General discussion

  • Creator
    Topic
  • #2094862

    OpenVMS SLS Backups Hang Tape Drive

    Locked

    by buckshot ·

    We have been having a problem with our OpenVMS VAX backups ever since we upgraded our systems to OpenVMS V7.2. In summary:

    -The problem only occurs on VAX systems.
    -The associated DLT drive shows in MOUNTED DISMOUNT CACHED ALLOC FOREIGN device status and a volume status of ?position lost? (in many of the cases, but not all).
    -When the batch entry is deleted it goes into ABORTING state and will not terminate.
    – To get rid of the entry we do a STOP/RESET on the queue, but we have to wait until other jobs in that queue have completed or we will terminate good backups as well
    – In some cases the associated batch job goes into RWAST or SUSP state. Although we have not seen that condition since the most recent patched TU and DU drivers were installed
    – If the batch job is deleted with STOP/ID, the drives remained allocated to a non-existent process which causes SLS problems during some drive selection activity. If this occurs we have to reconfigure SLS and restarted it.
    – We found early on with this problem that restarting the HSJ50 controlling the tape drive would sometimes release the drive, but in about 50% of the cases would crash the node where the backup was running.
    – After a reboot the problem returns almost immediately, but only consumes two or so drives per night with most drives being used between 20 to 50 times during the night (the average tape drive usage is 4.1 hours on a weekend night and 2.8 hours on a week night).

    ANY IDEAS? We have Compaq engineering looking at this, but surely someone else must be running OpenVMS 7.2 on a VAX and doing SLS backups.

All Comments

  • Author
    Replies
    • #3834778

      OpenVMS SLS Backups Hang Tape Drive

      by lo ·

      In reply to OpenVMS SLS Backups Hang Tape Drive

      Hi

      Been a long time since I did Sys Mgt on VMS. I will take your question to my Sys Mgr.

      On STOP/RESET, you can do STOP/NEXT to stop other jobs from starting. Executing jobs will complete. When the hung job is the only executing, do STOP/RESET

      There is also a /REQUE, not sure if it works with /RESET but that may stop loss of pending or executing jobs (job needs to have been queued with /RESTART I think)to not have to wait. lo

      • #3836374

        OpenVMS SLS Backups Hang Tape Drive

        by buckshot ·

        In reply to OpenVMS SLS Backups Hang Tape Drive

        Doesn’t seem to be a problem with the STOP/RESET. Actually, Compaq just today sent us a patched TUDRIVER we intend to try. They report one other customer in the world that had problems with a tape drive hanging like this. I feel so lucky to be amember of such an exclusive club!

    • #3849378

      OpenVMS SLS Backups Hang Tape Drive

      by buckshot ·

      In reply to OpenVMS SLS Backups Hang Tape Drive

      This question was closed by the author

Viewing 1 reply thread