I used the openvpn-install script located here: https://github.com/Nyr/openvpn-install to put together a test server - everything worked except none of the clients would connect. The logs on the server were:
Nov 17 16:26:17 openvpn ovpn-server[2233]: Options error: Bad protocol: '11'.
It appears that both the server.conf and client.conf file had the line
proto 11
, where both the clients (tested on iOS) and the server (linux) appeared to expect the text string
proto udp
rather than the numeric value.
When I manually edited the files to use "udp" instead of 11, restarted the server and rebuilt the client .ovpn files, everything seemed to work.
I haven't looked deeply into trying to determine how/why the config files were using the numbers instead of tcp/udp, but I figured it might be a common enough thing that others might benefit...
