traefik ingress is stuck in progress in ArgoCD, address field is empty by default #1048
Replies: 5 comments
-
Ah and of course my proposal is to add
by default, so that ingress has always a valid address. |
Beta Was this translation helpful? Give feedback.
-
Yes, this should actually be the default. Else services of type Ingress will point to an empty loadbalancer. You can just use instead of traefik_values Or if you prefer the HelmChartConfig
|
Beta Was this translation helpful? Give feedback.
-
@mysticaltech Your opinion? |
Beta Was this translation helpful? Give feedback.
-
@HendrikLevering @Silvest89 Yes, definitely best to add the field by default. PR most welcome 🙏 |
Beta Was this translation helpful? Give feedback.
-
Beta Was this translation helpful? Give feedback.
Uh oh!
There was an error while loading. Please reload this page.
-
If you use a Hetzner cluster with traefik and you deploy an app with ingress via ArgoCD, then ArgoCD is stuck in progress. The reason is: ArgoCD will check the ingress address to field. It will only switch to healthy, if the field is not empty. In the default Hetzner / Traefik setup this field is empty.
You can make traefik populate this field by adding this to traefik_values:
additionalArguments:
This will populate the address field with the load balancer address.
Ref:
argoproj/argo-cd#14607
https://community.traefik.io/t/traefik-ingress-not-getting-loadbalancer-ip/17445
However, I tried this with an existing running cluster. I added the additional argument. The result was, that a second load balancer was created and the first one was somehow in an orphaned state. So I had to manually delete the 2nd load balancer from Hetzner, delete the traefik service and create a new one, which reuses the old load balancer. (make a backup of the existing service before deletion!)
Ref for loadbalancer reusage:
https://github.com/hetznercloud/hcloud-cloud-controller-manager/blob/main/docs/load_balancers.md#reference-existing-load-balancers
I created the new one by putting the service in a yaml file and apply it to the cluster.
I put this information here, so that others with this problem may find it.
I have one question:
Do I have to be worried that my new traefik service is overidden? I read that traefik is redeployed on k3 start. I tested this by restarting a node, but so far it seems to continue working fine
Beta Was this translation helpful? Give feedback.
All reactions