.

Sys admin->bullet->foot. :-(  I have no conception of what I was
thinking when I decided to make use a proprietary file system for a
~400GB RAID.  I obviously need a cluebat to the head that day.

=Nadine=