Forum Discussion

abhas_mangal's avatar
abhas_mangal
Level 3
17 years ago
Solved

Slow Tape Drive Speed.

Hi,

Slow Tape Drive speed has been observed daily that is causing the Staging Images created using DSSU Staging policy to have much longer time to copy to the Tapes. Tape Drive speed only varies between 5-10 MB/sec that is below par of the normal Tape Drive Speed. I have tried to find out the tuning parameters and have applied the following changes to the following parameters:
SIZE_DATA_BUFFERS: 262144 (256k)
NUMBER_DATA_BUFFERS: 32
Still no better changes have been observed due to above tuning parameters.
The Tape Drive Details being used are of type HCART3. There are two Tape Drives being used that are HP.ULTRIUM.SCSI. The Tape Library used is C4 Tape Library.
The Netbackup version is 6.0 MP5.

Please let me know if there is any further information any one requires for understanding the problem.

Thanks,
Abhas
7.1.x and Earlier
Backup and Recovery
Configuring
Features
NetBackup
  • Marianne's avatar
    Marianne
    17 years ago
    TechNote http://seer.entsupport.symantec.com/docs/273532.htm :

    Slow performance on Disk Storage Units and Disk Staging Storage Units

    Details:
    1.  Increase the data buffers used by the VERITAS NetBackup (tm) disk manager process (bpdm) on master and media servers.
    The default buffer size for bpdm (disk backups) is 262144 (256K).

    Create  /usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS_DISK.
    Add the value 1048576 into the file.
    Create  /usr/openv/netbackup/db/config/NUMBER_DATA_BUFFERS_DISK.
    Add the value 16 into the file.

    No restart is needed for NetBackup after these files are created.
    If Status 89 is now seen on the backups, increase the system's shared memory value in the kernel settings and then reboot the system.

6 Replies

  • Marianne's avatar
    Marianne
    Level 6
    17 years ago
    Have a look at this TechNote:
    http://seer.entsupport.symantec.com/docs/273532.htm

  • techarch27's avatar
    techarch27
    Level 3
    17 years ago
    Check the bpkar output ..... if the throughput of bpkar and actual backup job differs immensly..... there is an issue with the hardware or connectivity.......
    Have the bpkar values in hand to getter a better picture.
  • Marianne's avatar
    Marianne
    Level 6
    17 years ago
    Have another look at original post - the problem is during staging from disk to tape....
  • Nicolai's avatar
    Nicolai
    Moderator
    17 years ago

    There are a few tings to check first:
    1. Are you running backup while staging ?. Writing and reading give very bad transfer performance.
    2. What's the disk/array capacity, have you verified the array can handle >80MB/sec. Try using vxbench,Bonnie++ or alike.
    3. Bad tape drive, seen this before - If the tape drive is close to be worn out, it will perform multiple re-write because the read after write failed. If more drive are limited to 10-15MB/sec, I don't think this is you're issue.
  • Marianne's avatar
    Marianne
    Level 6
    17 years ago
    TechNote http://seer.entsupport.symantec.com/docs/273532.htm :

    Slow performance on Disk Storage Units and Disk Staging Storage Units

    Details:
    1.  Increase the data buffers used by the VERITAS NetBackup (tm) disk manager process (bpdm) on master and media servers.
    The default buffer size for bpdm (disk backups) is 262144 (256K).

    Create  /usr/openv/netbackup/db/config/SIZE_DATA_BUFFERS_DISK.
    Add the value 1048576 into the file.
    Create  /usr/openv/netbackup/db/config/NUMBER_DATA_BUFFERS_DISK.
    Add the value 16 into the file.

    No restart is needed for NetBackup after these files are created.
    If Status 89 is now seen on the backups, increase the system's shared memory value in the kernel settings and then reboot the system.