Forum Discussion

rsm_gbg's avatar
rsm_gbg
Level 5
11 years ago

Tape mount problems

I got a NBU master server and a media server, the media server has a SL48 robot with 2 SCSI-LTO4 All my backups are running just fine, mounting tapes all the time, except my Vault_catalogbackup. ...
7.1.x and Earlier
Backup and Recovery
Error messages
NetBackup
Solaris
Troubleshooting
  • rsm_gbg's avatar
    rsm_gbg
    10 years ago

    Hi,

     

    This post got a bit sidelined...

    Yes I think I solved it, think...

    There was a stray sticker that had went off a tape inside the library.
    It was discovered when a tape was ejected without a sticker!
    After pulling the library apart and removed the sticker the problem dissappeared.

    For what ever reason the sticker must have been in a very odd spot to cause such strange behaviour.
    And why just the Vault?
    Maybe vault does a scan of all the tapes and thus hit the sticker "bug"  all the time?
    One could only speculate!

    - Roland

Marianne's avatar
Marianne
Level 6
11 years ago

All I can suggest is that you log a support call with Symantec.

"Something" is happening between the master and media server at this point that causes network failure:

22:00:31.027 [15297] <2> bpbrm main: ADDED FILES TO DB FOR ipndms_1403956823 1 /opt/openv/db/staging/DARS_DATA.db
22:01:26.406 [15297] <16> bpbrm main: from client ipndms: ERR - Cannot write to STDOUT. Errno = 32: Broken pipe

 

We see that a successful backup reports success within 3 seconds:

23:08:48.709 [5251] <2> bpbrm main: ADDED FILES TO DB FOR ipndms_1403960921 1 /opt/openv/db/staging/DARS_DATA.db
23:08:51.902 [5251] <2> bpbrm main: client ipndms EXIT STATUS = 0: the requested operation was successfully completed

 

So, unless there is a real network failure between master and media server, all processes will need to be logged and checked. 
I am not sure about all of them - Support will be able to tell you which logs are needed with increased logging levels.