Re: Applying logs to standby in 11g
Date: Mon, 22 Oct 2012 13:17:09 -0600
Message-ID: <CAJzM94C0OXja=rc2BrcwYqgsFWU90x0+WNDjNNf48usJy=89tw_at_mail.gmail.com>
Yeah, it would have helped if I had put the version in the email. I upgraded to 11.2.0.2. When I look at the Data Guard Performance page in EM it shows transport lag as 0.0 and apply lag of 1.0. That's pretty good so at this point I'm going to call it "changes in 11g" and move on. Thanks for the responses.
Sandy
On Mon, Oct 22, 2012 at 1:05 PM, David Fitzjarrell <oratune_at_yahoo.com>wrote:
> Is this 11.2.0.3? If so there is a new wrinkle to the redo mechanism --
> adaptive log file sync which uses foreground processes to 'poll' to see if
> the commit is complete, rather than using semaphores as earlier releases
> did. The semaphore route is still available in 11.2.0.3 and Oracle can
> switch between the two based on load and response time.
>
> This may explain the discrepancy you're seeing with v$archive_log
> 'reporting' log apply.
>
>
> David Fitzjarrell
>
>
> *From:* Sandra Becker <sbecker6925_at_gmail.com>
> *To:* oracle-l <oracle-l_at_freelists.org>
> *Sent:* Monday, October 22, 2012 12:32 PM
> *Subject:* Applying logs to standby in 11g
>
> This pas weekend I upgraded my production OLTP from 10g to 11g (SUSE11 on
> IBM zLinux), then I created an 11g standby. We moved the database to
> newer, faster, correctly configured storage at the same time. All went
> well and everything is functioning better than it did before the upgrade.
> My question is regarding the value in the applied column of v$archived_log.
> Under 10g the value was changed to 'YES' within a few minutes, or less,
> for smaller logs such as less than 1M. Under 11g I don't see that changing
> until a log switch so it appears I am a log behind. I can see the messages
> that media recover has started for one log followed by a message that
> another log is in transit so I am assuming that the first log has been
> successfully applied, but v$archived_log still shows it isn't applied.
> Perhaps that's not a valid assumption.
>
> .
> .
> .
> Media Recovery Log /u06/oradata/prod/arch/1_304_797150180.arc
> Media Recovery Waiting for thread 1 sequence 305 (in transit)
>
> Question: Is this normal behavior in 11g? If not, where should I look to
> see what the cause is? I've been searching MOS and Google but haven't seen
> this scenario addressed yet.
>
> Thanks in advance.
>
> --
> Sandy
> Transzap, Inc.
>
>
> --
> http://www.freelists.org/webpage/oracle-l
>
>
>
>
>
-- Sandy Transzap, Inc. -- http://www.freelists.org/webpage/oracle-lReceived on Mon Oct 22 2012 - 21:17:09 CEST