RE: Logical Standby Issues (cont.)

  • From: "ramick" <ramick@xxxxxxxxxxx>
  • To: <strickland.mark@xxxxxxxxx>, "'Ric Van Dyke'" <ric.van.dyke@xxxxxxxxxx>
  • Date: Wed, 26 Jul 2006 19:41:08 -0700

"some better settings for SQL Apply parameters"

Would you please post these and the reasons for them?

-----Original Message-----
From: oracle-l-bounce@xxxxxxxxxxxxx [mailto:oracle-l-bounce@xxxxxxxxxxxxx]
On Behalf Of Mark Strickland
Sent: Wednesday, July 26, 2006 2:33 PM
To: Ric Van Dyke
Cc: oracle-l@xxxxxxxxxxxxx
Subject: Re: Logical Standby Issues (cont.)

I've completed a test and verified that Logical Standby is not doing
full table scans.  Yes, it does convert the single update statement
(update table_name set newcolumn=existingcolumn) to individual
row-level update statements.  When I ran one of those update
statements through Explain Plan, it showed that it would do a full
table scan, but when I actually looked in v$sql_plan for the statement
currently executing, I could see that it is indeed doing an index
unique scan.  Oracle Support provided some better settings for SQL
Apply parameters and that seems to be helping, but it still takes
about 4 minutes for SQL Apply to chew through each 10-Mb archived log.
 At this rate, it would take about 30 hours to get through them all.
This would not be tolerable in Production if we were relying on our
Production Logical Standby for reporting.


Other related posts: