z/VM Presentations

In this session we will take an in depth look at just how large or wide or tall can you grow a z/VM system. The session will also discuss the type of data you can look at to evaluate how close your system is to those limits. As we look at different subsystems (processors, memory, etc.), we will discuss theoretical and practical limits. Included will be a brief look at some of the ways IBM tests these limits.

z/VM 7.4 and future releases adhere to the Linear Service model. Corrections to z/VM will be bundled into service fix packs and new features will be bundled into feature packs. Service fixes will be delivered on a monthly basis and feature packs will be delivered twice a year. The exceptions are critical fixes, which will ship as soon as possible, and hardware support which will be treated as a feature pack outside of the twice yearly cadence. By ordering any z/VM APAR (service fix pack, feature pack, hotfixes, or hardware support) you will receive all fixes and function that has been released up to the point when the APAR was closed.

For decades, z/VM has been known for its resilience and reliability. This presentation discusses the best ways to make sure your z/VM systems maintain the level of performance you have come to expect. Keeping current with service, having a monitoring strategy, and following documented procedures are key concepts that are covered in order to help avoid common issues.

z/VM's lineage is over fifty years old. A product with this much history means there is a wealth of information out on the internet. However, not all of it may be relevant or the exact information you're looking for. This presentation highlights the most important resources for someone interested in learning more about z/VM, or someone that needs to use z/VM for their day-to-day job. Find the best education and publication resources, learn how to effectively use the z/VM help facility, and find where you can get the latest news for z/VM.

z/VM 7.3 scales the maximum number of Single System Image (SSI) members from four to eight. This presentation reviews SSI concepts and goes into detail on migrating existing SSI clusters to eight members.

If you thought "regular" virtual machines were useful, wait until you get a load of adjunct virtual machines. These VMs run along side a principal VM and add a fun new twist to debugging. This presentation describes adjunct virtual machines, and how to set them up.

In configuring z/VM paging, you are faced with many choices, and the number of choices has increased as additional enhancements have been made. This presentation walks through the choices and gives guidance for making and verifying decisions. In order to discuss these choices, an introduction to the paging subsystem is given. This presentation focuses on paging to ECKD 3390 devices, but briefly mentions paging to EDEV SCSI devices. Topics include use of EAV, HyperPAV, number of volumes, etc.

Linux on IBM Z provides unprecedented value by allowing the consolidation of [hundreds / thousands] of Linux guests onto a single mainframe. However, just because you can't see the virtual machine doesn't mean you can't (or shouldn't) secure it. This talk enumerates the threats against virtualized environments ... and then explains how Linux on z Systems defeats each of those threats. Topics will include security from z at its most real (the hardware) up through Linux guest security and how to lock a LinuxONE providing cloud services.

The SFS Performance: Part 1 presentation discussed the basics of Shared File System (SFS) performance management. And if you.ve seen it, you learned you do not need to be an SFS expert to manage the SFS performance. Now we will go a step further and try and combine SFS internals with performance. What are those things inside SFS called agents? Are they spies or part of some covert action? If you choose to attend this session, you will learn to tell the good agents from the misbehaving agents, when VM data space exploitation is appropriate, and what all those counters from QUERY FILEPOOL REPORT really mean.

Most of the time we think about how a z/VM system looks when it is running. Two keys to well running systems are starting them up and shutting them down well. This session will suggest questions to ask and answer before you startup or shutdown your z/VM systems and guests, and will provide some suggestions to help you answer them, including: What order should I start virtual machines in and how many should I start at once? What do I put in AUTOLOG1's and AUTOLOG2's profiles? How do I determine when startup and shutdown procedures are complete and/or need action to be taken? Why do I need an "orderly" shutdown of both guests and the z/VM system and how do I do that?

In this presentation we use z/VM Perfkit screens to explore systems capacities and utilizations in a number of different categories of resource. Sample screen shots illustrate which Perfkit screens to consult, which columns are especially helpful, and what values in those column represent "full".

It sounds like an old real estate joke, but the three most important things to understand about Live Guest Relocation: Domains, Domains, Domains. Live Guest Relocation was introduced in z/VM 6.2 and many customers have grown to love it and depend on it. One of the key concepts associated with LGR is relocation domains. If you've ever wanted to really understand this part of the whole relocation construct, then come to this session. We'll describe what relocation domains are and how they affect LGR. You'll leave being able to understand how virtual architectures are influenced by relocation domains. In other words, you'll leave being able to amaze your friends and coworkers.

The z/VM TLS Server plays a key role in IBM Z virtualization security. Whether you're securing voluminous amounts of web traffic or just need to secure your hypervisor Telnet sessions, the TLS server is involved. This session will discuss the configuration and management of the server, as well as delve into the intricacies of protocol selection and certificate management.

We look at what happens when z/VM starts, and how your set up points the system down various paths. How does it know which configuration files to use? Which CP to load? What UserIDs are defined, and which should be the operator? Whether and where to leave the operator connected? What to automatically start? We walk through a z/VM system set up by a standard install, pointing out key configuration selections, their effects, and how to change them.

With z/VM 6.3's ability to use up to 1 TB of real memory, it is time to revisit the question "What type of disk should I use for z/VM Paging?" This session will give some background on z/VM paging, a review of devices options, and then look at real data from some work in this areas. Included will be discussion of how paging characteristics change under different workloads and circumstances.

This presentation provides an overview of the z/VM Virtual Switch and Virtual Network Interface Cards. Link Aggregation is explained along with the benefits it provides for system performance, reliability, and failover.

Capacity Planning can be a challenge in a stand alone environment, but how in the world do you do it in a virtualized environment such as z/VM? This session will describe the top focus areas along with key performance metrics that one should track when doing z/VM capacity planning. Additionally, we'll discuss at a high level some models for planning: natively on z/VM through home grown tools, through tools on z/OS, or in conjunction with other products. The companion document is a list of the top performance metrics to track.