1. Forum Rules (PLEASE CLICK HERE TO READ BEFORE POSTING) Click the link to access ADSM.ORG Acceptable Use Policy and forum rules which should be observed when using this website. Violators may be banned from this website. This message will disappear after you have made at least 12 posts. Thank you for your cooperation.

TSM backup Tape pool Error

Discussion in 'TSM Operation' started by nalinsl, Jul 11, 2012.

  1. nalinsl

    nalinsl New Member

    Joined:
    Nov 3, 2011
    Messages:
    20
    Likes Received:
    0
    Hi all,

    I'm trying to adjust the maintenance script of out TSM 5.5 server to backup all storage pools successfully but I'm getting interruptions when someone tries

    to restore some data.

    We have only 4 drives in 2 different libraries, COL and COPY tapes are residing in one library and NOCOL tapes are in the other library.

    This is the command in the maintenance script for backing up T_COL to T_COPY ( both in the same library and using 2 drives )


    ---------------------------------------------------------------------------------------------------------------------
    Command: backup stg T_COL T_COPY maxpr=1 WAIT=YES
    ---------------------------------------------------------------------------------------------------------------------

    My problem is even with the WAIT=YES parameter the backup process get interrupted when someone tries to restore data from T_COL


    This is the error message I'm getting
    ---------------------------------------------------------------------------------------------------------------------
    07/12/2012 11:38:58 ANR1440I All drives in use. Process 1817 being preempted
    by higher priority operation. (SESSION: 24301)
    07/12/2012 11:38:58 ANR1213I Backup process 1817 terminated for storage pool
    T_COL - process canceled. (SESSION: 23980, PROCESS: 1817)

    ---------------------------------------------------------------------------------------------------------------------

    Hope someone can help me to get out from this puzzle.

    Thanks

    Nalin
     
  2.  
  3. rmazzon

    rmazzon Senior Member

    Joined:
    Dec 6, 2004
    Messages:
    207
    Likes Received:
    17
    Occupation:
    TSM Managed Services Lead
    Location:
    NC
    The error results from the relative priorities TSM assigns to these tasks. The default priority of an admin schedule is 5. Restores are a slightly higher priority and thus will preempt the process using the drives the restore process needs. Are you running the backup via the TSM scheduler? If so you can update the admin schedule by adding "priority=1".

    If you are running this outside of the scheduler I think your only recourse is to use the NOPREEMPT option in the dsmserv.opt file.

    From the Admin Guide

    Mount Point Preemption
    The following are high priority operations that can preempt operations for a mount point:


    Backup database
    Restore
    Retrieve
    HSM recall
    Export
    Import
    The following lists operations that can be preempted and are listed in order of priority. The server selects the lowest priority operation to preempt, for example reclamation.


    Move data
    Migration from disk to sequential media
    Backup, archive, or HSM migration
    Migration from sequential media to sequential media
    Reclamation
    You can disable preemption by specifying NOPREEMPT in the server options file. When this option is specified, the BACKUP DB command is the only operation that can preempt other operations.
     
    nalinsl likes this.
  4. nalinsl

    nalinsl New Member

    Joined:
    Nov 3, 2011
    Messages:
    20
    Likes Received:
    0
    Hi rmazzon , Thank you for the detailed answer, anyway with 'WAIT=YES' parameter , shouldn't it pause the backup process for a while until the restore job is completed and then resume again ? or is there anyway of doing that ? - Thanks
     
  5. rmazzon

    rmazzon Senior Member

    Joined:
    Dec 6, 2004
    Messages:
    207
    Likes Received:
    17
    Occupation:
    TSM Managed Services Lead
    Location:
    NC
    Some processes that depend on a threshold. like migration, should restart automatically. But a backup stgpool command will not once TSM cancels it. If you encounter this issue regularly you might want to consider setting up a more complex backup stg schedule along these lines. Note this example is dependent on time of day windows and if the backup stgp processes is already running, not on the number of drives, so you would need to modify it to look at drive use, perhaps from the drives table looking for the required number of empty drives.

    Admin sched

    Schedule Name: BKSTG-TAPE01
    Description: Run automated BKSTG-TAPE01 Check and Execute script hour
    Command: run BKSTG-TAPE01 4

    script
    select server_name from status where hour(current_time) not in (12,13,14,15)
    if(rc_notfound) goto exit
    select * from processes where process='Backup Storage Pool' and status like 'Primary Pool TAPE01%'
    if(rc_notfound) backup stg tape01 copy01 maxpr=$1
    exit:
    exit
     
    nalinsl likes this.
  6. moon-buddy

    moon-buddy Moderator

    Joined:
    Aug 24, 2005
    Messages:
    6,230
    Likes Received:
    279
    Occupation:
    Electronics Engineer, Security Professional
    Location:
    Somewhere in the US
    Your issue will always be a pain on your back. Restores will always clobber your backups and you cannot do anything about it unless you give backups high priorities. But you don't want to do that.

    My two cents, increase the number of tape drives and assign unqiue tape drives per library. I know, easier said than done. However, this is the prudent way.
     
    Last edited: Jul 16, 2012
    nalinsl likes this.
  7. nalinsl

    nalinsl New Member

    Joined:
    Nov 3, 2011
    Messages:
    20
    Likes Received:
    0
    Thanks rmazzon and moon-buddy ! Your replies convinced me enough to go with more drives with a hardware upgrade.
    I only have 2 drives allocated both for Col and Copy pools and about 50 clients registered under the TSM server. So its high likely that someone will do a restore at least once every 2 days which means a failed Tape pool backup.

    I will push for a new tape library with more drives in our next hardware upgrade. Thanks Guys !!!
     

Share This Page