inefficient corefile

[miekg]

  .:53 {
        errors
        health
        kubernetes CLUSTER_DOMAIN REVERSE_CIDRS {
          pods insecure
          upstream
          fallthrough in-addr.arpa ip6.arpa
        }FEDERATIONS
        prometheus :9153
        forward . UPSTREAMNAMESERVER
        cache 30
        loop
        reload
        loadbalance
    }STUBDOMAINS

is ineffcient and should be split up into multiple servers.

r : Define a reverse zone for the given CIDR. You may specifcy this option more
         than once to add multiple reverse zones. If no reverse CIDRs are defined,
         then the default is to handle all reverse zones (i.e. in-addr.arpa and ip6.arpa)

will kill any reverse look ups.

See coredns/coredns#2593 (comment) for a better one

[chrisohaver]

We have no way of knowing what the pod cidr space is. We cannot determine this programmatically from the k8s api.

We can ask the user. But they don't typically know this. Hence the current solution.

[miekg]

.. with the tradeoff their performance is very bad. Sorry, this is a not a good generic solution.

[added]
If it's not possible to provide a generic solution, this directory should be deleted as its bad.

[miekg]

This is still true. Probably makes a lot of sense to remove fall through from kubernetes, as it only there to be abused

…

On Wed, 10 Apr 2019, 14:06 chrisohaver, ***@***.***> wrote: Closed #131 <#131>. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <#131 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAVkWwEDM_I1ow0g2N9o0dKt6e9XcXerks5vfeHdgaJpZM4bPL4c> .

[chrisohaver]

see duplicate issue and related discussion: coredns/coredns#2723