What happened:
The image in the metrics server deployment here is not getting pulled. I found this 5-6 hours ago.
What you expected to happen:
Image k8s.gcr.io/metrics-server:v0.3.7 should get pulled
Anything else we need to know?:
Command: docker pull k8s.gcr.io/metrics-server:v0.3.7
Output: Error response from daemon: manifest for k8s.gcr.io/metrics-server:v0.3.7 not found
docker Version: 18.09.7
arch: x86_64 GNU/Linux
Environment:
kubectl version):
/kind bug
Please use manifests that are attached to release.
You can download them by running
kubectl apply -f https://github.com/kubernetes-sigs/metrics-server/releases/download/v0.3.6/components.yaml
Thanks @serathius , is this temporary or should I use it for later as well?
This is long term. Rationale: https://github.com/kubernetes-sigs/metrics-server/issues/471
This will ensure that manifests will change only with release, leading to higher quality and stability.
Note: I think we can make this change more noticeable by moving "deploy" directory.
/cc @loburm
WDYT?
Should there be a warning now between the apply and paragraph that states which options to change in the metrics container? Not a big deal, but an abrupt change none the less.
@ronaldpetty PTAL https://github.com/kubernetes-sigs/metrics-server/pull/509
Thanks for feedback. We are preparing first release in long time. So we can definitely improve our process.
Thanks @serathius didn't see that.
Because I created it after your suggestion :P
Closing as https://github.com/kubernetes-sigs/metrics-server/pull/509 was merged making it obvious that manifests from "deploy" directory should not be used
Most helpful comment
This is long term. Rationale: https://github.com/kubernetes-sigs/metrics-server/issues/471
This will ensure that manifests will change only with release, leading to higher quality and stability.
Note: I think we can make this change more noticeable by moving "deploy" directory.
/cc @loburm
WDYT?