Beam::Minion - A distributed task runner for Beam::Wire containers


version 0.016


    # Command-line interface
    export BEAM_MINION=sqlite://test.db
    beam minion worker
    beam minion run <container> <service> [<args>...]
    beam minion help

    # Perl interface
    local $ENV{BEAM_MINION} = 'sqlite://test.db';
    Beam::Minion->enqueue( $container, $service, \@args, \%opt );


Beam::Minion is a distributed task runner. One or more workers are created to run tasks, and then each task is sent to a worker to be run. Tasks are configured as Beam::Runnable objects by Beam::Wire container files.



    Beam::Minion->enqueue( $container_name, $task_name, \@args, \%opt );

Enqueue the task named $task_name from the container named $container_name. The BEAM_MINION environment variable must be set.

\%opt is a hash reference with the following keys:


Number of times to retry this job if it fails. Defaults to 1.


Time (in seconds) to delay this job (from now). Defaults to 0.


The job priority. Higher priority jobs get performed first. Defaults to 0.

(These are the same options allowed in the Minion "enqueue" method)


Configure Minion

To start running your Beam::Runner jobs, you must first start a Minion worker with the beam minion worker.command. Minion requires a database to coordinate workers, and communicates with this database using a Minion::Backend.

The supported Minion backends are:

Once you've picked a database backend, configure the BEAM_MINION environment variable with the URL. Minion will automatically deploy the database tables it needs, so be sure to allow the right permissions (if the database has such things).

In order to communicate with Minion workers on other machines, it will be necessary to use a database accessible from the network (so, not SQLite).

Start a Worker

Once the BEAM_MINION environment variable is set, you can start a worker with beam minion worker. Each worker can run jobs from all the containers it can find from the BEAM_PATH environment variable. Each worker will run up to 4 jobs concurrently.

Spawn a Job

Jobs are spawned with beam minion run <container> <service>. The service must be an object that consumes the Beam::Runnable role. container should be a path to a container file and can be an absolute path, a path relative to the current directory, or a path relative to one of the paths in the BEAM_PATH environment variable (separated by :).

You can queue up jobs before you have workers running. As soon as a worker is available, it will start running jobs from the queue.


Beam::Wire, Beam::Runner, Minion


Doug Bell <>


Mohammad S Anwar <>


This software is copyright (c) 2018 by Doug Bell.

This is free software; you can redistribute it and/or modify it under the same terms as the Perl 5 programming language system itself.