Jeroen,
I guess there a two ways to look at this:
1). 2-4 hours log switch during the day is ok. every minute at night is a
bit costly. however, if your batch is completing in a timely manner (is it
*really* causing operational problems with the frequent log switches?) then
leave it alone.
2). if you want to create logs less frequently at night, you obviously need
larger redo logs. but that means that your daily activity would produce
fewer redo logs - thus you are accepting more risk for disk failure and loss
of data. you *could* increase your logs files and have a cron job run every
2 hours during the day that simple says "alter database archive log switch"
thus forcing a switch and producing an archive log file. this would give
you the performance at night and alleviate the risk during the day.
I would *never* drop and recreate log files every day as an operational
procedure. This to me is just too risky for what you are trying to do. I
would choose between #1 or #2 above.
hope this helps.
Tom Mercadante
Oracle Certified Professional
-----Original Message-----
From: Jeroen van Sluisdam [mailto:jeroen.van.sluisdam@(protected)]
Sent: Tuesday, April 06, 2004 4:28 AM
To: 'oracle-l@(protected)'
Subject: different redo needs same database
Hi,
I'm trying to improve my redo settings because I have logswitches during
daytime every
2 to 4 hours and at night during batches every couple of minutes and a few
times
within 1 minutes 2 logfile switches. With statspack (interval complete day
including night(
I also see
Avg
Total Wait wait Waits
Event Waits Timeouts Time (s) (ms) /txn
---------------------------- ------------ ---------- ---------- ------
--------
LGWR wait for redo copy 99,642 1,642 68 1 0.1
Details: Oracle 9.2.0.4, HP-UX 11.11,
I have 3 redo loggroups, each 2 members of 150Mb. All files are on a SAN
(EVA3000, raid 1)
SQL> show parameter log_check
NAME TYPE VALUE
------------------------------------ -----------
------------------------------
log_checkpoint_interval integer 5000
log_checkpoint_timeout integer 60
log_checkpoints_to_alert boolean TRUE
SQL> show parameter fast
NAME TYPE VALUE
------------------------------------ -----------
------------------------------
fast_start_io_target integer 0
fast_start_mttr_target integer 0
fast_start_parallel_rollback string LOW
I am thinking of maybe creating a script that creates small loggroups and
delete large groups for use during daytime
and creating another script with large loggroups for batch purpose. Any do's
and don't advice on these experiences are appreciated.
Tia,
Jeroen
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@(protected)
put 'unsubscribe' in the subject line.
--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to: oracle-l-request@(protected)
put 'unsubscribe' in the subject line.
--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
