Re: Status of online redo logs on the standby database
- From: Laimutis.Nedzinskas@xxxxxx
- To: oracle-L@xxxxxxxxxxxxx
- Date: Mon, 18 Jan 2010 10:21:35 +0200
You obviously went over this, didn't you:
http://www.oracle.com/technology/deploy/availability/pdf/MAA_WP_10gR2_RecoveryBestPractices.pdf
http://www.oracle.com/technology/deploy/availability/pdf/MAA_WP_10gR2_DataGuardNetworkBestPractices.pdf
?
from all that reading I got this bottom line:
Tune I/O subsystem, including some kernel parameters tweaking and
PARALLEL_EXECUTION_MESSAGE_SIZE = 64K
I personally set PARALLEL_EXECUTION_MESSAGE_SIZE = 64K by default on all
databases that have plenty of RAM.
Plus we had to tune SQLnet and TCP/IP related stuff cause LN.... waits were
taking their share.
---------------------------------------------------------------------------------
Please consider the environment before printing this e-mail
Christopher Reeve
<topher_reeve@me.
com> To
Sent by: oracle-L@xxxxxxxxxxxxx
oracle-l-bounce@f cc
reelists.org
Subject
Status of online redo logs on the
2010.01.17 17:02 standby database
Please respond to
topher_reeve@xxxx
om
Hello listers
My first post...
I have a data guard environment running on 10.2.0.4 on solaris 64. I have
had some performance issues, on the standby side of things, concerning the
rate at which redo is getting applied. Except for the disk subsystem, the
box itself is bored. So I'm trying to figure out how to reduce the
physical I/O. The environment itself is managed by the DG broker, and I
have it set up to do real time apply, which works great most of the time,
until we get to our batch processing periods, four times a month, then the
standby begins to lag. One way of reducing I/O, I have read, is to NOT
multiplex the standby redo logs. Certainly makes sense, with all the ways
of getting redo over. While that did improve performance and reduce I/O,
I'm still experiencing periods of lag. Is that enough background
information? OK, on to the question...
With my situation, real-time apply, etc., I expected to see activity in v
$standby_log, which I do. But I didn't expect to see activity in v$log,
which I do. I have not seen any status here besides "CLEARING" and
"CLEARING_CURRENT". So is this normal? If not, is this causing additional
I/O that I can reduce further?
Thanks,
Topher
--
//www.freelists.org/webpage/oracle-l
--
//www.freelists.org/webpage/oracle-l
Other related posts:
