Dashboard pod wont go into ready state
Has error CreateContainerConfigError in status
It doesnt happen
Install 0.7.0 release on docker
Kubernetes Platform:
docker
Kubernetes or OpenShift version:
Versions:
Tekton Dashboard:
0.7.0
Tekton Pipelines:
Dashbaord pod status
tekton-dashboard-95969f65b-fl9h5 0/1 CreateContainerConfigError 0 5m32s
Output from describe dashboard pod
Events:
Type Reason Age From Message
---- ------ ---- ---- -------
Normal Scheduled 84s default-scheduler Successfully assigned tekton-pipelines/tekton-dashboard-95969f65b-fl9h5 to docker-desktop
Normal Pulling 83s kubelet, docker-desktop Pulling image "gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:1a44d51696970ec16981eeada03c45e73b202d87a680c0a93f6fc3e27f8798b4"
Normal Pulled 73s kubelet, docker-desktop Successfully pulled image "gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:1a44d51696970ec16981eeada03c45e73b202d87a680c0a93f6fc3e27f8798b4"
Warning Failed 10s (x7 over 73s) kubelet, docker-desktop Error: container has runAsNonRoot and image will run as root
Normal Pulled 10s (x6 over 72s) kubelet, docker-desktop Container image "gcr.io/tekton-releases/github.com/tektoncd/dashboard/cmd/dashboard@sha256:1a44d51696970ec16981eeada03c45e73b202d87a680c0a93f6fc3e27f8798b4" already present on machine
Worth noting this is using the latest staged release candidate at https://console.cloud.google.com/storage/browser/tekton-releases/dashboard/koo/mytestbucket/previous/v0.7.0
Released images are built with ko right ?
Using the config in the repo below ?
https://github.com/tektoncd/dashboard/blob/c5e6a8d35b74ee608434c5d0f206f110240da71a/.ko.yaml#L1
Released images are built with
koright ?Using the config in the repo below ?
https://github.com/tektoncd/dashboard/blob/c5e6a8d35b74ee608434c5d0f206f110240da71a/.ko.yaml#L1
Yeah should be
@eddycharly do you think we need a
securityContext:
--
runAsUser: 1000
type stanza somewhere now? The now part is interesting though, I know you mention that image above and wondering if that has a new restriction (so we could fix with a non-root user, which we should do anyway IMO, or use an older version of that image if it is the problem)
We shouldn't need it at all as we're using a nonroot base image.
I will dive in the released image to see what happens.
I feel like it's an issue when building the release image.
kustomize build overlays/dev | ko apply -f - runs fine locally.
Do you see this line 2020/06/11 14:14:14 Using base gcr.io/distroless/base:nonroot for github.com/tektoncd/dashboard/cmd/dashboardin the release logs ?
I feel like it's an issue when building the release image.
kustomize build overlays/dev | ko apply -f -runs fine locally.Do you see this line
2020/06/11 14:14:14 Using base gcr.io/distroless/base:nonroot for github.com/tektoncd/dashboard/cmd/dashboardin the release logs ?
No, attaching logs here
releaserunbuild.txt
releaserunpublish.txt
I see
Using base gcr.io/distroless/static:latest for github.com/tektoncd/dashboard/cmd/dashboard
That's it, it's not a nonroot image
Do you know where is the dockerfile for the gcr.io/tekton-releases/dogfooding/ko-gcloud:latest image ?
Right, so it's like that file addition is doing nothing - any pointers @afrittoli @vdemeester?
^^
It's in the plumbing repo. I'm on my phone now, can give you a link later if you cannot find it.
From the logs, it's ko v0.5.0 so that's fine.
Dunno why but either .ko.yaml is not there or ko ignores it.
Got it 馃槃
It's not going to copy hidden files (those that begin with a .) !
Ping @a-roberts
Thank you @afrittoli we found the issue, in the end it was not related to the image.