The latest 134-page issue of Cheryl Watson’s Tuning Letter is now available on our publications website. We emailed all subscribers last night to let them know, and have already been receiving nice feedback from readers. The Table of Contents is now available on our public website for anyone to view.
Just about anyone that is using Db2 data sharing will be interested in our article about Asynchronous System-Managed Duplexing. This function has been available for over 2 years now, but because it requires Db2 V12, customer takeup has been a bit slow. However, one of our readers, Daniel Hamiel from Nedbank in South Africa, kindly shared Nedbank’s experiences with this new capability. The article starts with a reasonably detailed description of how this new flavor of structure duplexing differs from the System-Managed Duplexing capability that has been available for many years. It then describes the new performance metrics that are provided to help you evaluate it, and provides information about how Nedbank reduced their peak batch CPU consumption by over 700 MIPS (!) by switching to Asynchronous System-Managed Duplexing.
Regular Tuning Letter readers will recall that we published an article about zHyperWrite in Tuning Letter 2018 No. 2. Following that article, we were contacted by John Donoghue in AIB Bank in Ireland, who kindly offered to work with us on an article about their experiences with zHyperWrite. AIB’s data centers are roughly 20 km apart and they use synchronous disk mirroring, yet zHyperWrite has helped drop their response times for Db2 and IMS log writes to 0.4 milliseconds – and that is for the system that is remote to the primary disk subsystem! The article describes AIB’s experiences with implementing zHyperWrite for Db2 logs and IMS WADS, and their plans to expand that to include System Logger, IMS OLDS, and MQ logs in the future. If you use synchronous mirroring, this article will definitely be of interest to you.
We also have an excellent article by IBM’s Glenn Wilcock about ways to reduce HSM’s impact on your peak Rolling 4-Hour Average. Glenn covers recent and not-so-recent HSM enhancements, and we also have supporting information to help you get the best value from HSM’s SMF records. Some customers are moving nearly 200 TB of data per day with HSM, and it is not uncommon to see that HSM is one of the largest CPU consumers on a system. If you use HSM, take a few minutes to read this article – we can guarantee that you will find something in there that will be of interest to you.
Finally, we have articles with valuable tips for COBOL V6 compiles, using DFSORT to manipulate SMF data, understanding RMF SMF fields for zIIP processors, significantly reducing zSecure CPU consumption (from over 3000 seconds per day, down to under 10 seconds!), and our usual review of interesting APARs, Techdocs, Redbooks, and upcoming conferences. We hope you will enjoy it. Please let us know if you have any comments or questions, or if you have an experience that you would like to share with your peers.
If you are going to be at SHARE in Phoenix, please say hello. Cheryl will be there, as will Mario Bezzi, Brenda White, and myself (Frank), and we would love to hear your news and views, and suggestions for future Tuning Letter topics. Bye for now.