@lisa-altair - I understand the design and the change in behavior you are suggesting also looks good.
I just have one minor comment - it is not clear in the design that it would be the server that would recognize that ALL MoMs reporting the compute nodes/vnodes are offline, so it should go ahead and mark the compute nodes offline.
I’m not sure that we need to mention it’s the server that decides to offline the vnodes if there are no more moms, that seems like an internal implementation detail. It could be mentioned if others think it important but I wouldn’t consider it part of an external design.
The design looks good to me!
I have made a few modifications to the design. Please have a look.
The design looks good to me.