Re: Additional redo for supplemental logging at table level
From: David Mann <dmann99_at_gmail.com>
Date: Wed, 28 Dec 2011 09:59:05 -0500
Message-ID: <CAGazuyWMF_mSuH3YyNxCJ7H1YZR5vn05-=YX3s3rw3TraBWSYg_at_mail.gmail.com>
>Hi,
>I would like to know if anyone has done any testing on how much
>additional redo is generated if supplemental log groups are created on
>tables. I know that it will depend on the number of columns in a log
>group, column size, data type, etc., but I just want to get an idea
>about generally how much additional information (in terms of %age) is
>generated for DML statements.
Date: Wed, 28 Dec 2011 09:59:05 -0500
Message-ID: <CAGazuyWMF_mSuH3YyNxCJ7H1YZR5vn05-=YX3s3rw3TraBWSYg_at_mail.gmail.com>
>Hi,
>I would like to know if anyone has done any testing on how much
>additional redo is generated if supplemental log groups are created on
>tables. I know that it will depend on the number of columns in a log
>group, column size, data type, etc., but I just want to get an idea
>about generally how much additional information (in terms of %age) is
>generated for DML statements.
I did do formal testing of this for a previous Golden Gate implementation. I don't have exact numbers in front of me but our db was 600gb at the time, we were replicating 400 of 1000 tables (some of the busiest out of the 1000), and generated about 10gb redo a day. After going through all of this we saw our numbers go up by a couple of logs a day but I would definitely say in our case it was max 5% more if that. Of course your mileage may vary depending on the makeup of your DB and types of updates you are throwing at it. Basically everyone was freaking out over nothing :)
-Dave
-- Dave Mann www.brainio.us www.ba6.us - Database Stuff - http://www.ba6.us/rss.xml -- http://www.freelists.org/webpage/oracle-lReceived on Wed Dec 28 2011 - 08:59:05 CST