Hanging Out (Virtually) with IBM Champions, Upgrading an Older HMC and More

Originally published by TechChannel March 23, 2023

Rob McNelly explains how he recently updated an older HMC with the help of IBM Support and gives a reminder about technical debt

In January IBM announced its latest class of IBM Champions.

As is noted in that blog entry, “This year’s 839 IBM Champions come from 60 countries; 68% of them have been IBM Champions before and are returning to the program, including our 30 Lifetime IBM Champions.”

That I am one of those 30 Lifetime Champions is still humbling for me, but of course it’s an honor to be associated with all the Champions. And there is an ongoing association, via the IBM Champion Slack channels. I get on whenever I can. Being able to interact in near real time, to ask questions of and get advice from these very smart people located around the world, is pretty cool.

IBM Support to the Rescue

I was recently asked about upgrading an older HMC. Writing for this website, Jaqui Lynch has a lot of good general information about upgrading and maintaining your systems.

But once I found something relevant to the specific question, courtesy of IBM Support, I thought I had a handle on it. That’s because when I read this:

“This document provides instructions to upgrade vHMC for PowerVM and HMC 7063 machine type from Version 9 Release 1 with mandatory fix MH01858, or V9R2M950 to V10R1M1010.”

I felt confident that I could apply MH01858 and then go directly to V10R1M1010 without messing with any V9 versions.

Simply opening a ticket with IBM Support saved quite a bit of trouble, as they were able to confirm that my client could indeed make the leap. Unfortunately, a hardware issue unrelated to the upgrade threw us for a bit of a loop.

We proceeded through the steps. But once we ran:

            chhmc -c altdiskboot -s enable –mode upgrade

and then:

            hmcshutdown -t now -r

nothing happened.

Support had us go through the process again. When that didn’t work, they asked us to send a debug file.

At this point, they determined that somehow the RAID1 array kicked out one of the drives and sent us a procedure to resolve the issue. This involved getting the root password from Support and running various commands, including:

            /opt/hmc/bin/mvcli info -o vd
            /opt/hmc/bin/mvcli delete -o vd
            /opt/hmc/bin/mvcl rebuild

The point is, if you run into this type of issue, you will most likely need the help of IBM Support, so I’ll leave you to open your own ticket and get the procedure.

Once the array was rebuilt and Support confirmed there were no errors, we were able to continue with the firmware upgrade, the upgrade to V10R2M1030 and the managed system firmware.

Around that time, I pressed a button on my iPhone to upgrade the software. That process went just a bit more smoothly.

A Script to Collect HMC Profile Data

Also HMC-related: Kristian’s Blog: HMC Profile Diff

“More often than not, I find myself having to compare two Hardware Management Console (HMC) logical partition (LPAR) profile configurations. Sometimes this is to ensure that a profile in a disaster recovery site matches that of its production counterpart. Other times, it’s to make sure all members of a cluster have identical resource configurations.

“In a perfect world, I’d be managing the LPAR configurations using something like Terraform, but this currently isn’t an option. I can login to the HMC and manually verify the profiles (which is what I’ve been doing), but this is tedious and prone to error when you need to compare many profile pairs.”

A Reminder About Technical Debt

One of the interesting things about consumer software is that you have few choices when it comes to upgrades. While you can postpone updates on Windows machines, or your phone, the fixes will load eventually. Where am I going with this? The same place I always do. Please consider upgrading your machines on a regular schedule to avoid technical debt.

SXSW is Going On, But There’s an Even Cooler Scene in Austin

Many of us who work on AIX have been fortunate enough to have spent time at the IBM facility in Austin, Texas. But if you haven’t been on site, see it for yourself, and learn how the E1080 models were developed and tested.