Seems to me most of the advice being offered is predicated on being able to separate workload(s) to specific LPARs (CECs).
If you have a requirement to use service classes across disproportionate CECs (say prod on both/all), then WLM is deficient. Borked. Broken As Designed. Velocity, r/t, discretionary ... doesn't matter, it doesn't work as it should. I can construct WLM goals to achieve the results I want *MUCH* better on a ring of monoplexes than I can on a (non-homogeneous) parallel sysplex. IMHO doesn't say much for the WLM (sysplex) design. Maybe I just don't understand WLM. Maybe the people designing this stuff need to get out of the labs and into customer sites. Maybe I'm dreaming ... Shane ... ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [EMAIL PROTECTED] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
