Dear Tru64ers,
Yesterday our regular backup job failed, and I am unsure whether to blame
software or hardware. The tape was written successfully, but one of the
files causes errors when it is reread.
We write a whole series of backup files to DLT tape using commands like
ssh -n $remhost \
"/sbin/vdump -$level -Cu -b 60 -f - $fileset" \
| dd of=/dev/nrmt0h obs=60k
This part worked fine as usual with messages:
vdump: Dumped 221647992 of 221647992 bytes; 100.0% completed
vdump: Dumped 1 of 1 directories; 100.0% completed
vdump: Dumped 350 of 350 files; 100.0% completed
vdump: Dump completed at Mon Jul 17 23:37:46 2000
263200+0 records in
2193+1 records out
But the backup job also reads the tape
back to confirm it is readable, using commands
/sbin/vrestore -t -f /dev/nrmt0h >/dev/null
On one of the tape files this generated an error:
vrestore: unable to read from device '/dev/nrmt0h'; incomplete read 20480
< 61440
vrestore: unable to restore block <2193>; corrupt block
/usr/adm/dumps/users/weekly.dump[178]: 29510 Memory fault
This is repeatable: by repeating vrestore -t (or vrestore -l) on that file
of that tape the same thing happens.
But what puzzles me is that UERF shows no errors at all, and there are
also no errors in /var/adm/messages. I have cleaned the drive and still
cannot read back that file successfully.
So do I throw the tape away? Or is this an intermittent software problem
with vdump or dd or vrestore?
We are running Tru64 4.0E patch kit 4 on both the client and server
machine.
Thanks for any help,
Bob
==============================================================
Bob Vickers R.Vickers_at_dcs.rhbnc.ac.uk
Dept of Computer Science, Royal Holloway, University of London
WWW:
http://www.cs.rhbnc.ac.uk/home/bobv
Phone: +44 1784 443691
Received on Wed Jul 19 2000 - 09:55:45 NZST