More timeout problems :(

[Daniel M. Eischen]

> I just recently was doing some fairly disk intensive work (rebuilding the
> man pages, building kde, and just finished running make clean on
> /usr/src, and closing a mailbox in Pine), when Pine alerted me to a disk
> error, so I switched over to ttyv0 to see screenfulls of sd0 timeout.. sd0
> no longer timedout... sd0 timedout while recovering.. then sd1, then sd0
> again, etc, etc.  I'm running FreeBSD current sources were just
> cvsup'd last night, kernel is about a week old.  This is the first time
> I've ever had probelems with my disks (two Quantum narrow, one ultra, one 
> fast scsi on an aic-7880).  The longest the system has been up was 21
> days, no problems at all then.  After about twenty minutes, things started
> to want to use the disk, so the swapper faulted, and just about everything
> besides pine that wanted to write to the disk segfaulted.  I went into the
> kernel debugger, and panic'd the kernel.  Going into the SCSI Select
> "program" and disabling WIDE negotation worked, probing the scsi bus froze
> the computer.  Power cycle followed.  I do have SCB paging, tagged
> queuing, and memio enabled.  My question is this, is there any
> way to prevent this from happening again (luckily this time I didn't loose
> any data), if not how much does an NCR/Symbios 810 cost?

What kind of Quantum drives do you have?  You didn't notice any
QUEUE_FULL messages?

How's your bus terminated?

Dan Eischen
deischen at iworks.InterWorks.org