Deploy a NeonKUBE Cluster on AWS
Prerequisites
- AWS account
- Accept the NeonKUBE EULA
- Your AWS Access Key and Secret Key
Cluster Definition
First, you'll need to save this cluster definition as a YAML file on your machine. This provisions a single node cluster in us-west-2a using a c5.2xlarge based instance (currently, the NeonKUBE default instance size for AWS).
name: my-aws-cluster
hosting:
environment: aws
aws:
accessKeyId: [ACCESS_KEY_ID]
secretAccessKey: [SECRET_ACCESS_KEY]
availabilityZone: us-west-2a
defaultEbsOptimized: true
nodes:
control-0:
role: control-plane
Save the cluster definition somewhere on your disk, being sure to substitute your own AWS credentials for:
- [ACCESS_KEY_ID]
- [SECRET_ACCESS_KEY]
We plan on eventually integrating with AWS IAM Roles so AWS authntication will be more secure.
Cluster Deployment via NeonDESKTOP
-
Right click the NeonDESKTOP icon in the taskbar > Deploy cluster
-
Use the file picker to select the cluster definiton file you saved above
-
Grab a coffee and wait for NeonDESKTOP to deploy the cluster.
Cluster Deployment via NeonCLI
Cluster deployment happens in two phases:
-
Configuring the hosting infrastructure, which involves configuring the Resource Group, VPC, starting the Instances, configuring the LoadBalancer, IP Addresses, Disks, etc.
-
Setting up the cluster. This involves configuring Linux and Kubernetes and then installing the necessary cluster components via Helm charts.
Execute the commands to deploy this cluster, passing the PATH to the cluster definition file you saved above and CLUSTERNAME to the cluster name as defined in the definition:
Prepare the cluster infrastructure
neon cluster prepare /path/to/your/cluster-definition.yaml
Setup the cluster
neon cluster setup sysadmin@my-aws-cluster
Then wait for cluster setup to complete.
These operations are currently coordinated by NeonDESKTOP or NeonCLI running on your machine, so you'll need to keep your computer running and connected to the network.
We plan on supporting cloud based deployment in the future where these operations will be managed in the cloud, rather than from your computer.