It is going over UDP, and has a single TAP adapter joining incoming tunnels into a /26 network segment.
To minimize the work at remote sites (about 30, spread across the country) I would like to keep the configuration as close to the old as possible.
I am currently testing it by connecting from my laptop, Windows 10 running a recent version of OpenVPN.
The configuration file on my laptop allows me to connect to the old server as well as the new, just changing the remote IP and port.
It works perfectly with the old server.
The new server configuration differs only from the old one in the tunnel IP range and port number.
When I try to connect to the new server everything looks OK, the connection is established and the client gets the correct IP address etc. pushed for its own TAP, it says it's connected, the TAP interface goes from 'media disconnected' to working status, but there it stops.
I don't get ANY traffic through the tunnel. Can't even ping the server's TAP IP from the client.
When I raise the log level to verbosity 6 it shows there *is* something alive, there is constant traffic between client and server.
And when I change just one more line at both sides, from "proto udp" to "proto tcp", it starts working.
This is the configuration file at the server side. Does anyone see something that could explain it not working over UDP?
This configuration here works, but just changing the second line to 'proto udp' (and the same at the client) makes it stop getting any traffic through the tunnel.
port 1202
proto tcp
dev tap
dev-node VPN-UDP
ca OpenVPN-CA.ca.crt
cert server.crt
key server.key
dh dh1024.pem
server 10.255.255.128 255.255.255.192
ifconfig 10.255.255.129 255.255.255.192
ip-win32 dynamic
keepalive 20 120
cipher AES-128-CBC
#comp-lzo
max-clients 50
persist-key
persist-tun
ifconfig-pool-persist ipp-udp-1202.txt
status C:\\Logs\\OpenVPN\\1202-udp-status.log
log C:\\Logs\\OpenVPN\\1202-udp-log.log
verb 6
mute 30
comp-lzo is normally enabled, I just disabled it here to see if that made a difference. It didn't.