Forum Discussion

anil_kumar731's avatar
12 years ago

Backup consuming more time.

Hi,

 

  i have a Master server running on Solaris10 and Netbackup version is 7.0.1 at my Data centre. The master is also the media server . We have six drives configured for the server. All being  HP.ULTRIUM3-SCSI drives.

I have backup policy running with 32 multiplex with multiple windows clients. When I start the backup, I find the backup will start after 1 to 2 hrs.

The backup gets struck at "estimated 83902174 kbytes needed" .Kindly provide solution for the same.

Kindly find the below detail status for the backup:

24 Nov, 2012 7:37:35 AM - requesting resource T200001-hcart3-robot-tld-1
24 Nov, 2012 7:37:35 AM - requesting resource T200001.NBU_CLIENT.MAXJOBS.COLLECTIONSRV
24 Nov, 2012 7:37:35 AM - requesting resource T200001.NBU_POLICY.MAXJOBS.EVSQLbkp
24 Nov, 2012 7:37:37 AM - granted resource  T200001.NBU_CLIENT.MAXJOBS.COLLECTIONSRV
24 Nov, 2012 7:37:37 AM - granted resource  T200001.NBU_POLICY.MAXJOBS.EVSQLbkp
24 Nov, 2012 7:37:37 AM - granted resource  KB0087
24 Nov, 2012 7:37:37 AM - granted resource  HP.ULTRIUM3-SCSI.010
24 Nov, 2012 7:37:37 AM - granted resource  T200001-hcart3-robot-tld-1
24 Nov, 2012 7:37:37 AM - estimated 83902174 kbytes needed

  • Anil,

    first we need to sort out if multiplexing is the reason for the dealy or not,

    for testing , trigger one policy without multiplexing and see how it moves, if that moves fine..

    we need to concentrate on perfromance tuninng...

  • I was going to say, 32mpx is very high, why have you set it this high?

  • 32 multiplex is the Max limite, and you are using it on max limite..

    delay might be expected...

    do you really in need to using 32 multiplex? as per your info, its taking long time to allocate the resources for the job, (32 multiplex is one of the reasons)

    how about the jobs with no multiplexing...?

    if you dont have no mutiplexing jobs, try one for testing...

    your detail status not showing compleate log, can you post complete job detail stuats , it can show the job start times allocation times and all..

  • hi,

     

             We have around 150 servers out of which few are Solaris , AIX and the rest are Windows.

    We have 205 policies, out of which 17 policies are fired at once for different clients .All of them are using the same tape at a time for backup and the multiplex value is set to 32.

     

    Kindly suggest a best way to configure a policy using multiplexing.

     

  • Have you considered moving backups so they don't all 'fire' at the same time?

     

  • Anil,

    first we need to sort out if multiplexing is the reason for the dealy or not,

    for testing , trigger one policy without multiplexing and see how it moves, if that moves fine..

    we need to concentrate on perfromance tuninng...

  • Don't use 32 for a start ..

    Restore performance will be horribly slow.  Generally a MPX value of 4 is receomended as a good balance between backup performance and restore performance.

  • Hi,

    The backups without multiplexing are working fine.

    As we have two other media servers , each having 6 drives. The backup taken by them are non multiplexing and they are working fine.

  • Anil,

     

    you need to consider about the Tuning your master server.

    first sort out how much of data that you are taking from that media server.

    check avaliable backup window. 

    consider changing backup start times.. 

    make sure your proper exclude list, 

    try to reduce backup data size..

    it goes on..

  • Indeed, if backup performance is fine, then schedule your backups to run at different times. Multiplexing is good when client or network performance is stopping the tape drives from writing at optimum speeds by incorporating multiple clients data to keep the drive/tape moving along.