Queries on behavior of cluster.quorum-type on two-node RHS clusters

Issue

Is Client-side Quorum dependent on client-side code i.e. each client's glusterfs-fuse code? If so, does make it unsuitable for implementations sharing volumes via CIFS or NFS in addition to FUSE?
Can client-side and server-side quorum be used simultaneously? Should they be?
Using Client-side Quorum on a two-node cluster, once communication between nodes is recovered, will the primary brick overwrite the second brick in every case?
Assuming a two-node cluster with no dummy node, hosting a number of two-brick replica volumes, would be the best configuration I.e. create a single point of failure (removing the HA advantage of the cluster) but ensure data is always consistent?
Is the dummy node included in a 2-node "Starter Pack"?

A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.