> From: Michael McLagan [mailto:[EMAIL PROTECTED]]
> Sent: 20. januar 1999 17:48
> To: [EMAIL PROTECTED]; [EMAIL PROTECTED]
> Subject: Buslogic Flashpoint + RAID0 == LOCKUP
> I am looking for any comment, suggestions or thoughts about what may be
> causing this problem, and how I can go about resolving it.
>
> The PC is a 233MMX, 64M RAM, AMI Bios, EtherExpress Pro/100+, Buslogic
> Flashpoint BT930, 1 Microp 4221, 1 Microp 1991NT (restamped 3391NS), 3
Microp
> 3391NS, 1 Exabyte 8505XL, 1 Toshiba 3401B CDRom.
>
> The code is 2.2.0-pre8, raid0145-990108-2.2.0-pre5 (which doesn't quite
> patch clean, but 2 simple edits resolve the problem), raidtools-981214.
>
> The RAID0 partition is built with sdb1, sdc1, sdd1 and sde1.
chunk-size is
> 512, persistent-block is 1 and auto-recognise is compiled into the kernel.
>
> The machine locks up hard as a rock. No kernel dump, no error
messages, no
> indication of what's up. It reliably locks up if I start a restore off
the
> tape onto the RAID0 partition. Not always at the same point in the file,
but
> definately if I execute "cat /dev/tape > tape8-1.tar" onto the RAID0
partition,
> it inevitably fails. It has also locked up during the kernel boot
process,
> just at the point where the RAID0 autorecognise comes into play.
Interesting. I have a classic P-133MHz also with a BusLogic Flashpoint, and
doing striping (raid-0) on two Seagate Barracudas (2Gig drives). The
striping
was done under the old 2.0.x kernels, and I still use the aging 'mdtools'
that
I use to create the partitions with. I have four raid0 partitions, and have
so
far (in two years) never seen a problem with the striped partitions.
To me, it definitely sounds like a problem with the new raid code that was
introduced in 2.1.x, but I have never peeked into the RAID code, so I cannot
point to any particular area.
> I've tried different Buslogic cards. I've tried an AHA controller, but
it
> won't access the 9G drives properly :(. If I don't access the RAID, the
system
> seems to operate without incedent, so I don't believe that it's any other
> components.
Since we have so similar setups, it would be interesting if you tried using
a
2.0.x kernel, to see if the lockups persist. I can mail you the old mdtools
that I use for managing the md-partitions.
HTH,
Kai
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to [EMAIL PROTECTED]
