Re: how to perform system health check
January 11, 2002 09:04 AM
(in response to Eric Ng Yew Heng)
Eric, I do these ad hoc on a daily basis for large production
and a "formal" subsystem review about once every three months for
We use Omegamon DB2 and I look for EDM pool failures and % usage,
thresholds (current vs. historical and DMT) and "efficiency",
dedicated DSNDB07 VBP, HIPERPOOL read/write ratios, DBAT queing,
running threads, check for in-doubt threads, log write buffer
Page-ins for read/write, prefetch failures (VBP or engine), pages
I/O, etc. Also, note CPU utilization during the period in question.
necessary we drill down to look at volume contention and
Also check MSTR log for any unusual activity, display
may have become invalid but are still operable (need rebind), and
DATABASE looking for any objects not in R/W mode.
Pick your own poison...
Good luck, George