I just upgraded my Diversion version from 4.1.3 to 4.3.3 (I know, quit a leap). Doing so has broken a key part of my network setup: my dnsmasq queries need to be resolved through my vpn interface "tun12" to avoid DNS leaks on my VPN'd network. I can accomplish this via the following lines for NordVPN DNS servers in my dnsmasq.conf:
server=103.86.96.100@tun12
server=103.86.99.100@tun12
I was operating this way with no problem. However, since the upgrade, it ignores the @tun12 in my dnsmasq config, and I have a DNS leak.
I can do "/opt/etc/init.d/S56dnsmasq restart" which fixes my leak by applying my /opt/etc/dnsmasq.conf file which is set correctly, however doing so causes many DHCP issues where my statically assigned devices can no longer get IP addresses. Once I restart diversion or I do "service restart_dnsmasq", it fixes my DHCP issues, but re-introduces the DNS leak.
Any insights on how I should go about fixing this? I've found that hardcoding the correct server addresses with interfaces in my /tmp/resolv.dnsmasq file then "service restart_dnsmasq" does resolve the issue, but that seems like a hacky work around. Is there some problem with diversion that it digests the dnsmasq.conf file in an incomplete way (ignoring interfaces) then modifies conf files incorrectly when it starts? Frankly, I'm pretty unclear on a) what happens when I do service restart_dnsmasq, and b) what tweaking diversion does to my dnsmasq configuration. Any info on that would be much appreciated!
server=103.86.96.100@tun12
server=103.86.99.100@tun12
I was operating this way with no problem. However, since the upgrade, it ignores the @tun12 in my dnsmasq config, and I have a DNS leak.
I can do "/opt/etc/init.d/S56dnsmasq restart" which fixes my leak by applying my /opt/etc/dnsmasq.conf file which is set correctly, however doing so causes many DHCP issues where my statically assigned devices can no longer get IP addresses. Once I restart diversion or I do "service restart_dnsmasq", it fixes my DHCP issues, but re-introduces the DNS leak.
Any insights on how I should go about fixing this? I've found that hardcoding the correct server addresses with interfaces in my /tmp/resolv.dnsmasq file then "service restart_dnsmasq" does resolve the issue, but that seems like a hacky work around. Is there some problem with diversion that it digests the dnsmasq.conf file in an incomplete way (ignoring interfaces) then modifies conf files incorrectly when it starts? Frankly, I'm pretty unclear on a) what happens when I do service restart_dnsmasq, and b) what tweaking diversion does to my dnsmasq configuration. Any info on that would be much appreciated!