Also @creamy-pencil-82913 I have a deployment.yaml controlling the pod that runs my container. If the node dies where my pods are running, k3s will eventually restart them, but it doesn’t take some time. Do you have any suggestions or settings in my deployment.yaml so that my pods get spun back up faster after they are restarted into a new node? Or any docs or pro tips you can point me towards. Thanks

you might look at https://github.com/k3s-io/k3s/issues/1264

Hey Brandon, I hope all is well. I read the thread, and so in each of my server nodes, I updated each of their config.yamls to include: # /etc/rancher/k3s/config.yaml  (server nodes only)   kubelet-arg: ◦ node-status-update-frequency=4s   kube-controller-manager-arg: ◦ node-monitor-period=4s ◦ node-monitor-grace-period=16s ◦ pod-eviction-timeout=20s   kube-apiserver-arg: ◦ default-not-ready-toleration-seconds=20 ◦ default-unreachable-toleration-seconds=20 But after adding all of these flags into each of my server nodes, and then running “systemctl daemon-reload and then systemctl restart k3s” Now none of my nodes are responding when I run “kubectl get nodes” What did I do wrong here? Do I literally have to re-run the install script on each of my air-gapped setup nodes after updating their config.yamls ?

Sounds like you broke something and k3s isn't running?

All of my server nodes have this exact updated config.yaml with the exact syntax for the flags

Maybe start a new discussion thread?