At least three JIRAS relate: SOLR-4316, SOLR-4388, SOLR-5466 It seems to me, that if SOLR-5466 got done (don't see why it'd be hard) we'd have enough to make this work, particularly adding/deleting collections _and_ adding cores on specific nodes to specific shards.
In fact, we don't even need SOLR-5466, we have the cluster state from ZK. Although using the Collections api would insulate it from any changes to cluster state and would be easier to work with. Unfortunately I completely lack the UI skills necessary to work on this, but I'd be happy to collaborate on the rest of it. Anyone want to volunteer for the UI part? Should I cross-post on the user's list? I mean Stefan has done yeoman's duty here, maybe he'd like a rest :). I think this is a significant barrier to using SolrCloud, at least getting started. Most or all of the infrastructure is in place, it seems that this is mostly wrapping it in a UI. Thoughts?
