Cluster information:
Kubernetes version: micro-k8s 1.24
Cloud being used: (put bare-metal if not on a public cloud) vultr
Installation method: snap install microk8s
Host OS: ubuntu22.04
CNI and version:micro-k8s 1.24
CRI and version: micro-k8s 1.24
TLDR: How to bypass the CPU resource limit or allow overcommit?
I create an vultr 1core-8G instance for test kubesphere 3.3-RC. Then I ran into insufficient CPU issue. Clearly the requested CPU resource larger then node capacity.
So how to bypass CPU commit limit?
The only argument I found about CPU is cpu-cfs-quota but that’s no related with this question. And the server CPU usage is around 0.5% because only minio pod is running and even hostpath-provider is pending.
Because of all manifest wrote for real propose using, Not for testing so CPU request will be production level reasonable. So I can’t controll it: If I using raw manifest is find,I can just modify then, But for example kubesphere its define an installer and cdr-config. I can’t even see any yaml file, So don’t even mention modify them.