Forge Horizon deployment is not reliable

Posting here as i cannot find a support contact for Horizon. The Forge deployment service is awesome however from my experience it seems not reliable with Horizon.

25% of time, deployment through Forge, Horizon is stopped without being restarted. (In fact in Forge dashboard, checking daemon status, we can read 'RUNNING pid 18922, uptime 0:05:19' but when checking Horizon dashboard we can see it is in fact stopped)

It is necessary then to manually restart the Horizon daemon of server.

Is there an explanation?Something we do wrong?

Thanks for contrib

Robstar

7 years ago

Level 50

Have personally never had an issues via following the docs. I have Supervisord simply autorestart the service:

[program:horizon] process_name=%(program_name)s command=php /home/forge/app.com/artisan horizon autostart=true autorestart=true user=forge redirect_stderr=true stdout_logfile=/home/forge/app.com/horizon.log

Have you tried manually restarting queues as per the Laravel docs - https://laravel.com/docs/5.7/queues#queue-workers-and-deployment

ejdelmonico

7 years ago

Level 53

Did you set up the Forge daemons so that it restarts? Also, in your deploy script:

if [ -f artisan ]
then
    php artisan migrate --force
    php artisan view:clear
    php artisan cache:clear
    php artisan config:clear
    php artisan route:clear
    php artisan horizon:terminate
fi

Place that towards the end.

Please or to participate in this conversation.