Oracle AWR - Time Model Statistics / background elapsed time < background cpu time

From: Stefan Koehler <contact_at_soocs.de>
Date: Fri, 13 Mar 2015 12:46:33 +0100 (CET)
Message-ID: <795236161.206244.1426247193633.JavaMail.open-xchange_at_app01.ox.hosteurope.de>



Hi list,
today i received an AWR report from an Oracle 11.2.0.3 3-node RAC system and found something strange in the "Time Model Statistics" area. I personally don't like AWR reports very much (more a fan of the Method-R approach and detailed problem specific performance data), but right now this AWR is all i got.

The "Time Model Statistics" for background looks like this in the AWR: background cpu time = 47,342.32
background elapsed time = 33,676.89

The tree for background statistics looks like this regarding the Oracle documentation ( http://docs.oracle.com/cd/E11882_01/server.112/e40402/dynviews_3015.htm#REFRN30340 ):

  1. background elapsed time -- 2) background cpu time

The Oracle documentation also states that "The relationship between a parent and a child in the tree indicates containment only. Keep the following in mind with regard to the tree: Children do not necessarily add up to the parent.", but this does not make really sense in my opinion. The "background elapsed time" has to be larger than 33,676 seconds or even equal to 47,342 seconds, if the background process needed round about 47,342 seconds CPU time. I also searched MOS, but did not found any regarding bug.

However if i add up the "background cpu time" and "DB CPU" (for foregrounds), i got the Total CPU of the Instance CPU. So it seems like the "background cpu time" is the correct value.

Am i missing something here as the value of "background elapsed time" does not look logical to me?

Thank you guys.  

Best Regards
Stefan Koehler

Freelance Oracle performance consultant and researcher Homepage: http://www.soocs.de
Twitter: _at_OracleSK

--
http://www.freelists.org/webpage/oracle-l
Received on Fri Mar 13 2015 - 12:46:33 CET

Original text of this message