Leverage Cluster ID to help prevent joining discrete clusters

Now that we have a uniquely identifying Cluster ID (gh-6702), we could use it to prevent the joining of members participatory of another cluster. When a member is `[re]join`ed, it should either have no existing Cluster ID, or a matching Cluster ID to the cluster it is being joined with. A non-matching Cluster ID is an indication that the member was participating in another cluster, and will have a non-sensical raft log & state store, likely causing some problems.

Review gh-6702 to understand the Cluster ID and then trace the member [re]join path. The work is done when joining accepts members with no Cluster ID or a matching ID, and rejects a non-matching ID before it can create inconsistent Raft log or state-store data.

Written by the indexing model from the issue text.