kind: After some time kubernetes stops scheduling Pods all together
What happened:
Often after some time has passed the kubernetes cluster inside Kind stops scheduling pods all together, the pod doesn’t even show as Pending it simply doesn’t show, the solution commonly is to delete the kind cluster and create it again.
What you expected to happen:
If I delete a pod it should be re-scheduled (if it’s a statefulset or deployment)
Anything else we need to know?:
Also happens on macOS
Environment:
- kind version: (use
kind version): kind v0.9.0 go1.15.6 linux/amd64 - Kubernetes version: (use
kubectl version): client v1.19.4 , server v1.19.1 - Docker version: (use
docker info): Server Version: 20.10.1 - OS (e.g. from
/etc/os-release): 20.04.1 LTS (Focal Fossa)
About this issue
- Original URL
- State: closed
- Created 4 years ago
- Reactions: 2
- Comments: 21 (15 by maintainers)
/close https://github.com/kubernetes-sigs/kind/commit/3dbeb894e3092a336ab4278d3823e73a1d66aff7#diff-02aa2d22d0befe557afb283f8925c8a006bb5f0318dc33a10176fc78299e87d6R109
it will be available in 0.11.0
we can assume that though, since we control the podspec, and don’t support kindnetd otherwise. Let’s drop that check or just downgrade it to a warning log, and ensure we use the valid source of truth.