Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.carno.net.au!harbinger.cc.monash.edu.au!news.cs.su.oz.au!metro!metro!munnari.OZ.AU!news.ecn.uoknor.edu!solace!news.stealth.net!www.nntp.primenet.com!nntp.primenet.com!news1.best.com!nntp2.ba.best.com!not-for-mail
From: dillon@flea.best.net (Matt Dillon)
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: Re: Automatic reboots: Leaves no trace?
Date: 6 Feb 1997 16:56:12 -0800
Organization: BEST Internet Communications, Inc.
Lines: 51
Message-ID: <5ddujc$pmk@flea.best.net>
References: <32FA6F03.687A@cstone.net>
NNTP-Posting-Host: flea.best.net
Xref: euryale.cc.adfa.oz.au comp.unix.bsd.freebsd.misc:35336
:In article <32FA6F03.687A@cstone.net>, Paul Nguyen <pauln@cstone.net> wrote:
:>
:>------------5C77182A69E83
:>Content-Transfer-Encoding: 7bit
:>Content-Type: text/plain; charset=us-ascii
:>
:>Hi,
:>
:>I have a pentium 133 with an adaptec attach to a quantum atlas 4gig wide
:>drive. I am running FreeBSD 2.1.6.1-RELEASE. The machine stays up for
:>a month or so then it decides that it needs to reboot and it reboots.
:>But the only thing wrong is that it never leaves anything in the
:>messages file or anywhere so that we can figure what is going on. The
:>only thing it sez is that the / partition was not properly dismounted.
:>Any help would be greatly appreciated. Thankyou.
:>
:>-Paul
:>
:>------------5C77182A69E83
:...
Quantum drives have a firmware bug that blows the SCSI bus up when
FreeBSD really pushes them. It has something to do with the SCSI
command queueing.
We had to replace 10 quantums.. I think they were atlas's... We should
never have bought them in the first place but, silly us!, they were
cheaper then he new barracudas. We wound up throwing the quantums
into Windows and NT machines (which don't push the disks like FreeBSD
does) and bought barracudas (or whatever they are calling them now..
the really high speed half-height 4G drives that seagate came out
with a few months ago).
Do this from the console and see if the machine crashes:
# sync
# sync
# dd if=/dev/sd0c of=/dev/null bs=32k
(Now, if it does not crash, it is still possible that your problem
is the quantum firmware bug. If it does crash, then it's *definitely*
the quantum firmware bug).
On our machines, the above crashed the machine almost instantly...
and very consistantly. I was able to crash every machine we had
a quantum on. Note that the test is going through the buffered
disk driver. You can also try different block sizes to try to hit
the bug.
-Matt