Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: How to capture information to analyze high load peak?

Re: How to capture information to analyze high load peak?

From: Noons <wizofoz2k_at_yahoo.com.au>
Date: 7 Sep 2005 22:26:10 -0700
Message-ID: <1126157170.819669.19060@g47g2000cwa.googlegroups.com>


Thomas Kyte wrote:

> just to clarify - if you want to try and TUNE with statspack - a 15 minute
> snapshot or thereabouts it what you want. you do not want a 3 hour snapshot,
> that is too long, things are too averaged.
>
> If you want to TREND with statspack - hourly observations are good (the average
> transactions per second between 2pm and 3pm are ... type of trending)
> >
> >10g ... Learn to use AWR ... Use DBMS_WORKLOAD_REPOSITORY package to
> >alter default behavior to take snapshots every 15-20 minutes.

One more quick thing:
if you run statspack frequently and you have a busy system with lots and lots of concurrent SQL, watch out for disk space use. The default cleanup procedure does *not* clean up all rows created by statspack. You need to clean them manually. At least until 9.2.0.6. A common trap that can cause a lot of teeth-grinding. DAMHIKT... Received on Thu Sep 08 2005 - 00:26:10 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US