Re: "Checkpoint not complete" message in alert.log file

Ken Nichols wrote:
>
> Hi everyone,
>
> I see this message from time to time in alert_RCM.log (our database name
> is 'RCM'). Does the "Checkpoint not complete" message mean that
> anything is wrong? I have not been able to find a reference to this
> message anywhere in the docs.
>
> Thanks,
>
> Ken Nichols
>
> --------- log excerpt ------------
>
> Fri Sep 19 11:57:43 1997
> Thread 1 advanced to log sequence 1234
> Current log# 1 seq# 1234 mem# 0: /oracle/oradata/RCM/redoRCM01.log
> Thread 1 advanced to log sequence 1235
> Current log# 2 seq# 1235 mem# 0: /oracle/oradata/RCM/redoRCM02.log
> Thread 1 advanced to log sequence 1236
> Current log# 3 seq# 1236 mem# 0: /oracle/oradata/RCM/redoRCM03.log
> Thread 1 cannot allocate new log, sequence 1237
> Checkpoint not complete
> Current log# 3 seq# 1236 mem# 0: /oracle/oradata/RCM/redoRCM03.log
> Thread 1 advanced to log sequence 1237
> Current log# 1 seq# 1237 mem# 0: /oracle/oradata/RCM/redoRCM01.log
> Thread 1 advanced to log sequence 1238
> Current log# 2 seq# 1238 mem# 0: /oracle/oradata/RCM/redoRCM02.logWhat this means is that your redo logs are most likely too small or too
few. More often than not, it is the first. If you will notice, this error usually occurs when the database is in a high state of inserts, updates or deletes. This error means that all of the redo logs are writing and when trying to access the next in the order, it is still writing. I refer to this as spinning. Check in your v$sysstat parameter for redo log space requests. You will most likely find that this is a non-zero value, possibly quite high. Increase the size of your redo logs, and this should alleviate your problem.

-Sheilah Scheurich
DBA
to respond via email, remove the underscores. Received on Wed Oct 01 1997 - 00:00:00 CDT