"ccm add" creates node configurations with duplicated "remote_debug_port"

To recreate the issue, create a new cluster with 2 nodes then incrementally add 2 more nodes using "ccm add". The 1st added node will start properly but he 2nd added node will fail due to duplicate remote_debug_ports

[cqlsh 5.0.1 | Cassandra 3.0.7 | CQL spec 3.4.0 | Native protocol v4]

ccm create Cluster1 -v 3.0.7
ccm populate -n 2
ccm start
ccm add node3 -i 127.0.0.3 -j 7300 -b
ccm node3 start
ccm add node4 -i 127.0.0.4 -j 7400 -b
ccm node4 start

> ccm node1 show
> node1: UP
> cluster=Cluster1
> auto_bootstrap=False
> thrift=('127.0.0.1', 9160)
> binary=('127.0.0.1', 9042)
> storage=('127.0.0.1', 7000)
> jmx_port=7100
> remote_debug_port=0 <=== Port 0 for nodes created with "populate"
> byteman_port=0
> initial_token=-9223372036854775808
> pid=1964
>
> ccm node3 show
> node3: UP
> cluster=Cluster1
> auto_bootstrap=True
> thrift=('127.0.0.3', 9160)
> binary=('127.0.0.3', 9042)
> storage=('127.0.0.3', 7000)
> jmx_port=7300
> remote_debug_port=2000 <== port 2000 for nodes created with "add"
> byteman_port=0
> initial_token=None
> pid=1975
>
> ccm node4 show
> node4: DOWN
> cluster=Cluster1
> auto_bootstrap=True
> thrift=('127.0.0.4', 9160)
> binary=('127.0.0.4', 9042)
> storage=('127.0.0.4', 7000)
> jmx_port=7400
> remote_debug_port=2000 <== duplicate port causes node startup to fail
> byteman_port=0
> initial_token=None

Start with the `ccm add` entry point and the node configuration it creates, then reproduce the sequence in the issue with two populated nodes and two incrementally added nodes. Trace how `remote_debug_port` is assigned and verify with `ccm node3 show` and `ccm node4 show`; done means each added node receives a distinct port and starts successfully.

Written by the indexing model from the issue text.