updated readme

README.md

@@ -21,3 +21,56 @@ Then the tarball can be loaded into the docker registry, on the control plane no
 Verify the image is in fact the correct architecture:
 
 `docker image inspect <tag_name>`
+
+## Cluster Setup
+
+After the images are built, the next step is likely to bring up a cluster. The `clusterSetup.sh` script does this, with some assumptions about your application. The high level steps are as follows:
+
+- turn off swap: `swapoff -a`
+
+- start cluster, config is for pod subnet (may take time): `kubeadm init --config kubeadm-config.yml`
+
+- move config over:
+`mkdir -p $HOME/.kube
+sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
+sudo chown $(id -u):$(id -g) $HOME/.kube/config`
+
+- apply flannel networking: `kubectl apply -f flannel.yml`
+
+- remove taint from control plane: `kubectl taint nodes --all node-role.kubernetes.io/master-`
+
+- label node(s) if deployment calls for it (ones in this repo do via `nodeSelector: <tag>=<value>` in the deployment): `kubectl label nodes <node_name> <tag=<value>`
+
+- start applying deployments starting with master: `kubectl apply -f kubernetes/master-deployment.yml`
+
+- continue applying deployments for as many as needed
+
+- turn off cluster: `kubeadm reset` potentially followed by [clean-up][cleanupCluster]
+
+[cleanupCluster]: [https://kubernetes.io/docs/reference/setup-tools/kubeadm/kubeadm-reset/]
+
+## Handy Troubleshooting Commands
+
+There are a number of non-intuitive kubernetes commands that are super helpful for specific tasks:
+
+- drain a particular node so that it can be removed: `kubectl drain <node name> --delete-emptydir-data --force --ignore-daemonsets` then `kubectl delete node <node_name>`
+
+- get bash terminal inside pod: `kubectl exec -it <pod_name> -- bash`
+
+- get which pods are on which nodes (a travesty this one): `kubectl get pod -o=custom-columns=NAME:.metadata.name,STATUS:.status.phase,NODE:.spec.nodeName --all-namespaces`
+
+- another option to see pod location: `kubectl get pods --all-namespaces -o wide`
+
+- regenerate token join command: `kubeadm token create --print-join-command`
+
+## Links With More Information
+
+Debugging clusters can be somewhat difficult as everything is stuffed into pods, the links below have some steps to try:
+
+- if `CrashLoopBackoff` is thrown when running `kubectl get pods`, take a look [here][crashloop]
+- general steps for building images with buildx [here][docker-buildx]
+- there's some specific steps that got ARM64 builds working in my case, solution [here][docker-buildx-platform]
+
+[crashloop]: https://managedkube.com/kubernetes/pod/failure/crashloopbackoff/k8sbot/troubleshooting/2019/02/12/pod-failure-crashloopbackoff.html
+[docker-buildx]: https://collabnix.com/building-arm-based-docker-images-on-docker-desktop-made-possible-using-buildx/
+[docker-buildx-platform]: https://github.com/docker/buildx/issues/464

clusterStartup.sh

@@ -2,4 +2,11 @@
 # set -e
 
 # script that starts up kubernetes cluster, and does some basic qol
+
+swapoff -a # turn off swap memory
+kubeadm init --config kubeadm-config.yml
+
+mkdir -p ~/.kube
+sudo cp -i /etc/kubernetes/admin.conf $HOME/.kube/config
+sudo chown $(id -u):$(id -g) $HOME/.kube/config
 #tbd