Home » RDBMS Server » Performance Tuning » Latch Contention
icon4.gif  Latch Contention [message #177593] Thu, 15 June 2006 11:53 Go to next message
vikramjogi
Messages: 71
Registered: June 2005
Location: Scranton
Member
Hello All,

Please advice me on how to resolve latch contention. I have checked the application code, it is making use of the bind variables, I dont think the problem is with the shared pool sizing, I am not sure if the open_cursors and session_cached_cursors parameter values have any affect on the latch contention at all. Please shred some light on setting these parameters.

Thanks much,
Vikram
Re: Latch Contention [message #177595 is a reply to message #177593] Thu, 15 June 2006 12:32 Go to previous messageGo to next message
BlackSwan
Messages: 26766
Registered: January 2009
Location: SoCal
Senior Member
>Please advice me on how to resolve latch contention.
What evidence do you have that demonstrates latch contention is a problem in your database?

What OS & version?
Which version of Oracle?
Re: Latch Contention [message #177599 is a reply to message #177595] Thu, 15 June 2006 13:07 Go to previous messageGo to next message
vikramjogi
Messages: 71
Registered: June 2005
Location: Scranton
Member
Hello,

OS: AIX 5L
Oracle Version: 9.2.0.5

What evidence do you have that demonstrates latch contention is a problem in your database?

I have checked the statspack reports and reads like below

Load Profile
~~~~~~~~~~~~ Per Second Per Transaction
------------ ---------------
Redo size: 8,315.86 18,654.33
Logical reads: 23,967.23 53,763.87
Block changes: 58.65 131.57
Physical reads: 290.10 650.77
Physical writes: 38.80 87.03
User calls: 170.33 382.10
Parses: 108.19 242.70
Hard parses: 1.00 2.24
Sorts: 49.18 110.33
Logons: 0.16 0.36
Executes: 149.96 336.40
Transactions: 0.45

% Blocks changed per Read: 0.24
Recursive Call %: 39.70
Rollback per transaction %: 5.91
Rows per Sort: 50.17

Instance Efficiency Percentages (Target 100%)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Buffer Nowait %: 99.99 Redo NoWait %: 100.00
Buffer Hit %: 98.94 In-memory Sort %: 100.00
Library Hit %: 99.04 Soft Parse %: 99.08
Execute to Parse %: 27.85 Latch Hit %: 99.97
Parse CPU to Parse Elapsd %: 0.42 % Non-Parse CPU: 97.11

Shared Pool Statistics Begin End
------ ------
Memory Usage %: 94.18 95.71
% SQL with executions>1: 86.64 86.71
% Memory for SQL w/exec>1: 99.14 99.17

Top 5 Timed Events
~~~~~~~~~~~~~~~~~~ % Total
Event Waits Time (s) Ela Time
-------------------------------------------- ------------ ----------- --------
latch free 3,153,465 920,160 61.94
SQL*Net message from client 541,226 483,702 32.56
row cache lock 10,401 30,627 2.06
rdbms ipc message 10,117 25,749 1.73
buffer busy waits 7,407 5,919 .40


Looking at the statspack report as far as i know it is pretty clear that latch free is the culprit. Please let me know if you need further information to advise me on what can be done.


Thankyou,
Vikram
Re: Latch Contention [message #177604 is a reply to message #177593] Thu, 15 June 2006 13:28 Go to previous messageGo to next message
BlackSwan
Messages: 26766
Registered: January 2009
Location: SoCal
Senior Member
GOOGLE is your friend, but only when you use it!
Results 1 - 10 of about 48,400 for Oracle latch free waits
Re: Latch Contention [message #177610 is a reply to message #177604] Thu, 15 June 2006 13:59 Go to previous message
vikramjogi
Messages: 71
Registered: June 2005
Location: Scranton
Member
Thanks for the information. I have been googling and as a part of my searching I have posted this question here. And infact I thought that was the reason why we have forums. No one just posts it on the forum with out doing their homework.

Thanks,
Vikram
Previous Topic: queries on collecting schema statistics (oracle 9i)
Next Topic: More "SQL*Net message to/from client" than expected
Goto Forum:
  


Current Time: Fri Apr 26 10:28:50 CDT 2024