search

kubestellar / kubeflex

A flexible and scalable platform for running Kubernetes control plane APIs.
Apache License 2.0
49 stars 13 forks source link

bug: Incomplete Pod for KubeFlex #285

Closed milanahrae closed 1 month ago

milanahrae commented 2 months ago

Describe the bug

I am trying to get everything set for Kubernetes but I am having trouble setting up KubeFlex. I need all 3 of my pods running to continue with downloding Kubeflex and 1/3 of my pods constantly says it's running and not completed.

Steps To Reproduce

image (1) image Ngnix

Expected Behavior

My overall expected outcome is to create a Kubernetes cluster for Kubeflex and then continue the rest of the steps of the Getting Started guide on Kubestellar that requires Kubeflex. In order to do this, I need all 3 of my pods running for the Nginx Ingress Controller to be fully set up. I have tried accessing the ingress-nginx namespace to delete the pod and have a new one created, only for the same result to keep saying "running".

Additional Context

No response

pdettori commented 2 months ago

Hi @milanahrae could you please provide more details on your environment - OS, memory, CPU, docker version, kind version? At first sight, it seems that there might not be enough resources in your system.

milanahrae commented 2 months ago

Hi @milanahrae could you please provide more details on your environment - OS, memory, CPU, docker version, kind version? At first sight, it seems that there might not be enough resources in your system. memory docker This is what I have.

pdettori commented 1 month ago

@milanahrae seems like you may have enough resources, however, based on your screenshot above, it looks like the check_pre_req.sh fails for KubeFlex. Did you go over https://docs.kubestellar.io/release-0.24.0/direct/pre-reqs/ to install all pre-reqs?

It is normal to see those 2 pods (ingress-nginx-admission-create-rhh2c and ingress-nginx-admission-patch-jf42w) as completed (these are created by 2 jobs - ingress-nginx-admission-create and ingress-nginx-admission-patch). As for the 3rd one, that should stay in running state as that is a controller, and controllers in Kubernetes are long-running. It may take some time (up to 120 s) for that pod to get to 'READY' (the row for ingress-nginx-controller-xxx-xxx should show 1/1 for the READY column) - does it ever gets to that state ?

If not, could you please show the logs for that controller pod (if you restarted the install, that name may have changed)

kubectl logs -n ingress-nginx ingress-nginx-controller-dcf8d6f6d-4p4fb
milanahrae commented 1 month ago

This is a late reply but it eventually fixed itself after running the program so many times. Thank you for the help.

pdettori commented 1 month ago

@milanahrae : you are welcome. Closing the issue.