Archives of the TeradataForum

Message Posted: Fri, 31 Oct 2003 @ 12:32:10 GMT

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

Spool on development varies in our environment depending on the development efforts.

We make it flexible so the developers can develop the applications and work out the kinks before we go to test.

In our test environment, we allocate an estimated amount of spool based on the current queries and customer requirements that we have to work with, which is never enough information, to determine how much spool to allocate to mimic a production system.

The percentages of spool to CurrentPerm is 30% so if you have 100 gig on test the user gets 30 gig of spool.

This is based on power users and starts high once in production. We then use PeakSpool to determine what the users need to run their reports.

The ClearPeakDisk macro clears the users peak spool so we have a measure on whether to increase or decrease spool allocation for each user.

Remember, spool is the users work area and is also the speed bump for runaway queries. It is reallocated once the query has completed so you will have over allocation of spool on the system. This is not always a bad thing and cannot be helped.

Monitor spool usage on a daily basis (as if we have all the time in the world) but it is worth it. A healthy system with great performance is always going to have a strong DBA team that monitors every aspect of the system from spool allocation and usage to runaway queries and more.

Chris Coffing
Senior Vice President
Coffing Data Warehousing
www.coffingdw.com

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