[5.4] Queue prefix added twice with beanstalk

[it-can]

• Laravel Version: 5.4.20
• PHP Version: 7.1.4
• Database Driver & Version: MariaDB

Description:

In Laravel 5.4.20 there is a queue prefix added (#18860). But when I started adding a prefix it seems to be added twice in my beanstalk tube. I did not check of this happens with other queue drivers... I tested this with multiple laravel apps on multiple servers... The queue works fine without a prefix, but does not work when using a prefix...

without prefix:

portal

with prefix "server1"

server1server1portal

Steps To Reproduce:

Add queue prefix and list your beanstalk tubes

telnet 127.0.0.1 11300

list-tubes

[laurencei]

I cant replicate? I get server1portal

Can you post your full queue config?

[it-can]

config/queue.php

<?php

return [

    /*
    |--------------------------------------------------------------------------
    | Default Queue Driver
    |--------------------------------------------------------------------------
    |
    | The Laravel queue API supports a variety of back-ends via an unified
    | API, giving you convenient access to each back-end using the same
    | syntax for each one. Here you may set the default queue driver.
    |
    | Supported: "sync", "database", "beanstalkd", "sqs", "redis", "null"
    |
    */

    'default' => env('QUEUE_DRIVER', 'sync'),

    /*
    |--------------------------------------------------------------------------
    | Queue Prefix
    |--------------------------------------------------------------------------
    |
    | If you are running multiple sites on a single server, you may experience
    | crosstalk among sites if they use the same name for queue tubes. This
    | optional value defines a prefix that will automatically be applied
    | to queue tubes as a way to prevent this crosstalk.
    |
    */
    'prefix'  => env('QUEUE_PREFIX', ''),

    /*
    |--------------------------------------------------------------------------
    | Queue Connections
    |--------------------------------------------------------------------------
    |
    | Here you may configure the connection information for each server that
    | is used by your application. A default configuration has been added
    | for each back-end shipped with Laravel. You are free to add more.
    |
    */

    'connections' => [

        'sync' => [
            'driver' => 'sync',
        ],

        'database' => [
            'driver'      => 'database',
            'table'       => 'jobs',
            'queue'       => env('QUEUE_NAME', 'default'),
            'retry_after' => 90,
        ],

        'beanstalkd' => [
            'driver'      => 'beanstalkd',
            'host'        => 'localhost',
            'queue'       => env('QUEUE_NAME', 'default'),
            'retry_after' => 90,
        ],

        'sqs' => [
            'driver' => 'sqs',
            'key'    => 'your-public-key',
            'secret' => 'your-secret-key',
            'prefix' => 'https://sqs.us-east-1.amazonaws.com/your-account-id',
            'queue'  => env('QUEUE_NAME', 'default'),
            'region' => 'us-east-1',
        ],

        'redis' => [
            'driver'      => 'redis',
            'connection'  => 'default',
            'queue'       => env('QUEUE_NAME', 'default'),
            'retry_after' => 90,
        ],

    ],

    /*
    |--------------------------------------------------------------------------
    | Failed Queue Jobs
    |--------------------------------------------------------------------------
    |
    | These options configure the behavior of failed queue job logging so you
    | can control which database and table are used to store the jobs that
    | have failed. You may change them to any database / table you wish.
    |
    */

    'failed' => [
        'database' => env('DB_CONNECTION', 'mysql'),
        'table'    => 'failed_jobs',
    ],

];

in my .env file I set this:

QUEUE_DRIVER=beanstalkd
QUEUE_NAME=portal
QUEUE_PREFIX=server1

[laurencei]

well - that is sync and not beanstalkd?

[it-can]

@laurencei typo...

[laurencei]

ok - well I've got something similar and I dont see that.

however I think I found another bug relating to queue prefixes and "default" connections - so I'm looking into it now. It might be related - dunno...

[it-can]

@laurencei can you share some info on the possible "bug"?

[laurencei]

I'm able to replicate a prexfixprefixqueue with the database driver when running queue:work

I'm looking into it now and will do a PR - give me 30mins...

I'll check if affects other drivers once I worked out the exact source... nearly there...

[laurencei]

@it-can - I'm about to submit a PR.

I'm fairly sure it relates to your issue, because the worker will duplicate prefixes. My guess is when Beanstalkd is trying to access a non-existent queue it probably creates it first.

[it-can]

Nice sounds like it...

[laurencei]

@it-can - can you please test against the master branch. Taylor just merged my PR.

You might need to "flush" your beanstalkd tubes to remove the old double-prefixed from showing - then try again?

[themsaid]

Thank you @laurencei for the fix, I guess you meant to test against the 5.4 branch not master.

[laurencei]

Yeah - sorry - I meant the latest dev version of the 5.4 branch (since the fix is not yet tagged I dont think?)

[it-can]

This fix works for my case 👍 Thanks