On Fri, Apr 16, 2010 at 9:00 AM, Arthur Gutowski <[email protected]> wrote:
> On Thu, 15 Apr 2010 13:46:26, Ted MacNEIL <[email protected]> wrote: > > >Third, with the restructuring of the way initiators work, where draining > an init > actually terminates the offending address space underpinning that init, > rather > than a random one, we found most of the reasons for the exit disappeared. > > Not that JES flavor was specified, but with JES3, draining the init takes > more > effort. Quiesce the entire class/group (set INIT counts to zero and wait > for > jobs to complete), then restart the class/group. > > Is this still effective with WLM-managed INITs? We don't use them, but > ISTR > you lose the ability to manually drain/start inidividual initiators, and > therefore > would have to quiesce an entire job class. > > In these cases, IEFUSI seems like the easier path. CHECKREGIONLOSS looks > interesting, too... regardless of (ahem) kludges like the region loss thing, the important point to remember is that the system needs some space to be reserved at the high end of below the line and at the high end of below the bar for LSQA, ELSQA etc. The down-side if you don't leave enough space for the system areas can be quite disastrous. So if you're using IEFUSI don't just pin the respective region-below-16MB and above region-above-16MB sizes up against the top of their respective address ranges - even if the JCL specifies 0M. For below the line you should reserve at least a few hundred KiB and above the line you could reserve several MiB without hurting anyone's feelings. -- This email might be from the artist formerly known as CC (or not) You be the judge. ---------------------------------------------------------------------- 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
