Let's say the bufferpool hit ratio went down for index and/or data tablespaces.


Additional info:

- OLTP environment
- No configuration changes (ie. no change to tablespaces/bufferpools assignment, bp size, config parameters...)
- STMM is OFF
- No SQL / workload change
- Database has been active for some time
- No snapshot info is available before / after bp hit ratio change


What I think happened is db2 started fetching more data due to data change / runstats execution (should invalidate cached access plan) / access plan change. So, it started doing more async vs sync I/O / rows read vs. rows selected.


Does this sound right? Can you think of anything else?

Hi Bella,

if you really can exclude the SQL / workload changes, then the most likely reason is a access plan change. This could be due to data growth or index fragmentation.