I'm having trouble understanding all of this.  I found the documentation
on this for windows... http://support.veritas.com/docs/183702

I'm using NB5.1 MP1.  When I go through and look at these logs, I don't
see any waits or delays in the bpbkar log on the client.  

Then as I am going through this document, it says the default value for
buffer_size is 32K.  When I go onto my clients, they default to 16.

How in the world do I change all of my clients to 32k in one easy step.
Also, why is my default 16 instead of 32?  Could the person who
installed NB have done this on the install?  Is it because the media
server is set to 16?  I can't seem to find where that is set.  I see no
file that would have set it to 16.  Also, if I jump all of these up from
16 to 32, I shouldn't have any problems, right?
Thanks
Mike

-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu] On Behalf Of
Mark.Donaldson AT cexp DOT com
Sent: Wednesday, October 12, 2005 1:36 PM
To: jpiszcz AT servervault DOT com; veritas-bu AT mailman.eng.auburn DOT edu
Subject: RE: [Veritas-bu] Recommended number for NUM_DATA_BUFFERS?

The all-powerful document on tuning:

http://seer.support.veritas.com/docs/183702.htm

It tells you how to scan your bptm logs to see if your buffer settings
are
adequate.


Attached is my script that parses the bptm logs to add up some of the
wait
time numbers.  

It's not friendly.

Usage: policy_perf <policy_name> <logfile>

If you leave off the logfile, it defaults to today's logfile.  The
policy
name is required.

-M


-----Original Message-----
From: veritas-bu-admin AT mailman.eng.auburn DOT edu
[mailto:veritas-bu-admin AT mailman.eng.auburn DOT edu]On Behalf Of Piszcz,
Justin
Sent: Wednesday, October 12, 2005 11:23 AM
To: veritas-bu AT mailman.eng.auburn DOT edu
Subject: [Veritas-bu] Recommended number for NUM_DATA_BUFFERS?


On a box with 14GB of ram, mine is set to 32.
 
Would increasing this number to say 256 increase performance?
 
Justin.