Logical Standby Issues (cont.)

10.1.0.3 RAC on Solaris 9 w/ Physical and Logical Standby

Yep, SR is open.

Had another incident Friday night where a column was added to a
"large" table (not that large) in the primary database and then the
column was set to a value with a single update statement.  As before,
the DDL got applied to the logical standby immediately but then SQL
Apply slowed to a crawl.  I'm running a test in a test environment now
because I suspect that Oracle is taking that one update statement and
converting it into a separate update statement for each row and each
of those updates is doing a full table scan (yes, all the tables that
are being maintained by SQL Apply have primary/unique keys).  I'll
know more soon after my testing is complete, but am I insane in
thinking that Oracle should just run the single update statement
against the entire table rather than converting that into single-row
updates?  Or am misunderstanding how LogMiner creates Logical Change
Records?  I'm still learning how to query the various views related to
logical standby so I didn't do a great job digging into what was
really happening Friday night.

Regards,
Mark Strickland
Next Online Technologies
Seattle, WA
--
http://www.freelists.org/webpage/oracle-l


Other related posts: