SUMMARY --- read restore disk errors

From: Dan Williams (dan%ufnmr.uucp@eecs.nwu.edu)
Date: Tue Jul 16 1991 - 10:35:54 CDT


Sorry that this has taken so long but I was only recently able to try
the solutions suggested.

My original request was....
=========================================================================
Hello all,

We have a 3/260 work station that has been running fine for the last
two years. Just recently we have been receiving the following system
messages:

May 22 12:46:02 ufcia vmunix: xy0b: read restore (disk sequencer error)
         -- blk #20066, abs blk #36146
May 25 00:23:57 ufcia vmunix: xy0b: read restore (disk sequencer error)
         -- blk #78387, abs blk #94467
Jun 4 11:43:02 ufcia vmunix: xy0b: read restore (disk sequencer error)
         -- blk #1326, abs blk #17406
Jun 4 13:11:02 ufcia vmunix: xy0b: read restore (disk sequencer error)
         -- blk #57616, abs blk #73696
Jun 4 14:02:01 ufcia vmunix: xy0b: read restore (disk sequencer error)
         -- blk #78387, abs blk #94467

This partition (xy0b) is my machines swap space. Do I have a problem
developing with this disk? Are these messages just informative with
the system handling the problem?
==========================================================================

Many people suggested that the problem was control centered and I
should see about replacing the SMD controller. Another group
suggestted that the drive itself was in trouble (this was the way I was
leaning). However, there was a very large group that suggested that
the formatting just needed to be redone. This turned out to be the
solution. John Easton of Univ. of Minnesota had the best
explaination:

========================================================================
From: easton@atlas.socsci.umn.edu
We have had almost all hard disks "age" at around two years with a
phenomena we have dubbed "geometry drift" -- it seems to affect a disk
at least once in its life. This is probably distinct from developing
"bad spots". Our surmise is that the recording surface drifts
mechanically a tiny bit during the period of time immediately after
it's manufacture (1 to 3 years), and then settles down.

So we reformat disks as soon as they start showing the tendency to
fail -- difficulty in reading, followed sooner or later with write
failures. We have just enough disks/machines that we do not keep
records, but not so many that we completely loose track of failure
patterns.
=========================================================================

Thanks to all who replied with suggestions:

VINCE%UCONNVM@nervm.nerdc.ufl.edu
rodney@snowhite.cis.uoguelph.ca
sjh@helicon.math.purdue.edu
holle@ASC.SLB.COM
tonytran@leland.stanford.edu
sundev!ronin!kevin@Sun.COM
rwolf@dretor.dciem.dnd.ca
ken@shibuya.cc.columbia.edu
dfl@dretor.dciem.dnd.ca
seeger@manatee.cis.ufl.edu
easton@atlas.socsci.umn.edu
agcsun!asc!jerry@boulder.Colorado.EDU

and anyone else I forgot to mention.

--

Dan Williams | internet - dan@ufcia.health.ufl.edu University of Florida | uucp - Computerized Image Analysis Lab | voice - 904-392-0921 Box J-144, JHMHC | fax - 904-392-5145 Gainesville, FL 32610-0144 |



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:21 CDT