I’ve just upgraded from Core Update 202 to 203 and DNS resolving through zones doesn’t work any more.
I’m using 3 ipfire firewalls to simulate 3 datacenters. All of them are connected using Wireguard. I’ve set the zones dc01.local, dc02.local and dc03.local in Network → DNS Forwarding on all firewalls and it worked on Core Update 202.
After the update I can resolve all the hosts inside one zone (eg from host in dc01 to host in dc01) but I’m not able to resolve hosts in the other zones (from dc01 to dc02 or dc03 and other combinations). Have I missed something after the upgrade? Or wasn’t this meant to work like that?
Please note: Forwarded zones can no longer be specified as fully-qualified domain names. You now need to replace any entries on the DNS Forwarding page that use FQDNs with IP addresses.
I need the same feature and configured dns forwarding as the TO, but i don’t understand these required actions:
Forwarded zones can no longer be specified as fully-qualified domain names. You now need to replace any entries on the DNS Forwarding page that use FQDNs with IP addresses.
A conditional dns forwarding always needs the domain name (like domain.localdomain) and the ip-adress of the dns-resolver to what the dns-request is forwarded to.
I’m connected, as said, with Wireguard RoadWarrior into the network of the virtual data centers using firewall301.dc03.local. The one is in the hosts list under Network → Edit Hosts. Using nslookup I’m getting timeouts. Now I need first to know if DNS is listening just to the firewalls “native” (green) subnet? My RoadWarrior gets an IP from 192.168.203.0/24 and the green network is 192.168.103.0/24 on this firewall.
If this is the case, other firewalls can’t reach DNS from other networks as well and therefor can’t resolve the zones.
bbele@BB-School01:~$ nslookup firewall301.dc03.local
;; communications error to 192.168.103.254#53: timed out
;; communications error to 192.168.103.254#53: timed out
;; communications error to 192.168.103.254#53: timed out
Previously the DNS-Server entry was checked for either being a valid FQDN or IP. This worked with unbound.
Now that entry must always be an IP other wise it will show an error message saying Invalid IP Address as Knot Resolver only works with IPs for the forwarded server.
This looks just the zones are not forwarded at all but the nameservers - firewalls are reachable and are listening. I’ll try to find out where the zone caching or forwarding should be.
EDIT: I think zoning needs some more work. The /var/lib/knot-resolver/zones folder is not populated. Might be something but I’m not sure if this could be solved with the config.yaml in /etc/knot-resolver.