bptm hangs

[santhas]

Hi NB Gurus,

I have Solaris 10 NB Enterprise server connected to ST L700 that has 3 FC tape drives. I backup 100+ Win, 30 Unix/Linux and 30+ NDMP (tape drives are shared to a NAS filer through tape SAN).
NDMP and Media management backups run in separate backup windows. Every other day I see some bptm secessions hanging on the system trying to unload tapes, therefore, I have jobs queued up ( only one tapes drive was active). Today as soon as I killed the hung bptm secessions 3 tape drives started to process the queued jobs.
Any reason why bptm hangs while unloading tapes? How can I eliminate it?

Thanks
Santhas

 

[wileywombat]

what version of NBU ?

NBU will send periodic bptm -U's ( general unload ) in NBU 5.x to unload any unused tape drives.

Are the unloads for specific drives ? If you look at the pid of the unload did it occur when the drive was in use by a different process like a backup or restore in the bptm log ?

 

[santhas]

I am running NBU 6MP4. Every other day I have NBU bringing down the tape drives and I have to reboot the server to remove the hung bptm procs

 

[wileywombat]

So there are pending unloads for down tape drives ? sounds like a bug.

Get a copy of the bptm log with VERBOSE = 5 and a copy of the system log and call support.

 

[santhas]

Thanks for your inputs.

This week I had SUN engineer came on-site and upgraded the LTO 2 FW. Today I am going to replace one tape drive ( which has been failing nearly everyday)

I have enabled bptm log with VERBOSE=5, let see what is says.

Thanks again

 

[santhas]

We had SUN to replace one of the tape drive because we were seeing some error messages logged in messages file:

May 18 09:54:37 bk001 scsi: [ID 107833 kern.warning] WARNING: /pci@84,2000/lpfc@1/fp@0,0/st@w500104f0005e12d0,0 (st7):
May 18 09:54:37 bk001 transport rejected
May 18 09:54:41 bk001 svc.startd[7]: [ID 694882 daemon.notice] instance svc:/system/console-login:default exited with
status 1
May 18 09:54:47 bk001 scsi: [ID 107833 kern.warning] WARNING: /pci@84,2000/lpfc@1/fp@0,0/st@w500104f0005e12d0,0 (st7):
May 18 09:54:47 bk001 transport rejected
May 18 09:54:57 bk001 scsi: [ID 107833 kern.warning] WARNING: /pci@84,2000/lpfc@1/fp@0,0/st@w500104f0005e12d0,0 (st7):
May 18 09:54:57 bk001 transport rejected

But again we see the same messages today:

May 22 05:28:14 k001 scsi: [ID 107833 kern.warning] WARNING: /pci@84,2000/lpfc@1/fp@0,0/st@w500104f0005e12d0,0 (st7):
May 22 05:28:14 k001 transport rejected
n
May 22 05:37:04 k001 scsi: [ID 107833 kern.warning] WARNING: /pci@84,2000/lpfc@1/fp@0,0/st@w500104f0005e12d0,0 (st7):
May 22 05:37:04 bk001 transport rejected

I have to shutdown netbackup, reboot the server, unload the tapes using mt command and then re-start the netbackup to continue ( restart the backup jobs). Am I missing something here?. Anything wrong in my config? I am using SUN branded Emulex LP10000 card on Solaris 10.

Thanks