Archives of the TeradataForum
Message Posted: Mon, 02 Jul 2012 @ 14:58:57 GMT
| Subj: || || Re: DBQL - CPUTIME and CPUTIMENORM |
| From: || || Anomy Anom |
<-- Anonymously Posted: Monday, July 02, 2012 08:45 -->
CPUTIMENORM is scaled or "normalized" to be independent of the node type. On a system with coexistence / co-residence of different node types,
you should only use CPUTIMENORM. Even if all nodes are the same type, CPUTIMENORM is preferred so that you won't have to completely recalibrate
your thresholds when you upgrade.