Skip to main content Link Menu Expand (external link) Document Search Copy Copied

BlazingMQ in Action

The BlazingMQ project comes with an out-of-the-box docker cluster meant for experimentation. We’ll walk through bringing up a BlazingMQ broker, demonstrate some tooling, and highlight a few BlazingMQ features. This image is meant for demo purposes only.

To begin, first clone the BlazingMQ project and run the following commands from repo’s root directory:

$ docker compose -f docker/single-node/docker-compose.yaml up --build -d
$ docker compose -f docker/single-node/docker-compose.yaml run bmqtool
$ bmqtool -b tcp://bmqbrkr:30114

This command will bring up a BlazingMQ cluster with a default domain provisioned according to a config file. It will then drop you into a shell which can run a CLI program called bmqtool. Note that by default, BlazingMQ broker listens on port number 30114.

BlazingMQ CLI

BlazingMQ comes with a CLI tool to help test and interact with the broker called bmqtool. This program has several uses, including acting as a producer and/or consumer application. We will walk through its most basic workflow: the client REPL. The included docker compose workflow should get you started:

$ docker compose -f docker/single-node/docker-compose.yaml up bmqtool

> help

Creating your first queue

When you start up bmqtool, it should drop you into its default CLI mode. By default the tool does not begin a session with the broker in this mode. To open a connection with the BlazingMQ back-end, you’ll need to start a session:

> start

10MAY2023_15:27:19.305 (140195584415616) INFO m_bmqtool_interactive.cpp:140 --> Starting session: [ async = false ]
10MAY2023_15:27:19.311 (140195584415616) INFO m_bmqtool_interactive.cpp:151 <-- session.start(5.0) => SUCCESS (0)

Note: bmqtool may print log lines hiding the CLI prompt >. The CLI will still take commands you type into the CLI even if the > is not visible.

Now we may open our first queue! We will open the queue both as a producer and consumer. This means that any messages that we post will be echoed back to us. Note that a queue does not need to be created or declared upfront in BlazingMQ. It is automatically created under the hood by the system when it is opened for the first time.

> open uri

open uri="bmq://bmq.test.persistent.priority/my-first-queue" flags="read,write,ack"

The open command takes two parameters: the queue’s URI and permissions. Note that we are passing write and read flags when opening the queue. Producing messages to queues is done through the post operation. Try it out now:

> post uri="bmq://bmq.test.persistent.priority/my-first-queue" payload=["hello world"]

Since we opened the queue in the read mode as well, the message will be echoed back to us. This can be checked by the list command:

> list

12MAY2023_16:47:45.073 (139815379564416) INFO m_bmqtool_interactive.cpp:648 Unconfirmed message listing: 1 messages
  #  1  [40000000001068A4B4B275EDF7D3228F] Queue: '[ uri = bmq://bmq.test.persistent.priority/my-first-queue correlationId = [ autoValue = 2 ] ]' = 'hello world'

Messages are composed of a payload and an optional set of properties, which can act as metadata for the message. A property can be represented as a JSON object like so:

  "name": "string",
  "value": "any",
  "type": "string"

Queues are opened in namespaces called domains. If an opened queue doesn’t exist, then it will be created. Domains also help define default configurations for queues. We’ll look at how to configure custom domains later.

Once a message is delivered to a queue, it may be read by consumers and confirmed. In BlazingMQ, this is done with the confirm operation.

> confirm uri="bmq://bmq.test.persistent.priority/my-first-queue" guid="40000000001068A4B4B275EDF7D3228F"

Notice as part of the confirm operation we must pass a GUID, which matches the one from the output of the list command.

And that’s it! We just ran a BlazingMQ client application which acted as a producer and a consumer for a queue, posted a message and consumed it!

More hands-on examples of playing around with BlazingMQ can be found in this article, where we see some intermediate and advanced features of BlazingMQ in action.

We also encourage readers to visit these articles to learn more about BlazingMQ:

  • Concepts, which explains some terminology mentioned in this article, like domain, queue, queue uri, message, etc.

  • Our C++ and Java tutorials.

  • C++ and Java API reference documentation

  • BlazingMQ architecture


bin/ and bin/ build BlazingMQ and its dependencies, respectively, on Ubuntu 22.04.2 LTS and Darwin 22.6.0. They can serve as a basis to build BlazingMQ on other systems.