将运行Laravel队列的单个网站添加到Supervisor时,会按原样处理延迟的作业。
当我向Supervisor添加另一个网站时,不会为任何网站/队列处理延迟的作业。
我尝试重新读取,重新加载和重新启动服务。 supervisord.log显示正在运行的所有进程,但未触发作业。
如果我停止除1名工作者以外的所有工作人员
Laravel 5.7
Supervisord 3.3.1
redis-cli 3.2.6
于Debian / Nginx的
sudo nano /etc/supervisor/conf.d/website-a-worker.conf
[program:website-a-worker]
process_name=%(program_name)s_%(process_num)02d
command=php /var/www/website.a/artisan queue:work redis --sleep=3 --tries=3
autostart=true
autorestart=true
#user=laravel
numprocs=8
autostart=true
autorestart=true
stdout_logfile=/var/www/website.a/storage/logs/worker.log
stderr_logfile=/var/www/website.a/storage/logs/worker.err.log
sudo nano /etc/supervisor/conf.d/website-b-worker.conf
[program:website-b-worker]
process_name=%(program_name)s_%(process_num)02d
command=php /var/www/website.b/artisan queue:work redis --sleep=3 --tries=3
autostart=true
autorestart=true
#user=laravel
numprocs=8
autostart=true
autorestart=true
stdout_logfile=/var/www/website.b/storage/logs/worker.log
stderr_logfile=/var/www/website.b/storage/logs/worker.err.log
sudo supervisorctl reread
sudo supervisorctl update
sudo supervisorctl start website-a-worker:*
sudo supervisorctl start website-b-worker:*
// config/queue.php
...
'redis' => [
'driver' => 'redis',
'connection' => 'default',
'queue' => 'default',
'retry_after' => 600,
'block_for' => null,
],
...
<?php
namespace App\Jobs;
use Illuminate\Bus\Queueable;
use Illuminate\Queue\SerializesModels;
use Illuminate\Queue\InteractsWithQueue;
use Illuminate\Contracts\Queue\ShouldQueue;
use Illuminate\Foundation\Bus\Dispatchable;
class SomeJob implements ShouldQueue
{
use Dispatchable, InteractsWithQueue, Queueable, SerializesModels;
public function handle()
{
\Log::info('This only gets triggered if 1 supervisor worker is running.');
}
}
> sudo service supervisor restart
> cd /var/www/website.a
> php artisan tinker
>>> App\Jobs\SomeJob::dispatch()->delay(1);
# Job does not get triggered
>>> exit
这项工作永远不会被触发。但是,如果我停止使用website.b的主管,website.a将处理作业。
> sudo supervisorctl stop website-b-worker:*
website-b-worker:website-b-worker_02: stopped
website-b-worker:website-b-worker_03: stopped
website-b-worker:website-b-worker_00: stopped
website-b-worker:website-b-worker_01: stopped
website-b-worker:website-b-worker_06: stopped
website-b-worker:website-b-worker_07: stopped
website-b-worker:website-b-worker_04: stopped
website-b-worker:website-b-worker_05: stopped
> php artisan tinker
>>> App\Jobs\SomeJob::dispatch()->delay(1);
# Job gets triggered!
如何运行多个工作人员(每个网站至少1个)?
我在Laravel和Redis上使用过Supervisor,但一次只用于1个网站。但是我做了一些研究,发现你可以在Laravel https://laravel.com/docs/5.8/queues#connections-vs-queues中指定队列进程的名称。
Laravel qazxsw poi有一个setter方法qazxsw poi来指定队列的名称。 laravel docs示例:
Job
但是在你的代码onQueue()
中你可以用Job::dispatch()->onQueue('emails');
直接调用它;你的代码可以像:
SomeJob
在网站A的Supervisor流程中,将队列参数指定为$this->onQueue('queue_name')
:
<?php
namespace App\Jobs;
use Illuminate\Bus\Queueable;
use Illuminate\Queue\SerializesModels;
use Illuminate\Queue\InteractsWithQueue;
use Illuminate\Contracts\Queue\ShouldQueue;
use Illuminate\Foundation\Bus\Dispatchable;
class SomeJob implements ShouldQueue
{
use Dispatchable, InteractsWithQueue, Queueable, SerializesModels;
public function handle()
{
// on website a
$this->onQueue('websiteAQueue');
\Log::info('This only gets triggered if 1 supervisor worker is running.');
}
}
这样,当Laravel发送Job时,它会将它发送到指定的Queue而不是它找到的默认Queue。
编辑 - - - - - - - - - - - -
或者更好的是,您可以为每个网站配置使用不同的队列名称:
websiteAQueue
这样所有队列都会自动发送到[program:website-a-worker]
process_name=%(program_name)s_%(process_num)02d
command=php /var/www/website.a/artisan queue:work redis --queue=WebsiteAQueue --sleep=3 --tries=3
autostart=true
autorestart=true
#user=laravel
numprocs=8
autostart=true
autorestart=true
stdout_logfile=/var/www/website.a/storage/logs/worker.log
stderr_logfile=/var/www/website.a/storage/logs/worker.err.log
,因此您不必在每个作业上使用// Website A : config/queue.php
...
'redis' => [
'driver' => 'redis',
'connection' => 'default',
'queue' => 'websiteAQueue',
'retry_after' => 600,
'block_for' => null,
],
...
。