external-dns: Error with updating NodePort service in GKE
What happened: Cannot create DNS entry if NodePort service is present
What you expected to happen: Create the DNS entry
How to reproduce it (as minimally and precisely as possible): I create a GKE Kubernetes cluster with the default http load balancer, it creates the service: kube-system default-http-backend NodePort ... <none> 80:30231/TCP
In the external-dns log I find:
time=“2020-04-06T13:00:53Z” level=info msg=“Change zone: zone-prd-gcp-web-rcslan-it batch #0” time=“2020-04-06T13:00:53Z” level=info msg=“Add records: default-http-backend.kube-system.wgprxxxgke01.k8s.prd.gcp.web.rcslan.it. CNAME [. ] 300” time=“2020-04-06T13:00:53Z” level=error msg=“googleapi: Error 400: Invalid value for ‘entity.change.additions[0].rrdata[1]’: ‘’, invalid”
Note the empty CNAME “[. ]” that wants to create, that is wrong
Anything else we need to know?: As a workaround I can add to the NodePort service the entry: apiVersion: v1 kind: Service metadata: annotations: external-dns.alpha.kubernetes.io/hostname: DUMMY
So it is skipped and the DNS entries for other services are created
Environment: Google GKE 1.15.9-gke.22
- External-DNS version (use
external-dns --version): 0.7.1 - DNS provider: Google DNS
- Others:
About this issue
- Original URL
- State: closed
- Created 4 years ago
- Reactions: 3
- Comments: 15 (3 by maintainers)
I’ve run into the same problem, and notices that by default external-dns uses the nodes external IP for provisionning the DNS entry for node ports. I’ve fixed the issue by adding the
external-dns.alpha.kubernetes.io/access: privateannotation. I think there is still a bug, the google provider shouldn’t try to add an invalid entry (and maybe should provide a better error message)