[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: scsi controller or kernel failure?



On Wed, 2001-11-28 at 10:03, 'cduck' Chris Grierson wrote:
> i have received the following message(s) several times now, and once it
> happens, i have no choice but to reboot.  i had to manually reboot once
> before because i didn't have the SysRq magic built in my kernel, and
> severe filesystem damage (irreparable for the existing system) was
> sustained forcing me to reinstall.
> 
> scsi0: PCI error Interrupt at seqaddr = 0x180
> scsi0: Received Target Abort
> 
> (that 0x180 may have been something else, unfortunately)
> 
> the reinstall was on a smaller disk of the same series as the original
> (Quantum Atlas 10k).  unfortunately, it happened again, but thankfully
> i could sync the disk (not that i know that actually did anything
> useful, since the low-level SCSI layer was freaking out), and e2fsck
> managed to fix the filesystem problems on reboot.  this recoverable
> lockup has happened several times now, and i don't think it is going
> to stop.  has anyone had this and/or know what to do about it?  is it
> more likely to be a kernel (software) issue, or a SCSI controller
> issue (or the PCI controller!?) ?  the scsi controller is a Adaptec
> 7890, using the AIC7xxx kernel config option.

Built in to your motherboard then?

> also, eth0 was complaining (immediately after the scsi problem, and at
> no other time), but some google searching has me under the impression
> that it was related to some interrupt sharing that had gone awry since
> scsi0 was freaking out.

Perhaps this interupt sharing is the problem.  Try putting your SCSI
controller and your ethernet controller on different interrupts.  Since
these are your two most frequent interrupt sources, they really don't
belong on the same interrupt anyway.

-jwb



Reply to: