-
very Frequent redo switch
Hello Everybody,
see the time below, for each redo switch from server
MINUTES
----------
13.9
9.95
5.93
5.7
6.17
6.17
5.87
6.17
6.93
7.05
6.35
6.83
The size of each archive is also 1G.
1048574976 Nov 15 07:12 _8431.arc
1048574976 Nov 15 07:29 _8432.arc
1048568832 Nov 15 07:47 _8433.arc
1048574976 Nov 15 08:07 _8434.arc
1048572928 Nov 15 08:23 _8435.arc
1048574976 Nov 15 08:39 _8436.arc
1048574976 Nov 15 08:56 _8437.arc
1048572928 Nov 15 09:14 _8438.arc
My redo log size is 1G. The log block size is 1024 on HP. In my DB log_checkpoint_timeout is set to 86400, but Oracle recommends to set log_checkpont_interval which should be calculated as log block size.
1) I would like to set my checkpoint_interval = (1G*1024*1024*1024)/1024 is that fine?
2) I am thinking to increase the size of redo log size to 2G.
Please suggest me what should I do?
Regards
Nwcomer
Student
-
I'm not sure what "MINUTES" is, but your archive activity doesn't look bad to me. Do you get any "Checkpoint not complete" messages in your alert.log?
Jeff Hunter
-
Minutes = each redo switch time difference.Not always, but very often the time diff is only 3 -4 minutes.
No I did not get any "checkpoint not complete" error.
But I am thinking about Load of redo switch. Frequent redo switch has lot of overheads and performance can degrade since each datafile will be frozen during each redo switch.
What do you suggest?
Regards
Nwcomer
Student
-
tune your code to produce less redo - simple
-
The application has to do lot of updates.
Nwcomer
Student
-
Originally posted by newcomer
The application has to do lot of updates.
Then the database will have lots of redo.
Jeff Hunter
-
If it's a warehouse or something, you might want to look into NOLOGGING operations.
-
That does not look bad to me.
Is this consistant or were they running a large job?
Has DB performance suffered recently?
I have seen them less than 30 seconds apart, but it does not last for too long so no cause for alarm.
I remember when this place was cool.
-
Originally posted by Axr2
If it's a warehouse or something, you might want to look into NOLOGGING operations.
That's a pretty big decision to make -- wouldn't help with updates of course.
-
Originally posted by slimdave
That's a pretty big decision to make -- wouldn't help with updates of course.
I was just providing him/her food for thought, that's all.
Posting Permissions
- You may not post new threads
- You may not post replies
- You may not post attachments
- You may not edit your posts
-
Forum Rules
|
Click Here to Expand Forum to Full Width
|