NetBackup console is asking different tape (which is not in the tape drive)

[Guruh_Sulistiyo]

Hi,

I have several sites in NetBackup. Each site will use either single tape drive or TL2000 to backup the image to the tape. For some of the sites using single tape drive. I notice there are 1 or 2 site having problem with the backup image to the tape (D2T).

Let say about California site. For this site we have 4 tapes which are LTO5 tape. I soft label the tape with name as CA0001, CA0002, CA0003 and CA0004. I created new volume pools (named it as CA_Volume_Pool) for this site and move all those 4 tapes to this volume pool.

What I don't understand is, sometime in the blue moon. The D2T process is not working properly for this site. It's stuck in the process of writing and asking different tape altough there is a tape (which is less kilobytes written to the tape) inside the tape drive. Let say CA0002 is inside the tape drive. NetBackup console is keep asking for CA0001. Even after I denied the request or resubmit the request in the device monitor, it's still asking for CA0001.

Why NetBackup is asking for different tape altough there is a tape inside the tape drive (and still have much space)? Is there a way to force NetBackup to proceed with the D2T process with the tape we have inside the tape drive?

Due to this problem. Our tape rotation is a bit mess up. Because sometime NetBackup console is asking previous tape which we already sent out off site. 

We have around 15 sites using single tape drive. Rest of the site ares okey. NetBackup console never ask for pending tape to be inserted in the device monitor. It always write to the tape which is inside the tape drive, and the backup is always running smoothly everytime our field service do the rotation of the tape. But I am not sure why only these 2 sites having this problem.

Any advice will be much appreciate.

[mph999]

media/ drive selection is not logged in bptm, well, you see the result, but the decision is made in 'mds' log which logs into nbemm.

vxlogcfg -a -p 51216 -o 143 -s DebugLevel=6 -s DiagnosticLevel=6  (mds)

vxlogcfg -a -p 51216 -o 144 -s DebugLevel=6 -s DiagnosticLevel=6  (da)

... then after the job has run, gather the nbemm log

vxlogview -p 51216 -i 111 -d all -t 01:00:00  (this gets the last 1 hr of log, relative to when you run the command - so if the job started 2 hours ago, you'd need to adjust the time to go back to just before the job started)

If you then search for 'allocateTwin' in the file, you will be in approximately the right area in the log.

Example

08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [fill_bsr] MEDIA: id = TAPE03, key = 4000003
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [fill_bsr] DRIVE: name = HP.ULTRIUM1-SCSI.000, path = /dev/rmt/0cbn, secondary_path = , key = 2000002, ndmp_host_key = 0
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [fill_bsr] STU: name = womble-hcart, master server = womble
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [fill_bsr] DISKGROUP: id = *NULL*, key = 0
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [fill_bsr] DISKVOLUME: media_id = *NULL*, key = 0
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [fill_bsr] DISKMOUNTPOINT: mount_point = *NULL*, key = 0
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [allocateTwin] EXIT INFO:
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [allocateTwin] rbActions = 0
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [allocateTwin] alloc_med_list = 4000003
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [allocateTwin] alloc_drv_list = 2000002
08/31/16 07:33:00.327 [Debug] NB 51216 mds 143 PID:29461 TID:9 File ID:111 [jobid=814] 1 [allocateTwin] EXIT STATUS = 0 (EMM_ERROR_Success, Success)