Mainframe IT Geek Week !!

IBM ZR1 Decommission User Experience

Introduction

The week of April 26th through April 30th was recently announced as Mainframe IT Geek Week at Evolving Solutions on LinkedIn. The announcement was for good reason, that week, Evolving Solutions installed one of IBM’s recently announced business class mainframe servers known as the IBM Z Systems 8562 Model T02. This is the first in a series of articles that will discuss what took place at Evolving Solutions Headquarters in support of their new server. This article will focus on what is known as the IBM Decommission Process. 

Project Planning

The first step any large system client will undertake when investing in new server technology is to determine their server migration approach. Essentially, two approaches exist. The first approach is to perform a side by side install and as it’s name implies, both servers will coexist for a time on the same raised floor.  The second approach is known as a push-pull. As the name implies, there is only room for the new server. The push – pull approach places additional time pressures on both the client and the IBM installation team as they must first break down the old server and then turn around and build up the new server in it’s place.  

Clients typically choose the push – pull approach for reasons such as raised floor space within their data center is limited. Or, they may desire to reuse existing electrical connections to their new server in lieu of incurring electrician costs, or perhaps their server resides in a Co-Location (CoLo) facility and to ensure costs are held to a minimum, the existing server must be decommissioned prior to installing the new one as Cage Space is limited. 

Here, at Evolving Solutions, we chose the push – pull approach and we set aside two days during Mainframe Geek Week to first decommission our old server and then install our new T02 server.

The IBM Decommission Process

So what is this Decommission Process? Can’t you simply power off the old server and send it on it’s way back to IBM? The short answer – no. Decommissioning is the process the IBM System Support Representative (SSR) will follow when removing a mainframe server from the client’s raised floor and returning that server back to IBM. When this process is followed, the server is certified ready and capable of being installed at another client location, or used as a build platform as part of an IBM Hybrid Offering.  Your old mainframe has residual value! That value is taken into consideration as part of the purchase transaction and the decommission process guarantees that the returned server is in good working order.   

Why would a client decommission their server? One very good reason is their workload is growing and they require additional capacity. Couple that with the fact that when IBM announces a new mainframe, they include new features that are of interest to the client (e.g. System Recovery Boost). Clients also look for ways to improve their security posture and this is accomplished with advanced cryptographic capabilities found on mainframe servers. Lastly, Financial Reasons will spark reinvestment interest as Leases will come to term and/or Contracts centered on the Right to Use software will expire driving the need for the client to refresh their server platform. The reasons go on and on.  

Remember when

Before we discuss the steps involved on decommissioning our former ZR1 server, let’s revisit the installation process.  

It does not seem that long ago, but back in late 2018 we installed our IBM Z14 model ZR1 server.

First, we have the IBM ZR1 Server arrival from IBM. Note how the mainframe is shipped from manufacturing. It is in an ARBO Crate that is banded. Within the crate, the server is actually wholly contained within a humidity controlled and very large bag that looks similar to an electrostatic bag that is on steroids.

Figure 1 IBM ZR1 Arrival in ARBO Crate

Figure 2 IBM ZR1 additional packaging materials 4Q-2018

Second, IBM Servers also arrive with no shortage of additional assets found in the various boxes and pallets located on the right-side image. The very long boxes, for example hold the doors, front and back, to the new server. All materials first shipped from IBM will eventually find their way back to IBM when the server is decommissioned. Everything must be saved, including the packaging material. Most client’s do not have the room to store the Arbo Crate, pallets, boxes and the like. IBM recognizes this and as part of the server transaction, shipping material is made available to send back the old server. Likewise, the new server packaging material is returned to IBM. At Evolving Solutions, we have plenty of warehouse space – we retained our original ZR1 packaging materials. 

The IBM Server Specific Installation Guide

 Did you know that each Mainframe purchase is accompanied with an IBM Installation Guide? When Evolving Solutions purchased their ZR1 back in late 2018, we received the following Installation Manual.

 Figure 3 IBM ZR1 Installation Manual 

The actual installation is performed by the IBM SSR (System Support Representative) using the above manual that consists of well over 170 pages of instructions.  Mainframes offer up very high levels of availability and scale; installations are taken very seriously by IBM and the customer’s that invest in this server platform.

A glance into the Table of Contents from this manual reveals the IBM SSR installation activities.

 Figure 4a IBM ZR1 Installation Manual – Table of Contents

The installation process is not as simple as providing power as seen in chapter 6 above. It starts first with high level site readiness checks focused on building power. Also considered is frame location and proper frame clearance front to back and side to side. Once power is settled and believe me, this far into the install it will be, the IBM SSR focuses on installing the frame, e.g. the 19-inch rack as detailed in Chapter 3. 

Chapter 4 is where the IBM SSR installs the Human to Machine interface known as the HMC Console. As a side note, HMC’s historically were separate physical servers (Desktops, then Towers as well as Rack Mounted). Evolving Solutions’ new T02 will make use of the HMC Appliance – a first in the Twin Cities Area! Once installed, the power cords are physically attached to the frame and the server is powered up – Chapter 6. 

Chapters 7 and 8 focus on the creation of a private management network used to ensure the server remains 100% available as well as providing the client to machine interface known as the Hardware Management Console.  

Figure 4b IBM ZR1 Installation Manual – Table of Contents

This Human to Machine Interface will request a pair of Public IP Addresses that will allow system staff to administer their new server via their favorite web browser. Using our new T02 as an example, I remotely administer this server from my home in Cave Creek, Az. By default, the ability to POWER OFF and ON this server remotely is disabled. The IBM SSR will enable this capability upon request from the client. Needless to say, I asked our SSR to enable those options.

Figure 5 IBM Hardware Management Console

Prior to cabling, check out tests are executed as revealed in Chapter 9. These tests ensure all mainframe components, channels, internal cables, etc. are in proper working order.  

Chapter 10 calls out the need to physically cable to the clients storage and networking assets (e.g. I/O peripherals).  This is where the mainframe is physically connected to the outside world, e.g. Network as well as storage, Local and Remote consoles.

Chapter 11 will have the customer, or optionally the IBM SSR, run a standalone program that accepts as input a previously created I/O configuration file that depicts what cables were run from the frame to the attached I/O peripherals in the preceding chapter. Note here that the I/O connections are formally defined, that is how PRSM works.  As an aside, IBM’s Dynamic Partition Manager (DPM) eliminates the need to pre-define the base configuration. DPM “discovers” these assets dynamically.

Chapter 11 is all about the I/O Definition process. The input deck in support of this process looks something like the following….

Figure 6 Standalone IOCP input deck

Once the Standalone IOCP is loaded, a Power On Reset is performed to activate this new configuration. The IBM SSR, working alongside the client’s system programmer will validate the I/O Connections and if all looks good, the IBM SSR will turn the server over to the customer to load their operating system(s) and run their workload(s). That process took place late 2018 at our headquarters outside of Minneapolis.

Looking back at our ZR1 install from 2 ? years ago, we captured this photo showing both our mainframe compute and storage subsystem side by side.

Figure 7 IBM ZR1 and IBM DS8K Storage Subsystem

The IBM ZR1 (left-most image) proved to be an excellent addition to our Technology Lab and I am fortunate enough to be the one interacting with this server on a daily basis.  

Mainframe IT Geek Week - The Decommission Chapter

Now, it is time to use that same Installation Manual, locate the chapter that documents the Decommission process and perform the steps one at a time. Appendix A is the chapter that will decommission your server and allow certified return back to IBM. IBM refers to this process as Preparing for Relocation.  Studying Appendix A we find there are well over 160 steps that must be completed by the IBM SSR to decommission and IBM Certify our ZR1 server. This process takes the better part of ? a day to complete and is common across all mainframes offered from IBM.

 Figure 8 IBM ZR1 Installation Manual – Decommission

Restore to Factory Settings

The IBM ZR1 server that we decommissioned was originally purchased as a model A01. Did you know that by studying this model number, you can determine just how powerful our server was relative to the entire product line from IBM at that time. 

• The value 01 represents the number of General Purpose CPs available as shipped from the factory. The IBM ZR1 Server is allowed to have upto 6 General Purpose CPs. 
•  The alpha-character “A” depicts the capacity setting of the engine. On this class server, there are upwards of 250 potential capacity settings labeled Ann, Bnn, Cnn…. Znn.
• A model Z01 would represent a full-speed engine, a Z02 would represent a full-speed dyadic and a B04 would represent a knee-capped 4-way server.

Our machine, though shipped from IBM as an A01, actually ran as a Z04! 

We had a total of 12 cores available for personalization. Personalization is the process where the client selects the type of engine to be configured and directs the hardware to load the microcode in support of that personality, if you will. What this means in English is simply - IBM mainframes allow clients to turn on and off capacity non disruptively and that capacity supports traditional workloads, new workloads that run under Linux and clustering technology. 

Note the very first step to be performed prior to powering off the server – Removing and disabling On/Off Capacity on Demand, CIU, or CBU Features. If we failed to do this, the next recipient of this mainframe would be very happy as they would have picked up a total of 4 full speed engines. In addition to the General Purpose CPs, we also had configured a pair of zIIP engines used by z/OS to host database calls as well as network encryption. There was also a pair of Coupling Facility engines as our z/OS environment is clustered and we even had 4 Integrated Facility for Linux engines to host bare metal as well as virtualized Linux environments. All those engines were turned off during this first step. 

Clearly, this step is key! It’s turning the server back into a pumpkin. 

Sysplex and STP Considerations

Once the server has been dialed back down to it’s original state, the SSR will work with the client to ensure the server is neither part of a Sysplex nor actively feeding time pulses to members of the complex via Server Time Protocol, or STP (1). In the event this server is part of an STP network, the HMC would be used to remove the server from that active timing network. This is accomplished using the Manage System Time dialog found on the HMC, the Remove System from CTN option.

 Figure 9 Manage System Time

The ability to Carry-Forward hardware features

One of the steps found in the table of contents centers on Zeroizing Crypto Express5S (FC 0890) and Crypto Express6S (FC0893) adapter cards. One of the unique capabilities of Mainframe Server technology is you can leverage prior card investment and carry those cards forward to your next server. This step assumes the cards will return back to IBM. We will continue to keep our cards and this happens to be a capability that I have used since our first server, the IBM z13s purchased back in early 2017. As part of the Decommission process, now is the time to harvest hardware that will be carried forward.  

The image that follows reveals the cards we harvested from our ZR1 server.

Figure 10  IBM Mainframe Carry-Forward Capability

Working with the IBM SSR, we removed our Crypto Cards as well as a single Network Adapter Card known as an Open Systems Adapter as well as a FICON Channel Card used for I/O. The OSA Cards will allow Evolving Solutions to meet the IBM Dynamic Partition Manager (DPM) requirement of having an OSA card pair and at the same time improve my local console availability by defining the cards in a dual-purpose mode. DPM is a virtualization technology used by Linux images; there also exists a virtualization technology known as PRSM (Processor Resource Systems Manager), more on both in a follow-on blog.

Discontinue the System 

Our server will not be relocated to another client location. Rather, our system is being returned to IBM. As such, we did not deactivate the CEC at this point in time. Instead, we followed the procedure to discontinue the system.  This procedure will reinstall the original microcode that was shipped with this server and that code is located on a DVD known as the AROM (Activated Read Only Memory). Large System clients know that these assets are stored for safe keeping, ours was located and provided to the SSR for re-installation. The Discontinue process and use of the AROM DVD occurs both for the HMC and Support Element peers.

Out with the old and In with the new! 

Big Blue does not stand still, IBM announced the follow-on to our z14 ZR1 Server; the z15 model T02 April of last year (2020). Evolving Solutions invested in this new server platform first quarter of 2021, the primary focus of Mainframe IT Geek Week was to get that bad boy on our raised floor, IPL’d and ready to be freely explored in our lab. Not to mention gather practical experience on technology yet to be deployed within the Twin Cities such as IBM’s Hardware Management Appliance which we installed on our new server. 

Like the former ZR1, the T02 server from IBM is designed to deliver the scalability and granularity required for both small and medium sized enterprises across all industries.  As in the past, IBM accomplishes this by building off the enterprise class chipset IBM introduced in the Enterprise Class counterpart known as the z15 model T01. It is interesting to note, the 19” form factor found in both the z15 T01 and T02 was first introduced in the prior small and medium sized enterprise class mainframe server – our ZR1. 

The IBM SSR started the Decommission process at 9 AM, at 1:37 pm that same day we were rolling the old frame off the raised floor. Soon thereafter, the frame was shrink-wrapped and then placed back into the original ARBO Crate.   

Figure 11 IBM ZR1 Shrink-Wrapped and re... Arbo-Crated!

Introducing the IBM Z15 T02

With the IBM ZR1 Decommissioned and ready for transfer back to IBM in Poughkeepsie, it’s time to turn our attention to the IBM Z15 T02 Installation.

Just like the ZR1, the T02 arrival starts in an Arbo Crate and found it’s way to our raised floor. I connected the power cables to this new server and then called it a day. 

Of course, I did not leave the Innovation Center without first snapping a few pictures of our new mainframe.

Figure 12 The Z15 T02 ready for Installation on Day 2 of IT Geek Week!

Evolving Solutions IT Geek Week – Week 1 Wrap Up

IBM’s z15 T01 and T02 server design was the most extensive client partnering IBM has ever conducted for a Z release. I enjoyed having the opportunity to install our new T02 server. My next article will focus on the Installation of our T02. I will share experiences on the new HMC Appliance, demonstrate how the harvested cards were installed and discuss the various capacity options found on this new server. We will close it out with a couple of console log extracts that will reveal IBM’s System Recovery Boost in action.

For those of you that have made it this far, thanks for taking the time to give this article a read!

If you are interested in learning more about the IBM Z15 T02 server, or further explore breathing new life into your existing mainframe platform, feel free to reach out to the author via LinkedIn or send the author an email at [email protected] .

The Author:

Footnotes

(1)  Server Time Protocol (STP) is a server-wide facility that is implemented in the Licensed Internal Code (LIC) and presents a single view of time to Processor Resource/Systems Manager? (PR/SM?). STP is designed to provide the capability for multiple servers and Coupling Facilities to maintain time synchronization with each other. Server Time Protocol is required for Parallel Sysplex implementations that span more than one CEC. Evolving Solutions T02 Server includes the STP Feature, even though we operate as a single CEC Sysplex. There are advantages twice per year.