Strange logs -> losing OpenVPN connection every 20 - 120 seconds
-
I have the same problem, the problem is less since the 2.1.1-PRERELEASE (i386) built on Fri Feb 7 11:46:37 EST 2014
I had to change the "DOWN" time in gateway monitoring to 60 seconds.BUT:
OpenVPN does restart all services, even if there was no change to the WAN-ip:
Feb 10 12:46:24 php: rc.start_packages: Restarting/Starting all packages. Feb 10 12:46:22 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(opt1). Feb 10 12:46:22 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(wan). Feb 10 12:46:22 check_reload_status: Starting packages Feb 10 12:46:22 php: rc.newwanip: pfSense package system has detected an ip change -> 192.168.8.1 ... Restarting packages. Feb 10 12:46:22 php: rc.newwanip: rc.newwanip: on (IP address: 192.168.8.1) (interface: ) (real interface: ovpns1). Feb 10 12:46:22 php: rc.newwanip: rc.newwanip: Informational is starting ovpns1. Feb 10 12:46:19 check_reload_status: rc.newwanip starting ovpns1 Feb 10 12:46:19 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(opt1). Feb 10 12:46:19 kernel: ovpns1: link state changed to UP Feb 10 12:46:19 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(wan). Feb 10 12:46:19 check_reload_status: Reloading filter Feb 10 12:46:19 kernel: ovpns1: link state changed to DOWN Feb 10 12:46:19 php: rc.openvpn: OpenVPN: Resync server1 OpenVPN WAN Feb 10 12:46:19 php: rc.openvpn: OpenVPN: One or more OpenVPN tunnel endpoints may have changed its IP. Reloading endpoints that may use GW_WAN. Feb 10 12:46:16 check_reload_status: Reloading filter Feb 10 12:46:16 check_reload_status: Restarting OpenVPN tunnels/interfaces Feb 10 12:46:16 check_reload_status: Restarting ipsec tunnels Feb 10 12:46:16 check_reload_status: updating dyndns GW_WAN Feb 10 12:45:02 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(opt1). Feb 10 12:45:01 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(wan). Feb 10 12:30:39 php: rc.update_urltables: /etc/rc.update_urltables: pfBlockerBadIPList does not need updated. Feb 10 12:30:39 php: rc.update_urltables: /etc/rc.update_urltables: Starting URL table alias updates Feb 10 12:30:02 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(opt1). Feb 10 12:30:01 php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(wan). Feb 10 12:30:01 php: rc.update_urltables: /etc/rc.update_urltables: Sleeping for 37 seconds. Feb 10 12:30:01 php: rc.update_urltables: /etc/rc.update_urltables: Starting up.Also I keep getting```
php: rc.filter_configure_sync: Could not find IPv6 gateway for interface(opt1). -
The OpenVPN clients are restarted if a WAN seems to go down/up even if the new WAN IP turns out to be the same as the old one. I believe this is because there were (are?) issues with OpenVPN continuing on after such a WAN transition. Restarting them makes sure they re-establish happily.
Wouldn't be such an event logged under "Gateways"? If so, that's not the case.
The DNS translations of Aliases that contain FQDNs are re-checked every 5 minutes. But that happens on my (and many people's) systems without triggering a WAN transition and all the resultant processing.
Just for the lols, how would one disable this?
What is in cron? (Install the Cron package to look at that easily) Any jobs that specify "3-59/5" that would make it start at 00:03 and every 5 minutes thereafter?
Nope.```
minute hour mday month wday who command
1,31 0-5 * * * root /usr/bin/nice -n20 adjkerntz -a
1 3 1 * * root /usr/bin/nice -n20 /etc/rc.update_bogons.sh
*/60 * * * * root /usr/bin/nice -n20 /usr/local/sbin/expiretable -v -t 3600 sshlockout
1 1 * * * root /usr/bin/nice -n20 /etc/rc.dyndns.update
*/60 * * * * root /usr/bin/nice -n20 /usr/local/sbin/expiretable -v -t 3600 virusprot
30 12 * * * root /usr/bin/nice -n20 /etc/rc.update_urltables@phil.davis: > Or is there a 5-minute egg timer somewhere on a connected device or cable that glitches it every 5 minutes ;) I hope not, but it seems so… :'( Edit: I just tested this scenario using Mikrotik RouterOS. My OpenVPN tunnel is up and running fine for the last two hours or so. No hiccups at all. I really don't want to switch back to RouterOS tho. I'm already starting to miss pfSense's amazing WebUI. So, please continue help me to get this fixed. :-[ -
I think you have assigned your ovpn interface.
That activates the monitoring and probably causing a lopping of sorts there.Can you check if you can ping the monitoring ip by default or disable it entirely?
One of those i think will fix your issues. -
@ermal:
Can you check if you can ping the monitoring ip by default or disable it entirely?
One of those i think will fix your issues.- I turned off gateway monitoring altogether. This didn't solve the problem either.
No, it did not. :-\ And I disabled monitoring on all the gateways. They always appear as online, which is why I don't understand most of the logs.
Btw. I had to assign it, because I wanted to transparently route all traffic of one specific workstation.
-
So find a machine to ping on the other side then :)
Though disabling monitoring should have avoided this unless you have other problems in the tunnel.
-
I reverted back to
2.1-RELEASE (amd64)
built on Wed Sep 11 18:17:48 EDT 2013and this oddity is gone for good. Guess I'm going to stick to that for a some more time!
-
The system log isn't all that telling on its own, what does the OpenVPN log show?
-
Still happening even with latest snapshot. The WAN IP is in fact static, it never changes. These newwanip detections are a pile of BS.
rc.newwanip: pfSense package system has detected an ip changeSame thing happens with IPsec:
php: rc.newipsecdns: IPSEC: One or more IPsec tunnel endpoints has changed its IP. Refreshing.As for OpenVPN log, nothing useful there either:
Mar 6 12:25:37 openvpn[47673]: /sbin/ifconfig ovpns1 10.0.8.1 10.0.8.1 mtu 1500 netmask 255.255.255.0 up Mar 6 12:25:37 openvpn[47673]: do_ifconfig, tt->ipv6=1, tt->did_ifconfig_ipv6_setup=1 Mar 6 12:25:37 openvpn[47673]: TUN/TAP device /dev/tun1 opened Mar 6 12:25:37 openvpn[47673]: TUN/TAP device ovpns1 exists previously, keep at program end Mar 6 12:25:37 openvpn[47673]: Control Channel Authentication: using '/var/etc/openvpn/server1.tls-auth' as a OpenVPN static key file Mar 6 12:25:36 openvpn[47673]: Initializing OpenSSL support for engine 'cryptodev' Mar 6 12:25:36 openvpn[47673]: NOTE: the current --script-security setting may allow this configuration to call user-defined scripts Mar 6 12:25:36 openvpn[47673]: OpenVPN 2.3.2 i386-portbld-freebsd8.3 [SSL (OpenSSL)] [LZO] [eurephia] [MH] [IPv6] built on Jan 23 2014 Mar 6 12:25:36 openvpn[97470]: SIGTERM[hard,] received, process exiting Mar 6 12:25:36 openvpn[97470]: /usr/local/sbin/ovpn-linkdown ovpns1 1500 1558 10.0.8.1 255.255.255.0 init Mar 6 12:25:36 openvpn[97470]: event_wait : Interrupted system call (code=4) -
Every time a connection/reconnection occurs you see the logs for newwanip.
Do you have anything similar to below in your logs?
log_error("DEVD Ethernet attached event for {$iface}"); log_error("HOTPLUG: Configuring interface {$iface}");Give this a try again as well:
diff --git a/etc/rc.linkup b/etc/rc.linkup index 1994336..43607b1 100755 --- a/etc/rc.linkup +++ b/etc/rc.linkup @@ -100,7 +100,7 @@ if (!file_exists("{$g['varrun_path']}/booting") && empty($g['booting'])) { break; } $interface = convert_real_interface_to_friendly_interface_name($argv[2]); - if (!empty($interface)) + if (!empty($interface) && substr($argv[2], 0, 4) != "ovpn") handle_argument_group($interface, $action); } -
No, no hotplug/devd events in the log. As for the patch, well yes I can try that (should obviously stop triggering the OVPN restart), but it's not really like it'd happen for any good reason at all. It seems like it just decides the WAN IP "changed" a couple of times a day, randomly.
-
Well if your Dyndns or any other monitoring for hostname changes triggers the event you have to find which is doing that.
Also find out why its triggering the event.To me from the logs it looked like devd was doing that since it was not clear on dyndns/hostnames on vpns being used from the logs.
Maybe the config.xml can help to validate the options here.
-
@ermal:
Well if your Dyndns or any other monitoring for hostname changes triggers the event you have to find which is doing that.
Also find out why its triggering the event.To me from the logs it looked like devd was doing that since it was not clear on dyndns/hostnames on vpns being used from the logs.
Now that you mention dyndns…
Mar 7 08:22:53 check_reload_status: Restarting OpenVPN tunnels/interfaces Mar 7 08:22:53 check_reload_status: updating dyndns HEIPV6_TUNNELV6 Mar 7 08:22:53 check_reload_status: Restarting OpenVPN tunnels/interfaces Mar 7 08:22:53 check_reload_status: updating dyndns WAN_DHCP Mar 7 08:22:52 check_reload_status: Reloading filter Mar 7 08:22:52 check_reload_status: Restarting OpenVPN tunnels/interfaces Mar 7 08:22:52 check_reload_status: Restarting ipsec tunnels Mar 7 08:22:52 check_reload_status: updating dyndns WAN_DHCP Mar 7 08:22:52 check_reload_status: Reloading filter Mar 7 08:22:52 check_reload_status: Restarting OpenVPN tunnels/interfaces Mar 7 08:22:52 check_reload_status: Restarting ipsec tunnels Mar 7 08:22:52 check_reload_status: updating dyndns HEIPV6_TUNNELV6Obviously, this again makes no sense since there is nothing dynamic - except for the gateway being statically "dynamic". The GW IP never changes, it is configured as static IP on the GIF interface. What's "updating" there? Again, the WAN_DHCP is a static DHCP lease, never changes.
-
Can you check the relevant files for this to see why it triggers?
-
@ermal:
Can you check the relevant files for this to see why it triggers?
Maybe… if only I knew the relevant files. :D
-
/conf/dyndns*.cache