I am also having this exact same issue. Here is my setup:
m0n0wall has TWO (2) physical interfaces, one on the internet, the other as 10.1.1.1 which is the gateway for our LAN subnet (10.1.1.0). The PPTP server address is 10.1.1.254 with a remote IP range of 10.1.1.80/28. We use Windows Server 2008 RADIUS for authentication. In the firewall, ALL traffic is allowed on both the LAN and PPTP interfaces. On the WAN interface, ALL traffic is allowed from PPTP clients to LAN net and from LAN net to PPTP clients.
VPN clients can connect and authenticate and everything works but, 1:10 after connecting, the m0n0wall log reports:
Sep 8 13:49:22 mpd: 10.1.1.254 -> 10.1.1.80 Sep 8 13:49:22 mpd: [pt0] IFACE: Up event Sep 8 13:50:32 mpd: [pt0] LCP: no reply to 1 echo request(s) Sep 8 13:50:42 mpd: [pt0] LCP: no reply to 2 echo request(s) Sep 8 13:50:52 mpd: [pt0] LCP: no reply to 3 echo request(s) Sep 8 13:51:02 mpd: [pt0] LCP: no reply to 4 echo request(s) Sep 8 13:51:12 mpd: [pt0] LCP: no reply to 5 echo request(s) Sep 8 13:51:12 mpd: [pt0] LCP: peer not responding to echo requests Sep 8 13:51:12 mpd: [pt0] LCP: state change Opened --> Stopping
And then disconnects the user. If we setup a continuous ping from the VPN client to the LAN gateway (10.1.1.1), then the link does *NOT* get disconnected and we don't see any echo request failures in the log. I'm guessing mpd monitors the line, if it sees that it is idle, it then attempts to contact the client and if it doesn't get any response, disconnects. Initially, the client didn't have ICMP allowed in it's firewall and therefore couldn't be pinged from the LAN. Once enabling ICMP, I was then able to ping the client's IP but mpd still reported echo request failures and disconnects.
Previously, we had mpd running on a FreeBSD server on the LAN and forwarded PPTP into it from the m0n0wall and never had this problem. My thought is that there is some option in mpd on m0n0wall which may be causing this to happen.
Any ideas?
Thanks in advance, Philip
|