Michael,

This is Mike from the SDS team , I apologize as I cannot fully answer this question from a Disk Caching standpoint but I will do my best to address this from a Hyperscale perspective.

I reviewed how the Disk Caching is set up and apparently it is simply stating a scheduling criteria for database log backups.

This would imply it is aimed at a target storage ( the hyperscale distributed file system disk) as opposed to any one individual node.

There are some components in Commvault, like the index cache, that will be effected by an offline node since there are Index signatures that are unique to the cache on that node. 

In this Disk Caching instance, since we are pointing to a library as opposed to an individual node, there should be no interupption in service if you need to bring down a node for maintenance while the other cluster nodes are online.