Home » RDBMS Server » Server Administration » LOG MEMBER??
LOG MEMBER?? [message #55724] Wed, 12 February 2003 07:49 Go to next message
raman
Messages: 66
Registered: February 2000
Member
Hi

I see the follwing statement at several places, in alert log file(bdump).

Thread 1 cannot allocate new log, sequence 16478
Checkpoint not complete.

Do I have to add a log member or what?

thanks,
-ramanujam

Report message to a moderator

Re: LOG MEMBER?? [message #55726 is a reply to message #55724] Wed, 12 February 2003 08:33 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10708
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
you need larger redolog files.
and also
check your LOG_CHECKPOINT_INTERVAL
LOG_CHECKPOINT_TIMEOUT
and tune them

Report message to a moderator

Re: LOG MEMBER?? [message #55728 is a reply to message #55724] Wed, 12 February 2003 10:08 Go to previous messageGo to next message
raman
Messages: 66
Registered: February 2000
Member
Could you please tell me how can I increase the size of the redolog file? Did you mean adding a member?

At present log_checkpoint_interval:1000
and log_checkpoint_timeout:180

thanks,
-Ramanujam

Report message to a moderator

Re: LOG MEMBER?? [message #55738 is a reply to message #55724] Thu, 13 February 2003 04:33 Go to previous messageGo to next message
seshachalapathi andukuri
Messages: 2
Registered: February 2003
Junior Member
We have Oracle 8.1.6 , with 4 redo log members each with 100 MB size. We are also having same problem.
Instead of adding more redo log members we changed no of db writer process from 1 to 2. And we observe time taken for completing a checkpoint seems to be main bottleneck. Which means system I/O is the bottleneck, and system I/O is not coping with high volume of DML generated in the system.
Please comment on this

Report message to a moderator

Re: LOG MEMBER?? [message #55743 is a reply to message #55724] Thu, 13 February 2003 05:39 Go to previous messageGo to next message
Mahesh Rajendran
Messages: 10708
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
please post the exact message as appearing in alertlog or trace.
what is the frequency of logswitch??
if it is too less, then u may need to tune this.
You can apply the same policies like ramanujam.
if you have issues, pls let us know.

Report message to a moderator

Re: LOG MEMBER?? [message #55745 is a reply to message #55743] Thu, 13 February 2003 09:03 Go to previous messageGo to next message
seshachalapathi andukuri
Messages: 2
Registered: February 2003
Junior Member
The following message appears in alert file.
--------
Thread 1 cannot allocate new log
Checkpoint not complete.
----------
This message is used to appear 2-3 times a day (some time more) before we changed Db_writer_processes from 1 to 2 and checkpoint interval from 300,000 to 400,000 os blocks. Redo log size is 100Mb.And also observed checkpoint is taking long time to complete. So we thought by increasing Redo logs we can only stop the the above messages but actual problem, is not solved. We are basically looking why Dbwr process is taking too long to write dirty blocks to disk?
(we are on Solaris 8, oracle 8.1.6).

Report message to a moderator

Re: LOG MEMBER?? [message #55747 is a reply to message #55743] Thu, 13 February 2003 09:35 Go to previous message
Mahesh Rajendran
Messages: 10708
Registered: March 2002
Location: oracleDocoVille
Senior Member
Account Moderator
since you have already increased db_writer
try
decreasing log_buffers (say around 200 to 250k)

Report message to a moderator

Previous Topic: Install 9i and 9iAS on unix machine.
Next Topic: View has the same name of another schema
Goto Forum:
  

[ Syndicate this forum (XML) ] [ RSS ]

Current Time: Sat Dec 28 09:15:40 CST 2024
.:: Forum Home :: Blogger Home :: Wiki Home :: Contact :: Privacy ::.