Hosts not removed from host file on clean exit

[deadwards90]

If for any reason the application loses connection to the pod, it will not get removed from the hosts file.

I've seen this happen in a few different ways it can lose connection but the problem is always the same.

Example below where both services were left in the file after exit.

INFO[12:10:36]  _          _           __             _
INFO[12:10:36] | | ___   _| |__   ___ / _|_      ____| |
INFO[12:10:36] | |/ / | | | '_ \ / _ \ |_\ \ /\ / / _  |
INFO[12:10:36] |   <| |_| | |_) |  __/  _|\ V  V / (_| |
INFO[12:10:36] |_|\_\\__,_|_.__/ \___|_|   \_/\_/ \__,_|
INFO[12:10:36]
INFO[12:10:36] Version 1.22.5
INFO[12:10:36] https://github.com/txn2/kubefwd
INFO[12:10:36]
INFO[12:10:36] Press [Ctrl-C] to stop forwarding.
INFO[12:10:36] 'cat /etc/hosts' to see all host entries.
INFO[12:10:36] Loaded hosts file C:\Windows\System32\Drivers\etc\hosts
INFO[12:10:36] HostFile management: Original hosts backup already exists at C:\Users\user/hosts.original
INFO[12:10:36] Adding custom domain prod01 to all forwarded entries
INFO[12:10:37] Successfully connected context: prod01
INFO[12:10:37] Port-Forward:       127.1.27.1 ledgers:80 to pod ledgers-7dbb48d674-sqmsf:8080
INFO[12:10:37] Port-Forward:       127.1.27.2 portfolios:80 to pod portfolios-59c5989f54-2ntdb:8080
ERRO[12:12:10] Runtime: error copying from local connection to remote stream: read tcp4 127.1.27.1:80->127.0.0.1:63192: wsarecv: An existing connection was forcibly closed by the remote host.
ERRO[12:12:10] Runtime: error copying from local connection to remote stream: read tcp4 127.1.27.1:80->127.0.0.1:63191: wsarecv: An existing connection was forcibly closed by the remote host.
ERRO[13:08:11] Runtime: lost connection to pod
WARN[13:08:11] Stopped forwarding pod ledgers-7dbb48d674-sqmsf for ledgers.prod01.prod01
ERRO[13:08:28] Runtime: lost connection to pod
WARN[13:08:28] Stopped forwarding pod portfolios-59c5989f54-2ntdb for portfolios.prod01.prod01
INFO[13:53:50] Received shutdown signal
INFO[13:53:50] Stopped watching Service events in namespace prod01 in prod01 context
INFO[13:53:50] Clean exit

---

This is quite similar to this issue but this does not seem to be a race condition.