Home » RDBMS Server » Performance Tuning » RedoLog Contention
RedoLog Contention [message #179764] Wed, 28 June 2006 11:23 Go to next message
Kavitha Suresh
Messages: 22
Registered: September 2005
Junior Member
According to Metalink Note:147471.1, it is specified that if the value obtained from query below is not near to zero it is an indication of RedoLog Contention

SELECT name, value
FROM v$sysstat
WHERE name = 'redo log space requests';

We have a DB where I see that this value is 259. Also, we have a Data warehouse batch process in which we have an INSERT INTO ..SELECT which takes eternal and does not return back.

However when the SELECT is run independent of the INSERT, the query returns back in about 10 mins.

Does this mean that we have a RedoLog Contention? Can anybody help me in understanding why we could see such a problem. Any help is highly appreciated.

Thanks,
Kavitha
Re: RedoLog Contention [message #179768 is a reply to message #179764] Wed, 28 June 2006 11:46 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10707
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
>> we have an INSERT INTO ..SELECT
Are you using this along with nologging/append and disabled indexes?
Re: RedoLog Contention [message #179771 is a reply to message #179768] Wed, 28 June 2006 11:50 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10707
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
Also look into alert log file.
Look for "checkpoint not complete".
How frequent is this?
And post any other important event that was reported in log.
How big is your redo log file?
Re: RedoLog Contention [message #179773 is a reply to message #179768] Wed, 28 June 2006 11:54 Go to previous message
Kavitha Suresh
Messages: 22
Registered: September 2005
Junior Member
Indexes are not disabled.
Also, we are not using Append hint or nologging option..
We tried dropping the index and gave it a shot ...it did not help.
The DBA claims we donot have any fragmentation ..
The module runs perfectly fine in all other higher/lower environments with volumes higher than what is present in this environment.

We did not have any wait events either ...
Previous Topic: slow insert/update - same schema on two servers
Next Topic: NORMALIZATION AND DENORMALIZATION
Goto Forum:
  


Current Time: Fri Apr 26 21:46:57 CDT 2024