Deploying HTC-Grid

We are now ready to deploy HTC-Grid using the terraform stack provided by the project. The terraform stack uses ~/environment/aws-htc-grid/generated/grid_config.json to drive the configuration of the deployment.

Before we deploy the project, we need to initialize the terraform state. Remember we will be using the $S3_TFSTATE_HTCGRID_BUCKET_NAME bucket to hold the state. You can read more about the S3 terraform backend here.

make init-grid-deployment TAG=$TAG REGION=$HTCGRID_REGION

All the dependencies have been created and are now ready. We are now ready to deploy the HTC-Grid project. There is one last thing to note. HTC-Grid deploys Grafana running inside EKS which is authenticated using Amazon Cognito. While you can modify and select which passwords to use in Amazon Cognito, the Grafana internal deployment still requires an admin password. Select a memorable password and use the value instead of the placeholder <my_grafana_admin_password> below (make sure this password follows the Cognito Default Policy):

make apply-custom-runtime TAG=$TAG REGION=$HTCGRID_REGION GRAFANA_ADMIN_PASSWORD='<my_grafana_admin_password>'

The execution of this command will prompt for yes to continue. Just type yes and click Enter for the command to proceed. The installation should finish in about 20 Mins.

Validating HTC-Grid Deployment

If terraform apply is successful then in the terraform folder two files are created:

  • kubeconfig_htc_$TAG: this file will give you access to the EKS cluster through kubectl
  • agent_config.json: this file contains all the parameters, so the agent can run in the infrastructure

Let’s validate that the Compute Plane has been set up accordingly. First of all, we need to configure our environment with Kubectl configuration pointing to our cluster. We will also read from the terraform output the Agent config file and prepare our environment to select the configuration on newly created terminals.

cd ~/environment/aws-htc-grid/deployment/grid/terraform
export AGENT_CONFIG_FILE=$(terraform output -raw agent_config)
echo "export AGENT_CONFIG_FILE=$AGENT_CONFIG_FILE" >> ~/.bashrc

With this done, we can get the number of nodes in the cluster using the following command:

kubectl get nodes

You should have one or more nodes. If not please review the configuration files and particularly the variable eks_worker_groups

To check that the HTC-Agent is running on the system, you can run the following command:

kubectl get pods

Finally to check all the deployments are running as expected,

kubectl get pods --all-namespaces

You should have all your pods in a running state, this might take a couple of minutes.

Site Info