Testing The Customer Installation

A step-by-step guide of what the experience will be like installing the application

Part 3 Of A Series

This is part 3 of a guide that walks through creating a sample application in Replicated Ship. If you haven’t followed the previous sections of this guide, go back to deploying an application before following this guide.

Installing the Release

Now that we’ve created a customer and a release, we can produce an installation script for our test customer.

Get Install Script

In the modal, you’ll want to navigate to “Ship OSS”.

Get Install Script

Installing Ship

This guide assumes you’ll run Ship with Homebrew using brew install ship, or downloading the binary, but you can also run Ship with the replicated/ship docker container.

Copy the command from this screen, we’ll run it in the next step. When your end customer installs your application, you’ll give them this command to run on their workstation.

Run the command

Run the command from the modal. It should look something like:

ship init "replicated.app/superci-enterprise?license_id=MoYBbdP4sy6qomD9WdAWjAsWPsKF3YZ1"

This command will read the release we created in https://vendor.replicated.com and then prompt us to open a browser:

Reading replicated.app/superci-enterprise?license_id=MoYBbdP4sy6qomD9WdAWjAsWPsKF3YZ1 ...
Determining application type ...
Detected application type replicated.app

Please visit the following URL in your browser to continue the installation


From here, you’ll work through the lifecycle we defined previously, until you get to the “You’re all done” page:

All Done

Review Outputs

Once you reach the end of the workflow, the ship init CLI command will exit, and you can review the files creeated:

$ find .

Our inline assets from above have been created under the installer directory. There’s also a .ship/state.json which holds metadata about the installation. We’ll use this state file later in the Shipping an Update chapter of this guide.

Installing to a cluster

Executing this installer script generated everything that’s needed to deploy our application to a Kubernetes cluster. But we need to bring a Kubernetes cluster, and should already have kubectl set up and configured to continue with the next steps.

Deploy resources

As instructed in the lifeycle, we can

kubectl apply -f installer/k8s

and we’re off to the races. If you’re on Docker for Desktop, the LoadBalancer service should get your nginx listening on port 80 on localhost. If you’re on a cloud-provided cluster, you’ll need to inspect kubectl get svc to find the load balancer IP.

Next Steps

Now that we’ve shipped our first release, its time to work through Shipping an Update.