*BSD News Article 60458


Return to BSD News archive

Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!news.ecn.uoknor.edu!news.eng.convex.com!hermes.oc.com!news.unt.edu!cs.utexas.edu!howland.reston.ans.net!newsfeed.internetmci.com!EU.net!peer-news.britain.eu.net!ukc!crane.ukc.ac.uk!K.R.Marshall
From: K.R.Marshall@ukc.ac.uk (Keith Marshall)
Newsgroups: comp.unix.bsd.freebsd.misc
Subject: The wonders of power cuts
Date: Wed, 24 Jan 96 13:49:41 GMT
Organization: University of Kent at Canterbury, UK.
Lines: 65
Sender: K.R.Marshall@ukc.ac.uk
Message-ID: <804@crane.ukc.ac.uk>
Reply-To: K.R.Marshall@ukc.ac.uk (Keith Marshall)
NNTP-Posting-Host: crane.ukc.ac.uk

Hi all,

I'm just wondering if anyone has any extra ideas on this little
problem I have. As far as I can see it is probably the Adaptec card
that is slightly damaged, but I can't prove it.

We run a Viglen 486 DX4 75 PCI computer with 8 Mb RAM, 250Mb western
digital hard disk, SMC Elite ethernet card, Diamond Stealth 64
graphics card (but I'm not running X - it just came with the PC) and
an Adaptec 1542CF controller with 7 Toshiba CD-ROM drives chained.

This acts as a CD-ROM and file-server across a Sun PC-NFS network.

It's been running FreeBSD 2.0.5-R fine for several months now. It's
been extremely stable, we've had no trouble whatsoever. A couple of
weeks ago we had a power cut when some workmen cut through a set of
power cables in Canterbury. Ever since it's been crashing regularly
(every couple of hours or so) with the following error messages:

Fatal trap 12: page fault while in kernel mode
Fault virtual address	= 0x4a
Fault code		= supervisor read, page not present
instruction pointer	= 0x8:0xf01244ec
Code segment		= base 0x0, limit 0xfffff, type 0x1b
			= DPL 0, pres 1, def32 1, gran 1
processor eflags	= interrupt enabled, resume, IOPL=0
Current process		= 79 (nfsd)
Interrupt mask		= bio
Panic: page fault

syncing disks... 2 2 2 2 2 2 2 2 2 (maybe some more 2's) giving up.


Now, the process isn't always nfsd, and those sync numbers aren't
always "2". I can force a crash by downloading huge amounts of data
from a CD-ROM (i.e. by copying a huge file across NFS), but it
sometimes crashes just when doing large amounts of hard disk I/O.

I re-installed the OS to start with, using 2.1.0-R, since there was
very little stuff on there that wasn't easily replaceable and I wanted
to check that the swap space hadn't been corrupted. This had no
effect.

I then thought of the memory, so I tried changing the SIMMS with an
identical Viglen, and that didn't solve it. I disabled the external
cache, but that didn't solve it. I tried running without the Adaptec
card in, and it didn't crash, but I have no sure way of testing it
other than trying huge amounts of disk I/O so I don't know if that was
comprehensive. The adaptec internal diagnostic checks run fine, but
they only test DMA transfers. 

The only other SCSI cards I have around are Future Domain ones, which
won't work with FreeBSD, so I can't swap the card. I'm just about at
the stage of ordering a new Adaptec card to see if that's the problem,
but I was wondering if anyone else could suggest something to try.

Sorry about the length of this post - it's turned out a trifle longer
than I was expecting ;)

Keith.
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"Mummy was an asteroid, Daddy was a  | Keith Marshall 
small, non-stick kitchen utensil..." | Computing Officer, Templeman Library
  - Quiet Sun, 1975                  | University of Kent at Canterbury.