The whole process will take ~30m. I'm assuming Linux client (Required: awscli, jq, terraform)
- Configure your aws credentials using "default" profile (~/.aws/credentials)
- I'm assuming you already have an SSH key generated in the default ~/.ssh directory.
- If not use the following, "ssh-keygen -t rsa -b 4096" or however you'd like to generate the keys.
- Clone this repo
- Be sure to adjust the domains in the following files for your environment.
- "public_domain" in ./terraform.tfvars
- "baseDomain" in ./okd/ignition/install-config.yaml
- Download latest okd "client & "install" from github https://github.com/openshift/okd/releases
- For corp account use "4.7.0-0.okd-2021-05-22-050008", newer version don't like session tokens.
- Untar both files in root of cloned repo
- Move "oc" & "kubectl" to "/usr/local/bin"
- Run "./scripts/deploy_okd.sh" (may need to insall libvirt-clients)
- Run "terraform init --upgrade"
- Run "terraform validate" #validates code
- Run "terraform plan" #validates AWS connectivity and object createion
- Run "terraform apply -auto-approve"
- "export KUBECONFIG=$PWD/ignition/auth/kubeconfig"
- Monitor process for control nodes to go active. (Time ~15m)
- oc get nodes
- oc get csr
- Once worker nodes are up you'll need to approve their csr. Wait to see
"Pending" and run the following command. This will need to be done twice.
- oc get csr -o go-template='{{range .items}}{{if not .status}}{{.metadata.name}}{{"\n"}}{{end}}{{end}}' | xargs --no-run-if-empty oc adm certificate approve
- Watch for cluster operators to deploy (Time ~30m)
- watch -n3 oc get co