1. Cheryl Watson’s TUNING Letter 2004 No. 6
2. Corrections to TUNING Letter 2004 No. 5
3. Redesigned IBM Web site
1. Cheryl Watson’s TUNING Letter 2004 No. 6
The forty-eight page 2004 No. 6 TUNING Letter was emailed to electronic subscribers yesterday. The print issues will be mailed within a week. You can purchase a printed copy of the current TUNING Letter for $95 at http://www.watsonwalker.com. The following is a summary of just some of the contents of this latest TUNING Letter:
VSAM RLS Enhancements
Even though a good deal of your production data probably resides in VSAM data sets, it has always been a complicated matter to share those data sets with applications running on multiple systems. The VSAM Record Level Sharing (RLS) feature provides the ability to simplify this sharing, but RLS has not been without its problems. The RLS development team at IBM hopes to correct these problems by enhancing RLS to be more resistant to errors. In this issue, one of the IBM developers involved with this change presents some of the improvements that are already being made available through normal software maintenance. She also shares some of the tools that are already available for researching RLS problems and reporting them to software support. Make sure to read this important article starting on page 41.
Resources for zAAP Processors
Unlike the past few issues, you will not find any articles this month regarding IBM’s new zSeries Application Assist Processor (zAAP). But IBM is starting to make more reference information available related to these new processors, and we have included information about those resources in this issue. Starting on page 6, we provide a list of the APAR maintenance that applies to these processors. The people at the Washington Systems Center have also been making some materials available, including some new presentations (see page 14) and some white papers (see page 15). Finally, on page 17 you can read about a new Redbook that recently became available.
Great Diagnostic Tool
Those who research problems for a living have probably lost count of the number of times a problem was caused by using the wrong version of a program or module. This problem often occurs in the ISPF environment, where hundreds of libraries are often allocated to the session and the concatenation of multiple libraries is a common occurrence. As we note in our article on page 37, there is a great free ISPF command that can be used to help diagnose these problems. Although this tool is not new, there may be some features that are not as well known. Some of these obscure functions make the tool even more valuable, because they allow you to diagnose system-wide problems that are outside the scope of the current ISPF session.
Better System Performance
In our User Experiences section on page 25, we mention an important fix that has improved IPL speed dramatically for some installations. In that article we also mention a little-known facility for measuring IPL speed that may be very useful if your IPL times are becoming longer than expected. The use of Parallel Access Volumes (PAV) can improve performance, but measuring that performance is not always easy. See ourNuggets from the Forums item on page 8 for some tips shared by readers. Finally, in our Q & A section on page 28 we discuss Multi-Period Batch and what to do when Workload Manager misses goals. These are two of the areas that we receive many questions about, so we know there should be a lot of interest in the tips we give here.
2. Corrections to TUNING Letter 2004 No. 5
Many thanks to Pat Dixon and Gilles Lachassagne of Computer Associates for pointing out some errors in the SMF Support section of our zAAP Update article in the last issue (see TUNING Letter 2004 No. 5, page 41). We stated that the first three CPU consumption values listed in Table 6 (SMF30CPT, SMF30ENC and SMF30DET) represent milliseconds (.001), when they actually represent hundredths of a second (.01). This is unchanged from previous operating system releases. This means that the constant 16,000 used in later calculations in that section should be replaced by 160,000 to account for this difference in scale. This constant should also be the dividend in the calculation rather than the divisor. This means that whenever you see the statement ‘(SMF30SUS/16000)’ in a calculation, you should replace it with ‘(160000/SMF30SUS)’. Thanks very much Pat and Gilles!
This error only appears in the electronic and print version of the TUNING Letter and will be corrected on the CD-ROM.
Gilles also noted that, although the formula for calculating a normalization factor is correct in theory, their testing has shown a wide variation in the value being produced. They have opened a problem related to this, and are pursuing it with IBM.
At the bottom of page 36 in the same issue, Jerry Urbaniak of Acxiom noted a transposition in one of the APAR numbers listed. We mentioned RMF zAAP APAR OA07531, when we actually meant OA05731. This will also be corrected on the CD-ROM. Thanks Jerry!
3. Redesigned IBM Web site
Most of us probably have dozens of bookmarks to various IBM online resources, but how often do we visit their home page (http://www.ibm.com)? IBM recently announced a major redesign of this site. You can read a short announcement of the changes by browsing http://www.ibm.com/ibm/sjp/new-web-site/us/, but it’s probably just easier to visit the home page and see how you like it. IBM has so many online resources that finding what you want is often a challenge. We hope that this new design will help.
Stay tuned!