Configuring and Provisioning Containers

Now that we’ve created our images, we can provision our application with a network and containers. First, we’re going to add configuration to our Pulumi program. Pulumi is a tool to configure your infrastructure, and that includes being able to configure the different stacks with different values. As a result, it makes sense to include the basic configurations as variables at the top of your program.

Configure the application

Add the following configuration variables to your Pulumi program below the imports:

config = pulumi.Config()
frontend_port = config.require_int("frontend_port")
backend_port = config.require_int("backend_port")
mongo_port = config.require_int("mongo_port")

Your Pulumi program should now match this code:

import os
import pulumi
import pulumi_docker as docker

# get configuration
config = pulumi.Config()
frontend_port = config.require_int("frontend_port")
backend_port = config.require_int("backend_port")
mongo_port = config.require_int("mongo_port")

stack = pulumi.get_stack()

# build our backend image!
backend_image_name = "backend"
backend = docker.Image("backend",
                        build=docker.DockerBuild(context=f"{os.getcwd()}/app/backend"),
                        image_name=f"{backend_image_name}:{stack}",
                        skip_push=True
                        )

# build our frontend image!
frontend_image_name = "frontend"
frontend = docker.Image("frontend",
                        build=docker.DockerBuild(context=f"{os.getcwd()}/app/frontend"),
                        image_name=f"{frontend_image_name}:{stack}",
                        skip_push=True
                        )

# build our mongodb image!
mongo_image = docker.RemoteImage("mongo", name="mongo:bionic")

Try and run your pulumi up again at this point. You should get an error like this:

Diagnostics:
  pulumi:pulumi:Stack (my-first-app-dev):
    error: Missing required configuration variable 'my-first-app:frontend_port'
        please set a value using the command `pulumi config set my-first-app:frontend_port <value>`
    error: an unhandled error occurred: Program exited with non-zero exit code: 1

This is because we have specified that this config option is required. Remember how we can use the same program to define multiple stacks? Let’s set the ports for this stack, which the Pulumi command line knows already from when you first initialized the project (it’s the dev stack by default):

pulumi config set frontend_port 3001
pulumi config set backend_port 3000
pulumi config set mongo_port 27017

This set of commands creates a file in your directory called Pulumi.dev.yaml to store the configuration for this stack.

Now, try and rerun your Pulumi program.

Your Pulumi program should now run, but you’re not actually using these newly configured ports just yet! That’s because we don’t have any container resources that use the ports; we only have image resources.

Create a Container resource

In the last topic, we built Docker images. Now we want to create Docker containers and pass our configuration to them. Our containers will need to connect to each other, so we will need to create a Network, which is another resource. Add the following code at the bottom of your program:

# create a network!
network = docker.Network("network", name=f"services-{stack}")

Define a new Container resource in your Pulumi program below the Network resource, like this:

# create the backend container!
backend_container = docker.Container("backend_container",
                        name=f"backend-{stack}",
                        image=backend.base_image_name,
                        ports=[docker.ContainerPortArgs(
                            internal=backend_port,
                            external=backend_port)],
                        envs=[
                            f"DATABASE_HOST={mongo_host}",
                            f"DATABASE_NAME={database}",
                            f"NODE_ENV={node_environment}"
                        ],
                        networks_advanced=[docker.ContainerNetworksAdvancedArgs(
                            name=network.name
                        )],
                        opts=pulumi.ResourceOptions(depends_on=[mongo_container])
                        )

It is important to note something here. In the Container resource, we are referencing baseImageName from the Image resource. Pulumi now knows there is a dependency between these two resources and will know to create the Container resource after the Image resource. Another dependency to note is that the backend_container depends on the mongo_container. If we tried to run pulumi up without the mongo_container running or present somewhere in state, Pulumi would let us know that the resource didn’t exist and would stop.

The backend container also requires environment variables to connect to the mongo container and set the node environment for Express.js. These are set in ./app/backend/src/.env. Like before we can set them using pulumi config on the command line:

pulumi config set mongo_host mongodb://mongo:27017
pulumi config set database cart
pulumi config set node_environment development

Then, we need to add them to the top of our program with the rest of the configuration variables.

mongo_host = config.require("mongo_host") # Note that strings are the default, so it's not `config.require_str`, just `config.require`.
database = config.require("database")
node_environment = config.require("node_environment")

We also need to create Container resources for the frontend and mongo containers. Put the mongo_container declaration just above the backend_container one, and the frontend_container declaration at the end of the file. Here’s the code for the mongo container:

# create the mongo container!
mongo_container = docker.Container("mongo_container",
                        image=mongo_image.latest,
                        name=f"mongo-{stack}",
                        ports=[docker.ContainerPortArgs(
                          internal=mongo_port,
                          external=mongo_port
                        )],
                        networks_advanced=[docker.ContainerNetworksAdvancedArgs(
                            name=network.name,
                            aliases=["mongo"]
                        )]
                        )

And the code for the frontend container:

# create the frontend container!
frontend_container = docker.Container("frontend_container",
                                      image=frontend.base_image_name,
                                      name=f"frontend-{stack}",
                                      ports=[docker.ContainerPortArgs(
                                          internal=frontend_port,
                                          external=frontend_port
                                      )],
                                      envs=[
                                          f"LISTEN_PORT={frontend_port}",
                                          f"HTTP_PROXY=backend-{stack}:{backend_port}"
                                      ],
                                      networks_advanced=[docker.ContainerNetworksAdvancedArgs(
                                          name=network.name
                                      )]
                                      )

Let’s see what the whole program looks like next.

Put it all together

Now that we know how to create a container we can complete our program.

import os
import pulumi
import pulumi_docker as docker

stack = pulumi.get_stack()
config = pulumi.Config()

frontend_port = config.require_int("frontend_port")
backend_port = config.require_int("backend_port")
mongo_port = config.require_int("mongo_port")
mongo_host = config.require("mongo_host")
database = config.require("database")
node_environment = config.require("node_environment")

backend_image_name = "backend"
backend = docker.Image("backend",
                       build=docker.DockerBuild(context=f"{os.getcwd()}/app/backend"),
                       image_name=f"{backend_image_name}:{stack}",
                       skip_push=True
                       )

frontend_image_name = "frontend"
frontend = docker.Image("frontend",
                        build=docker.DockerBuild(context=f"{os.getcwd()}/app/frontend"),
                        image_name=f"{frontend_image_name}:{stack}",
                        skip_push=True
                        )

mongo_image = docker.RemoteImage("mongo", name="mongo:bionic")

network = docker.Network("network", name=f"services-{stack}")

mongo_container = docker.Container("mongo_container",
                                   image=mongo_image.latest,
                                   name=f"mongo-{stack}",
                                   ports=[docker.ContainerPortArgs(
                                       internal=mongo_port,
                                       external=mongo_port
                                   )],
                                   networks_advanced=[docker.ContainerNetworksAdvancedArgs(
                                       name=network.name,
                                       aliases=["mongo"]
                                   )]
                                   )

backend_container = docker.Container("backend_container",
                                     image=backend.base_image_name,
                                     name=f"backend-{stack}",
                                     ports=[docker.ContainerPortArgs(
                                         internal=backend_port,
                                         external=backend_port
                                     )],
                                     envs=[
                                         f"DATABASE_HOST={mongo_host}",
                                         f"DATABASE_NAME={database}",
                                         f"NODE_ENV={node_environment}"
                                     ],
                                     networks_advanced=[docker.ContainerNetworksAdvancedArgs(
                                         name=network.name
                                     )],
                                     opts=pulumi.ResourceOptions(depends_on=[mongo_container])
                                     )

frontend_container = docker.Container("frontend_container",
                                      image=frontend.base_image_name,
                                      name=f"frontend-{stack}",
                                      ports=[docker.ContainerPortArgs(
                                          internal=frontend_port,
                                          external=frontend_port
                                      )],
                                      envs=[
                                          f"LISTEN_PORT={frontend_port}",
                                          f"PROXY_PROTOCOL=http://",
                                          f"HTTP_PROXY=backend-{stack}:{backend_port}"
                                      ],
                                      networks_advanced=[docker.ContainerNetworksAdvancedArgs(
                                          name=network.name
                                      )]
                                      )

With Docker networking, we can use image names to refer to a container. In our example, the React frontend client sends requests to the Express backend client. The URL to the backend is set via the setupProxy.js file in the app/frontend/src directory with the HTTP_PROXY environment variable.

Run pulumi up to get the application running. However, the store is empty, and we need to add products to the database.

Populate the database

Now we can populate MongoDB and set up our Pulumi program to autopopulate the next time we deploy. First, copy the products.json file into the same directory as your index.js index.ts __main__.py main.go MyStack.cs Program.fs MyStack.vb file.

cp app/data/products.json .

Then, we’ll mount the file to an ephemeral seed container, and then use mongoimport to transfer that data into the database. The ephemeral container imports the data we need into the initialized database, all in a container that does work and shuts down and removes itself after the import is completed. This kind of workflow with temporary containers is fairly common when updating databases intermittently with data pulled from another system. Generally, however, you would typically wire the database to an API and update it that way. This example demonstrates, however, that you can do this workflow with Pulumi, just like you could with Docker.

Add this snippet after the backend_container declaration:

data_seed_container = docker.Container("data_seed_container",
                                       image=mongo_image.latest,
                                       name="data_seed",
                                       must_run=False,
                                       rm=True,
                                       opts=pulumi.ResourceOptions(depends_on=[backend_container]),
                                       mounts=[docker.ContainerMountArgs(
                                           target="/home/products.json",
                                           type="bind",
                                           source=f"{os.getcwd()}/products.json"
                                       )],
                                       command=[
                                           "sh", "-c",
                                           "mongoimport --host mongo --db cart --collection products --type json --file /home/products.json --jsonArray"
                                       ],
                                       networks_advanced=[docker.ContainerNetworksAdvancedArgs(
                                           name=network.name
                                       )]
                                       )

Note the mounts part, which allows you to use a bind mount storage type to add the necessary file. If you’re not familiar with mounts compared to volumes in Docker, see the docs on bind mounts. In this case, we’re using a bind mount over a volume for simplicity.

Once you’ve added this snippet, run pulumi up to refresh the data in the database. From here, you can choose to comment out or remove that seed container as long as you have the infrastructure up and running. If you were to rebuild this environment elsewhere, though, or to tear it down and rebuild it from scratch, you’ll need to have this resource in your code.

Open a browser to http://localhost:3001, and our application is now deployed.

Cleaning up

Whenever you’re working on learning something new with Pulumi, it’s always a good idea to clean up any resources you’ve created so you don’t get charged on a free tier or otherwise leave behind resources you’ll never use. Let’s clean up.

Run the pulumi destroy command to remove all of the resources:

$ pulumi destroy
Previewing destroy (dev)

View Live: https://app.pulumi.com/<org>/<project>/<stack>/previews/<build-id>

...
Do you want to perform this destroy? yes
Destroying (dev)

View Live: https://app.pulumi.com/<org>/<project>/<stack>/updates/<update-id>

...

The resources in the stack have been deleted, but the history and configuration associated with the stack are still maintained.
If you want to remove the stack completely, run 'pulumi stack rm dev'.

Now your resources should all be cleared! That last comment you see in the output notes that the stack and all of the configuration and history will stay in your dashboard on the Pulumi Service (app.pulumi.com and will be stored in the Pulumi Service that maintains the history. For now, that’s okay. We’ll talk more about removing the project from your history in another pathway.

Congratulations, you’ve now finished Pulumi Fundamentals! You learned to create a Pulumi project; work on your Pulumi program to build Docker images, containers, and networks; and deploy the infrastructure locally with your first resource provider. Now, head back to the main page and explore some other tutorials to understand more about Pulumi. The best next step to take is to explore the Building with Pulumi pathway.

Return to Pulumi Fundamentals →

See Also

Configure the application

Create a Container resource

Put it all together

Populate the database

Cleaning up