Re: Logical Standby Issues (cont.)

  • From: Carel-Jan Engel <cjpengel.dbalert@xxxxxxxxx>
  • To: strickland.mark@xxxxxxxxx
  • Date: Tue, 08 Aug 2006 08:12:59 +0200

Mark,

A few remarks.

First: 10 Mb redo logs seem rather small to me. Try bigger ones.
Frequent logswitching requires a lot of checkpointing etc. Do you get
any information on this from the alert.log? 

Second: a script with 1,000,000 statements circumvents the fact that
every single SQL in a Logical Standby has to be generated in place. That
is faster whatsoever.

Third: Does your script the same update? In other words, does it contain
the same set of unique statements, in the same order? If the original
update processes the table in physical order, and your script ordered
the statements differently, that can affect the order blocks have to be
read from disk, and therefore generate another type of load.

Fourth: Did you consider to perform a Cary (trace 10046) on the SQL
Apply process? Maybe you get some insight at that point. I'm curious to
see the results.

Just my EUR 0.02



Best regards,

Carel-Jan Engel

===
If you think education is expensive, try ignorance. (Derek Bok)
===

Other related posts: