[Gllug] Hard drive problems

[Chris Bell]

On Thu 31 May, Rich Walker wrote:
> 
> Geo <caparo.g at gmail.com> writes:
> 
> > On Thursday 31 May 2007 16:00, Rich Walker wrote:
> >> I don't know if anyone else has them, but yesterday and today I had a
> >> Samsung Spinpoint SP2514N fail each day, with lots of IDE busy errors.
> >>
> >> I've just removed them and now a Maxtor drive is doing the same thing.
> >>
> >> Kernel version 2.6.20 SMP - any ideas?
> 
> > Hi,
> >  Some A / B testing with new cables and another IDE based machine would 
> > isolate where the error lies 
> 
> Original situation: machine has a bunch of drives and controller
> cards. Each drive is on one cable. Each controller card has two drives.
> 
> Drives: hde hdg hdi hdk hdm hdo make up /dev/md1 (One spare, RAID-5)
> 
> 
> hde (Spinpoint) starts giving errors. Then I get errors from ext3 on one
> partition on /dev/md1
> 
> hde gets pulled from the array, and the spare added. When resync
> finished, I reboot and do a fsck on the faulty partition.
> 
> Next day: hdi (Spinpoint) gives same errors, followed by file system
> errors. I pull hdi from the array, shut the machine down, remove hde and
> hdi from the machine, and replace hdi with a handy 250GB
> Seagate. Reboot, go through fsck hell again, add hdi to md1 and let it rebuild.
> 
> On a Windows machine, the two Spinpoint disks are formatted
> happily. We're running chkdsk on there now.
> 
> Meanwhile a Maxtor drive (hdk) has started giving the 
> 
> [10059.640000] hdk: drive_cmd: status=0xd0 { Busy }
> [10059.640000] ide: failed opcode was: 0xea
> 
> error that the other two did. No file system errors *yet*.
> 
> 
> I have ordered a couple more 250GB disks "in case" for tomorrow, and
> have just compiled a 2.6.21.3 kernel to replace the 2.6.20 I was
> running.
> 

   I have had similar unexplained problems in the past which turned out to
be a dying PSU, (probably failing capacitors). Other things to look out for
are faulty series connectors in the power leads for splitting off additional
power outlets and case fans.
   I have had several power problems recently, with a known fault in the
main feeder outside intermittently shorting between phases and putting huge
splats on my supply. The fault has been located under a large group of ducts
carrying fibre optics, at the side of the A40, so I expect the electricty
company will damage some of the fibres, then the various fibre companies
will dig up the road again, then the water and gas pipes will be damaged,
then......

-- 
Chris Bell

-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug