On Fri, 5 May 2006 12:26:38 -0700, Edward Jaffe wrote: >Thomas H Puddicombe wrote: >> One way would be to define a third LPAR. Assign a weight to the new LPAR >> equivalent to 4/32 of the total of all the LPAR weights. Cap this new >> LPAR. Put a looping task into this LPAR. PR/SM will then ensure that no >> more and no less than 4/32 of the total CPU is wasted running a silly loop. >> Whatever's not used by the loop is available to the other LPARs. Warn >> capacity planners and resource billing people that there's a "soaker" >> workload. >> > >I don't get it! Is this "soaker" some sort of stand-alone IPLable >program you've written? Or do you need to IPL z/OS to run it? If the >latter case, how does it keep your z/OS VWLC peak from reaching 32 MSU? >Is there some sort of VWLC exemption for running a z/OS "soaker"? Does >IBM allow you to exclude that LPAR's SMF records from your SCRT input?
Your point is well taken, Ed, but Mr. Puddicombe COULD be running the 'soaker' under a zLinux system, couldn't he? (Wouldn't that allow him to absorb the CPU and not have that LPAR participate in the VWLC licensing? I'm not sure about that myself, but they wouldn't be z/OS MSUs...) I wouldn't ordinarily advocate use of zLinux for the purpose of mopping up 'excess' MSUs but then again this mess is due more to the burdensome software charges than anything else (as far as I've been able to see in my customer-world experience). -- Tom Schmidt Madison, WI (zLinux doesn't REQUIRE the use of an IFL, but every place I've been WANTED to use an IFL for it, for mostly the opposite reasons.) ---------------------------------------------------------------------- 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
