How can I diagnose this?

From: Rainer Mager (rvm@gol.com)
Date: Tue Feb 15 2000 - 09:54:43 EST


Hi all,

    I'm facing a problem that I just can't make any progress against.
The problem is that my system just
spontaneously freezes. There is no information in the syslog, no core,
panic (actually there might be a panic but since I'm usually in X I
can't see it), anything. How can I possibly try to fix this?

    I think that the problem may be SCSI related. I used to sometimes
get strange SCSI message in my syslog a little while before the problem.
For a while I even thought that the problem was introduced after the
2.2.9 kernel since I never had the problem in 2.2.9 but I did in every
2.2.x kernel after that. I'm currently running 2.2.14 and I do see the
problem (but no strange SCSI messages). Unfortunately, I can't reliably
reproduce the problem so even if I go back to 2.2.9 and have a 6 month
uptime that doesn't mean the problem is solved.
    Another reason I believe it is SCSI related is because I don't think
I've ever had the problem if SCSI wasn't installed (i.e., my aic7xxx
modules wasn't loaded). Since my main disk is currently an IDE drive
this is reasonably easy to do.

    In terms of frequency, the problem sometimes happens twice in quick
succession. For example, today I started up Netscape (just as I always
do) and started of VMware at the same time, and then while using NS it
froze. I reset my machine and during the fsck of my big IDE drive it
froze again. During that fsck it was also resyncing my SCSI RAID discks
(yes, I have all the latest 2.2.14 RAID patches in). This happened again
when I rebooted again. I then rebooted a third time, disabled my RAID
disks before they started syncing and could boot fine. I now have all
SCSI unloaded and seem ok.
    On the other hand, I've had week long uptimes with SCSI and RAID
running the whole time.

    I'm running a Dual PIII 600E system. I have a built-in AIC-7895 SCSI
chip which gives me 2 SCSI busses.

    Any ideas or suggestions on how to debug this?

Thanks,

--Rainer

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Feb 15 2000 - 21:00:29 EST