Re: SCSI Kernel Problem - BAD

Steven S. Dick (ssd@nevets.oau.org)
Tue, 19 Mar 1996 11:37:54 EST

On Mar 18, 7:19pm, "Eric Youngdale" wrote:
} I have something I would like people to try if they have been
} experiencing data corruption problems with SCSI. The patch is pretty
} simple, it just disables resets being sent because of timeout - someone
} else tried something like this and found that it helped.
}
} Note that there may be error conditions which will no longer be
} recoverable with this set, but I am not sure how often this comes up.
} Right now I would like to know if this improves system reliability.

I tried your patch.
It made my error condition unrecoverable...

Of course, considering that before, when my system recovered, it usually
died minutes later with horrible memory corruption, I'm not sure that this
is worse.

Mar 19 03:20:49 nevets kernel: Configuring Adaptec (SCSI-ID 7) at IO:330, IRQ 11, DMA priority 5
Mar 19 11:19:32 nevets kernel: SCSI reset not performed...
Mar 19 11:19:32 nevets kernel: SCSI reset not performed...
[etc, until reset]

My system was pretty much locked up completely during this--at least, I
couldn't log in, and all my existing sessions were locked up.
Obviously, syslogd was working, as these messages were recorded. I
could hear it syncing as each one was written. I suppose this is a big
improvement, as in the past, the messages were NOT saved, as syslogd
died early on (memory corruption?). Syslogd writes to an IDE drive.
It's possible my system locked up 'cause the entire scsi system locked
up, and my swap is primarily on the scsi drive. (I forgot to swapoff
before testing.)

I triggered the error by mounting a cdrom and typing
dd if=/dev/sr0 of=bigfile

and it died after copying about 6M onto my scsi hd.

Steve