MojoX::Redis: an asynchronous Redis implementation for Mojolicious

Author of this post is und3f, author of MojoX::Redis. You can follow him on Twitter and/or GitHub

Today's topic is MojoX::Redis module for Mojolicious and asynchronous commucation with Redis.

Essentially Redis is a lightweight but advanced key-value store with great possibilities. It is available from Redis is similar to memcached but the dataset is not volatile, and values can not only be strings, like in memcached, but also lists, sets and ordered sets. All data types can be manipulated with atomic operations: push/pop elements, add/remove elements, perform server side union, intersection, difference between the sets, and so forth. Redis supports different kinds of sorting abilities.

Redis has quite a simple but effective communication protocol, with its limitations and benefits. It is described in Redis documentation Commands within single connections are executed sequentually - remember about that if you need to write some results to database and main connection is blocked by long-time operation like blpop.

When to use Redis? Redis is for sure very advanced key-value store, but you must remember that all operation related to creation and maintenance of complicated relations are on your own. It is hard to use Redis for applications with a great amount of relations between data (for example social networks like FaceBook), but if you need to store unrelated data - Redis is perfect.

MojoX::Redis is quite simple in usage. First of all we need to create a MojoX::Redis instance.

use MojoX::Redis;
my $redis = MojoX::Redis->new;

Now execute commands, MojoX::Redis will take care about the tcp connection.

    ping => sub {
        my ($redis, $result);

        # In case of error $result will be undefined
        die "Redis error occured: " . $redis->error
            unless defined $result;

        # Profit
        print $result->[0], "\n";

Some Redis commands expect one or more arguments. Also you can ignore callback if you don't want to ensure that command was executed successfully.

        set => [test => "testok"]
        get => 'test' => sub { ... }

MojoX::Redis is asynchronous, that's why to start it you have to start Mojo::IOLoop:


Here are some examples of using MojoX::Redis from a real life application. It is a daemon that uses Redis for communication. It inputs data from key "hosts" (which is a List, see Redis data types) and stores results in a key named "host:$host".

For this example to work you need Redis 2.0 or higher.

#!/usr/bin/env perl

use strict;
use warnings;

use Mojo::Client;
use MojoX::Redis;

my $redis = MojoX::Redis->new;

# We need new redis connection for results
my $redis_res = MojoX::Redis->new;

# ioloop is not running yet, so we need to install singleton loop
my $client = Mojo::Client->new(ioloop => $redis->ioloop)->async;

# Just simulate the first command execution to avoid copy-pasting
&request_host($redis, []);

# Here the magic starts to work
sub request_host {
    my ($redis, $result) = @_;

    # Possibly we got some error from Redis
    die $redis->error unless defined defined $result;

    # Everything is ok - just check the result
    my ($key, $host) = @$result;

    # If host is not defined - command were interrupted by a timeout
    if (defined $host) {
            $host => sub {
                my $client = shift;
                    set => ["result:$host" => $client->res->body],
                    sub {
                        print @{$_[1]};

    # Execute "blocking pop" command on array "hosts" with timeout of
    # 200 seconds
    $redis->execute(blpop => [hosts => 200] => \&request_host);


blog comments powered by Disqus