Huck is a small, cross-platform Ruby application to make sharing information about hosts, applications, or anything else easy between machines on a network.

Huck's current intended functionality is:

• Generate some information using scripts, files, or something like facter or ohai
• Serialize and submit the information to a messaging queue
• Daemon process to read information from the queue and execute something for each received message

Extensibility is a major design goal in Huck. If you can generate some data in ruby, you can make a new Huck data generator. If you can connect to a different queueing service, you can make a new data sender/receiver, and if you have a function that handles a string argument, you can make a new handler.

These generate information on client machines to submit into a queue for a server machine to read. Huck ships with receivers for Ohai and Facter, as well as a few simple generators to read files, collect hostname and platform information, or execute arbitrary scripts.

These run on clients and submit generated data into a queue. They use a generic interface so that more messaging systems may be added on later. Currently Amazon SQS (default) and RabbitMQ are supported.

These poll messages out from a queue for processing, and also use a generic interface to support more messaging systems. Currently Amazon SQS (default) and RabbitMQ are supported.

Handlers run on Huck servers and take action based on received messages. They are arbitrary pieces of code which parse messages and then make other calls using the data. Huck ships with two handlers: echo (for demo/debugging purposes), and exec, which invokes an arbitrary script on the filesystem and passes data in via stdin, and satisfies most use cases for Huck.

$ gem install huck
$ gem install aws-sdk  # if you want to use SQS
$ gem install bunny    # if you want to use RabbitMQ
$ gem install facter   # if you want to use Facter
$ gem install ohai     # if you want to use Ohai

NOTE Huck does not depend directly on provider-related gems to avoid installing them on every machine.

This gem ships with a command-line interface to make things easy. You can run them like this:

Start a server:

$ huck serve

Run the client once:

$ huck run

Take a look at the config example to get an idea of how to configure Huck for your environment.

If you don't want to put your config in ~/huck.conf (the default), you may pass the -c command-line option to both the client and server to specify a different place.

You may wish to use Huck as a library instead of using its CLI. You can run the Huck server with the following code:

Huck.serve

While using Huck as a library, it is possible to pass arbitrary code blocks to use as message handlers. This looks something like:

Huck.serve do |msg|
  puts msg
end

The client can also be used from the library easily.

Huck.run

The run method can accept a block, very similar to the serve method.

Huck.run do
  "This is a test message"
end

Both serve and run will accept a :config_file option to specify the config path. It is also possible to pass the configuration in as a hash using the :config option.

Concept and prototypes brewed with @xorl