Archives of the TeradataForum

Message Posted: Tue, 05 Mar 2002 @ 15:21:53 GMT


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

Subj:		Priority Scheduler Question:

From:		Garcia Vior Eduardo

We are trying to implement a PSF.

We did a stress test with this configuration:

RP1 wgt 80 (Only one PG active during the test - $F).

RP2 wgt 20 (Only one PG active during the test - $V).

RP0 as default (Only system process active during the test).

Under this scenario ran moderate, simple and complex queries in background at $F, and ran moderate queries at $F and $V in order to measure the impact on response times.

The results weren't what we expected... 'cause the response time at $V was higher than $F, the only query that was running in this PG was a canary...the opposite we expected. Is this a normal situation, is it convenient to implement PSF with these results?

Anybody has overall positive results using PSF??

Thanks!