Symptom: Node node status NotReady PELGis not Heathly Kubelet failed to report the POD status (timeout: 3 minutes). The system was loaded and a zombie process was detected. Solution: Restart the server. The CNI component fails to obtain network status. Solution: Delete the CNI component on the main control node, restart pod. Failed to start cAdvisorinotify_add_watch/sys/fs/cgroup/cpuset kubepods/processing methods: Run sysctlfs.inotify. Max_user_watches =1048576 to restart Kubelet