-
Notifications
You must be signed in to change notification settings - Fork 2.5k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
0.14.1 causing crash loopback related to v1.Gateway #4366
Comments
I believe my assumption is true. I can see in "gateway-api: fix wildcard matching" #4124 v1beta1 was removed for v1. |
Experiencing the same issues with GKE Gateway API |
Same issue here I was getting nuts :D For now I will downgrade a chart version |
Just in case using chart in version 6.38.0 and
seems to do a trick temporaily on GKE |
After reviewing more carefully the changes in 0.14.1, I think this is the culprit #4019
|
I did the initial implementation for Gateway API in External DNS. I wasn't the one who migrated from v1beta1 to v1, but it seemed like a reasonable change to me... I'll take a look at what would be necessary to support both v1 and v1beta1. |
@abursavich Looks like everyone upgraded to v1, does anyone know when Google plans to upgrade to v1?, maybe supporting both CRD versions wont be needed if its planned sometime soon. |
Technically (pedantically), the docs say:
If you install a newer version of the CRDs then the resources will be auto-converted by the Kubernetes API server to the new versions, but I don't know if GKE will stomp the change. There's a SIG-Network Gateway API meeting this afternoon, which I plan to join to get input on this. As an added benefit, there's usually someone from Google there that will probably care about the GKE implementation. |
The CRD Management guidelines seem to imply that if GKE rolled back the newer CRDs then it would be a bug in GKE:
|
Thank you for your help @abursavich , much appreciated. Ping me if you need any type of support. |
Still an issue in 0.14.2 |
I tested that 0.14.1 and later work with newer Gateway API version in GKE. Maybe External-DNS documentation should now be changed to reflect that Gateway API isn't experimental phase anymore since the 1.0.0 GA release and subsequent breaking changes should be labeled as such. |
same problem here using 0.14.2 on gke |
@abursavich you clarified that this is an issue in GKE, correct? Can this issue be closed? |
@candita This is not just a GKE issue, this is an issue for everyone not using GA version of Gateway API. GKE just bundles the Gateway API CRDs within the GKE versions, that's why mostly all the reported issues are manifesting there. Change that caused this is the same as removing support for betav1 from Ingress in the middle of transition period. But as said already in this thread, external-dns stated
This issue can be wiped under the rug with this documentation comment, as users should be able to upgrade the Gateway API version by themselves. But maybe the documentation should now reflect that Gateway API is GA in the documentation |
What happened:
After updating to 0.14.1, the pod goes into a crash loop with the error:
failed to sync *v1.Gateway: context deadline exceeded
This does not happen with 0.14.0.
What you expected to happen:
For the pod to run without crashing.
How to reproduce it (as minimally and precisely as possible):
Updated to 0.14.1.
Anything else we need to know?:
I believe this is due to the use of a gateway-api source. Our source configuration looks at ingress and http-route resources.
My assumption is that the changes to gateway-api in 0.14.1 are looking for v1 resources \ CRD's. That said, at least in GKE, gateway-api resources are still being deployed using v1beta1. The GKE documentation also currently references v1beta1 CRD's.
I imagine that external-dns changes to gateway-api need to support both v1beta1 as well as v1.
The text was updated successfully, but these errors were encountered: