we are having some 4 hrs constant load performance tests going on where the final readings comes some what like these

db2 9.1.3
windows
------------

Buffer pool data writes = 21341
Asynchronous pool data page writes = 15711
Buffer pool index writes = 5134841
Asynchronous pool index page writes = 152345

LSN Gap cleaner triggers = 10899
Dirty page steal cleaner triggers = 574
Dirty page threshold cleaner triggers = 99589

-------------

sum of all Bufferpool ~ 2GB
softmax 1000
logfilsz 12000
CHNGPGS_THRESH 25
NUM_IOCLEANERS AUTOMATIC
-------------


even though the % steal is quite low but isn't in our case the page cleaners
are getting fired too aggresively.
can someone please recommend some better cfg

regards
Rahul

increase ur CHNGPGS_THRESH in increment of 5 and test..may be 30-40 would be a good range

thanks,
but dont you think synch write is happening too often than asynch writes.
i am really bothered because of that. wont delaying cleaning would increase it further ?

check how many page cleaners are active at a time, so u would get an idea..then if u see more steal cleaner triggers..try to increase it...its just a thought..test it thoroughly...

Set DB2_USE_ALTERNATE_PAGE_CLEANING registry variable on..That will help you

This will cause the page cleaners to fire more agressively.

Regarding the original post, I think you are worrying about a problem that does not exist.