z/VM performance tips

Contents

Note that some items are duplicated under different categories.

Capacity Planning
Cheat Sheets
Configuring for Performance
Connectivity
Hardware-Related Items
Monitor and Data Collection
Problem Resolution
Subsystem Performance
Tuning Guidelines
Tutorial, Reference, Doc

---

Capacity Planning

• Using the CPU Measurement Facility
• z9 to z10 workload scaling
• IBM Z Large System Performance Reference Page - LSPR processor sizing information.
• Linux for zSeries Performance Hints and Tips
• SRM constants and processor version ID codes
• VM Performance and Capacity Sizing Considerations - Various hints about capacity planning as it pertains to z/VM.
• Processor Utilization in LPAR - reconciling utilization numbers from the VMPRF LPAR report with RTM.
• N-way, CMOS, and master processor performance considerations - Some of the areas to be aware of when looking at larger N-way or CMOS configurations.
• VSE/ESA performance when running under VM/ESA - A short description of the potential performance effects of running VSE/ESA as guest of VM/ESA. Also describes the factors behind "It depends".
• Performance Considerations for Linux Guests - A few things to watch out for when running Linux as a guest of VM.
• Configuring Processor Storage - How much central vs. how much expanded.
• VM 64-bit Considerations - Background on z/VM's 64-bit usage and limitations. This has changed greatly with z/VM 5.2.0 so there are two separate discussions covering z/VM prior to z/VM 5.2.0 and z/VM 5.2.0.
• Preferred Guest Migration Considerations - Tips on migrating preferred guest environments to z890 and z800 processors.

 

Cheat Sheets

• Performance-related APARs - This contains a list of APARs that are related to performance. Most improve performance. Some are listed as a warning that they degrade performance and some affect performance management.
• Processor Utilization in LPAR - reconciling utilization numbers from the VMPRF LPAR report with RTM.
• CPU Utilization in an SMT World - Understanding what all those Perfkit utilization reports mean when your LPAR is running SMT-2.
• CMS saved segment management - Describes the purpose, use, and location requirements for various segments shipped with CMS.
• Performance Considerations for Linux Guests - A few things to watch out for when running Linux as a guest of VM.

 

Configuring for Performance

• Configuring Processor Storage - How much central vs. how much expanded.
• N-way, CMOS, and master processor performance considerations - Some of the areas to be aware of when looking at larger N-way or CMOS configurations.
• Minidisk cache information - This is an overview of minidisk cache and some tuning information, including some migration guidelines from older releases.
• Using VM Shared Linux Kernel Support - Guidelines for setting up an NSS (named saved system) for the Linux kernel. This allows Linux to be IPLed (booted) from a shared copy of the kernel. Boot time can be improved with this method and there can be savings to VM system memory requirements since the pages associated with the kernel are shared amongst multiple Linux guests.
• Excessive PR/SM Overhead - How to recognize it and how to abate it.
• Controlling Vertical-Low Logical Cores - Things you can do to control use of those pesky vertical-lows.
• An Unparking Example - A breakdown of a real-world SMT-2 unparking scenario.
• LPAR Entitlement Calculator - A calculator you can use to explore the relationship between core count, weight, entitlement, and numbers of VH, VM, and VL logical cores.
• Dispatch Heuristics vs. Single-Engine Speed - A discussion of the tradeoff between dispatch latency and single-engine speed and of the knobs one can twist to affect those factors.
• Channel Busy: Widely Misunderstood - A discussion of channel utilization, sometimes also called chpid utilization.

 

Connectivity

• Connectivity options - What are my connectivity choices and which one is best?
• Common performance problems and solutions - Brief description of problem and solution (tuning advice, hardware service required, or simply an explanation).

 

Hardware-Related Tips

• Using the CPU Measurement Facility
• About the MT-diagnostic counters
• Some notes on SIIS
• zSeries Large System Performance Reference Page - LSPR processor sizing information.
• SRM constants and processor version ID codes
• Processor Utilization in LPAR - Reconciling utilization numbers from the VMPRF LPAR report with RTM.
• N-way, CMOS, and master processor performance considerations - Some of the areas to be aware of when looking at larger N-way or CMOS configurations.
• Configuring Processor Storage - How much central vs. how much expanded.

 

Monitor and Data Collection

• Performance and Monitor Data Collection - What data to collect and tips on how to do this.
• MONWRITE Data and Migrations - What to collect before and after a migration.
• Monitor MONDCSS Segment - What is that segment for and how do I set one up?
• z/VM performance products - Information on the IBM VM/ESA performance products.
• Tools for Mining MONWRITE Data - A list of tools we know about.
• Processor Utilization in LPAR - reconciling utilization numbers from the VMPRF LPAR report with RTM.
• VM Data Spaces - A discussion of how to track paging I/O associated with VM data spaces.
• Monitor Application Data Registration - Information on proper use of application data product ids and how to register one with IBM.
• Using the CPU Measurement Facility

 

Problem Resolution

• Common performance problems and solutions - Brief description of problem and solution (tuning advice, hardware service required, or simply an explanation).
• Performance-related APARs - This contains a list of APARs that are related to performance. Most improve performance. Some are listed as a warning that they degrade performance and some affect performance management.
• Understanding spool file initialization performance - This is a discussion of the factors involved in the performance of spool file initialization when z/VM is IPLed.
• Understanding poor performance due to paging increases - A brief explanation of what can cause a system to page more and how to determine that cause.

 

Subsystem Performance

• TCP/IP Performance Information - Tips, presentations, and links related to VM/ESA TCP/IP performance.
• VM Scheduler Basics - A discussion of some of the basic concepts of scheduling and dispatching in z/VM.
• VM/ESA data-in-memory techniques - A discussion of the relative merits of saved segments, VM data spaces, minidisk caching, virtual disks in storage, and NUCXLOAD/EXECLOAD for improving performance.
• Minidisk cache information - This is an overview of minidisk cache and some tuning information, including some migration guidelines from older releases.
• VM Data Spaces - A discussion of how to track paging I/O associated with VM data spaces.
• Understanding poor performance due to paging increases - A brief explanation of what can cause a system to page more and how to determine that cause.
• Understanding spool file initialization performance - This is an discussion of the factors involved in the performance of spool file initialization when z/VM is IPLed.
• SPXTAPE Performance - This is a valuable excerpt from the VM/ESA 1.2.2 Performance Report on SPXTAPE performance.
• Tuning Dirmaint Performance
• VM 64-bit Considerations - Background on z/VM's current 64-bit usage and limitations.

 

Tuning Guidelines

• Minidisk cache information - This is an overview of minidisk cache and some tuning information, including some migration guidelines from older releases.
• Tuning Dirmaint Performance

 

Tutorial, Reference, Doc

• SMT Vocabulary - Cores, processors, and utilization, oh my! Use the correct words and phrases, win friends, and influence people.
• Virtual CPU Time - Raw time, MT-1 equivalent time, and prorated core time.
• Processor Utilization in LPAR - Reconciling utilization numbers from the VMPRF LPAR report with RTM.
• z/VM performance products - Information on the IBM z/VM performance products.
• Performance references and documents - List of manuals that are of interest to VM/ESA performance.
• VM Data Spaces - A discussion of how to track paging I/O associated with VM data spaces.
• VM/ESA data-in-memory techniques - A discussion of the relative merits of saved segments, VM data spaces, minidisk caching, virtual disks in storage, and NUCXLOAD/EXECLOAD for improving performance.
• CMS saved segment management - Describes the purpose, use, and location requirements for various segments shipped with CMS.
• VSE/ESA performance when running under VM/ESA - A short description of the potential performance effects of running VSE/ESA as guest of VM/ESA. Also describes the factors behind "It depends".
• Minidisk cache information - This is an overview of minidisk cache and some tuning information, including some migration guidelines from older releases.
• Understanding spool file initialization performance - This is an discussion of the factors involved in the performance of spool file initialization when z/VM is IPLed.
• Understanding poor performance due to paging increases - A brief explanation of what can cause a system to page more and how to determine that cause.
• Performance-related APARs - This contains a list of APARs that are related to performance. Most improve performance. Some are listed as a warning that they degrade performance and some affect performance management.
• Configuring Processor Storage - How much central vs. how much expanded.
• Performance Considerations for OS/390 Guests - A few things to watch out for when running OS/390 as a guest of VM/ESA.
• Performance Considerations for Linux Guests - A few things to watch out for when running Linux as a guest of VM.