> > Is OG orphaned storage (Memory leak) or still owned by someone? > Our list of OG storage contains a lot of entries and I > believe they are still owned by someone.
"Owner Gone" might be a storage leak, or it might be storage that contains code, data, control blocks yada yada, for products that are still alive and well. It is not uncommon for subsystems to use a "throw-away" address space to get going. Any common storage obtained during that process would show up later as OG even though its really still in use. That's a good reason not to use the FREEMAIN option of products (like ours and Tivoli/Candle) that offer a point and shoot shortcut to your next unplanned IPL. > How did you handle them? Did you ask the ISV or IBM about them? > I've seen a large for RRS and IBM raise FIN APAR OW57028. On a case by case basis I would think. If you can find eyecatchers and fingerprints you can ask the vendor WTF? Sometimes it is "working as designed" and sometimes its "oops". Either way you want to know the answer, particularly if you see evidence that the unaccounted areas get bigger over time. > I now have a similiar issue with DB2 and Websphere MQ, It > seems most of them happen if the memory allocation is done > while HASN <> PASN. The GETMAIN and STORAGE OBTAIN contains > the OWNER parameter. Is there a performance issue to omit > the OWNER=PRIMARY keyword? Not that I am aware of, but OWNER doesn't really do anything other than storage tracking. If the address space identified as the owner is really gone then it doesn't matter who it was. At least with storage tracking on you can tell who it was. Take a look at the IGVCAUB macro for clues about how its all strung together. 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
