How to tune waits as per given statspaack output here

Question

Hi
Following is the out put of statspack.
can anywone help to resolve this
Top 5 Timed Events
~~~~~~~~~~~~~~~~~~ % Total
Event Waits Time (s) Ela Time
CPU time 3,131 95.03
log file sync 18,477 55 1.67
log file parallel write 18,981 49 1.50
db file parallel write 648 29 .89
db file sequential read 1,617 13 .38
^LWait Events for DB: MPLSDB Instance: mplsdb Snaps: 7 -8
-> s - second
-> cs - centisecond - 100th of a second
-> ms - millisecond - 1000th of a second
-> us - microsecond - 1000000th of a second
-> ordered by wait time desc, waits desc (idle events last)

Answer 1

Hmmm.....
Sorry Sybrand, I have to disagree with you. I would say there's too little information to determine if there's a problem.
It's quite possible that the posted load profile is from a system that's humming along, with no concerns. A system with a majority of the data cached, and SQL statements run quickly.
It's also possible that the system is suffering from CPU shortage, and there are SQLs that could stand to be tuned, to take advantage of better access paths, and do less logical I/O, there by burning less CPU. Or, it could be an application design problem, where the individual SQLs are all well tuned, but the application design needs to be modified to call those SQLs less frequently.
I'd argue that either of the above scenarios could be true, and based on the scant amount of information provided in the original post, it's impossible to tell.
For a clearer picture, we'd need to know:
- What's happening at the O/S level, in terms of CPU usage, runqueues, etc
- How is the system responding to various users' business tasks? Are they happy? Are they feeling pain?
-Mark