You're probably best to run the configuration advisor and or the design
advisor from the Control Center (right click on the database object) or
the db2advis command from the command line to get recommendations.
With no explanation of your workload or data, the answer in this
newsgroup can only be "it depends" - totally random access to data may
get little benefit from buffering data in memory, unless the amount of
memory if close to the amount of data, or the access is truly not random
- there is some pattern DB2 can exploit to pre-fetch hot data into the
bufferpool.
Predictable access can take advantage of well-configured bufferpools,
such as assigning a hot table to its own bufferpool, with enough memory
to keep the entire table in memory, or a transactional workload where
the same rows are frequently accessed.
Jagdip Singh wrote:
Hi all,
Is there any recommended bufferpool size while designing database
I mean some equation which governs bufferpool size
regards,
Jagdip
