Skip to content

Latest commit

 

History

History
60 lines (38 loc) · 1.8 KB

README.md

File metadata and controls

60 lines (38 loc) · 1.8 KB
Raw

provider-build-scripts

Example cluster installation

172.18.* is the internal network

  1. Initialize the control-plane node

provider.h100.sdg.val.akash.pub is just the example

./k3sAndProviderServices.sh -d traefik -s provider.h100.sdg.val.akash.pub -g -n 172.18.

If the cluster has a dedicated ephemeral storage, specify the location using -o and -k.

Example

./k3sAndProviderServices.sh -d traefik -s provider.h100.sdg.val.akash.pub -g -n 172.18. -o /data/containerd -k /data/kubelet

IMPORTANT: Note down the line K3s control-plane and worker node token: as it'll contain the token you'll need to join further nodes. If you forget to save this info, you can always get it by running cat /var/lib/rancher/k3s/server/node-token command.

  1. Join the 2nd and 3rd control-pane nodes

IMPORTANT: Make sure to have at least 3 control-plane nodes. If you only have two, then it means you only have two etcd (embedded into k3s) Kubernetes databases. When less than half are down, etcd won't work.

172.18.140.11 is the internal IP of the first control-plane node

echo -n "Enter K3s control-plane and worker node token: "
read -s TOKEN
echo
./k3sAndProviderServices.sh -s provider.h100.sdg.val.akash.pub -e $(curl -s ident.me) -m 172.18.140.11 -c $TOKEN -g -n 172.18.

If the cluster has a dedicated ephemeral storage, specify the location using -o and -k.

Example

./k3sAndProviderServices.sh -s provider.h100.sdg.val.akash.pub -m 172.18.140.11 -c $TOKEN -g -n 172.18. -o /data/containerd -k /data/kubelet
  1. Join the worker nodes
./workerNode.sh -m 172.18.140.11 -t ${TOKEN} -g

If the cluster has a dedicated ephemeral storage, specify the location using -o and -k.

Example

./workerNode.sh -m 172.18.140.11 -t ${TOKEN} -g -o /data/containerd -k /data/kubelet