kubelet breaks CRI standards by running containers in 'k8s.io' containerd namespace.
Kubernetes pods run in the 'k8s.io' namespace of containerd. This 'runtime namespace' concept is containerd specific and not a part of CRI (at least I could not find it).
If kubelet truly follows the CRI standard, then it must not be able to run containers in such containerd namespaces.
crictl does not support this 'runtime namespace' thing and therefore can only communicate to containerd in the default namespace -- kubelet must do the same.
The Kubernetes project currently lacks enough contributors to adequately respond to all issues.
This bot triages un-triaged issues according to the following rules:
- After 90d of inactivity,
lifecycle/staleis applied - After 30d of inactivity since
lifecycle/stalewas applied,lifecycle/rottenis applied - After 30d of inactivity since
lifecycle/rottenwas applied, the issue is closed
You can:
- Mark this issue as fresh with
/remove-lifecycle stale - Close this issue with
/close - Offer to help out with Issue Triage
Please send feedback to sig-contributor-experience at kubernetes/community.
/lifecycle stale
/remove-lifecycle stale