Drawbacks to using real time apply?

From: Newman, Christopher <cjnewman_at_uillinois.edu>
Date: Fri, 23 May 2008 08:40:35 -0500
Message-ID: <565F609E6D736D439837F1A1A797F34101981844@ADMINMAIL1.ui.uillinois.edu>


We have 3 standby databases running 10gR2, 2 physical and 1 logical in maximum performance mode. We also have a downstream CDC database using streams which registers its logs from one of the physical standbys. We are looking at switching to real time apply with our dataguard configuration. I don't see any problems with the streams because, in addition to shipping the redo real time, it appears that the archive logs always get shipped as well.

Has anyone encountered any problems with using real time apply? IE, LGWR ASYNC to their standbys? Redo getting backed up under heavy load? Network issues?

What happens if there is a long disconnect between the primary and a standby using real time apply? Will the standby revert to reading the archive log if the real time gap is large enough?

Also - Currently when I query ' select dest_name, status, recovery_mode from V$ARCHIVE_DEST_STATUS;' on our primary DB, the real time apply destination shows as valid, but the recovery_mode shows 'unknown', however the standby shows ' STANDBY_ARCHIVE_DEST VALID MANAGED REAL TIME APPLY'. In fact, it shows that for *all* archive locations on the standby. Is this normal behavior?

Thanks in advance - Chris

Received on Fri May 23 2008 - 08:40:35 CDT

Original text of this message