STL Time Machine Report #48 - Monday 05 July 1999 (1999-07-05) Sunday was a very hot Fourth of July. After watching the local fireworks display with my extended family, I drove in to work at 11:30 PM for the July 5th deathwatch. July 5th is our first Year 2000 fail date. We have a large VSAM file that is updated online. Records added or changed have a "purge date" calculated as 180 days from the current date. July 4th is 99/185 in IBM Julian date format, so the purge date is 99/365. But on July 5th (99/186), the calculated purge date is 00/001. During our overnight batch maintenance a COBOL program examines the purge date, and if it is less than the current date then the record is deleted. So early in the morning of July 5th, any record with a purge date of 99185 will be deleted because 99185 is less than 99186. Unfortunately, without remediation 00001 is also less than 99186, which is not what's wanted. This "fail date" was identified three years ago, during the initial, non-automated assessment. Curt was working by himself. His duty was to monitor the file and report to management. I ran a batch job before midnight to list the purge dates on the big VSAM file, just for me. Curt had already run a couple for himself. No purge dates in the year 2000, as expected. One worry was that there wouldn't be any new update activity between midnight and 1:00 AM, due to the holiday weekend, but the network monitor was showing us update activity. At five minutes after midnight I ran the batch job again, and it showed 40 records with purge dates in the year 2000. We ran this diagnostic two more times before 1:00 AM. Now all we had to do was wait for the scheduled batch purge. By 2:50 AM, we knew. Purge dates of 00/001 are windowed to 2000/001 and are therefore greater than the current date of 99/186, windowed to 1999/186. Records with purge dates of 99185 were deleted. We passed our first Year 2000 fail date with no problems. And I was confident we would, because this has been tested many times in the time machine. I also learned how to send messages to alphanumeric pagers. One disaster recovery manager called in from a hotel in Chicago. Message received. Break out the champagne. Our override file also survived the planned June 30th purge of all active records. No incidents and all applications are running normally. TESTING Testing will continue through the end of the year. June was a very hectic month for me. I installed a CICS project to add routers to my primary system. Lots of JCL changes. It looks like I won't be getting a hotel room for the rollover, unless there is a snow or transportation emergency. In previous years we have put Operations staff up in nearby hotels who were snowed in at work. I live close enough to the office that I have walked to work in a snow storm before. As far as I'm concerned, we are ready for rollover now. SIGNS, OMENS, PORTENTS Meanwhile, back in comp.software.year-2000, the debate about bank runs, meltdowns, oil shortages, water shortages, budget surpluses, sewage surpluses, and infrastructure collapse - is hotter than ever. It's particularly sad that the new rage seems to be hoaxes perpetrated against Paul Milne and the doom crowd. While it may be entertaining to some, I don't think it helps the overall situation any. If we can't get reliable information, we can't determine what's happening right now, much less what's likely to happen at rollover. PREDICTIONS In the last TMR, I included a couple of quotes I liked, one from Harlan Smith, to the effect that storing food and water would not preserve our lifestyles. The other quote was from Hoff_meister, asking, "how do you prepare for a recession?". I'm feeling adventurous enough to go out on a limb and post another prediction or two, maybe even log it with Bradley K. Sherman so I can cover myself in infamy if I'm wrong. For December 1st, 1999, no Bank Runs in the USA, and no nationwide martial law. Supposedly Harry Truman once said he wanted to meet a one-armed economist, because he was tired of hearing them say, "on the other hand...". So here's my "on the other hand". I found hoff_meister's quote interesting, because I believe a recession is likely in the next two years, even without Y2k. And any real consequences of Y2k are likely to trigger or exacerbate a recession. I expect problems to occur after January 1st, 2000. Watch for earnings declines in the PC industry, H-P, Dell, Intel, Microsoft. Compaq is already experiencing problems. There was an interesting article on this subject in the "Economist" just recently. PC prices are dropping, forcing profit margins lower. With increased internet access, users are less interested in new releases of Windows or Office. This year's cheap PC has more than enough juice to surf the web next year. So how do you prepare for a recession? Reduce debt, reduce expenses, build up cash reserves, update your resume, update your skills. Be ready to change jobs or relocate. Storing food might help, but it's not guaranteed to be the best strategy. There really aren't any guarantees at all. I've fallen behind on updating my web page, but recently I added a Hamasaki page, hosting some recent WRP's plus direct links to Cory's pages. I added a link to Dave Eastabrook's T-D page, and there's a new picture on the bio page. Hadn't realized how old the other picture was. I have gone gray in the service of my geekdom. As of 1999-07-05, my countdown now reads: 47 days until 1999-08-21 (GPS Rollover) 66 days until 1999-09-09 (Another date) 180 days until 2000-01-01 (Rollover) Previous Year 2000 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!"