[CX] Intermittent server-add failure due to renaming of previously explicitly renamed node

Analytics regression tests are intermittently failing due to ns_server causing a rename of n_0 during a server add of a subsequent node. This rename causes the server add to fail, which fails the test. As n_0 was explicitly (user) renamed upon starting the cluster, the later rename is unexpected.

See attached logs, which demonstrate the issue:

• cluster initialized (cluster_run)
• n_0 (user) renamed to 127.0.0.1
• n_1, n_2, n_3 added to cluster
• cluster rebalanced to n_0, n_1, n_2, n_3
• n_1, n_2, n_3 rebalanced out
• n_1 attempts to add to cluster
• server-add fails due to concurrent (system) rename of n_0 to machine IP