Replication retries to connect to master indefinitely

During the first call box.cfg() now waits for an eternity for all
cluster members to start up. All subsequntial box.cfg() calls will
timeout after 10 seconds if one or more hosts are unreachable.

Fixes #1511