Aller au contenu

KubeDaemonSetRolloutStuck#

Troubleshooting#

  • Identify which pod are not running :
    • kubectl -n <namespace> get pod may give you the failing pods quickly
    • or find them with a precise investigation : kubectl get daemonset <failing daemonset> and kubectl get daemonset <failing daemonset> -o json | jq '.spec.selector' then kubectl get pod -l <the labels of the selector>
  • Check why the failing pods are not scheduled or not ready with common techniques (kubectl describe node...)
  • Because a daemonset is specific to a node, check the node too :
    • kubectl describe node <the node where the pod is running>
    • Cpu, Memory, Network metrics of the node; Kubernetes logs and system logs of the node.
    • If you decide to reboot the node to solve the problem, think twice and follow the procedure to reboot the node