Re: Problem wih Oracle reads from the redo logs.

From: Vitaliy Mogilevskiy <vit100gain_at_earthlink.net>
Date: Tue, 13 Apr 1999 17:55:27 -0700
Message-ID: <3713E77E.F76307DD@earthlink.net>

Ravi Kommalapati wrote:

> Hello,
> We have created a volume for redo logs by striping across 6 disks
> using the Veritas volume manger, and we have 4 redo log files 250M each,
> there are no other files on this volume other than the redo logs.
> Currently the redo logs switch every 7 minutes during peak usage.

===> Striping REDO LOGS is not a good idea since LGWR writes sequencially

• Tuning Redo Log Buffer ** - to see when check points occure set INIT.ORA you'll get it in the ALERT FILE !!! LOG_CHECKPOINTS_TO_ALERT=TRUE
  • to turn EXTRA checkpoints OFF set: LOG_CHECKPOINT_INTERVAL=<LARGEST LOG SIZE> LOG_CHECKPOINT_TIMEOUT=FALSE

Run this select statement:

select (req.value*5000)/entries.value "Ratio" from v$sysstat req

,      v$sysstat entries
where  req.name        = 'redo log space requests'

and entries.name = 'redo entries'

=========> "Ratio" should be < 1 !!!!!!!!!

+--------------------------------------------------------------------------+
| Vitaliy Mogilevskiy
| Senior Consultant
| CORE Technology Group, Inc.
| E-mail:    vit100gain_at_earthlink.net
| Web Page:  http://home.earthlink.net/~vit100gain/index.html

|            *** Free DBA Script Library at my Web Page ***
+--------------------------------------------------------------------------+

Received on Tue Apr 13 1999 - 19:55:27 CDT