forward . /etc/resolv.conf not work sometimes when use local dns

[wwgfhf]

I use coredns 1.5.0 in k8s 1.14.3 , the config of coredns is

.:53 {
        errors
        health
        ready
        kubernetes cluster.local in-addr.arpa ip6.arpa {
          pods insecure
          upstream /etc/resolv.conf
          fallthrough in-addr.arpa ip6.arpa
        }
        prometheus :9153
        forward . /etc/resolv.conf {
          prefer_udp
        }
        cache 30
        loop
        reload
        loadbalance
    }

i write my local dns in /etc/resolv.conf, but it not work sometimes in pod.
But when i change the /etc/resolv.conf to local dns address

forward . 10.110.18.115

it worked always

[chrisohaver]

Can you show us what is in /etc/resolv.conf?

Does it work more reliably if you remove prefer_udp? #3110

[wwgfhf]

@chrisohaver Thanks for your reply, this is my /etc/resolv.conf

nameserver 10.110.18.115
nameserver 114.114.114.114

And it seeams same as before when remove prefer_udp.

[chrisohaver]

You have two nameservers in /etc/resolv.conf, which is why you see the behavior difference when you specify just one of them. "114.114.114.114" presumably not working.

By default the forward plugin uses the "random" policy, see https://github.com/coredns/coredns/blob/master/plugin/forward/README.md. You can use "sequential" instead so the "10.110.18.115" server is always tried first. e.g.

        forward . /etc/resolv.conf {
          policy sequential
          prefer_udp
        }