There is some chatter (I am ignorant of the details) about z/VM high availability clusters. I am not sure how they work. My understanding is that the Z/VM folks are trying to catch up with the concept of a parallel sysplex.
<snip> John Summerfield wrote: Presumably if you're updating the slushware, you have to shut down everything anyway, the question is whether it's all at once or staged. Unless (and I don't know this) you can migrate the running system to different hardware (real or virtual) without IPL. We could almost do this in the 70s on the 3168 (one could take out half an MP, but only one half, not one at a time). </snip>
