Local Kyma deployment fails with pending Pods

Symptom

Local Kyma deployment with k3d fails with one or more Pods in state Pending. Describing the Pod reveals such an error message:

Click to copy
0/2 nodes are available: 2 node(s) had taint {node.kubernetes.io/disk-pressure: }, that the pod didn't tolerate.

Cause

The underlying Docker environment ran out of resources (memory/CPU/disk). Thus, k3d marked all Kubernetes nodes with a taint disk-pressure.

Remedy

Verify the cause:

  1. Find out which Pods are pending:
    Click to copy
    kubectl --all-namespaces get pods
  2. For the pending Pods, verify which error message you get:
    Click to copy
    kubectl -n {POD_NAMESPACE} describe pod {POD_NAME}

Fix the issue:

  1. In the preferences of your Docker Desktop, adjust the Docker resource assignment.
  2. Check whether there are sufficient resources (memory/CPU/disk) on the device.
  3. Adjust the resources as needed. If you're not sure, try 8 GB for the evaluation profile.