Laravel queues with Supervisor - Entered FATAL state, too many start retries too quickly

Question

I'm trying to use Laravel queues with supervisor but the service is not working properly. My /var/log/supervisor/supervisord.log is:

2018-06-18 10:56:07,441 INFO spawned: 'laravel-worker_00' with pid 20838
2018-06-18 10:56:07,446 INFO spawned: 'laravel-worker_01' with pid 20839
2018-06-18 10:56:08,021 INFO exited: laravel-worker_01 (exit status 255; not expected)
2018-06-18 10:56:08,033 INFO gave up: laravel-worker_01 entered FATAL state, too many start retries too quickly
2018-06-18 10:56:08,033 INFO exited: laravel-worker_00 (exit status 255; not expected)
2018-06-18 10:56:09,034 INFO gave up: laravel-worker_00 entered FATAL state, too many start retries too quickly

My config /etc/supervisord.d/laravel-worker.conf is:

[program:laravel-worker]
process_name=%(program_name)s_%(process_num)02d
command=/usr/bin/php /var/www/my-project/artisan queue:work --sleep=3 --tries=3
autostart=true
autorestart=true
user=root:root
numprocs=8
redirect_stderr=true
stdout_logfile=/var/www/my-project/worker.log

When I try to restart all programs:

$ sudo supervisorctl restart all
$ laravel-worker:laravel-worker_00: ERROR (abnormal termination)
$ laravel-worker:laravel-worker_01: ERROR (abnormal termination)

I'm a newbie with supervisor so someone can guide me?

Answer 1

You need to add startsecs = 0 to your laravel-worker config like so:

[program:laravel-worker]     
process_name=%(program_name)s_%(process_num)02d
command=/usr/bin/php /var/www/my-project/artisan queue:work --sleep=3 --tries=3
.....
startsecs = 0

startsecs is default to 1s, if the program does not stay up for 1s it will see the startup as failure. Set it to 0 so the program needn’t stay running for any particular amount of time. You can check this github issue for more info: https://github.com/Supervisor/supervisor/issues/212

It would be better if you run the queue worker in daemon mode, use the --daemon flag: command=/usr/bin/php /var/www/my-project/artisan queue:work database --daemon --sleep=3 --tries=3

After you change the config file, you may need to run supervisorctl reload to allow the changes to take effect.

Answer 2

Just want to share my case.

First, the user I used didn't have access to write to the log files so I added it as a sudo user.

Second, now that Supervisor can write to the error log file, I saw that there was a parse error when trying to run the PHP script. I fixed the error and now Supervisor runs the script perfectly.

Hope this helps someone too.