I have an intermittent problem when running a nightly backup.
The job uses tar to copy oracle database files to tape.
Basically we use an HP Netserver LH 3 system with megaraid drivers. (I have updated the driver)
HP DLT 40
I am on kernel version 2.4.2-2smp.
The only message that may be related is "ISR called reentrantly!!" This message appears when ever we use our DLT.
The symptoms are.
(a) The backup stops for no apparent reason with no messages.
(b) Everything else runs fine.
(c) Anything that access the /proc file system in association with the job hangs.
ie ps or top will hang or you cat the backups pid.
(d) The DLT appears not to be used any more. ie "mt -f /dev/st0 reoffl" works.
(e) Sometimes after many hours it may all of a sudden come back to life.
(f) Usually a reboot is required to clear the problem.
Is it likely to be a scsi or a kernel problem?
Hopefully there is enough information to start with.
Thanks in appreciation of your helpful replies.
Barry
