STL Time Machine Report #17 - Tue 19 May 1998 (1998-05-19) Several times in the last few weeks I have downloaded more than 300 new messages in comp.software.year-2000. It's getting much harder to keep up. COBOL and LE/MVS (Language Environment) The Time Machine is running on current dates. As part of this, Tech Services installed LE/MVS as the default runtime in the Y2K Test LPAR. Several jobs abended, and some configuration changes were made that required re-booting the Time Machine. I don't know how to set up MVS/ESA 5.2.2, but the explanation I received was that there were problems with LPA and LINKLIST. It's a little like having duplicate DLL's in Windows with a bad path. Interestingly enough, all the problems showed up in old OS/VS COBOL programs, or COBOL programs which used a statically linked assembler subprogram to dynamically call another assembler sub- program. I think somebody referred to this as "Hartford Disease". Was it Bill Lynch? I've been telling people to junk it for years. Now it's not working and they have to rip it out wherever it's left. COBOL II programs did not have any problems with LE/MVS. The Group Four re-test will include running all the previously tested applications concurrently. Group Two testing included running all Group One applications as well. CICS 4.1/ACF2 6.1 Performance problems Finally there's some good news to report here. In April we tried to run a performance stress test of a CICS 4.1 application, and it failed miserably. Tech Services applied more than 20 vendor patches to ACF2/MVS, ACF2/CICS, and IBM CICS 4.1 to resolve performance problems. Several Tech Services guys participated in the stress test for the first time in years. Tech Services identified and resolved two additional problems during the stress test. This is good news because it means we will be able to upgrade all our CICS regions to 4.1 this year. We have more than 80 of them. Just two months ago we were still converting the last CICS 2.1.2 regions to CICS 3.3. It's nice to get to talk to Tech Services. They tell me we're getting two more engines for our MVS mainframe, for a total of six. And in July they start working on the upgrade to OS/390 for installation in 1999. I had an interesting discussion with our capacity guy. I was whining about not being current on our patch levels, and he explained that TS never likes to be on the bleeding edge. If you have the most current version/release of any product, there are no published patches for it. Any problems and you have to send huge storage dumps and trace files off to the vendor, and it might take weeks or months to get a problem resolved. If there's a published patch for a known problem you can get it resolved in one day. A Personal Remediation Anecdote I have a small system I wrote: Four COBOL programs to pull CICS transaction dumps off SAR, count them by abend code, and report them in various formats with descriptions of our user assigned dump codes. Years ago I had a vice president who had me count these statistics manually and give him a report every Monday morning. This is not a production system. Altogether it's about 3000 lines of COBOL. I stayed late a couple of nights last week and fixed it for Y2K. Extracted dump dates had to be windowed to 4-digit years. The history file had to be expanded for dates in sort keys and to allow for duplicate dumps within the same second on the time stamp. Just for grins I converted it from COBOL II to COBOL/MVS, so I could easily display compile date and current date with 4-digit years. I fixed it for about 12 cents a line. But I know this system, didn't have to write any project plans or test plans, didn't have to do any peer review or get nine signatures to install it. If I had to do it as a project it would have been much closer to a dollar per LOC and would have taken two or three months. Previous Time Machine Reports are available at: http://home.att.net/~arnold.trembley/tmr.htm STANDARD DISCLAIMER: I am NOT an official corporate spokesperson. My opinions should not be held against my benevolent employer. -- Arnold Trembley http://home.att.net/~arnold.trembley/ "Y2K? Because Centuries Happen!"