A client asked: "Do I need downtime to reboot the HMC?"
Short answer: No (but good luck finding that short answer)
Finding that answer (with --verbose enabled):
Of course, a reboot of the HMC involves downtime of the HMC itself, but what about the logical partitions (LPARs) and the IBM Power Systems they belong to?
First, let's check the official documentation. The Hardware Management Console V7 Handbook has a section on HMC Shut Down or Restart:
2.5.2 Shut Down or RestartThis task enables you to shut down (turn off the console) or to restart the console. To shut down the console, make sureRestart the HMC is not selected, then click OK to proceed with the shutdown. To restart the console, make sure Restart the HMC is selected, then click OK to proceed with the shutdown.
This has explained how to shut down (and restart) the HMC, but what we don't know is what the impact might be to the Managed Systems and their LPARs.
The IBM Systems Hardware information provides similar information: how to shut down the HMC, but not when to do it (or, more importantly, when not to do it).
However, we happen to find this little snippet in the section of the HMC V7 Handbook on upgrading the HMC's firmware:
11.3 HMC firmware maintenance
The HMC is completely independent from the server. The server and all partitions can remain active while maintenance is performed on the HMC, allowing you to easily keep your HMC at the latest maintenance level.
Well, any HMC-specific activities will be interrupted by a reboot of the HMC. So that would include
- Dynamic LPAR (DLPAR) operations, which allow you to add or remove LPAR resources
- Console (virtual terminal) sessions, where you're connected to the LPAR, for example via the vtmenu command
- LPAR profile configuration
- HMC Backups
- other HMC activities.
So, what turned out to be a simple, and very common question: "what's the business impact of an HMC reboot?" is not such a simple one to find the answer to. Of course, like all things, it's easy if you've done it before.