Archives of the TeradataForum

Message Posted: Wed, 09 Feb 2005 @ 20:48:45 GMT

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

I didn't see where anyone responded to this, so let me give you my two cents as someone who's spent a lot of time over the last few years staring at CPU and IO utilizations.

An SQL statement that is executed in exactly the same way from execution to execution might very slightly from execution to execution, but usually not by an appreciable amount.

However, queries that give the same explain plan give different results based on what actually happens during the execution of the plan. The most common reasons that I can explain are:

1) FSG cache - The small tables in your query are placed into FSG cache.

2) Full Cylinder Reads - If eligible and there is an available slot.

3) Syncronized Scans - If eligible and someone else is participating.

4) Memory Paging - I think I tied addition resource utilization to high paging counts some time back, but I haven't been at a site where paging has been an issue for awhile.

Other observations that I can't really explain are:

Your system is reaching some saturation point. Either flow control or AWT exhaustion seems to increase the amount of resources that are required to fulfill the requests. Don't know why this is, but I've observed this at multiple sites.

Robert Meunier

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