On 20 September 2010 02:21, Barbara Nitz <nitz-...@gmx.net> wrote: > Tony, > a bit after the fact, but what strikes me is that Jim uses the commands > directly while you apparently use the 'analysis' panels. Option 2.2 does NOT > really correspond to the IPCS command STATUS CPU REGISTERS that Jim > talked about.
In my case they produced the same result. However I neglected to post the resolution of the problem here, though I did mention it to Jim directly. It was a simple "apply gun A to foot B" situation. I had reused a dump dataset on the very constrained system in question, and failed to delete its records in IPCS, so it was showing me registers from a previous dump, but storage from the current one. Perhaps an easy trap to fall into, but I should know better. It might be nice if IPCS would recognize and perhaps warn in this situation, e.g. if the dump timestamp doesn't match the saved info. That wouldn't take a lot of processing, presumably. > I just checked in one dump (I hate the analysis panels, they > never do what I want them to), and there is a lot of contention analysis going > on before the actual registers are displayed, at least the first time around. > > That makes me suspect that some sort of overlay has occured on your system > (in your IPCS session) *before* IPCS ever got around to doing the st cpu re > output/formatting. I have chased a number of those overlays in the past, and > they were never really reproducible (only occured on first issuance of the > command), so I was never able to find and report the actual bug. But I > distinctly remember that it *always* had something to do with some sort of > contention. I also got abend0c4 and other stuff, so my userids take the odd > dump because I still have slips set in the system. The trick was that an identifiable bogus value being displayed by STATUS CPU REGS (the R1 content at SLIP) occured in the dump dataset only in dumped storage, and not in the dump header record. So I spent some time (much more than I should've) on the theory that IPCS was (contrary to what Jim said) finding register information in control blocks in dumped storage under some circumstances. It wasn't, of course. > Could you try to drop all records (all awareness) from that dump on the > original > system and re-initialize the dump? Heh - you found the problem... Thanks. Tony H. ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to lists...@bama.ua.edu with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
