search

kubernetes-sigs / external-dns

Configure external DNS servers (AWS Route53, Google CloudDNS and others) for Kubernetes Ingresses and Services
Apache License 2.0
7.69k stars 2.56k forks source link

AzureDNS, records not removed when DNSEndpoint is deleted #2080

Closed Justin-DynamicD closed 3 years ago

Justin-DynamicD commented 3 years ago

What happened:

using a helmchart, deployed 0.8.0 and configured the following key values:

registry: txt
txtOwnerId: myCluster
txtSuffix: _autoDNS

I then created a DNSEndpoint resource with a simple spec:

Spec:
  Endpoints:
    Dns Name:     coolapp.example.com
    Record TTL:   180
    Record Type:  CNAME
    Targets:
      randomaws-alb-endpoint.com

Logs show both records create successfully (CNAME and TXT).

Then, I delete the DNSEndpoint:

What you expected to happen:

Both records to get deleted. Instead, nothing. Not even a log implying it tried.

Deleting the DNSEndpoint doesn't appear to trigger any kind of delete opperation. poilling is set to defautl value of 1min ... nothing.

Is there some kind of setting I should be using I'm missing?

bitnami chart I'm using: https://github.com/bitnami/charts/tree/master/bitnami/external-dns

Justin-DynamicD commented 3 years ago

Update:

changing the policy to "sync" from "upsert-only" seems to have fixed the behavior. I'm not certain what this value actually controls (not seeing it referenced in the readme), but I guess I can go through the Go code and figure it out.

Assuming this is expected behavior.