regaring wait events

[mahajanakhil198]

Do we need to tune the wait event if total_timeouts column from v$system_event corresponding to that wait event is 0 or not increasing? Though the value in time_waited is high. The instance has been rebooted 3-4 days ago.

[PAVB]

it depends on the wait event.

[mahajanakhil198]

Any comments after having a look at this:

                                                                     Avg
                                                     Total Wait    wait       Waits
Event                                  Waits   Timeouts      Time (s)   (ms)     /txn
---------------------------- ------------ ---------- ---------- ------ --------
db file sequential read           15,425           0        121           8         2.1
direct path read                  190,330          0        109           1        25.5
log file parallel write               13,674          0         34           2          1.8
log file sync                           5,922          0         30           5          0.8

Don't be lazy and use the code tags

[PAVB]

You can copy this on a textpad to have a better look of it.

Why would I do that when you can post it properly with wrap code tags?

[tamilselvan]

The avg wait time (ms) for the top events are 8, 1, 2, 5. When these numbers are not very high, should I need to worry about the events? No.
Did any user complaint about performance?

[PAVB]

For an instance running for four days wait events suggest very little activity.
"db file scatered read" doesn't even show on your top wait events.
Is anybody complaining about performance? - if this is the case address the specific query user is complaining about.

[mahajanakhil198]

Yes, there are complaints. All I am supposed to do is to find something wrong if it is there. I checked out the memory structures, etc. Everything seemed to be alright. I had a doubt about wait events. So just felt like asking it. I would collect more data on Monday and then share it.


And Yes, I wanted to represent the above output in the righteous manner but after so many attempts, I just couldn't find the right option. Thanks anyways!

[PAVB]

Yes, there are complaints.

1- Identify the top five complains - start for the ones bothering management.
2- Identify offending queries.
3- Fine tune offending queries.
4- Once offending queries are fixed document execution plan, cardinality of involved tables and, response time - tkprof-ed trace wouldn't hurt.
5- Go to #1