Cheryl’s List #138 – February 3, 2010

by | Feb 3, 2010 | Cheryl's List

1.  About Cheryl Watson’s Tuning Letter 2009 No. 6
2.  Important Notice About SHARE in Seattle
3.  Latest Red Alert
4.  Mean Time to Recovery Redbook

1.  About Cheryl Watson’s Tuning Letter 2009 No. 6

The forty-two page 2009 No. 6 Tuning Letter was emailed to paid subscribers on December 23nd. You may visit our Web site at www.watsonwalker.com to obtain subscription information. The following is our Management Summary page from that issue, talking about some of the contents of this latest Tuning Letter:

How to Influence IBM
The SHARE organization is the largest IBM user group that has as one of its goals the influencing of z/OS development. They accomplish this by allowing SHARE members to create, discuss, and vote on “requirements” to be sent to IBM. Within the core z/OS project alone, IBM has implemented over half of the 1,000 requirements submitted in the last few years. If your company is a member of SHARE, your employees can help write and vote on these suggestions to IBM. Our article, starting on page page 22, shows how they can do that. If your company isn’t a SHARE member, then I highly recommend that you consider it. For $500 a year (or by just sending one person to one of the yearly conferences), a company can become a member. This not only allows you to participate in requirements, but also, more importantly, provides all of your employees access to all of the past SHARE conference proceedings. We consider these proceedings to be some of the best educational material available, and there is no additional charge for any of it.

Understanding Parmlib
The controls for z/OS reside in one or more parmlib data sets, such as SYS1.PARMLIB. Three of the members of parmlib (LOADxx, IEASYMxx, and IEASYSxx) determine how most of the other parmlib members are used for each LPAR. A mistake made in any of these members can be critical to how well (and whether) each LPAR is initialized. In most installations, at least one senior systems programmer understands the relationship between these members. But, unfortunately, this relationship is often not well understood by the majority of new systems programmers and performance analysts. Mistakes can result. Our Focus article (starting on page 3) describes the IPL process from the operator IPLing a system to the IPL process initializing and processing these three parmlib members.

Elsewhere in this Issue
You’ll find many other useful items throughout this newsletter:  New features of     ALLOCxx. • An update on capacity planning for Coupling Facility (CF) links. • Many New Function APARs, including more DFSORT facilities and faster SVC dumps. • Lots of warnings about invalid SMF data. • A great new Redbook on modernizing your current batch applications.

2.  Important Notice About SHARE in Seattle

I’m really looking forward to SHARE in Seattle (March 14-18), and I hope to see many of you there. In trying to reduce costs without increasing registration fees, SHARE has shortened from a 5-day event to a 4-day event (Monday through Thursday). But I’d like to issue a warning about Thursday. In the past, the conference finished on Friday at noon, and most people left for home in the afternoon. SHARE had scheduled some of the most highly attended z/OS sessions (e.g. Bit Bucket, z/OS Goody Bag, and my Hot Flashes) on Friday morning to encourage people to stay for the morning session. But those three sessions (and many more) are now scheduled for Thursday afternoon. And the MVS Free-For-All is scheduled for 6 on Thursday. So if you spend the money to travel to SHARE, it’s very important for you to spend Thursday night in Seattle in order to gain the most from your trip. SHARE is still some of the best and least expensive training you’ll find. For registration and more information, go to www.share.org.

3.  Latest Red Alert

On January 12 and January 21, IBM issued new Red Alerts. You can sign up to automatically receive Red Alerts at http://www14.software.ibm.com/webapp/set2/sas/f/redAlerts/home.html.

  • 2010.01.12 – DB2 9 for z/OS potential data corruption and DB2 crash after using the CATMAINT UPDATE SCHEMA SWITCH utility, which is new in DB2 9

After running the CATMAINT UPDATE SCHEMA SWITCH utility, object’s DBDs were corrupted causing various overlay type ABENDs when accessing the row. This problem only affects databases that contain tables that have columns with user defined default values. As a preventative action, install the PTF for PK90557.

  • 2010.01.21 – z/OS Potential DB2 Data Inconsistencies Following a CF Failure

When a CF terminates the structures on the CF are physically deleted and duplex structures will fail over to the other instance of the structure and simplex structures will rebuild into a new instance of the structure. If the CF comes back up quickly after a CF failure, it is possible, and even likely, that the rebuilding structures will rebuild into the CF that just terminated and came back. While the new instances of the structures are being allocated, XCF may still be cleaning up for the prior instances of structures that were allocated in the CF before the CF Failure. As part of the cleanup XCF may incorrectly delete record data from the new instance of a lock structure.

4.  Mean Time to Recovery Redbook

The Redbook that I worked on in May is now available in draft form – SG24-7816-00– System z Mean Time to Recovery Best Practices. Even if I hadn’t worked on it, I would have told you that it’s a fantastic book! I’ll have more to point out in our next Tuning Letter issue, but for now, I’d like to point out four things to make you rush and get it:

  1. Chapter 3 describes some z/OS tools, one of which is the IPLDATA block that contains timings during IPL. There is a link in the book to download a program to display that information. In order to get the most out of the Redbook, you should download the program now and run it.
  2. Chapter 4 describes the z/OS IPL sequence in more detail than you will find in any public documentation. If you want to understand the IPL logic, this is the place to come.
  3. Chapters 6 through 9 describe techniques to reduce the start up and shut down times for CICS, DB2, IMS, and WebSphere. You won’t find this type of information in a single place, and the suggestions can greatly reduce the down times for these applications.
  4. You’ve probably seen separate Workload Manager recommendations for CICS, DB2, IMS, and WebSphere, but have you seen recommendations for when you run all of the subsystems at once? Try Chapter 5 and you will.

Stay Tuned!

Subscribe to Cheryl's List