[Question] Cannot completion installation in K8s provisioned by Rancher rke

[scheung38]

I have 3 Droplets in Digital Ocean, each of
Ubuntu 20 2 GB Memory / 50 GB

Describe the bug*

To Reproduce
Steps to reproduce the behavior:

1. rke up
2. kubectl apply -f strimzi/sample.yaml -n kafka

➜ ~ kubectl get pods -n kafka -w

Expected behavior
A clear and concise description of what you expected to happen.
Completion

Environment (please complete the following information):

• Strimzi version:

• Client Version: v1.20.1

• Server Version: v1.19.4

• Infrastructure: Rancher rke up

Is this related to sufficient node resources?

[scholzj]

It is a bit weird when you open both issue and talk about it on Slack. What is your preferred channel?

[scholzj]

If the pod is pending, one of the usual reasons for it is missing storage. From your configuration, the storage needs to be created by the default storage class - do you know if you have one? It could be also no capacity in your cluster etc. Can you run the following commands and share the output? That should help us find out more.

• kubectl get pv
• kubectl get pvc
• kubectl get storageclass
• kubectl describe pod my-cluster-zookeeper-0
• kubectl describe statefulset my-cluster-zookeeper

On your screenshot, you seem to have quite a lot of pods which seem to be in the same state, but judging by the name they do not belong to Strimzi.

[scholzj]

PS: 2GB RAM / 1 cpu seems to be very small to hold the worker node including Kafka. But that should not cause the pods pending. So that might be an issue you run into later.