In this issue, I’ll cover the following:
- High CPU in SMS
- JES2 Checkpoint Problems With Fast Coupling Facilities
- OS/390 R2 Double-SYSRES
- Our Booth at SHARE in San Francisco
1. High CPU in SMS
Jim Purdie and Bob Maw of a large Midwestern bank found a rather important SMS APAR recently. APAR OW23856, “Excessive Connect Time for SMS ACDS”, describes what appears to be a fairly innocuous SMS problem of excessive connect time when handling a large number of volume statistics updates. The APAR notes that the problem has been observed in environments with OEM products such as MIM running. Jim noted that when they migrated to SP 5.2.2 and DFSMS/MVS 1.3, their average daily CPU time in the SMS address space seemed rather high (5.6 hours of CPU time a day on a 9X2!). After applying the fixing PTF, the CPU time went to .1 hour a day. And this APAR isn’t even marked HIPER. (I would think anything that can eat up 5.6 hours of CPU time a day would be considered HIPER!)
2. JES2 Checkpoint Problems With Fast Coupling Facilities
If you’re considering putting your JES2 checkpoint data set on a fast coupling facility, this is for you. Jerry King, the instructor and co-developer of our new Application Tuning class, called me about a problem he heard of that occurs when putting the JES2 checkpoint data set on a fast coupling facility. JES2 can abend with a S6C5. Turning on CTRACE (component trace) doesn’t identify the problem because the problem disappears. It seems that due to a timing problem with IARVSERV shared pages processing, JES2’s use of shared pages results in the abend. When CTRACE is turned on, JES2 is slowed down enough that the timing problem is resolved. Most sites have solved the problem in the past by turning on CTRACE, but one site that couldn’t afford the overhead in production is trying out a new PTF. See APAR OW23564, “An ABEND6C5 RCIE023124�”, for the PTFs (yes, I know there’s no mention of JES2 or checkpoint in the APAR description, but Kathy Walsh from IBM’s Washington Systems Center (WSC) indicates that this IS the corresponding APAR for the problem). If you’re at SHARE next week, Kathy will discuss this in her “WSC Hot Topics”, session 2500.
3. OS/390 R2 Multi-volume SYSRES
OS/390 R2 is causing a problem in some installations because the number of data sets that need to reside on SYSRES no longer fit on a single volume. IBM has provided new support, called “Extended Indirect Catalog,” to resolve this problem and provides support back to MVS SP 5.2 APAR OW25115 describes both the problem and the solution. The APAR also references other APARs needed for product support (OW23082 – original APAR, OW24928 – DFSMShsm, OW24889 – DFSMSdss, OW23105 – ISPF, and OW24881 – RACF). If you’re thinking of installing OS/390 R2, be sure to check out this support before you begin. If you’re attending SHARE next week, see Session 2817, “Bit Bucket X’0D'” on Friday 11am for a description of this new facility by Carl Youngren.
4. Please Visit Our Booth at SHARE in San Francisco
Tom and I will be at SHARE in San Francisco all next week. We’ll have a booth at the Technology Exchange, so please stop by and say hello, pick up our latest brochures, our brand new SMF reference card, and hear about our software plans. You can also catch me at the following sessions, as well as at many other of the EWCP (Enterprise Wide Capacity and Performance) sessions (#2500-2599):
Session 2502, Sun 4:30pm (EWCP closed session) – Presenting: “CMOS and COBOL”
Session 2568, Tue 4:30pm (not announced yet – replacement session) – Presenting: “New Key OS/390 Capacity & Performance Indicators”
Session 2558, Tue 6pm – On panel: “Author, Author”
Session 2511, Wed 11am – Chairing: “WLM User Experiences” by Tim VanderHoek
(If you’re considering running WLM in a sysplex, be sure to come hear Tim’s good and bad experiences)
Session 2505, Thu 4:30pm – (EWCP closed session) – Presenting: “Hot APARs” and “MVS Object Experiences”