| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
 |
 | |||
| Oracle FAQ | Your Portal to the Oracle Knowledge Grid | |
|
| |||
Home -> Community -> Mailing Lists -> Oracle-L -> RE: checkpoint incomplete issue
Even though a logfile has been archived, it's contents still need to be
written to disk via a checkpoint before it can be used again. This insures
that instance recovery is possible (you would have to do media recovery
after shutdown abort in some cases if this was not so).
So, what is happening is that your disks are not keeping up with the checkpoint volume. Until the checkpoints can complete, your database will just stall, waiting.
Look at your disks and see if there is something you can do to spread the IO around better. This can mean more disks, or perhaps just moving some datafiles around. I don't know if you are sititng on a SAN or JBODS, but this is the area that you are most likley going to need to address.
Cheers,
Robert
Robert G. Freeman
Management Consultant
TUSC - The Oracle Experts www.tusc.com
OCP - 7, 8, 8i, 9i, 10g
Author:
Portable DBA Oracle
Oracle Database 10g New Features
Oracle9i RMAN Backup and Recovery
Oracle9i New Features
-----Original Message-----
From: oracle-l-bounce_at_freelists.org
To: oracle-l_at_freelists.org
Sent: 11/18/2004 3:29 PM
Subject: checkpoint incomplete issue
I'm running a batch updates on my database and I've
been getting the following messages in my alert file:
"cannot allocate new log
check point incomplete."
My job hangs waiting on log file switch (checkpoint incomplete) event.
I have 8 redo log groups (300MB members). When I did a query on v$log, all of the log members' archived status ='YES' except for the current member. So it appears that the rest of the seven groups have been archived but yet the database could not advanced to the next redo log.
Furthermore, the alert.log does not show frequent log
switches. It switches log every 10mins and about every
5mins, the database complains
about "cannot allocate new log". Does not appear to be
a problem with the number or size of the redo logs.
Anyone has any idea what may be a possible problem?
checkpoint?
thanks.
susan
--
http://www.freelists.org/webpage/oracle-l
--
http://www.freelists.org/webpage/oracle-l
Received on Thu Nov 18 2004 - 16:43:19 CST
 |
 |