Alfred is a command line tool writen in Perl to manage server tasks as Perl modules. It also provides a daemon wich stays listening for new scheduled tasks to be executed and takes care of the job.

First, you'll ned to clone the repository and instal the dependencies:

$ git glone https://github.com/danillosouza/alfred.git
$ cd alfred
$ carton install

In order to install one of the Perl modules used in Alfred(Net::SSH::Perl), you'll have to install the following package:

# with apt-get
$ sudo apt-get install libgmp3-dev

# with yum
$ sudo yum install libgmp3-dev

Alfred dependencies are managed by Carton, a Perl module to isolate the packages you need to your project, without affecting your global Perl installation. To get Carton you only need to:

# if you have cpanminus installed
$ cpanm Carton

# otherwise search by Carton directly into CPAN
$ perl -MCPAN -e install

Alfred is very simple to use, it provides a few commands:

• install - Install itself to the current user path.
• daemon - Starts Alfred daemon.
• dismiss - Stops Alfred daemon.
• list - List all available tasks.
• queue - List all tasks currently in the queue.
• help - Show help information for the given task.
• create - Create a new task.
• purge - Destroy the given task.
• run - Execute the given task.
• run:remote - Execute the given task in a remote server.
• schedule - Add the given task to the daemon queue.
• cron - Add a new crontab job to run the given task.

After cloning the repository and installing the dependencies, you'll want to install it to your local path, you only need to run:

$ ./alfred install
$ cd ~

Alfred comes with a dummy task named Foo::Bar that only shows a massage, you can use it to see how to interact with your tasks:

$ alfred run Foo::Bar
# Heeeeeey!

$ alfred help Foo::Bar
# Hooooooo!

Your tasks can also accept parameters, so if it was the case, you could call it like that:

$ alfred run My::Task --my --custom --params

Naturally, you'll want to create your own tasks, and it's simple:

$ alfred create My::Task
# He then will show you the path where the Task file was created, so you can code it

If you need to see the tasks currently available to you, just run:

$ alfred list

And to remove an existent task:

$ alfred purge My::Task

Beyond this, Alfred can also run as a daemon, so in case you need larger Tasks to be executed, you only need to stack them in the queue so when he's done with the previous Tasks, he will run it:

# Starts the daemon
$ alfred daemon

# Stack a new task to the queue
$ alfred schedule My::Task

# List all scheduled Tasks
$ alfred queue

# Stops the daemon, if it's running
$ alfred dismiss

You can even create a cronjob to run your tasks, let's say you have a task managing your database backups and you need to run it every day at 1:20am, is simple as that:

# Create a cronjob to run with the current user
$ alfred cron "20 1 * * *" Database::Backup

# If you want to create a cronjob for another user to run, pass the username like this
$ alfred cron "@bruce 20 1 * * *" Database::Backup

Alfred also can run tasks in a remote installation. In his root directory you'll find a file called remotes.yaml, containing the names and connection data for your remote servers running alfred instances, the file looks like this:

---
host1:
    host: 192.168.0.10
    port: 22
    user: your-username
    pass: your-password

host2:
    host: 192.168.0.20
    port: 22
    user: another-username
    pass: another-password
---

So when you feel the need to run a task in one of these remote instances, you would call Alfred like this:

$ alfred run:remote host1 Foo::Bar --my --params