"cmccorma" <[email protected]> wrote in message news:<c8fe61be-9f22-495c-9248-cf24e29ab...@n11g2000yqb.googlegroups.com>... On Aug 4, 2:58 am, [email protected] (Vernooij, CP - SPLXM) wrote: > "Pommier, Rex R." <[email protected]> wrote in message > <news:[email protected]>.. > . > > Kees, > > > Since nobody else chimed in here, I will. I think the system is > working > > as advertised. If you check the PAGEDEL or PAGEADD command in the > > system commands reference, it says (at least my really old hardcopy) > > > <quote> > > > When you delete a page data set, the system migrates the in-use slots > to > > other data sets before it deletes the data set. > > > The system keeps track of the in-use slots on both the old or deleted > > data set and the new data set until the owner references the pages. > > Thus, when you issue a PAGEADD command to allocate a new data set, the > > system might indicate that some slots on the newly allocated data set > > are already in use. As soon as the owner references a page, the > system > > frees the slot both from the newly allocated data set and from the > data > > set to which the page was migrated. > > > </quote> > > > IIRC, the reason for this is that the system keeps track of where the > > in-use pages are located by page data set. When you delete a data > set, > > the system physically moves the pages to other data sets, but keeps > the > > page table for the original data set in memory, with just a pointer to > > the new location of the pages. Since these pages are in use within > the > > table, when you allocate and activate a new data set, even though the > > physical pages in your new data set aren't in use, the system can't > use > > them because the page table is still referencing the old pages by the > > location where the pages were before you did the PAGEDEL. (or > something > > like that) > > > HTH > > > Rex > > Hi Rex, > > Yes, that explains what I see. It is indeed what I feared, the system > keeps track of frames that were in the old pagedataset, only I was > afraid it would start looking for them again in the new pagedataset. You > can imagine what chaos that would cause in a production system. But > luckily it knows what it is doing. > > It is in my 1.8 system commands manual too. However I never thought of > look for 'system internals' in the 'system commands' manual. > > Thanks a lot, > Kees. > ********************************************************************** > ********************************************************************** > > ---------------------------------------------------------------------- > For IBM-MAIN subscribe / signoff / archive access instructions, > send email to [email protected] with the message: GET IBM-MAIN INFO > Search the archives athttp://bama.ua.edu/archives/ibm-main.html- Hide quoted > text - > > - Show quoted text -
Rex's explanation is correct. This is not a new "problem". It has worked the same since way back in the days of MVS/XA. As the page frames eventually get referenced then the counts will become closer to correct. Two things to watch out for: 1- if do many pagedel and pageadd's, such as to reallocate all of your local pages in a round robin fashion, you might run into a situation where the datasets appear artificially "too full" to allow more pagedel's. In that case you might have to wait a while for the counts to correct themselves before continuing. 2- some tasks (ie: TCPIP, Websphere, ZFS) can create a large number of page frames early on in their life, and then never touch them again. If these frames are written out to the local page you tried to delete, the in-use counts might never be corrected until the task is recycled. Chris Thanks Chris, In a way they are 'artificially' full, but as ASM is seeing it, the slots are really in-use and contiguous slot allocation will start working less efficiently on these page datasets. So this 'virtual' problem will also be a 'real' problem. Kees. ********************************************************************** For information, services and offers, please visit our web site: http://www.klm.com. This e-mail and any attachment may contain confidential and privileged material intended for the addressee only. If you are not the addressee, you are notified that no part of the e-mail or any attachment may be disclosed, copied or distributed, and that any other action related to this e-mail or attachment is strictly prohibited, and may be unlawful. If you have received this e-mail by error, please notify the sender immediately by return e-mail, and delete this message. Koninklijke Luchtvaart Maatschappij NV (KLM), its subsidiaries and/or its employees shall not be liable for the incorrect or incomplete transmission of this e-mail or any attachments, nor responsible for any delay in receipt. Koninklijke Luchtvaart Maatschappij N.V. (also known as KLM Royal Dutch Airlines) is registered in Amstelveen, The Netherlands, with registered number 33014286 ********************************************************************** ---------------------------------------------------------------------- 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
