All jobs abort when "I/O bus was reset" occurs

[Guest_imported]

My environment is purely Windows environment with mixed of NT 4.0 and W2K servers, total over 20 server. We are using SAN and NetBackup 3.41 as our backup solution. The solution implemented few months but never get smooth. We encountered a big problem on "I/O bus reset" issue in a SAN environment.

We found all running jobs was abort at the same time with "Media Write Error #84". After examine the bptm debug log, I found a strange messages "WriteFile failed with: The I/O bus was reset.". The same error messages found on all effected media servers.

After look at all server event log, we suspect the I/O bus reset may init by one of NT server then we replace the HBA.

However, in a large scale SAN environment, we could not avoid I/O bus reset since it is normal recovery action in SCSI technology. Any system, device attached to SAN and even the tape drive will trigger bus reset.

I would like to know are there any tuning parameters in W2K/NT or NetBackup can increase the tolerence level such as timeout value that could help to min. the impact. Or can we disable/ignore the "I/O bus reset" ?

 

[sdeb]

I've already encountered that kind of problems in nearly the same SAN configuration. Problem was a conflict between Netbackup and an other backup software who tried to take control of Netbackup tape devices.

Be sure that you've disabled "RSM" under win2K (native backup tool) and that there's not any other backup software (I personally met this problem with a MEDIA server that was running Arcserve ) running.

Hope this helps.