Admin Alert: A Skeleton Checklist for Performing Power i Upgrades

March 10, 2010 Joe Hertvik

A checklist is a great foundation for performing a complex job. Properly constructed, checklists allow you to quickly perform all necessary tasks to accomplish a goal. This week, I’ll present a skeleton checklist for performing Power i upgrades, listing off universal tasks that must be completed for a successful hardware or software upgrade. Use this checklist as a starting point for your own upgrades, adding specific tasks as needed.

The Skeleton Checklist

The skeleton checklist was created to ensure that whenever my shop undertook a hardware upgrade–whether it was a small upgrade or a complete system replacement–that we didn’t forget to perform the common tasks along with our specific upgrade task. My iSeries, System i, and Power i checklist consists of the following steps, representing common tasks to be accomplished during a variety of upgrade situations.

Complete and file all the necessary paperwork.
Make sure everyone knows what’s going on.
Publish the upgrade schedule (or schedules) to all interested parties.
Inventory any hardware components or software media that will be installed in your machine.
Place any scheduled jobs on hold that will be affected by the upgrade.
If necessary, put daily or weekly backups on hold.
If your machine has a Capacity BackUp (CBU) system, ensure that the CBU will have comparable capabilities with the source machine after the upgrade is finished.
Make sure your CBU is ready to stand in for your production box, if necessary.
Make sure any outside vendors participating in the upgrade are authorized to enter the building.
Make sure there’s an active System Service Tools (SST) user ID and password that can be used by outside installers.
Make sure one of the installers knows the Hardware Management Console (HMC) user ID and password.
Ensure you have a system shutdown plan in place.
Have a post-installation test plan ready to go.

The Detailed Checklist

However, while you can use my checklist to stage your upgrades, sometimes you need more information about exactly what you’re checking off to begin with. Here’s the checklist again, but this time I fleshed out its details to demonstrate the rationale behind each checkpoint.

1.Complete and file all the necessary paperwork. This includes in-house paperwork (system change documentation, test plan, etc.) as well as external paperwork for regulatory or certification requirements (Sarbanes-Oxley, PCI Data Security Standard, etc.). Remember to file your paperwork with the proper authorities.

2.Make sure everyone knows what’s going on. This includes the following people:

IT Management–Do your boss and your staff know what you’re doing?
Data owners–Are critical users aware of the change and do they know what to do when it’s finished?
Vendors performing the work–Do outside contractors working on the project know exactly what you want them to do?

3.Publish the upgrade schedule to all interested parties. Do your installers know when to show up? Do your users know when the machine won’t be available? If Web sites interface with your upgrading i/OS partition, make sure to put a notice on the Web page telling your customers about Web site availability.

4.Inventory hardware components or software media that will be installed in your machine. Make sure all ordered parts have arrived undamaged well in advance of the installation date. It’s better to find out if something is missing before the install, rather than during the install.

5.Place scheduled jobs on hold that will be affected by the upgrade. If your upgrade maintenance window conflicts with a processing window, hold or reschedule any critical jobs that would normally run during that window.

6.If necessary, put daily or weekly backups on hold. Conflicting scheduled backups must be accounted for if you’re going to take the machine down for an upgrade. Reschedule accordingly.

7.If your machine has a Capacity BackUp (CBU) system, ensure that the CBU will have comparable capabilities with the source machine after the upgrade is finished. Make sure you have enough CBU resources (especially disk) to allow it to adequately replace your production box after an upgrade. If not, you may have to upgrade the CBU alongside the production box.

8.Make sure your CBU is ready to stand in for your production box, if necessary. If a production system upgrade will take more than a few hours, consider switching processing over to a CBU for continued operation. If your CBU isn’t ready to stand in for your production, check out my two-part article on preparing a CBU for live fail over. Part one can be read by clicking here, and part two can be found here.

9.Make sure any outside vendors are authorized to enter the building. If your iSeries, System i, or Power i box resides at a hosting center, call and place the installer on the site’s Let In and Let Out (LILO) list so they can enter and leave the building. If the machine is located in your on-site data center, make any necessary arrangements for the vendor to enter and leave the building.

10.Make sure there’s an active System Service Tools (SST) user ID and password that can be used by outside installers. Most hardware upgrades will require the installer to enter SST, which will require someone to sign into the SST function to implement hardware changes or to look up information. You can either give your vendor the QSECOFR SST user ID and password or you can create a separate SST user profile for temporary users who need to get into SST.

11.Make sure someone on site knows the Hardware Management Console (HMC) user ID and password. Just like SST, someone may need to get into the HMC to check information or make changes relating to your upgrade.

12.Ensure you have a system shutdown plan in place. Unlike Windows servers, which can be rebooted many times per year, many i/OS boxes can go several months or even a year or more without IPLing. During an upgrade process, you will probably have to IPL your machine at least once so make sure you have an IPL pre-planning and post-planning checklist available.

13.Have a post-installation test plan ready to go. Create a short document detailing out your test plan for making sure that the upgrade worked. Use the plan to test the upgrade and to document the results for later usage.

While no checklist is perfect, if you follow the items presented here, they will help you remember to perform the common tasks that support all system upgrades. If you have any suggestions for other items to flesh out this skeleton even more, feel free to email me via the IT Jungle Contact Web page.