During any backup job/policy that I have configured I notice that both of the drives in my tape library are going "down" at least once every half hour and stay that way until an operator goes into Device Monitor to "up" the drive. This is queuing up all of the concurrent jobs. This is especially painful on a weekend backup with multiple servers to backup. Has anyone experienced this? I have spoken to HP regarding a possible hardware/firmware issue and they have found no errors. Symantec has not responded to my case as of yet.
Here is my environment: 1 - Netbackup 6.0 MP4 master/media server on an Server 2003 Enterprise Edition platform 1 - HP MSL 4048 Fiber Channel tape library
Yep most likely drive mapping use shutdown netbackup or make sure no backups are running then use robtest to move a tape into the and then mt -f drivepath status to verify it is movig the tape into the correct drive. Also check that both drives and tapes are of the same types ie AIT(1) pay attention to the number after the drive type. Drive and Tapes have to match. use scan and sgscan to verify confige.
The process to absolutely verify that the drive path mapping is correct on ... Also reference technote 193280 for more details on the robtest utility and ... seer.support.veritas.com/docs/264193.htm
Dennis and Bob make some excellent points, here. Hint* =) I know I've had this same issue before and it was a mapping issue. Robtest will help you out. In my case, however, I ended up removing and reinstalling the drives in both netbackup and in windows on my master media server, on which the drives were attached. The issue arose when I had replaced some tape drives on the backup servers.
The gentlemen are probably right as your problem comes in pair of drives. Do consider their suggestion and keep us posted if possible. Another quick way to verify whether your drive mapping is mixed up is to down all other "good" drives except these two (during your quiet time) and initiate a backup. If you see that the backup job try to mount a tape and ended with mounting timeout (meanwhile you can verify physically whether the media mounted to another drive instead), then you can be sure that's the problem.