On Sat, 28 Apr 2012 09:47:29 +0200, R.S. <[email protected]> wrote:
>W dniu 2012-04-27 23:03, John McDowell pisze: >> I'm trying to get a feel for problems that occur in the early stages of z/OS >> system start up (e.g. IPL/NIP). Generally problems in these stages result >> in a non-restartable wait state, for example wait state x'0B1' (e.g. LOADxx >> or IODF problem). >> >> Questions: >> 1. FREQUENCY: How often do they occur ? >> 2. DURATION: How long does it take to resolve them (e.g. minutes, hours, >> etc.) ? >> 3. IMPACT: What are the consequences (e.g. missed SLAs, etc.) ? >> 4. CAUSE: What are the underlying sources (e.g. hardware, software, etc.) ? >> 5. RECOVERY: How do you recover from them ? > >1. Rarely. IPL is performed rarely. In my case I haven't noticed such >problem *on production systems* for years. Such problems do happen >during tests, like new system, PTFs applied (and IPLTEXT not refreshed), >new CPC, new LPAR, DR test, etc. >BTW: I *hate* looking at last 3 digits, then previous digits... ;-) >Since the numbers are available on HMC, it would be nice to have button >Explain which could (under the covers) open the book and perform the >analysis for me. > >2. The time depends on two-three elements: >a) time to open the book. It can be few seconds when I'm on my PC (HMC >accessed remotely), it can be minutes when I do it on real HMC and I >have to use another PC for documentation access. >b) time to write down the digits, extract wait state code and reason code. >c) (optional) sometimes I need to check whether description is accurate >or maybe fix something (like LOAD member). I usually logon to TSO on >another system and view/modify the things. It could take 5 min. > >3. Lost time, some stress. From business point of view it doesn't affect >my SLA. > >4. IODF in multiple extents, OS config with bad offline/online device >set (i.e. IODF device is described as OFFLINE YES), mistakes in LOADxx, >not refreshed IPLTEXT (after PTF APPLY), typo in LOAD window on HMC. > >5. See 2. > I think R.S.'s response is typical for most of us. Although I can't remember the last time I had a wait due to nucleus or IODF in multiple extents, occasionally I have typo'd a loadparm when IPLing one of my sandbox LPARs. Loadparms never change in production except for a brief period when migrating to a new OS release and the sysprog will do that initial IPL/change and the HMC remembers it of course. Load addresses change often, but no one ever seems to type those in wrong. I really like the idea of the HMC being able to quickly open a reference to the correct wait state code. Mark -- Mark Zelden - Zelden Consulting Services - z/OS, OS/390 and MVS mailto:[email protected] Mark's MVS Utilities: http://www.mzelden.com/mvsutil.html Systems Programming expert at http://expertanswercenter.techtarget.com/ ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
