My query is about checkpoints. Ia m running oracle 7.3.4 in Sun Solaris. According to oracle checkpotints occures primiraily when (While database working normally)
1. When a logswitch occures
2. based on the parameter value LOG_CHECKPOINT_INTERVAL
3. based on the parameter value LOG_CHECKPOINT_TIMEOUT
Well, now my redo log size is 100MB
LOG_CHECKPOINT_TIMEOUT = 0
LOG_CHECKPOINT_INTERVAL=20000000 (Which is grater than the size of the redolog file)
Normally, there are 15 - 20 arcgive log files created every day. ( 15 - 20 logswitches). Bur I found the value of checkpoints is 700 for today. LOG_CHECKPOINT_TIMEOUT=0, ie Time based checkpoint is disabled. LOG_CHECKPOINT_INTERVAL is higher than redo log size. Again 15 -20 logswitches.
My question, What is the reason for this many number of checkpoints.??
Please hel me.
How did you find the checkpoint number? and what is your redo log buffer size?
Is there any possibility that following commands are being issued manually or by some jobs..
ALTER SYSTEM SWITCH LOGFILE;
ALTER SYSTEM CHECKPOINT;
nagarjuna How did you find the checkpoint number? and what is your redo log buffer size?
u can find checkpoint number from the following query
select distinct checkpoint_change# from v$datafile;
we can get from check point details from alert log file also if u have set the
LOG_CHECKPOINTS_TO_ALERT = TRUE .
just count how many check points are happening in a single day .
checkpoint from log switch and LOG_CHECKPOINTS_TO_ALERT are so called official checkpoints or interval checkpoints, internally Oracle performs many incremental checkpoints (from Oracle 8.x) which is quite difficult to follow (at least I dont know where are these incrmenetal checkpoints recorded)
Only way I know if incremental checkpoint has occured is by looking v$sysstat DBWR checkpoint buffers written
Thanks for the replies.
Well, I think my calculation was wrong. I used the value DBWR checkpoints in t V$SYSSTAT to calculate. But later I understand that these are the number of times DBWR requested for a checkpoint, not an actual checkpoint.
According to Mr. Pando, These are the incremental check points. I have no idea about that..!
I appreciate your replies.
hey note i said from Oracle 8.x
Checkpoints still not ends here!
Now I found big values (compared to the number fo checkpoints) in 'DBWR checkpoints' in v$sysstat. These are the requests by the DBWR for a checkpoint. Why the DBWR is requesting for so many checkcpoints? Is this high value is a sysmptom of a problem? How DBWR decide to request for a check point? If this is a problem, how can I correct it?
Thomas P S
Click Here to Expand Forum to Full Width