Killing a UDP connection

[retrovertigo]

I'm connecting to my work VPN using OpenVPN, but every so often my wireless connection flat out dies (while still connected to the access point). This is thanks to the 10.4.10 update, which has absolutely murdered my wireless connectivity.

The problem is this: I can't re-connect to my VPN after the connection dies, because when I try I get the following error in my OpenVPN log:

Fri 07/13/07 09:58 AM: TCP/UDP: Socket bind failed on local address [undef]:1194: Address already in use

I got the following information from netstat:

Proto Recv-Q Send-Q  Local Address          Foreign Address        (state)
udp4       0      0  *.1194                 *.*

Currently, the only way I've been able to re-connect is by rebooting my laptop. This is not feasible, though. Is there any way to kill that connection/socket so that OpenVPN is able to proceed?

 

[tyr2]

This means the openvpn process is still running, you need to stop it (with 'kill' or similar) before you can start it again.

ps -aux, or activity monitor will get you the process id (pid), then do 'kill -KILL pid' in the terminal.

 

[retrovertigo]

Yeah I figured that, but before I posted I ran "ps -aux | grep openvpn" and didn't see anything. However, I just remembered that I'm using the new version of TunnelBlick to manage my openvpn connection, and the new version comes with the openvpn binary bundled in the application. I re-ran "ps -aux" without the pipe and found the following line:

root       246   0.0  0.1    27784    900  ??  Ss   10:44AM   0:00.88 /Applications/Tunnelblick.app/Contents/Resources/o

That line was truncated in the ps output, it was actually "/Applications/Tunnelblick.app/Contents/Resources/openvpn".

So, if and when this happens again I will look for that process and kill it.

Thanks!