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: Proving or Debunking the need for rebuilding

Re: Proving or Debunking the need for rebuilding

From: <hasta_l3_at_hotmail.com>
Date: 15 Nov 2006 10:39:53 -0800
Message-ID: <1163615993.616586.162730@h54g2000cwb.googlegroups.com>


Richard Foote wrote :

> <hasta_l3_at_hotmail.com> wrote in message
> news:1163528916.396146.307250_at_m7g2000cwm.googlegroups.com...
> > hpuxrac wrote :
> >
> >> hasta_l3_at_hotmail.com wrote:
> >> > >
> >> > > What conclusions did you reach based on your approach?
> >> >
> >> > On my system - and I stress : on my system
> >> >
> >
> >> What selects saw a 30% improvement? Was that in response time?
> >> Any OLTP application that is retrieving 4000 rows repeatedly is going to
> >> vary in response time anyway based on transient conditions, at least in
> >> my experience. Again the phrase "An occasional" makes me wonder if
> >> this is based on test cases or what exactly.
> >
> > Yes, the test case is well specified. It happens that we open lists,
> > and the underlying select goes through a huge index but answers
> > only 4000 rows or so...
> >
> > The time to open a list - to execute the test plan - is "occasionaly"
> > important, probably when the cache has to be filled. We are here
> > talking of times on the order of 20 secs. When the test is retried,
> > response time is of course dramatically better.
> >
> > Now, the 30% was measured between an "occasional" slowdown
> > before rebuild, and an "occasional" slowdown after the rebuild.
> >
> > I realize that the conditions varies wildly. I'm reporting the notes I
> >
> > took.

>
>

> However, my money is on it having nothing to do with just an index rebuild
> ...

>

Could very well be, Richard. Perhaps you are reading too much in my words ? (which were admittedly very terse in the first report :-)

I was actually very careful not to voice any conclusions till now. For the record, the conclusions are :

  1. Did the rebuild impact the display of screens ? NO.
  2. Did the rebuild impact the night job ? YES, definitely yes.
  3. Did the rebuild impact opening lists ? Definitely a suspect area.

(I am overstating for clarity. Please add the customary caution :-)

Point (2) would be the main focus point, and deserves further investigations to understand what's going on.

Point (3) should be handled afterwards, and requires a more controlled test, for example flushing systematically the cache before execution of the test case.

Going through your questions :

> This is all rather vague and ambiguous isn't ? I have no idea what
> "occasionally important" is suppose to mean for example.

Certainly, it is vague... It happens after the system has run for a while... Recall that I am on purpose simulating full activity.

> I refer back to my earlier post where I suggest it's rather difficult
> to attribute these figures (whatever they actually mean) back to
> rebuilding an index.

Certainly again... I wouldnt dare to attribute or disattribute in this specific case. The point of a global test is to pinpoint with few efforts the troublesome areas from the user point of view .

I find interesting that this exercise provided all three possible answers : yes, no, perhaps :-)

>

> You mention that "the underlying select goes through a huge index but
> answers only 4000 rows or so" and that "we are talking of times on the order
> of 20 secs". With no SQL, no execution plans, no trace files etc. to base
> any meaning out of these comments, 20 secs sounds like an extraordinary long
> time to select just 4000 rows. Even with a huge index, a range scan should
> only visit a small number of index blocks and the 4000 table blocks at most.
> And this takes 20 secs ? So one can only assume there being various tables
> and joins being performed

Yes.

> and again it suggests that the execution plan
> could differ between good and bad execution times.

We are using rule.

> Some information on the SQL, the execution plan (both good and bad if
> different), some table/index stats and some summary execution statistics
> (both good and bad) would be useful in determining why you're getting this
> "occasional" 30% improvement.

I know, I know... But the first step would be to make "occasional" more
reproducible... Unfortunatly, I lack time right now, like everybody I would guess :-)

Have a nice day, Richard

Received on Wed Nov 15 2006 - 12:39:53 CST

Original text of this message

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