Acceldata Torch uses Kubernetes for deployment and execution. This is guided by Replicated Kots which is responsible for deployment providing single click experience to end customer. This document describes how to deploy in EKS. Similar setup can be done for other cloud provider.
Creating the cluster
eksctl user can create a Amazon managed Kubernetes cluster.
Upon completion, user will be able to see the nodes in the cluster using
kubectl get nodes command.
Next step is to install Replicated
After the installation of kots, you need to install Acceldata Torch.
This command would ask for the namespace. When prompted, please provide the namespace where Torch has to be deployed. The command will also ask for the Kots Admin password. When prompted, provide the password. This password is required to log into the admin console.
At the end, the command will start a port forward and the admin console can be accessed at
http://localhost:8800 in a browser.
To start the tunnel again, execute
kubectl kots admin-console --namespace <namespace>. Replace
<namespace> with the
namespace provided during installation.
In the browser, the first screen will ask for the password.
Next, it will ask to upload the license file provided by acceldata.
Then, the configuration screen loads. The process is same for both types of installation.
After the configuring the system, user should click on the deploy button on the next screen. This would take few minutes to complete.
After few minutes, the following services should be visible.
Accessing the Torch Application
nginx-ingress-nginx-controller service shown above, the system will assign an external LoadBalancer dns.
The Torch UI can be accessed from the DNS provided for the nginx ingress.
For example: Open
http://ad87fba36d77xxxxxxdf539cdcaba49b-489574617.us-east-2.elb.amazonaws.com to start the Torch UI.