OpenVPN point-to-point: connection up but remote access eventually fails from LAN
I'm hitting an odd issue on 2.4.5-RELEASE that seems to have crept up only in the past several months. I have a working point-to-point OpenVPN connection between two pfSense systems (System A and System B). After a few days, machines on the System A LAN are unable to access machines on the System B LAN (including pinging the gateway).
When I check OpenVPN status on System B, i see the following set of messages once a minute (oddly, this is the case whether or not the traffic over the network is working or not):
Jan 27 06:31:37 openvpn 53821 MANAGEMENT: CMD 'status 2' Jan 27 06:31:37 openvpn 53821 MANAGEMENT: Client connected from /var/etc/openvpn/server4.sock Jan 27 06:31:37 openvpn 27421 MANAGEMENT: Client disconnected Jan 27 06:31:37 openvpn 27421 MANAGEMENT: CMD 'quit'
(Re: the above log entries, this post seems to indicate this is normal, my logging verbosity is set to 3 on the server.)
The only thing that seems to fix this issue is to restart both the server and client OpenVPN connections, then all returns to working for a few more days. I can't seem to track down how it's getting into this state, or where the issue lies.
Anecdotally, I will add that I'm streaming MP3 files over this connection using Sonos (System A) and Subsonic (System B), which seems to be when the VPN fails -- that's likely correlation but not causation.
I just installed syslog-ng on both systems to hopefully gather more forensic data, but I thought I'd post here to see if this behavior might have some obvious places to check.