We emailed Tuning Letter 2016 No. 4 to our subscribers on February 8. This issue provides all you wanted to know (and probably some stuff you didn’t want to know!) about z Systems Collocated Application Pricing (zCAP). It also discusses the recently-announced z Systems Workload Pricing for Cloud (zWPC), and how you can use Workload Manager to gather the CPU time for your MWP-qualifying and zWPC-qualifying workloads and feed that directly to SCRT without having to process millions of transaction-level SMF records and mess around with CSV files. Even if you currently don’t use MWP or zWPC, there are some really nice additional functions in WLM (like reporting of CPU time at the service class and report class level) that apply to nearly all z/OS users.
Our Tuning Letter 2016 No. 3 contained an introduction to Spark (available in the new Spark section of our website) and showed you how to create and test an SQL query against a sequential data set containing SMF records. In this issue, we took readers a step further, providing tips about getting the Spark infrastructure up and running quickly. And, in line with our objective that every Spark article should result in some tangible outcome, we showed readers with zero Java skills (like me!) how to take the SQL queries that you created using Data Studio, turn them into Java code, and run that in Spark. The SMF query that we used as an example is very basic, but it was created by someone that doesn’t know any SQL or any Java (me!), it proves that all the pieces of your Spark environment work and communicate with each other, and it gives you a working example that you can then build upon.
IBM introduced the CPU Measurement Facility (the source of SMF type 113 records) on z10, but it was only with the z13 generation that it started getting more interest from customers and vendors. One of the biggest determinants of the amount of capacity that you will get from your z13 is the Relative Nest Intensity (RNI) of your workloads, and the only way to see and tune your RNI is using the type 113 records. And with a new generation IBM mainframes expected later this year, it is more important than ever that you are familiar with your RNI and how you can influence it. To help you in that endeavor, we are producing a 3-part series of articles about CPU MF in collaboration with Todd Havekost from IntelliMagic, and David Hutton from IBM. The first installment explains the importance of understanding the information in the type 113 records and shows you how to start collecting them. Our next installment will show the type of insight that you can extract from your type 113 records.
This issue’s User Experiences describes the experiences of a large zEDC user (“We first began to experiment with zEDC during 4Q2014 and we never experienced a single related issue or problem with it on any sysplex.”), and provides tips about using RMF XCF and CF reports to see what is happening in your sysplex, using the RMF XML reports, and where to get information on WLM best practices. We also have all the normal information about important New Function and performance-related APARs, white papers, flashes, Redbooks, blogs, and so on. You can find the table of contents at http://watsonwalker.com/publications/tuningletter/table-of-contents/.
If your site is a Tuning Letter subscriber, make sure that your site’s contact passes it along. We are also encouraging subscribers to give us group email IDs (e.g. zOS-Support@greatcompany.com) so we can send you issues directly. We had several email rejections from subscribers, so if you are an active subscriber and you didn’t receive your issue, please send an email to firstname.lastname@example.org. If you particularly liked some article, please let us know – our job is to provide you with the most helpful information possible. And the best way to help us do that is to send us your feedback about what you like or don’t like.
That’s all for now. Stay Tuned for more mainframe news and views.