Shalom, Managers !
Thank you very much, Managers, for your help and advices. My original message
was:
> We have a network of many SS-1, SS-1+, SS-2, SLC and ELC. I have an
> EXAByte-8200 connected to one of the SLCs and run every night backup
> procedure for all SUNs using
>
> rsh -n hostname /etc/dump 1ucsf 6000 myhost:/dev/nrst1 /filesystem
>
> in the backup script. Cron helps me every night to run it.
> Everything was suspiciously good, but I have upgraded my dump host to
> 4.1.1. From this day a very strange (from my unexpirienced point of view)
> thing started. For ONLY THREE HOSTS of my backup list I get almost every time
> unpredictible error messages in the form:
>
> st1: Error for command 'write', Error Level: 'Fatal'
> Block: 49 File Number: 1
> Sense Key: Media Error
> st1: Error for command 'write file mark', Error Level: 'Fatal'
> Block: 49
> Sense Key: Media Error
>
> I have checked everything I only can: tapes, drive, procedure etc. The only
> thing which can help me is simply to take this three hosts out from the list.
> Then everything again is OK for the same tapes, which gave me the message.
> This three hosts are usual SS-2 running 4.1.1 without some special points
> I can mention. All other SS-2 can be processed without problems. The same
> command works OK for these three SS-2 when I've tried to write onto cartridge
> tape instead of EXAByte. I have also no visible network problem.
>
> Maybe someone can help me. Thank you very much.
There was a lot of messages with your good advices, Managers. The ideas could be
subdivided onto several groups:
A. "Something not good with your backup procedure, Greg !"
Maybe it's my fault, I have not given the full description of the
procedure and the script. Nevertheless, I've got several backup
scripts and several good ideas how to improve my procedure. It is
not really my current problem, because my script used to work fine
before my problem started, but sure there was a lot of very
interesting points.
B. "Something not good with 4.1.1 (or SS-2) !"
There were advices to put some timing delay in the backup script and
some other points like these. Sorry, nothing helps. Now I understand,
that 4.1.1 has nothing to do with the problem, because now 4.1 gives
me the same results. Maybe this is only coincidence, that problem
has started after upgrading to 4.1.1.
C. "Something not good with your SCSI cables !"
I have checked it by replacing cables three or four times, but with
the same result, sorry.
D. "Something not good with your EXAByte !"
It looks like this. After reading the mail and even thinking (!) a little
I've understood that my 8200 should be sent to our local EXAByte
representative as it needs head checking (and maybe repairing) and
also for firmware upgrade. My special thanks to Kevin McElearney from
MIT Lincoln Laboratory for the good idea to check the situation
with the same data but with dd, tar etc. and also with the same data
but from SS-1 or from 4.1 host. THE SAME RESULT ! So it looks
absolutely like the situation described by Sydney S. Weinstein
from Datacomp Systems, Inc.:
"I have seen these many a time, and the only cure I have found is to
have the 8200 replaced. Some data patterns cause the drive to not
write the tape correctly as the heads wear. It is data related..."
So let's see the results after repairing the 8200.
A lot of thanks to all of you, Managers, who kindly helped me (this is a very
long list to put it here).
Gregory.
============================================================================
Gregory Gutarts greg@orbot.co.il
System Administrator voice: 972-8-433533
Orbot Systems Ltd. fax: 972-8-438769
Industrial Zone, P.O.Box 215
70651 Yavne, Israel
============================================================================
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:06:34 CDT