1. HFS and PDSE Update
2. Cheryl Watson’s TUNING Letter, No. 6 Summary
1. HFS and PDSE Update
In the last Cheryl’s List, I indicated that the PTFs for HFS and PDSE files on non-SMS-managed volumes was available, but the WSC Flash had not been issued yet. Flash10007, HFS & PDSE Support on non-Managed SMS Volumes, was finally issued on 01/24/00. This flash documents all of the APARs and PTFs which provide support for HFS and PDSE files on on-SMS-managed volumes. It also provides information on how to define them and the differences between managed and unmanaged PDSEs, such as the fact that non-SMS-managed PDSEs do not have to be cataloged.
2. Cheryl Watson’s TUNING Letter, No. 6 Summary
The 40-page 1999, No. 6 TUNING Letter issue was just completed, along with the December 1999 CPU Chart. Electronic subscribers received their Tuning Letter on 01/31/00 and received their CPU Chart today. Printed issues should be mailed before February 11. The CD-ROM containing all of 1999 is in progress and we’ll let you know when it begins shipping.
Congratulations!
Congratulations to all of the sites who successfully fought and won the Y2K battle. We do not believe it was a lot of hype. We believe, rather, that the crisis was avoided through your hard work and the money invested by your management.
Data Compression
Several items on hardware data compression, HDC, are included in this issue. HDC is used to reduce DASD space, reduce run times, and eliminate restrictions (such as the 4 GB VSAM restriction). But it comes at a cost in CPU time. With DASD space going at 25 cents a MB, is it worth it to implement compression and use extra CPU cycles? We provide several user experiences in this issue showing that for many sites, compression is the logical way to go. At one site, compression resulted in the number of EXCPs per job dropping from 65,000 to 14,000, recovery of 46GB of DASD, and reduction of a batch window by 50%. On the other hand, compression rates can vary and some hardware implementations are so poor that you would never want to try compression.
GOLC Pricing
IBM introduced Growth Opportunity License Charge, GOLC, with the announcement of their new Multiprise servers. How well has it been accepted by ISVs (Independent Software Vendors)? Not as well as IBM had hoped. Some vendors are helping people find considerable savings when moving from smaller, outdated machines to this newer hardware and software. Others continue to charge extremely high upgrade charges and license fees for this machine. It means that you simply can’t bring in one of these new machines without analyzing ALL of the software costs. We also include Amdahl’s announcement of their 1000 Family Servers, which is expected to be approved for GOLC pricing, and a concern from some sites when they moved to a single CPU on a Multiprise machine. See page 22.
HFS and PDSE
IBM had originally announced, but only now has provided, support for HFS and PDSE files residing on non-SMS-managed volumes. See our notes on pages 11, 28, and 33. This should provide a great deal of help to non-SMS sites who have wanted to implement these facilities.
Elsewhere
In our S/390 News on page 5, we include feedback on an important WLM APAR, an APAR for a CPU spike in ADSM, how to locate the format of Domino SMF records, several APARs for successful HFS implementation, some last minute bugs from Y2K, free publications (including a free online LE course), recommendations for CA-MIM, a change to the Catalog command, an APAR for dynamic linklist, techniques for an extremely large sort, and help for interpreting CPU reports when crypto is activated. The WSC flashes on page 10 describe some last minute Y2K problems and a useful flash providing LE performance considerations for IMS. SHARE proceedings are now online (page 11), a forum is available for customers to discuss ISV software assets, and an S/390 Bulletin has an article showing the relative costs per user of providing service on mainframe, UNIX, and PC systems. Which platform is the cheapest? The mainframe, of course!
Our Q & A on page 37 provides recommendations when running CA-Dispatch in goal mode, a technique for detecting looping jobs (including a program to monitor them), a discussion of monitored address spaces, and considerations for estimating the overhead of upgrades. Finally, you’ll find some very good New Year’s Resolutions for Year 2000 for CICS folks on page 35.
Stay tuned!