kubernetes-training

• Point of contact from master node to worker node
• load/unload container to their node as instructed by scheduler
• send back report at reg. interval for nodes and containers

Main Duty

• it commands the container runtime (i,e: Docker) to start container as instructed by the scheduler.
• it monitors state of pod & container and report to api-server

Installing

• Must be installed manually on each worker
• must be tls bootstrap

Helpful Troubleshooting Command

• To query kubelet’s status

	systemctl status kubelet

• To restart:

    systemctl restart kubelet
  

• Using journalctl

  journalctl -e -u kubelet
  

• If query node’s health using kubectl isn’t returned, most likely the problem is from the kubelet
• Steps to troubleshoot:
  • ssh into the node
  • crictl ps to see what containers are up
  • journalctl and /var/log/syslog for logs

  • query all config files(i.e: ps aux

    grep should show mostly everything) : etc/kubernetes/kubelet.conf, /var/lib/kubelet/config.yaml

Resource

• Kubelet Config File
• Configure Cluster using Kubeadm