On Wed, 7 Jul 2010 17:14:10 -0400, Don Williams <[email protected]> wrote:
>It's been quite a while since we tried GRS star w/DYNDISP=YES (I think it >was z890 or z9). We had one ICF engine that we shared between a >production sysplex (DYNDISP=NO) and development sysplex (DYNDISP=YES). >We tried to use GRS star on our development sysplex (2-systems). GRS star >ran too slow under DYNDISP=YES. Our phones rang off the hook until we fell >back to GRS ring (which requires a sysplex restart). Of course, GRS star >under DYNDISP=NO runs great. > >> -----Original Message----- >> From: IBM Mainframe Discussion List [mailto:[email protected]] On >> Behalf Of R.S. >> Sent: Wednesday, July 07, 2010 4:18 PM >> To: [email protected] >> Subject: Re: Could two parallel sysplex share one CF LPAR >> >> (my $0.02) >>... BTW: GRS star should always >> perform better than ring, even when shared CF engines are in use, >> shouldn't it? >> -- Should, yes, but our experience is more like Don's. Back when we started aggreagating our systems for pricing, we made conscious decision to implement Star with the first two systems. From the start, Star outperformed Ring even in that 2-way. When we added systems from our second CEC (where another self-contained 2-way sysplex ran), we got hints of performance issues. Yadda, yadda, several systems and a couple of months later, we found the right tuning knobs, and life is mucho bettero. I cannot abide managing more than three systems in a Ring. In previous lives I was responsible for configurations, and shared CTC's for Ring got to be a bear at three. Regards, Art Gutowski Ford Motor Company ---------------------------------------------------------------------- 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
