Gee hasn't this turned into a long running thread? I have worked in many places where the sysprog's job was to say no loudly and often. I was a sysprog myself. BTDTGTS. I confess to being largely on the side of the user community in these cases. But that's just my own bias showing.
Yes, you do need to ensure you have enough paging space to support the workload. Not some abstract wild-eyed theoretical "oh my god they're eating my storage" workload, but the one you really have. And yes, it is indeed possible to get a WAIT03C by just filling AUX storage. But you could do that before 64-bit and you still can with badly behaved 31 bit programs. Easy. By way of a refresher... allocating virtual storage costs you almost nothing. A few fixed frames for the region (64-bit only) page and segment tables. Using the virtual storage costs real frames, but only the ones you "touch" (clearing it, putting data in it etc.) and, should the OS ever need to reclaim that storage for a more deserving workload, the storage has paged to aux or discarded if unchanged. If they are doing their jobs effectively, DBAs will size their buffer pools to support their actual workload and data volume, plus some reasonable overage. That's a good thing and there are tools to help them do that. We even sell a few. If they go nuts and overallocate... so what if its not used? There are still very very few 64-bit exploiters. I doubt they are going to go postal on you (in general) and there's ZERO chance an existing non-64-bit program is going to do anything at all above the bar. Other than the unarguable truth that "above the bar" is a really big place, why are we obsessing about it? CC ---------------------------------------------------------------------- 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
