external-dns
external-dns copied to clipboard
kubernetes-sigs
DNS not updated when load balancer changes address
What happened: Contour logs
time="2021-09-20T17:37:52Z" level=info msg="received a new address for status.loadBalancer" context=loadBalancerStatusWriter loadbalancer-address=a09eed6b7d4769a328a1dc38396843-b3258fd74268025a.elb.us-east-1.amazonaws.com
external-dns logs:
time="2021-09-21T16:51:58Z" level=info msg="All records are already up to date"
Envoy service data:
AME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE SELECTOR
envoy LoadBalancer 10.100.146.240 a09eed6b7d4769a328a1dc38396843-b3258fd74268025a.elb.us-east-1.amazonaws.com 80:31483/TCP,443:32088/TCP 23h app=envoy
What you expected to happen:
Update of the new CNAME in Route53
Essentially that's the only thing I expect of external-dns nothing else.
How to reproduce it (as minimally and precisely as possible):
Use external-dns with projectcontour on EKS
It's much much worse than that.
It always updates the address in Route53 to a wrong CNAME!
That CNAME no longer exists in the cluster, yet external-dns will insist on setting it in Route53. Making the whole cluster unreachable.
Only totally removing external-dns deployment can restore cluster functionality.
The same happens when changing ingress class on a service (new ingress class has different lbip than the last). I was changing ingress class by deleting ingress for my service and creating new one with new ingress class. I expected ip to change. Provider: powerdns with txt records and traefik to traefik2 migration. On premise kubernetes 1.19, external dns from helm chart (5.4.10)
Reproduced again, and again, and again.
It seems like external-dns just doesn't really handle load balancer address change at all.
I.e. the one and only functionality it's needed for
The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.
This bot triages issues and PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue or PR as fresh with
/remove-lifecycle stale - Mark this issue or PR as rotten with
/lifecycle rotten - Close this issue or PR with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.
This bot triages issues and PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue or PR as fresh with
/remove-lifecycle stale - Mark this issue or PR as rotten with
/lifecycle rotten - Close this issue or PR with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.
This bot triages issues and PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue or PR as fresh with
/remove-lifecycle rotten - Close this issue or PR with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle rotten
The Kubernetes project currently lacks enough contributors to adequately respond to all issues and PRs.
This bot triages issues and PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue or PR as fresh with
/remove-lifecycle stale - Mark this issue or PR as rotten with
/lifecycle rotten - Close this issue or PR with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough active contributors to adequately respond to all issues and PRs.
This bot triages issues and PRs according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue or PR as fresh with
/remove-lifecycle rotten - Close this issue or PR with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle rotten
verified that external-dns:v0.12.2 does not update route 53 records when you change ingress from one ingress class to another one which has completely different load balancer hostname. I had to delete existing route 53 records to let external-dns create it again.
@k0da added --events but did not help. Updated ingress class to new one and 30 min passed. external-dns did not update route 53 yet. external-dns pod arguments:
--log-level=info --log-format=text --interval=1m --events --source=service --source=ingress --policy=upsert-only --registry=txt --provider=aws
@ismailyenigul did you happen to resolve this issue or figured out any work around ? i'm having the same issue and was wondering if there is any work around.
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
@ismailyenigul did you happen to resolve this issue or figured out any work around ? i'm having the same issue and was wondering if there is any work around.
No solution by changing settings of external-dns but If I delete related records in route 53. External-dns created new records correctly.
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
We have the same issue with the CloudFlare provider. Is there any plans to give this a look? In other words making external-dns support updating DNS records on providers in the case that the source get its IP value changed/updated?
Thank you very much.
