Mike, <quote> Consider this real-world (i.e. this actually happens) scenario: 1. Customer restricts the region size and it works 2. Due to application or other perhaps minor changes, the normal JVM usage now exceeds the region size gets 878-10, OOM etc 3. Customer raises PMR and sends dump 4. Level 2 look at the dump and then send to level 3 5. Level 3 are busy with higher priority problems and can't handle it right now 6. Level 3 see that it is a customer setup problem, pass it back to L2 7. L2 tells customer (nicely) it is not the JVM it is the way they have set things up
Now count how many days have elapsed. How is this perceived by the customer? Many customers do not know how to debug these problems. <unquote> Unfortunately the scenario you painted is all too likely, and the question that came up for me first is: Why can't IBM handle customer's problems in a timely manner? Maybe because within IBM level2 doesn't know how to debug these problems? Back when I did IBM level2, we were chewed out regularly when we didn't detect these things ourselves and dared to escalate them to level3. The task was to present level3 with the cause of the error, not have level3 analyze the problem in the first place. <new quote> 2. Why not raise a PMR if you have a problem? 3. Many of the problems that I see are the result of an abend, e.g. 878-10 <unquote> We did raise a PMR with the application (written by IBM), only to be told that they don't know what the heck they should do with the dump we gave them and after two weeks of a productive application terminating their main thread out of the blue without taking down the address space at unpredictable times they still have no clue themselves how to analyze it (or so it appears to me). The application did not write a dump or throw any recognizable abend. When I opened an ETR first with OMVS (where this applcation has their feet as it is another ported thing), to be forwarded to LE (who got Java involved), the only tangible thing (and admittedly they could not do more than that) was that this application had forced us to install some JZOS Java code that is unsupported and had been unsupported for quite a while when the installation instructions were written. Great. Given all that, I'll have to look into how to debug Java some more in the future. May I contact you directly with anything I'll find or that is unclear to me? Be warned that that might be a lot, as I don't come from the java programmer side and look at diagnosis from a deep system level, so expect lots of complaints, if I set my mind to it. I did download the diagnosis books that I found following the link in your signature. Regards, Barbara Nitz -- Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kann`s mit allen: http://www.gmx.net/de/go/multimessenger ---------------------------------------------------------------------- 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
