STL Time Machine Report #2 - 1998-01-08 I'm on a short-term deathmarch project to get a single application running in the Time Machine. 12-hour days, possible weekend work, you know the drill. The basic LPAR works and can be IPL'ed forward. We're currently on true date while we get the application built. So what kinds of problems are we running into? Incomplete infrastructure. We need file X from system Y, but system Y isn't ready yet and can't create file X with post-2000 dates. We have to hand-build the file. JOB Q needs the Database libraries, but nobody migrated them and we don't have security to copy them over. Only Gary can do that and he went home after putting in 11 hours. How do I set up my terminal emulator to connect to the other LPAR? Critical tapes were scratched but not uncataloged. Not all Time Machine JCL was tested before the DASD volumes were migrated to the LPAR until 2001. Not enough time. Even the tested JCL still has JCL errors because dataset names have changed, library names have changed. Control card members are missing. That model statement won't work on your IEBGENER. Can I have security access to fix this proc? Damn, I gotta recreate that file. The test network is not working, and needs to be looked at by the right people. You mean they didn't have to stay past 5:00 PM? Who knows how to work the new job scheduler? Why didn't you get your schedules straight before you moved the jobs in? Did anybody realize there wouldn't be any 3420 tape drives in the Time Machine for that old assembler program? Sure you can print jobs from the Time Machine. We gave you one HP LaserJet 4. You mean that's not enough? How much do you guys need to print anyway? These kinds of problems are typical for setting up a new test facility. There were lots of good posts on Time Machine problems these past few days. I'm an applications guy, not an operating systems geek. I don't know how long it takes to set up a Time Machine, but I can tell you every application system has its own set of problems even if tech services has the operating systems software installed and ready to go. The next group to put their application in will have the same problems, because it's another whole set of programs, JCL, and data files. It will take you longer than you think to set up a time machine. I heard a rumor from a contractor who knows another contractor that Anheuser-Busch (the big brewery company in St. Louis) is testing all their programs with TicToc, and apparently has no plans to do any time machine testing. You'd better buy all your beer with good "born-on" dates before the singularity. We tested all our CICS apps with IBM's SupportPac CH15 to simulate future dates. We would never have learned that CICS 3.3 dies on 01/01/2000 without the Time Machine. And our Third Party database will work in the Time Machine with COBOL II runtime libraries but not with LE/MVS. Last I heard, they're looking into LE patches or configuration as opposed to waiting for the new release from the vendor. The vendor's schedule is not soon enough for our schedule. Another digression -- I know a geek at another shop whose horn-hair has been sitting on the auditor's report, because they have no plans to address Year 2000 issues. The CEO just saw it. Ooopsie, maybe we'd better start thinking about this. What is your group doing about this problem? "We're migrating our application system to client/server." Good, there's one system we don't have to worry about. (This is a True Story, and the names have been changed to protect the guilty). -- Arnold Trembley http://home.att.net/~arnold.trembley/ "Y2K? Because Centuries Happen!"