This discussion is now locked
Posted by Hari Shanmugadhasan on 31 December 2012 at 8:42
It would be helpful if the use of DB2 accounting data for the analysis was illustrated in the Redpaper.
Page 7 says, "Target DB2 elapsed time for Q apply, obtained from IBM Tivoli OMEGAMON® XE for DB2 Performance Monitor on z/OS."
However, that usage is not illustrated and there is no appendix with the DB2 Accounting reports for the key test scenarios, with the important numbers for Q Apply identified along with the appropriate minimum traces to focus on Q Apply alone.
Posted by Hari Shanmugadhasan on 31 December 2012 at 8:52
It would be helpful if the Redpaper illustrated the identification and resolution of performance bottlenecks in DB2.
Page 6 includes among its potential bottlenecks, "3. Q Apply serialization: It can be caused by DB2 transactions that update the same row. 4. DB2 lock contention . . ."
Yet, these cases are not discussed. DB2 lock contention isn't even reported on in the Redpaper. There only appear to be reports of just the gross level number "Q Apply DB2 class 2 elapsed time (ms)".