Forum Discussion

Scott_Lundberg's avatar
Scott_Lundberg
Level 4
16 years ago
Solved

Tape is unloaded after every job

Environment:
NB 6.5.3.1
Tandberg Robotic library  Magnum 224
Exabyte LTO3 drive in the library
SCSI U-320 connection
Windows 2003 Std R2 with latest Windows updates (as of 2009-03-25)

Problem:
During an MSSQL backup where several databases on the same server are backed up, the tape drive is repeatedly unloaded and loaded after each and every database is backed up.  Looking in the Tandberg interface shows the same tape being moved from the Drive, to the robot slot and then back to the drive again.
Netbackup appears to wait the media unmount delay (I have tried 180 and 300 seconds, no change) before reloading the drive, but it unloads the drive immediately after the current database is written.  
This sounds similar to the issue in bug http://support.veritas.com/docs/283280, but this was resolved...

Anyone else seeing this problem?

Thanks,
Scott
7.1.x and Earlier
Backup and Recovery
Configuring
Database
NetBackup
  • Scott_Lundberg's avatar
    Scott_Lundberg
    16 years ago
    After dealing with a failed robot and a failed drive and 4 weeks of shipping bad products, we have finally got a working library again...

    In the meantime, I opened a support call with Symantec and after checking several settings, it appears that the issue is related to the Global Attributes -> Maximum Jobs per client.  After changing this from 1 to 99 and then using the client settings to limit other clients besides the SQL server to 1, the issue with multiple rewinds, unloads and loads during the same SQL server backup is solved. 


  • J_H_Is_gone's avatar
    J_H_Is_gone
    Level 6
    16 years ago

    maybe it is an issue with tape 186APZ - freeze that tape so netbackup will not try to use it, and see if it goes for the other tapes quicker.

  • Andy_Welburn's avatar
    Andy_Welburn
    Level 6
    16 years ago

    "...maybe it is an issue with tape 186APZ - freeze that tape so netbackup will not try to use it, and see if it goes for the other tapes quicker.

    ..."

    I agree with Judy. It appears that there are issues with this tape. Eject the tape & check it out for damage or maybe incorrectly positioned in slot (had one or two upside down before now, think our oppo's were sleepwalking ;) )
  • Scott_Lundberg's avatar
    Scott_Lundberg
    Level 4
    16 years ago
    Well this morning I arrived and checked on the backups and the servo in the robot has failed...  Not sure if it is coincidence or the cause of the problems.  I will have to wait until a new library shows up to further test.
  • Scott_Lundberg's avatar
    Scott_Lundberg
    Level 4
    16 years ago
    After dealing with a failed robot and a failed drive and 4 weeks of shipping bad products, we have finally got a working library again...

    In the meantime, I opened a support call with Symantec and after checking several settings, it appears that the issue is related to the Global Attributes -> Maximum Jobs per client.  After changing this from 1 to 99 and then using the client settings to limit other clients besides the SQL server to 1, the issue with multiple rewinds, unloads and loads during the same SQL server backup is solved.