Re: wait events

In article <8s4296$2t5$1_at_nnrp1.deja.com>,   fly13_at_my-deja.com wrote:
> hi girls and boys,
> I use the statspack of oracle to tune my database, and i can see in
 the
> table concerning waiting events :
>
> Avg
> Total Wait wait
> Waits
> Event Waits Timeouts Time (cs)
> (ms) /txn
> ---------------------------- ------------ ---------- ----------- -----
 -
> -----
> control file sequential read 33 0 6
> 2 33.0
> control file parallel write 50 0 4
> 1 50.0
> refresh controlfile command 5 0 2
> 4 5.0
> direct path read 832 0 1 0
> 832.0
> direct path write 644 0 0 0
> 644.0
> log file parallel write 9 0 0
> 0 9.0
> SQL*Net break/reset to clien 4 0 0
> 0 4.0
> file open 4 0 0
> 0 4.0
> LGWR wait for redo copy 1 0 0
> 0 1.0
> log file sync 1 0 0
> 0 1.0
> SQL*Net message from client 84 0 3,042
> 362 84.0
> SQL*Net message to client 84 0 0
> 0 84.0
> -------------------------------------------------------------
>
> Background Wait Events for DB: DBTEST Instance: dbtest Snaps:
> 1 -
>
> Avg
> Total Wait wait
> Waits
> Event Waits Timeouts Time (cs)
> (ms) /txn
> ---------------------------- ------------ ---------- ----------- -----
 -
> -----
> control file parallel write 50 0 4
> 1 50.0
> control file sequential read 15 0 2
> 1 15.0
> log file parallel write 8 0 0
> 0 8.0
> LGWR wait for redo copy 1 0 0
> 0 1.0
> rdbms ipc message 171 150 8,514 498
> 171.0
> pmon timer 50 50 2,764
> 553 50.0
> -------------------------------------------------------------
>
> SQL ordered by Gets for DB: DBTEST Instance: dbtest Snaps:
 1 -
> 2
>
> So i would like to know if anybody can tell what i could to do to
> decrease waiting times.
>
> Thanks
>
> See you
>
> Fly13
>

Oracle always thinks it is waiting on something. It is the nature of the beast. I admit to only giving these numbers a quick look, and I am not familiar with the tool. But nothing jumped out as being a problem. What makes you think you need to tune something?

The importance of Waits are relative to the number of times the the wait event was attempted. That is if you get 1000 waits for something you try 1000 times that may well indicate a problem, but if you get a 1000 waits for an event that you perform 100,000 times then this is probably not a problem though it does depend on the event.

I do not believe that you can make good tuning decisions based on the isolated information above. Also it does not appear to me that the above system was carrying much of a load when these numbers were generated. If you have a database level tuning problem it generally takes a load to truely reflect the real problem areas.

This is just my opinion and I am not a db tuning guru. There are a couple of regular posters who I think qualify as such, and I will be reading to see if they respond.

--
Mark D. Powell  -- The only advice that counts is the advice that
 you follow so follow your own advice --


Sent via Deja.com http://www.deja.com/
Before you buy.