Setting both the bind and advertise addresses to the same address causes erroneous logging

Nobody has claimed this yet.

#### Overview of the Issue

Setting both the bind and advertise addresses to the same address causes erroneous logging.
I realize its redundant but im curious as to why this would cause errors in the logs, i.e.:

#### Reproduction Steps

Run Consul with this configuration:

```{"bootstrap_expect":1,"client_addr":"0.0.0.0","datacenter":"datacenter1","data_dir":"/some/dir","domain":"consul","enable_script_checks":true,"enable_syslog":true,"leave_on_terminate":true,"log_rotate_duration":"24h","log_rotate_max_files":30,"log_level":"INFO","rejoin_after_leave":true,"server":true,"ui":true,"bind_addr":"240.0.0.3","advertise_addr":"240.0.0.3"}```

### Consul info for both Client and Server
Consul 1.6.2-dev

### Log Fragments

```Jan 30 16:38:58 node-0 consul[7398]: consul: error getting server health from "node-0": context deadline exceeded

Jan 30 16:38:58 node-0 consul[7398]: 2020/01/30 16:38:58 [WARN] consul: error getting server health from "node-0": context deadline exceeded

Jan 30 16:38:59 node-0 consul[7398]: 2020/01/30 16:38:59 [WARN] consul: error getting server health from "node-0": rpc error getting client: failed to get conn: dial tcp 127.0.0.1:0->240.0.0.3:8300: connect: connection refused

Jan 30 16:38:59 node-0 consul[7398]: consul: error getting server health from "node-0": rpc error getting client: failed to get conn: dial tcp 127.0.0.1:0->240.0.0.3:8300: connect: connection refused

Jan 30 16:39:00 node-0 consul[7398]: consul: error getting server health from "node-0": context deadline exceeded

Jan 30 16:39:00 node-0 consul[7398]: 2020/01/30 16:39:00 [WARN] consul: error getting server health from "node-0": context deadline exceeded

Jan 30 16:39:01 node-0 consul[7398]: consul: error getting server health from "node-0": rpc error getting client: failed to get conn: dial tcp 127.0.0.1:0->240.0.0.3:8300: connect: connection refused```

Start by reproducing the reported Consul 1.6.2-dev configuration with identical bind_addr and advertise_addr values, then trace the server-health RPC and logging path that produces the connection-refused and deadline-exceeded messages. Done means the cause is identified and the erroneous logging behavior is corrected or explicitly clarified for this configuration.

Written by the indexing model from the issue text.