OpenWrt Forum Archive

Topic: VPN drops every 15 minutes

The content of this topic has been archived on 19 Apr 2018. There are no obvious gaps in this topic, but there may still be some posts missing at the end.


I'm using "OpenWrt Chaos Calmer 15.05.1 / LuCI 15.05-149-g0d8bbd2 Release (git-15.363.78009-956be55)" on a WD MyNet N600 Router (yeah, I know, I have a D-Link DIR-860L on order) with OpenVPN and a NordVPN connection.  It drops regularly - I wrote a script to monitor it and log the data as a band-aid (I could restart the connection a few seconds faster by sending OpenVPN a SIGUSR1 at the right time), and I noticed something weird:

The drop-outs occurred every 15 minutes.  Not only that, but on the 00, 15, & 45 minutes of each hour - or just before that time.  When the VPN started had no effect on this time at all.

I've checked the crontab, and found nothing whatsoever.  What am I missing?

Thanks in advance!


Can you post the system and kernel logs which cover at least 30-45 minutes of activity, so we can see what kind of messages there are in the logs regarding this.

I recall that the VPN implementations sometimes had problems with changing external IPs (i.e. a very short DHCP lease interval from your ISP), which did cause connection drops. Not sure if your problem has anything to do with that, but in my opinion it's good to point out issues that other people have had.

(Last edited by Antek on 13 Jan 2018, 11:24)


Thanks for the reply.  I was putting together the logs you requested - "clean" ones without me fussing with the router while they ran - when I found another problem - one of my LAN networks couldn't ping their gateway, but they could get out and  the gateway could ping them.  I finally traced  the problem to an "ifup" script for an OpenVPN connection that had two lines:

ip route add dev tun0 table vpn
ip route add dev tun0 table vpn

"dev tun0" is, of course, my VPN connection's interface.  I looked at that a million times - the lines above and below it use the "dev tun0" correctly - and I missed it each time  Changing that to "dev eth0" solved the ping problem, and - so far - the drop problem has not recurred.

I appreciate your help!


The discussion might have continued from here.