RE: Logical Standby Issues (cont.)

  • From: "Rich Amick" <rAmick@xxxxxxxxxxx>
  • To: "'Mark Strickland'" <strickland.mark@xxxxxxxxx>
  • Date: Thu, 10 Aug 2006 16:01:59 -0700

Ahh, only one apply process monopolizing one CPU.

Having more than one when you split up the work makes sense to me.  If you
committed after every batch of 100,000 rows, each batch being a separate
transaction, they can be parallelized.

What are the top 5 timed events near the end of each test (mining and
apply)?  (looking for an increasing latching problem causing CPU spin taking
away CPU from the single apply process)

-----Original Message-----
From: Mark Strickland [mailto:strickland.mark@xxxxxxxxx] 
Sent: Thursday, August 10, 2006 3:49 PM
To: Rich Amick
Cc: oracle-l@xxxxxxxxxxxxx
Subject: Re: Logical Standby Issues (cont.)

Oh, also, the Applier process was driving one CPU to 100% (4-cpu box).
 Only one Applier process was used.  Interestingly, when I tested with
10 batches of 100,000 rows each, two Applier processes were used.
Don't know why.  I guess that as Log Miner worked its way through the
archived logs, each time that it got through one batch and started
another one, there were probably two streams available for SQL Apply
to work on.  Guessing...



Other related posts: