Question

A user reports that a newly created Pod in the 'batch' namespace has been stuck in the 'ContainerCreating' state for several minutes. Running 'kubectl describe pod' shows the events: 'Successfully assigned batch/job-runner to worker-3' followed by repeated warnings 'FailedCreatePodSandBox ... failed to set up sandbox container ... plugin type="bridge" failed (add): failed to allocate for range 0: no IP addresses available in range set'. Other Pods scheduled to worker-3 show similar errors. What is the most likely root cause?

Accepted Answer

The CNI IP address pool (podCIDR range) allocated to worker-3 is exhausted, so the CNI plugin cannot assign an IP to new Pods.

Answer

The container image is too large and the kubelet is timing out while pulling it from the registry.

Answer

The Pod's resource requests exceed the allocatable CPU and memory on worker-3.

Answer

The CoreDNS service is down, preventing the new Pod from resolving the registry hostname.