Edge-to-edge chaining

If your organization is geographically dispersed, you might want all your users to have an edge server nearby.

An edge server can get files and metadata from another edge server rather than a distant commit server. You can have any number of such edge servers in the chain. Among the many possible scenarios are:

1. edge3 -> edge2 -> edge1 -> commit server

2. edge3 -> edge2 -> edge1 -> forwarding replica -> commit

3. edge3 -> edge2 -> edge1 -> forwarding replica 2 -> forwarding replica 1 -> commit

If you include a forwarding replica, it must connect directly with another forwarding replica or the commit server. Do NOT put an edge server between a forwarding replica and the commit server

Configure each server so that its P4TARGET is the closest inner server. For example, edge2's target is edge1.

Configure each server with a service user with ticket-based authentication to ALL of the servers that are closer to the commit server. For example, in scenario 1 above, edge2 needs to be authenticated with both edge1 and the commit server.