RE: Tracking Temp Segment Usage and Event 10046

2002-09-19 Thread Cary Millsap
Title: Tracking Temp Segment Usage and Event 10046









For more information about 10053, see Wolfgang Breitlings http://www.hotsos.com/dnloads/1/10053/Breitling2002.pdf.
Wolfgang will
present this and another new 10053 paper at our Symposium in Dallas next February (http://www.hotsos.com/events/symposium).






Cary Millsap
Hotsos Enterprises, Ltd.
http://www.hotsos.com

Upcoming events:
- Hotsos Clinic, Oct
13 San Francisco, Oct 1517 Dallas, Dec 911 Honolulu
- 2003 Hotsos Symposium on
Oracle System Performance, Feb 912 Dallas
- Next event: Miracle Database Forum, Sep
2022 Middlefart Denmark



-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]] On Behalf Of Jamadagni,
Rajendra
Sent: Wednesday, September 18,
2002 10:33 AM
To: Multiple recipients of list
ORACLE-L
Subject: RE: Tracking Temp Segment
Usage and Event 10046





I spent last two
days working on that SQR and finally nailed it. Developers were using DISTINCT
as a rule and CBO was choosing incorrect indexes. 











I had to do
following to make it work ...











1.we put
hints in appropriate places





2.created 3
new indexes to help the queries





3. Created 1 MV





4. Change the
pl/sql code to be more efficient





5. Remove
unnecessary distinct conditions.











After all this
said and done, I finally have the SQR taking around 375 seconds compared to
about 1500 seconds. Even if I run all 8 of them in parallel it doesn't blow the
TEMP segments.











Now comes the
difficult part, to explain the development the things that we had to do and why
we had to do them.











I think I am
really starting to see the power of event 10046 ... I just wish I should learn
more on event 10053 as well, but that will come too eventually.











I am still going
to track the temp segment usage though ...











Thanks everyone
for all your kind suggestions, I am not done yet, there is more work to be
done.











Raj





__



Rajendra Jamadagni
 MIS, ESPN Inc.

Rajendra dot Jamadagni at ESPN dot com

Any opinion expressed here is personal and doesn't reflect that
of ESPN Inc. 

QOTD: Any clod can have facts, but having an opinion
is an art!



-Original Message-
From: Khedr, Waleed
[mailto:[EMAIL PROTECTED]]
Sent: Wednesday, September 18,
2002 11:03 AM
To: Multiple recipients of list
ORACLE-L
Subject: RE: Tracking Temp Segment
Usage and Event 10046



I'd start looking at the execution plans
first and examine if there is any Cartesian joins and the order the tables get
joined.











Waleed












RE: Tracking Temp Segment Usage and Event 10046

2002-09-19 Thread Jamadagni, Rajendra
Title: Tracking Temp Segment Usage and Event 10046



Thanks Cary

One thing I noticed is some parameters referenced in that document are 
different in 9iR2 ... I guess Wolfgang is probably working on the 9iR2 version 
...

It is a very nice paper ...
Raj
__
Rajendra 
Jamadagni 
 MIS, ESPN Inc.
Rajendra dot Jamadagni at ESPN dot 
com
Any opinion expressed here is 
personal and doesn't reflect that of ESPN Inc. 
QOTD: Any clod can have facts, but 
having an opinion is an art!

  -Original Message-From: Cary Millsap 
  [mailto:[EMAIL PROTECTED]]Sent: Thursday, September 19, 2002 
  5:24 AMTo: Multiple recipients of list ORACLE-LSubject: 
  RE: Tracking Temp Segment Usage and Event 10046
  
  For more information 
  about 10053, see Wolfgang Breitlings http://www.hotsos.com/dnloads/1/10053/Breitling2002.pdf. 
  Wolfgang will present this 
  and another new 10053 paper at our Symposium in Dallas next February (http://www.hotsos.com/events/symposium). 
  
  
  
  Cary 
  MillsapHotsos 
  Enterprises, Ltd.http://www.hotsos.comUpcoming 
  events:- Hotsos 
  Clinic, Oct 13 San Francisco, Oct 1517 Dallas, Dec 911 Honolulu- 2003 Hotsos Symposium on 
  Oracle® System Performance, Feb 912 Dallas- Next event: Miracle Database Forum, Sep 2022 
  Middlefart Denmark
  

This e-mail 
message is confidential, intended only for the named recipient(s) above and may 
contain information that is privileged, attorney work product or exempt from 
disclosure under applicable law. If you have received this message in error, or are 
not the named recipient(s), please immediately notify corporate MIS at (860) 766-2000 
and delete this e-mail message from your computer, Thank 
you.*2



RE: Tracking Temp Segment Usage and Event 10046

2002-09-18 Thread Khedr, Waleed
Title: Tracking Temp Segment Usage and Event 10046



I'd 
start looking at the execution plans first and examine if there is any Cartesian 
joins and the order the tables get joined.

Waleed

  -Original Message-From: Jamadagni, Rajendra 
  [mailto:[EMAIL PROTECTED]]Sent: Tuesday, September 17, 
  2002 10:34 AMTo: Multiple recipients of list 
  ORACLE-LSubject: Tracking Temp Segment Usage and Event 
  10046
  I have a problem with a process and its temp 
  segment usage. 
  Previously, in RBO (8161) we used to run 8 reports 
  with slightly different parameters in parallel and it used to work fine with 
  all other load on the system and the total TEMP space was 6GB
  Since we migrated to CBO (all tables  indexes 
  are analyzed using estimate), these processes fail with unable to extent temp 
  segment error. Currently we have allocated 12GB to temp tablespace.
  I also have performance issues so I am doing a 
  10046 trace at level 8 on this one. Is there a way to identify when this 
  process used temp space, how much and when does it de-allocate?
  Would this information be available in 10046 trace 
  or is there something else that I could do to track that ... 
  Thanks in advance Raj __ Rajendra Jamadagni 
   MIS, ESPN Inc. Rajendra dot Jamadagni at ESPN dot com 
  Any opinion expressed here is personal and 
  doesn't reflect that of ESPN Inc. QOTD: Any clod can have facts, but having an opinion is an art! 
  


RE: Tracking Temp Segment Usage and Event 10046

2002-09-18 Thread STEVE OLLIG

aren't Cartesian joins are unlikely since all was well in the RBO world?
sounds to me like the RBO was simply selecting better query plans than the
CBO is.
 
Raj - have you tried playing around with the OPTIMIZER_MODE parameter or
OPTIMIZER_INDEX_CACHING; and OPTIMIZER_INDEX_COST_ADJ as described in Tim
Gorman's paper?
http://www.evdbt.com/SearchIntelligenceCBO.doc
http://www.evdbt.com/SearchIntelligenceCBO.doc 

-Original Message-
Sent: Wednesday, September 18, 2002 10:03 AM
To: Multiple recipients of list ORACLE-L


I'd start looking at the execution plans first and examine if there is any
Cartesian joins and the order the tables get joined.
 
Waleed

-Original Message-
Sent: Tuesday, September 17, 2002 10:34 AM
To: Multiple recipients of list ORACLE-L



I have a problem with a process and its temp segment usage. 

Previously, in RBO (8161) we used to run 8 reports with slightly different
parameters in parallel and it used to work fine with all other load on the
system and the total TEMP space was 6GB

Since we migrated to CBO (all tables  indexes are analyzed using estimate),
these processes fail with unable to extent temp segment error. Currently we
have allocated 12GB to temp tablespace.

I also have performance issues so I am doing a 10046 trace at level 8 on
this one. Is there a way to identify when this process used temp space, how
much and when does it de-allocate?

Would this information be available in 10046 trace or is there something
else that I could do to track that ... 

Thanks in advance 
Raj 
__ 
Rajendra Jamadagni  MIS, ESPN Inc. 
Rajendra dot Jamadagni at ESPN dot com 
Any opinion expressed here is personal and doesn't reflect that of ESPN Inc.

QOTD: Any clod can have facts, but having an opinion is an art! 

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: STEVE OLLIG
  INET: [EMAIL PROTECTED]

Fat City Network Services-- 858-538-5051 http://www.fatcity.com
San Diego, California-- Mailing list and web hosting services
-
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).



RE: Tracking Temp Segment Usage and Event 10046

2002-09-18 Thread Jamadagni, Rajendra
Title: Tracking Temp Segment Usage and Event 10046



I spent last two days working on that SQR and finally nailed it. 
Developers were using DISTINCT as a rule and CBO was choosing incorrect indexes. 


I had to do following to make it work ...

1.we put hints in appropriate places
2.created 3 new indexes to help the queries
3. Created 1 MV
4. Change the pl/sql code to be more efficient
5. Remove unnecessary distinct conditions.

After all this said and done, I finally have the SQR taking around 375 
seconds compared to about 1500 seconds. Even if I run all 8 of them in parallel 
it doesn't blow the TEMP segments.

Now comes the difficult part, to explain the development the things that 
we had to do and why we had to do them.

I think I am really starting to see the power of event 10046 ... I just 
wish I should learn more on event 10053 as well, but that will come too 
eventually.

I am still going to track the temp segment usage though 
...

Thanks everyone for all your kind suggestions, I am not done yet, there 
is more work to be done.

Raj
__
Rajendra 
Jamadagni 
 MIS, ESPN Inc.
Rajendra dot Jamadagni at ESPN dot 
com
Any opinion expressed here is 
personal and doesn't reflect that of ESPN Inc. 
QOTD: Any clod can have facts, but 
having an opinion is an art!

  -Original Message-From: Khedr, Waleed 
  [mailto:[EMAIL PROTECTED]]Sent: Wednesday, September 18, 2002 
  11:03 AMTo: Multiple recipients of list ORACLE-LSubject: 
  RE: Tracking Temp Segment Usage and Event 10046
  I'd 
  start looking at the execution plans first and examine if there is any 
  Cartesian joins and the order the tables get joined.
  
  Waleed

This e-mail 
message is confidential, intended only for the named recipient(s) above and may 
contain information that is privileged, attorney work product or exempt from 
disclosure under applicable law. If you have received this message in error, or are 
not the named recipient(s), please immediately notify corporate MIS at (860) 766-2000 
and delete this e-mail message from your computer, Thank 
you.*2



RE: Tracking Temp Segment Usage and Event 10046

2002-09-18 Thread Connor McDonald

v$sort_usage and v$sort_segment are always a good
start

hth
connor

 --- Khedr, Waleed [EMAIL PROTECTED] wrote: 
I'd start looking at the execution plans first and
 examine if there is any
 Cartesian joins and the order the tables get joined.
  
 Waleed
 
 -Original Message-
 Sent: Tuesday, September 17, 2002 10:34 AM
 To: Multiple recipients of list ORACLE-L
 
 
 
 I have a problem with a process and its temp segment
 usage. 
 
 Previously, in RBO (8161) we used to run 8 reports
 with slightly different
 parameters in parallel and it used to work fine with
 all other load on the
 system and the total TEMP space was 6GB
 
 Since we migrated to CBO (all tables  indexes are
 analyzed using estimate),
 these processes fail with unable to extent temp
 segment error. Currently we
 have allocated 12GB to temp tablespace.
 
 I also have performance issues so I am doing a 10046
 trace at level 8 on
 this one. Is there a way to identify when this
 process used temp space, how
 much and when does it de-allocate?
 
 Would this information be available in 10046 trace
 or is there something
 else that I could do to track that ... 
 
 Thanks in advance 
 Raj 

__
 
 Rajendra Jamadagni  MIS, ESPN Inc. 
 Rajendra dot Jamadagni at ESPN dot com 
 Any opinion expressed here is personal and doesn't
 reflect that of ESPN Inc.
 
 QOTD: Any clod can have facts, but having an opinion
 is an art! 
 
  

=
Connor McDonald
http://www.oracledba.co.uk
http://www.oaktable.net

Remember amateurs built the ark - Professionals built the Titanic

__
Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts
http://uk.my.yahoo.com
-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: =?iso-8859-1?q?Connor=20McDonald?=
  INET: [EMAIL PROTECTED]

Fat City Network Services-- 858-538-5051 http://www.fatcity.com
San Diego, California-- Mailing list and web hosting services
-
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).



RE: Tracking Temp Segment Usage and Event 10046

2002-09-17 Thread K Gopalakrishnan
Title: Tracking Temp Segment Usage and Event 10046



Raj:

You 
will be able to identify the processes (along with the SQLs and number of 
extents) using the 10046 level 8 trace. 
If you 
use the temp files you see the file# ( in the direct path read/write events) as 
db_files+1 or MAXDBFILES+1
or 
similar values. DOn't try to map the FILE# with the dictionary or X$KCFIO. THis 
is the only catch.


KG

  -Original Message-From: [EMAIL PROTECTED] 
  [mailto:[EMAIL PROTECTED]]On Behalf Of Jamadagni, 
  RajendraSent: Tuesday, September 17, 2002 7:34 AMTo: 
  Multiple recipients of list ORACLE-LSubject: Tracking Temp Segment 
  Usage and Event 10046
  I have a problem with a process and its temp 
  segment usage. 
  Previously, in RBO (8161) we used to run 8 reports 
  with slightly different parameters in parallel and it used to work fine with 
  all other load on the system and the total TEMP space was 6GB
  Since we migrated to CBO (all tables  indexes 
  are analyzed using estimate), these processes fail with unable to extent temp 
  segment error. Currently we have allocated 12GB to temp tablespace.
  I also have performance issues so I am doing a 
  10046 trace at level 8 on this one. Is there a way to identify when this 
  process used temp space, how much and when does it de-allocate?
  Would this information be available in 10046 trace 
  or is there something else that I could do to track that ... 
  Thanks in advance Raj __ Rajendra Jamadagni 
   MIS, ESPN Inc. Rajendra dot Jamadagni at ESPN dot com 
  Any opinion expressed here is personal and 
  doesn't reflect that of ESPN Inc. QOTD: Any clod can have facts, but having an opinion is an art! 
  


RE: Tracking Temp Segment Usage and Event 10046

2002-09-17 Thread Jamadagni, Rajendra
Title: RE: Tracking Temp Segment Usage and Event 10046





Dennis,


I did, but the problem is this is a SQR report, and it ain't one sql. SQR reports are like giant cursor loops but un structured. There are about 100 different sql statements (not including the recursive ones). 

In RBO, eight such reports could run parallel and not affect anything and finish in approx 25 minutes. In CBO they crash in about 14 minutes into the report and out of eight only about 3/4 manage to finish.

Performance is my second priority, reducing its temp space usage is my first priority. The trace files is about 400MB+ in size.

Raj
__
Rajendra Jamadagni  MIS, ESPN Inc.
Rajendra dot Jamadagni at ESPN dot com
Any opinion expressed here is personal and doesn't reflect that of ESPN Inc. 
QOTD: Any clod can have facts, but having an opinion is an art!



-Original Message-
From: DENNIS WILLIAMS [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, September 17, 2002 10:11 AM
To: '[EMAIL PROTECTED]'
Cc: Jamadagni, Rajendra
Subject: RE: Tracking Temp Segment Usage and Event 10046



Raj - I personally favor using the simpler diagnostics first. Why not run
your SQL through EXPLAIN PLAN and see which SQL is performing sorts? My
immediate guess is that under RBO an index was being used that isn't now.


Dennis Williams
DBA
Lifetouch, Inc.
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED] 



-Original Message-
From: Jamadagni, Rajendra [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, September 17, 2002 9:34 AM
To: Multiple recipients of list ORACLE-L
Subject: Tracking Temp Segment Usage and Event 10046




I have a problem with a process and its temp segment usage. 


Previously, in RBO (8161) we used to run 8 reports with slightly different
parameters in parallel and it used to work fine with all other load on the
system and the total TEMP space was 6GB


Since we migrated to CBO (all tables  indexes are analyzed using estimate),
these processes fail with unable to extent temp segment error. Currently we
have allocated 12GB to temp tablespace.


I also have performance issues so I am doing a 10046 trace at level 8 on
this one. Is there a way to identify when this process used temp space, how
much and when does it de-allocate?


Would this information be available in 10046 trace or is there something
else that I could do to track that ... 


Thanks in advance 
Raj 
__ 
Rajendra Jamadagni MIS, ESPN Inc. 
Rajendra dot Jamadagni at ESPN dot com 
Any opinion expressed here is personal and doesn't reflect that of ESPN Inc.


QOTD: Any clod can have facts, but having an opinion is an art! 




This e-mail 
message is confidential, intended only for the named recipient(s) above and may 
contain information that is privileged, attorney work product or exempt from 
disclosure under applicable law. If you have received this message in error, or are 
not the named recipient(s), please immediately notify corporate MIS at (860) 766-2000 
and delete this e-mail message from your computer, Thank 
you.*2



RE: Tracking Temp Segment Usage and Event 10046

2002-09-17 Thread DENNIS WILLIAMS

Raj - I personally favor using the simpler diagnostics first. Why not run
your SQL through EXPLAIN PLAN and see which SQL is performing sorts? My
immediate guess is that under RBO an index was being used that isn't now.
 
 
Dennis Williams
DBA
Lifetouch, Inc.
[EMAIL PROTECTED] mailto:[EMAIL PROTECTED] 
 

-Original Message-
Sent: Tuesday, September 17, 2002 9:34 AM
To: Multiple recipients of list ORACLE-L



I have a problem with a process and its temp segment usage. 

Previously, in RBO (8161) we used to run 8 reports with slightly different
parameters in parallel and it used to work fine with all other load on the
system and the total TEMP space was 6GB

Since we migrated to CBO (all tables  indexes are analyzed using estimate),
these processes fail with unable to extent temp segment error. Currently we
have allocated 12GB to temp tablespace.

I also have performance issues so I am doing a 10046 trace at level 8 on
this one. Is there a way to identify when this process used temp space, how
much and when does it de-allocate?

Would this information be available in 10046 trace or is there something
else that I could do to track that ... 

Thanks in advance 
Raj 
__ 
Rajendra Jamadagni  MIS, ESPN Inc. 
Rajendra dot Jamadagni at ESPN dot com 
Any opinion expressed here is personal and doesn't reflect that of ESPN Inc.

QOTD: Any clod can have facts, but having an opinion is an art! 

-- 
Please see the official ORACLE-L FAQ: http://www.orafaq.com
-- 
Author: DENNIS WILLIAMS
  INET: [EMAIL PROTECTED]

Fat City Network Services-- 858-538-5051 http://www.fatcity.com
San Diego, California-- Mailing list and web hosting services
-
To REMOVE yourself from this mailing list, send an E-Mail message
to: [EMAIL PROTECTED] (note EXACT spelling of 'ListGuru') and in
the message BODY, include a line containing: UNSUB ORACLE-L
(or the name of mailing list you want to be removed from).  You may
also send the HELP command for other information (like subscribing).