Scot covered most of this - most importantly the need to inspect the workloads on the LPAR(s) of interest.
Despite being for DEVL, the latter report is better as the machine is now busy. So let's turn the question around: Q: Is PROD getting all the CPU it wants ?. A: YES. The weights indicate you (somebody) thought PROD should be using ~73% of the box at full load. In this period it was achieving ~62%. If it wanted more it would drive the other LPARs down to achieve at least that proportion (long term averages). If work is perceived as being delayed, that says to me WLM goals need (serious) attention. But there are other factor to be considered: - is this a base sysplex or separate monoplexes - how much sharing between LPARs; is GRS active; are the rules appropriate - is/are the WLM policy(s) appropriate for DEVL vs PROD - CPU is not the only delay reason. - why do you have a capacity group There is no "silver bullet", just lots of legwork to be done. Shane ... (and yes, uniprocessors are the pits) ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
