RE: Configure SDU and TDU parameters to improve Performance

Stephen,

Thanks for the info.  I have lot of soft parsing.  I have setup CURSOR_SHARING 
to SIMILAR.  I am planning to increase  the value for SESSION_CACHED_CURSORS to 
solve soft parsing problem.

Thanks
Muqthar

-----Original Message-----
From: Stephen Andert [mailto:StephenAndert@xxxxxxxxxxxxxxx] 
Sent: Thursday, April 29, 2004 3:21 PM
To: Muqthar.Ahmed@xxxxxxxxxxxxxxxxx; oracle-l@xxxxxxxxxxxxx
Subject: Re: Configure SDU and TDU parameters to improve Performance


Muqthar,

I have done some work with configuring SDU and TDU, but not with RAC. 
To tune SQL*Net-related waits, you can take one of two approaches (or
both):

Approach 1: Make SQL*Net communication more efficient.  This can be
done the way you are suggesting.  In my experience, the time invested in
this resulted in a very minor (almost unmeasurable) improvement that
might have been caused by other factors.  Increasing network bandwidth
and similar activities also fall into this category.

Approach 2: Reduce the SQL*Net traffic.  For example, if you are
parsing a lot, look at cursor sharing and cursor utilization (Steve
Adams has some good scripts for this).  Reducing parsing will reduce
network traffic and make remaining traffic more efficient.  Reducing
unecessary network traffic would also help this as well.

Good luck.
Stephen Andert

>>> Muqthar.Ahmed@xxxxxxxxxxxxxxxxx 04/29/04 08:08AM >>>
Hi,
 
I have a RAC system of 3 nodes with 7 webservers (clients).  I would
like to configure SDU/TDU parameters at client level to improve the
performance.
 
Reference:
Note:69591.1 "SQL*Net Performance Tuning in Applications"
Note:44694.1 "SQL*Net Packet Sizes (SDU & TDU Parameters)"
 
The STATSPACK output is:
 
                                                                       
                                                         Avg
                                                                       
                         Total Wait                 wait            
Waits
Event                                                      Waits       
Timeouts         Time (s)                (ms)               /txn
----------------------------                         ----------------- 
    ----------------     ----------------     -----------------     
---------------
NODE1
 
SQL*Net more data to client                  172,456                   
0                   6                     0               13.6
SQL*Net break/reset to clien                       685                 
  0                    0                     0                 0.1
SQL*Net message from client                728,623                    0
           34,884                   48               57.3
SQL*Net more data from clien                 34,055                   
0                    5                     0                 2.7
SQL*Net message to client                   728,626                   
0                    1                     0               57.3
 
NODE2
SQL*Net more data to client                  208,574                   
0                  19                      0               15.5
SQL*Net break/reset to clien                         10                
   0                    0                      0                 0.0
SQL*Net message from client                852,229                   0 
           36,419                    43               63.5
SQL*Net more data from clien                 41,220                   
0                    7                      0                 3.1
SQL*Net message to client                   852,234                   
0                    2                      0               63.5
 
NODE3
SQL*Net more data to client                  112,425                   
0                    8                     0                16.3
SQL*Net break/reset to clien                         18                
   0                    0                     0                  0.0
SQL*Net message from client               462,005                     0
           20,244                   44                66.9
SQL*Net more data from clien                 22,418                   
0                    5                     0                  3.2
SQL*Net message to client                   462,009                   
0                    1                     0                 66.9
 
Statistic                                                              
  Total                          per Second                      per
Trans
---------------------------------                          
------------------------                     ---------------------      
       ----------------------
NODE1
SQL*Net roundtrips to/from client                         725,847      
                           769.7                              57.1
bytes received via SQL*Net from c                 188,354,007          
                 199,739.1                        14,822.9
bytes sent via SQL*Net to client                     546,604,076       
                   579,643.8                        43,016.0
 
NODE2
SQL*Net roundtrips to/from client                         852,882      
                            916.1                              63.5
bytes received via SQL*Net from c                 224,478,043          
                  241,115.0                        16,722.1
bytes sent via SQL*Net to client                    654,527,562        
                    703,037.1                        48,758.0
 

NODE3
SQL*Net roundtrips to/from client                         462,472      
                            496.8                              67.0
bytes received via SQL*Net from c                 122,075,296          
                  131,122.8                        17,671.6
bytes sent via SQL*Net to client                    353,129,083        
                    379,300.8                        51,118.9
 
Any recommendations??????
 
Cary,
 
Last week I have received your book in the mail.  I started reading and
I am in chapter 4........the concept about performance is completely
different.  Some of the examples / scenarios about managers / developers
/ end users you mentioned are so true.
 
Thanks
Muqthar Ahmed
Database Administrator
DecorateToday.com
 
 

                *       *       *       *       *       *       *       *       
*

The information contained in this E-mail message is privileged,
confidential, and may be protected from disclosure; please be aware that
any other use, printing,copying, 
disclosure or dissemination of this communication may be subject to
legal restriction or sanction. If you think that you have received this
E-mail message in error, please 
reply to the sender and delete it from your computer. Thank you. 

----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com 
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx 
put 'unsubscribe' in the subject line.
--
Archives are at http://www.freelists.org/archives/oracle-l/ 
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html 
-----------------------------------------------------------------

                *       *       *       *       *       *       *       *       
*

The information contained in this E-mail message is privileged, confidential, 
and may be protected from disclosure; please be aware that any other use, 
printing,copying, 
disclosure or dissemination of this communication may be subject to legal 
restriction or sanction. If you think that you have received this E-mail 
message in error, please 
reply to the sender and delete it from your computer. Thank you. 
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Other related posts: