Run a Wallaroo Go Application in Docker

In this section, we're going to run an example Wallaroo application in Docker. By the time you are finished, you'll have validated that your Docker environment is set up and working correctly. If you haven't already completed the Docker setup instructions, please do so before continuing.

There are a few Wallaroo support applications that you'll be interacting with for the first time:

  • Our Metrics UI allows you to monitor the performance and health of your applications.
  • Data receiver is designed to capture TCP output from Wallaroo applications.
  • Giles sender is used to send test data into Wallaroo applications over TCP.

You're going to set up our "Celsius to Fahrenheit" example application. Giles sender will be used to pump data into the application. Data receiver will receive the output, and our Metrics UI will be running so you can observe the overall performance.

The Metrics UI process will be run in the background. The other three processes (data_receiver, sender, and Wallaroo) will run in the foreground. We recommend that you run each process in a separate terminal.

NOTE: If you haven't set up Docker to run without root, you will need to use sudo with your Docker commands.

Let's get started!

Since Wallaroo is a distributed application, its components need to run separately, and concurrently, so that they may connect to one another to form the application cluster. For this example, you will need 6 separate terminal shells to start the docker container, run the metrics UI, run a source, run a sink, run the Celsius application, and eventually, to send a cluster shutdown command.

Shell 1: Start the Wallaroo Docker container

Windows Powershell
Windows Command Prompt
docker run --rm -it --privileged -p 4000:4000 \
-v /tmp/wallaroo-docker/wallaroo-0.5.4/wallaroo-src:/src/wallaroo \
--name wally \
docker run --rm -it --privileged -p 4000:4000 `
-v c:/wallaroo-docker/wallaroo-0.5.4/wallaroo-src:/src/wallaroo `
--name wally `
docker run --rm -it --privileged -p 4000:4000 ^
-v c:/wallaroo-docker/wallaroo-0.5.4/wallaroo-src:/src/wallaroo ^
--name wally ^

Breaking down the Docker command

  • docker run: The Docker command to start a new container.

  • --rm: Automatically clean up the container and remove the file system on exit.

  • -it: Allows us to work with interactive processes by allocating a tty for the container.

  • --privileged: Gives the container access to the hosts' devices. This allows certain system calls to be used by Wallaroo, specifically mbind and set_mempolicy. This setting is optional, but by excluding it there will be a performance degradation in Wallaroo's processing capabilities.

  • -p 4000:4000: Maps the default port for HTTP requests for the Metrics UI from the container to the host. This makes it possible to call up the Metrics UI from a browser on the host.

  • -v /tmp/wallaroo-docker/wallaroo-0.5.4/wallaroo-src:/src/wallaroo: Mounts a host directory as a data volume within the container. The first time you run this, an empty directory needs to be used in order for the Docker container to copy the Wallaroo source code to your host. If an empty directory is not used, we are assuming it is prepopulated with the Wallaroo source code from this point forward. This allows you to open and modify the Wallaroo source code with the editor of your choice on your host. The Wallaroo source code will persist on your machine after the container is stopped or deleted. This setting is optional, but without it you would need to use an editor within the container to view or modify the Wallaroo source code.

  • --name wally: The name for the container. This setting is optional but makes it easier to reference the container in later commands.

Starting new shells

For each Shell you're expected to setup, you'd have to run the following to enter the Wallaroo Docker container:

Enter the Wallaroo Docker container:

docker exec -it wally env-setup

This command will start a new Bash shell within the container, which will run the env-setup script to ensure our environment is set up properly.

Shell 2: Start the Metrics UI

To start the Metrics UI run:

metrics_reporter_ui start

You can verify it started up correctly by visiting http://localhost:4000.

If you need to restart the UI, run:

metrics_reporter_ui restart

When it's time to stop the UI, run:

metrics_reporter_ui stop

If you need to start the UI after stopping it, run:

metrics_reporter_ui start

Shell 3: Run Data Receiver

We'll use Data Receiver to listen for data from our Wallaroo application.

data_receiver --listen --no-write --ponythreads=1 --ponynoblock

Data Receiver will start up and receive data without creating any output. By default, it prints received data to standard out, but we are giving it the --no-write flag which results in no output.

Shell 4: Run the "Celsius to Fahrenheit" Application

First, we need to build the application.

cd /src/wallaroo/examples/go/celsius

Now that we are in the proper directory, and the Metrics UI and Data receiver are up and running, we can run the application itself by executing the following command:

cd /src/wallaroo/examples/go/celsius
./celsius --in \
  --out --metrics --control \
  --data --name worker-name --external \
  --cluster-initializer --ponythreads=1 --ponynoblock

This tells the "Celsius to Fahrenheit" application that it should listen on port 7000 for incoming data, write outgoing data to port 5555, and send metrics data to port 5001.

Shell 5: Sending Data with Giles Sender

We will be sending in 25,000,000 messages using a pre-generated data file. The data file will be repeatedly sent via Giles Sender until we reach 25,000,000 messages.

You will now be able to start the sender with the following command:

sender --host --messages 25000000 --binary \
--batch-size 50 --interval 10_000_000 --repeat --no-write \
--msg-size 8 --ponythreads=1 --ponynoblock \
  --file /src/wallaroo/examples/go/celsius/celsius.msg

If the sender is working correctly, you should see Connected printed to the screen. If you see that, you can be assured that we are now sending data into our example application.

Check Out Some Metrics

First Look

Once the sender has successfully connected, if you visit the Metrics UI, the landing page should show you that the "Celsius to Fahrenheit" application has successfully connected.

Landing Page

If your landing page resembles the one above, the "Celsius to Fahrenheit" application has successfully connected to the Metrics UI.

Now, let's have a look at some metrics. By clicking on the "Celsius to Fahrenheit" link, you'll be taken to the "Application Dashboard" page. On this page you should see metric stats for the following:

  • a single pipeline: Celsius Conversion
  • a single worker: Initializer
  • three computations: Add32, Decode Time in TCP Source, Multiply by 1.8

Application Dashboard Page

You'll see the metric stats update as data continues to be processed in our application.

You can then click into one of the elements within a category to get to a detailed metrics page for that element. If we were to click into the Add32 computation, we'll be taken to this page:

Computation Detailed Metrics page

Feel free to click around and get a feel for how the Metrics UI is set up and how it is used to monitor a running Wallaroo application. If you'd like a deeper dive into the Metrics UI, have a look at our Monitoring Metrics with the Monitoring Hub section.


Shell 6: Cluster Shutdown

You can shut down the cluster with this command at any time:


You can shut down Giles Sender and Data Receiver by pressing Ctrl-c from their respective shells.

You can shut down the Metrics UI with the following command:

metrics_reporter_ui stop

Wallaroo Container

To shut down the Wallaroo container, use the docker stop command in a shell on your host:

docker stop wally

This command will also terminate any active sessions you may have left open to the docker container.

For tips on editing existing Wallaroo example code, have a look at our Tips for using Wallaroo in Docker section.

results matching ""

    No results matching ""