What is MCHILDS saying about Mainframe OS's

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

One word: "Oopsie."

Executive Summary: There are 50,000 IBM-Style mainframes ranging from smaller one like my 2+ MIPS deskside MVS system, up to hundred+ million dollar processor complexes. These things run the world. IBM normally issues monthly consolidated fix kits, these are grouped into semi-annual (roughly) maintenance releases. This summer, IBM issued a collection of YR2000 fixes for "High Impact, Pervasive" problems. At this late date, some firms will have a difficult time installing and testing the fixes.

Analysis: I don't know what will fall out from this. This is a bad sign. Even worse, I am concerned that other system and software vendors have not issued similar fix kits.

Future: I am looking into this and will present my findings in the next print edition WRP(this is the quarterly pub). It will go on the web a few weeks later to all.

-- cory (kiyoinc@ibm.XOUT.net), August 29, 1999

Answers

Cory (or anyone who might know), I understand that whereas some manufacturers issue patches that are supposed to always be installed as soon as received, because they always assume that all previous patches have been (i.e., "up to current revision level"), other manufacturers design specific patches for specific problems, and discourage installation unless one has the specific problem. The reason that this is discouraged is because sometimes these patches that are designed to fix one specific problem can conflict with patches designed to fix another specific problem.

Is this of relevance to the IBM mainframe world? And, in particular, would these "best and final" Y2K patches possibly conflict with older patches that took a very narrow approach towards patching the OS?

-- King of Spain (madrid@aol.com), August 29, 1999.

Cory,

Oh, is that what MCHILDS was trying to say. It was well concealed. :-)

Jerry

-- Jerry B (skeptic76@erols.com), August 29, 1999.


King,

Absolutely. Conflicts (Fix B, in the presence of Fix A, doesn't work) , regression (Fix B undoes Fix A), fix introduces new problems are all possible and frequently occur.

That's why it takes time to work this all out. In addition, some firms have modified the OS so that they have to make final adjustments on site.

Yes, MCHILDs was presenting the background info. It was a lot of techno-talk to say, "Oopsie."

-- cory (kiyoinc@ibm.XOUT.net), August 29, 1999.


That's an interesting interpretation of MCHILD's posts, Cory.

Me, I see a fairly detailed document describing why you won't be able to do rolling IPL's to the current release of OS/390 after 2000, if you stay on release 4 or below.

But hey, I'm not a mainframe geek. Maybe you can elaborate.

-- Hoffmeister (hoff_meister@my-deja.com), August 29, 1999.


I looked beyond the specific document that MCHILDs pulled. I guessed that he followed the path to BOULDER.IBM that I detailed the other day picked the most readable of the hundreds of kilobytes of text on YR2000, HIPER, etc.

For that one report of the hundreds at BOULDER.IBM, you are correct. If you have trouble sleeping, take a look at others and consider that this is the end of August 1999 and fixes are still coming out of IBM. (It is to IBM's credit that they are so forth coming about YR2000 issues.)

For what it's worth, I don't think that MCHILDs is an OS/390 internals person, but that's not important. What is important is that he has taken the time to pull documents down from IBM.

-- cory (kiyoinc@ibm.XOUT.net), August 29, 1999.



Cory,

I think most people who are reading this do not really understand what this means. Very few people have any experience with mainframe operations other than receiving a bill from a credit card company. Could you please describe what a HIPER patch is all about, how much work and what kind goes into installing and testing these patchs and what can go wrong when you do this kind of thing with only a few months left (not enough time, etc).

Thanks

-- ..- (dit@dot.dash), August 30, 1999.


Moderation questions? read the FAQ