Forum Discussion

unkn0wnn's avatar
unkn0wnn
Level 4
2 years ago

Tape is marked as FULL everytime tape drive cleaning is requested

I have spotted that each time tape drive cleaning is requested then tape currently in the drive is marked  as FULL and only written like 1TB of various other data.

Can anybody advise why would that be? Unsure if this is an issue so this is something that I've spotted, at a time cleaning is required netbackup automatically marks tape as full prior to perform drive cleaning.

Also each time cleaning is needed I get the below alert from tape library:

Attention : Drive Warn or Crit Tape Alert flag

Event Code: 0x84 - tape alert
Element number: 0x1E, 30
Drive number: 0x04, 4
Tape Alert Flag: 0x15, 21

 

 

  • Tape Alert 0x15,21 is a cleaning required message from the drive. How often does the tape drive request cleaning ?

    Cleaning request should be rare - if you see a lot of request, the drive may be faulty.

    To be on the safe side - inspect the tape in use - is the same tape repeatedly being requested, but never written ? 

    remember you can always freeze a tape until a reasonable suspicion is confirmed or not.

    Does the tape marked full also get frozen/suspended by Netbackup ?

    Same issues posted back in 2015:

    https://vox.veritas.com/t5/Backup-Exec/Drive-Warning-or-Crit-Alert-flag-help/td-p/741123

     

    • unkn0wnn's avatar
      unkn0wnn
      Level 4

      Cleaning request is very rare, no tapes frozen etc. I am just concerned why tapes become Full each time cleaning is requested.

  • I don't think those two events are linked.

    When writing a tape, the OS see the tape as a continuous number of sectors. When the tape drives detect the EOM (end of media) marker on the tape, the tape drives send a signal to the OS that the tape is full.

    Netbackup as such doesn't measurer remaining capacity on a tape, it can't. Netbackup only measure the number of sectors written to tape (consumed capacity).

    To narrow down what is happening, look out for SCSI sense codes in the event log. A code consist of a sense key, a additional sense code (ASC) and additional sense code qualifier (ASCQ). What they mean can ben looked-up at this Wikipedia page: https://en.wikipedia.org/wiki/Key_Code_Qualifier

     

    • mph999's avatar
      mph999
      Level 6
      To follow on from Nicolais excellent reply. NetBackup does not have any understanding of tape capacity. It would, if left to its own devices, try and write to a tape forever.

      So, how do tapes become full ???

      The tape contains an ‘logical end-of-tape’ mark that was written just before the physical end of the tape during manufacturering.

      Once the tape gets to this point, the tape firmware detects it and knows there is only a small amount of tape left, so it updates the tape driver with this information. There is enough tape left to complete writing the current block, but the tape driver refuses any further data and sends a signal to the backup software informing it the tape is full.

      Put another way, NetBackup has no ability to mark a tape full by itself. So tapes marked full before they should be, is either a tape drive firmware or driver issue.