1. Correction for BPXPRMxx Recommendation
2. Caution on APAR for DB2 V10
3. Feedback on z114 Machines
4. Request for Experiences with PFA
5. Cheryl Watson’s Tuning Letter 2011 No. 5
6. Last Update on zPrime
1. Correction for BPXPRMxx Recommendation
In our Tuning Letter 2010 No. 6 (page 11), we made a recommendation that you should set the value of MAXFILEPROC to 1024, because that’s what was added to the IBM Initialization and Tuning Reference manual in z/OS 1.9. It wasn’t until z/OS 1.13 removed that statement from the manual, that I realized that elsewhere the recommendation AND the default is 64000. We will correct this for the next DVD, but it’s important to change if you had used our 1024 recommendation. MAXFILEPROC is the maximum number of files, sockets, directories, and pipes that can be active or allocated at one time for each process. There is a z/OS Health Check, USS_MAXSOCKETS_MAXFILEPROC, to check this setting, and you can determine the current setting with ‘D OMVS,O’. Some large sites with a large number of TN3270 connections have had to increase this value to 128000 or even the maximum of 524286.
2. IBM’s New z114 Processors
After applying a z/OS APAR OA35885, one site experienced an increase in the CPU time of the DB2MSTR (DB2 10) address space. It increased by a factor of 100. They did not see the same effect on DB2 9. Here are the related APAR descriptions:
OA35885 (z/OS 1.10-1.13, 22Jun2011)) – Current Real and Auxiliary Storage Usage Information is not Available for High Virtual Storage. This APAR adds a new IARV64 REQUEST option of REQUEST=COUNTPAGES, which provides the number of 4K pages with a copy in real, the number of pages with a copy on aux, and the number of pages with a copy in real or aux storage. This is in support of APAR PM24723.
PM24723 (DB2 10, 28Jun2011) – IFCID 225 Real Storage Statistics Enhancements IFCID225. This was a HIPER APAR to provide real storage monitoring for above the bar storage, and uses the facility provided by OA35885.
The problem is still being investigated by IBM, but we’d like to know if anybody else who is running DB2 10 has experienced this same problem after applying APAR OA35885. Please let us know at technical@watsonwalker.com. Thanks.
3. Feedback on z114 Machines
We’ve been hearing great feedback about the new z114 processors. A typical comment comes from Rick Ohara:
“We love our z114. We are able to do 30-40% more work with this box vs our old z9! And with a 5% price increase coming on January 1, 2012, I don’t see how you can afford not to!”
Others tell me that is was a no-brainer after looking at the comparative costs for maintenance and energy consumption.
4. Request for Experiences with PFA
PFA, Preventive Failure Analysis, is a z/OS facility that can predict possible failures and provide early notification. We described PFA in our Tuning Letter 2010 No. 2, pages 27-33, and provided some user experiences in our Tuning Letter 2010 No. 5, pages 3-7. Not many people were using PFA at that time. So now, we’d like to find out if anyone is using it successfully, or whether people are still having problems with it. Please send an email to technical@watsonwalker.com to share your experience.
5. Cheryl Watson’s Tuning Letter 2011 No. 5
I am still working on Tuning Letter No. 5. My back surgery has put me behind schedule. But the good news is that I’m able to sit at my desk again, which makes writing much more comfortable. And ALL pain is now gone. Subscribers should see 2011 No. 5 this month but 2011 No. 6 may not be published until January 2012. We really appreciate the patience our readers have shown us.
6. Last Update on zPrime
Most people are aware that I recommended that customers not embrace zPrime when it came out. See Cheryl’s Lists #148 and #149. I suspected that the courts would not support Neon Software’s zPrime product to move workload to zIIPs and zAAPs outside of IBM’s stated intention. A few people have disagreed with my position. But a little known court document reveals what might be behind the court’s decision (http://www.courthousenews.com/2011/07/22/Order%20on%20sealing%20documents.pdf). From page 2 – “the documents include discussion of evidence that Neon employees admitted to reverse-assembling IBM’s code, hiding evidence of this reverse-assembly, and lying under oath about whether Neon engaged in reverse-assembly.”
Stay Tuned!