Archives of the TeradataForum

Message Posted: Fri, 19 Jun 2009 @ 11:20:53 GMT

	<Prev	Next>		<<First	<Prev	Next>	Last>>

Hi,

No. If a query is doing any work then CPU, IO etc will be increaing and spool may or may not be changing. By themselves they don't show whether or not the query is good or bad (or ugly !).

If you're only looking at CPU and IO then you can use the cpu:disk ratio (aka the "Larry Higa ratio"). This is calculated as:

CPU milliseconds per disk io

So:

CPU * 1000 / disk io count

Different people use different good/bad numbers but I think that you're generally looking for values < 10. Very high values for this are indications of a large product join.

Also look for parallel efficiency (or 'skew' as it's shown in pmon). The skew numbers shown in pmon should be low - conversely if you're looking at parallel efficiency then the numbers should be high !

Hope that helps.

Dave

Ward Analytics Ltd: Information in motion ( www.ward-analytics.com )

	<Prev	Next>		<<First	<Prev	Next>	Last>>