| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: 8.1.7.4 migration from 32 bit to 64 bit problem - solved!
Ok, the finally answer is....
I had to increase by Hash_AREA_size to 50M!!!! from 2M.
Then the explain plan looks like it should. Guess they should add this to the docs besides increasing the share pool size.
It make sense to me now....especially since we use view within views for this hybernate stuff.
Thanks for all the suggestions.
Kathy
-----Original Message-----
From: Karen Morton [mailto:karen.morton_at_hotsos.com]
Sent: Friday, November 19, 2004 6:04 PM
To: kduret_at_starkinvestments.com; oracle-l_at_freelists.org
Subject: RE: 8.1.7.4 migration from 32 bit to 64 bit problem
Kathy,
If you're getting a cartesian join in the "bad" plan, check to see if the
cardinality of
the first row source of the cartesian thinks one row is being returned by
that row source
operation. At times the optimizer will choose to do a cartesian join
thinking that it can
save some effort by buffering the 2nd row source into memory (treating it
similarly to how
it would treat a sort). If the 2nd row source is big, then the problem
you're seeing
could occur.
My guess is that before the section of the plan that is now doing the
cartesian was being
accomplished with a nested loops or possibly hash join. See if you can
identify the
portion that matches up for both the old plan and the new plan. If you can
isolate the
one portion of the plan that is different, you can then focus your efforts
on the part of
the predicate that is involved. Check to see that all indexes are there
that were there
before as well. A missing index could cause this shift in plan as well.
I know this may sound a bit vague, but without the plan excerpts and query
text, it's a
bit difficult to do more than speculate.
Hope it helps some.
Karen Morton
Hotsos Enterprises, Ltd.
http://www.hotsos.com
Upcoming events at http://www.hotsos.com/education/schedule.html
-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org]
On Behalf Of
Duret, Kathy
Sent: Friday, November 19, 2004 1:33 PM
To: 'Ron.Reidy_at_arraybiopharma.com'; mark.powell_at_eds.com;
oracle-l_at_freelists.org
Subject: RE: 8.1.7.4 migration from 32 bit to 64 bit problem
The explan plans are very long but it boils down to now the bad explain plan
is doing a
merge cartesian join.
I have tried various things, including re:
alter session set optimizer_mode='RULE';
alter session set optimizer_index_cost_adj=50; alter session set optimizer_index_caching=90;
alter session set optimizer_index_cost_adj=1; alter session set optimizer_index_caching=100;
putting ordered and use_nl in all the views, setting the _complex_view parameter and bouncing the database.
INcreasing the SGA.
I am willing to trying anything at this point...
Kathy
-----Original Message-----
From: Reidy, Ron [mailto:Ron.Reidy_at_arraybiopharma.com]
Sent: Friday, November 19, 2004 3:05 PM
To: kduret_at_starkinvestments.com; mark.powell_at_eds.com; oracle-l_at_freelists.org
Subject: RE: 8.1.7.4 migration from 32 bit to 64 bit problem
Can you post both plans?
Maybe someone here can give you an idea.
-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org]On
Behalf Of
Duret, Kathy
Sent: Friday, November 19, 2004 1:48 PM
To: 'mark.powell_at_eds.com'; 'oracle-l_at_freelists.org'
Subject: RE: 8.1.7.4 migration from 32 bit to 64 bit problem
Nothing was changed..... I tried to increase the shared pool on the new =
dev version and
bounced it.... no good.
I am trying various hints in all the views and sub views to no avail. I
finally got a
descent tech now for my tar....=20
Yes I just updated the stats again for the new tech. I am going to give =
him a 10053
trace as well. The explain plan for the new database is very strange for
this query.
Kathy
-----Original Message-----
From: Powell, Mark D [mailto:mark.powell_at_eds.com]
Sent: Friday, November 19, 2004 2:06 PM
To: 'oracle-l_at_freelists.org'
Subject: RE: 8.1.7.4 migration from 32 bit to 64 bit problem
Sometimes even a small change in plans can result in a large change in
performance for the
query depending on what the change is.
Were the statistics updated on the new version?
Was the shared pool increased in size to compensate for the additional 4
bytes in every
address pointer used. The 64 bit version of 8.1.7 needs about a 20%
increase in the
shared pool just to run the same load in our experience, but then we have a
lot of stored
code (pl/sql in the = database).
Were any database parameters changed?
HTH -- Mark D Powell --
-----Original Message-----
From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org]On
Behalf Of
Jeremiah Wilton
Sent: Friday, November 19, 2004 2:09 PM
To: Oracle L (E-mail)
Subject: RE: 8.1.7.4 migration from 32 bit to 64 bit problem
Sorry if this has been explored, but it sounds like the difference in the
plans is the
problem. Can you elaborate on WHAT is different about the plans?
Is something else hogging temp?
-- Jeremiah Wilton Independent Oracle Professional Oracle Certified Master Disaster Recovery - Seminars - Technical Interviews http://www.speakeasy.net/~jwilton On Fri, 19 Nov 2004, Duret, Kathy wrote:Received on Mon Nov 22 2004 - 08:33:08 CST
> two database set up is EXACTLY the same (init.ora, files size etc)
>
> Query using two views in OLD 32 bit runs in 2 seonds and with the =
WHOLE
> company running on the database uses less than 1/2 G of temp =
tablespaces.
>
> Query ONLY running on new production database get ora-1652 (out of =
temp
> space) on 1G of temp space after 81 seconds.
>
> Once again same set up, same data, BUT the explain plans are =
different but
> are fairly similiar.
-- http://www.freelists.org/webpage/oracle-l -- http://www.freelists.org/webpage/oracle-l This transmission contains information solely for intended recipient and = may be privileged, confidential and/or otherwise protect from disclosure. = If you are not the intended recipient, please contact the sender and delete = all copies of this transmission. This message and/or the materials = contained herein are not an offer to sell, or a solicitation of an offer to buy, = any securities or other instruments. The information has been obtained or derived from sources believed by us to be reliable, but we do not = represent that it is accurate or complete. Any opinions or estimates contained in this information constitute our judgment as of this date and are subject = to change without notice. Any information you share with us will be used = in the operation of our business, and we do not request and do not want any material, nonpublic information. Absent an express prior written = agreement, we are not agreeing to treat any information confidentially and will use = any and all information and reserve the right to publish or disclose any information you share with us. -- http://www.freelists.org/webpage/oracle-l This electronic message transmission is a PRIVATE communication which = contains information which may be confidential or privileged. The information is = intended=20 to be for the use of the individual or entity named above. If you are = not the=20 intended recipient, please be aware that any disclosure, copying, = distribution=20 or use of the contents of this information is prohibited. Please notify = the sender of the delivery error by replying to this message, or notify us = by telephone (877-633-2436, ext. 0), and then delete it from your system. -- http://www.freelists.org/webpage/oracle-l This transmission contains information solely for intended recipient and may be privileged, confidential and/or otherwise protect from disclosure. If you are not the intended recipient, please contact the sender and delete all copies of this transmission. This message and/or the materials contained herein are not an offer to sell, or a solicitation of an offer to buy, any securities or other instruments. The information has been obtained or derived from sources believed by us to be reliable, but we do not represent that it is accurate or complete. Any opinions or estimates contained in this information constitute our judgment as of this date and are subject to change without notice. Any information you share with us will be used in the operation of our business, and we do not request and do not want any material, nonpublic information. Absent an express prior written agreement, we are not agreeing to treat any information confidentially and will use any and all information and reserve the right to publish or disclose any information you share with us. -- http://www.freelists.org/webpage/oracle-l This transmission contains information solely for intended recipient and may be privileged, confidential and/or otherwise protect from disclosure. If you are not the intended recipient, please contact the sender and delete all copies of this transmission. This message and/or the materials contained herein are not an offer to sell, or a solicitation of an offer to buy, any securities or other instruments. The information has been obtained or derived from sources believed by us to be reliable, but we do not represent that it is accurate or complete. Any opinions or estimates contained in this information constitute our judgment as of this date and are subject to change without notice. Any information you share with us will be used in the operation of our business, and we do not request and do not want any material, nonpublic information. Absent an express prior written agreement, we are not agreeing to treat any information confidentially and will use any and all information and reserve the right to publish or disclose any information you share with us. -- http://www.freelists.org/webpage/oracle-l
 |
 |