Expand your Orion cluster¶
Cloud¶
If you are using a cloud deployment, we recommend you install an Autoscaler for your cluster. With an autoscaler, whenever you request a workstation, Kubernetes will check for available compute resources.
If your current cluster cannot run the workstation you had requested, it will automatically bring online another node. Once you configure your autoscaler, it will also bring down any unused nodes, making sure you are not being charged for them and are utilizing only the necessary resources.
For AWS deployments, we highly recommend:
On-Premise (k3s)¶
For k3s-based clusters, we offer you the ability to expand your cluster using our playbooks. We assume that the cluster settings match our QuickStart deployment, meaning that:
- Flannel is the CNI and the wireguard backend is enabled
- by default, k3s uses the less secure unencrypted backend. Our platform currently forces controlplane nodes are set up with encryption enabled
Offline/air-gapped considerations¶
For offline installations, make sure that:
- the
juno-fx/ansible-ee:latest
image can be pulled from your nodes via the configuration in /etc/rancher/k3s/registries.yaml - this configuration is set up automatically when using our install helper script found in the Quick Start Guide
- if you adjust it, please make sure to do so across all your control plane nodes
- when expanding the cluster, the file will be copied from your control plane node to any newly added nodes
- when you have multiple nodes, this will be performed on one of them. That is why it's important this file is consistent across the board.
How-to¶
If you have installed Orion using our Oneclick Installer on a single node, you can expand your cluster by adding more nodes at any time. You can do this also with an existing k3s cluster, as long as it meets our System Requirements.
- Go to the "Networking" tab in Genesis
- Select "Credentials" from the top navbar.
-
Create a credential for node provisioning.
-
Go to the "Provisioning Nodes", open the node creation form and submit a provisioning job.
Fill in the hostname/IP you want to provision, select the credential you'd like to use and the node roles for it.
You can review the status and logs of the node provisioning in the table below - each provisioning job is listed and logs for it are accessible. The jobs are cleaned up automatically after 24h.
Once the node is up, you can see it in the nodes tab: