I just observed something that I believe is an instance of this bug in production, though it's a bit unclear to me exactly what happened.

We have the following config, running CoreDNS 1.6.6:

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

The problem I observed was after the CoreDNS pods had been running for a while, suddenly some pods started getting NXDOMAIN for a forwarded DNS response. Pushing a new config flushed the cache and resolved it, at least for a while.

Oddly enough, it only affected some pods, other pods that tried to resolve the same domain didn't have the same issue. It also continued to affect the pod even after we deleted it and it got rescheduled on a different node.

We've disabled the autopath plugin for now, so I'm not sure I'll be able to reproduce it again, unfortunately.

Description (*)

We are no longer using this provider, and there's an official provider in Terraform instead that can be used.

Contribution checklist (*)

• [x] I have read the Contributing Guidelines
• [x] Commits have been made with meaningful commit messages
• [N/A] All automated tests have passed successfully

Description (*)

We are no longer using this provider, and there's an official provider in Terraform instead that can be used.

Contribution checklist (*)

• [x] I have read the Contributing Guidelines
• [x] Commits have been made with meaningful commit messages
• [N/A] All automated tests have passed successfully