Netbackup buffer tuning question

[MickBoosh]

Hi,

If I don't have the SIZE_DATA_BUFFERS & NUMBER_DATA_BUFFERS files does that mean I am using the default values?

If so, what are they? I am experiancing slow duplication from disk to tape, and I am trying to troubleshoot. See snippet below, I don't know if it's normal for these delays.

Windows 2003 NBU 7.5.0.3

22/01/2013 13:03:59 - awaiting resource XXXXXSTU-3Drives - No drives are available

22/01/2013 13:06:31 - granted resource PJ4762

22/01/2013 13:06:31 - granted resource HP.ULTRIUM3-SCSI.000

22/01/2013 13:06:31 - granted resource XXXXX-STU-3Drives

22/01/2013 13:06:33 - Info bptm(pid=7424) start            

22/01/2013 13:06:33 - started process bptm (7424)

22/01/2013 13:06:35 - Info bptm(pid=7424) start backup           

22/01/2013 13:06:36 - Info bpdm(pid=11036) started            

22/01/2013 13:06:36 - started process bpdm (11036)

22/01/2013 13:06:36 - Info bpdm(pid=11036) reading backup image          

22/01/2013 13:06:36 - Info bpdm(pid=11036) using 30 data buffers         

22/01/2013 13:06:36 - begin reading

22/01/2013 13:06:38 - Info bptm(pid=7424) media id PJ4762 mounted on drive index 0, drivepath {2,0,1,0}, drivename HP.ULTRIUM3-SCSI.000, copy 2

22/01/2013 13:06:38 - Info bptm(pid=7424) INF - Waiting for positioning of media id PJ4762 on server XXXXX for writing.

22/01/2013 13:21:21 - end reading; read time: 00:14:45

22/01/2013 13:21:21 - begin reading

22/01/2013 13:22:59 - Info bptm(pid=7424) waited for full buffer 47068 times, delayed 61934 times    

22/01/2013 13:23:00 - end reading; read time: 00:01:39

22/01/2013 13:23:01 - Info bpdm(pid=11036) completed reading backup image         

22/01/2013 13:23:04 - Info bpdm(pid=11036) using 30 data buffers         

22/01/2013 13:23:04 - begin reading

22/01/2013 13:37:15 - end reading; read time: 00:14:11

22/01/2013 13:37:15 - begin reading

[V4]

Below must help you in setting sizes

Default Shared Buffers

http://www.symantec.com/docs/HOWTO56152

Check Mark's solution on how to 

https://www-secure.symantec.com/connect/forums/how-data-buffer-size-and-buffer-number-effect-netbackup-backup-and-restore-performance#comment-6479631

[watsons]

"waited for full buffer 47068 times, delayed 61934 times"

When you see the above messages, it means media server is waiting for full buffer to write into for that many times, increasing your buffer size does not always help in this case. You will have to find out the bottleneck of why it's slow to free up the full buffer (from the client end). It could be the network issue or client has a very poor I/O.

Some guidelines on troubleshooting this: http://www.symantec.com/docs/TECH18422

[Marianne]

Have a look at these links:

http://www.symantec.com/docs/TECH35968

Backup Planning and Performance Tuning Guide http://www.symantec.com/docs/DOC4483

[MickBoosh]

Thanks I've created the two files as suggested.

We have 2 staging and 2 tape libraries, one is running ok. I've noticed in the All Entries Log that the problomatic tape library has very slow transfer speeds to tape (100kb/sec or less) but I've also noticed that the file sizes it is backing up are usually very small, they bigger they are the faster the transfer rate it (obviously, this is as expected) Could someone take a look at this and give their opinion? 

I have attached the file. The media server XXXXBKP is the one that runs ok and XXXX75 is th slow one.

[Marianne]

bptm log on problematic media server will be more useful than All Log Entries report.

[MickBoosh]

I added the two buffer size files about 09:30 this morning for your information.