puma systemd 脚本不启动 puma

Question

I've been copy and pasting snippets of the following puma systems start script from multiple sources but it doesn't work. The system is a Debian 9.6 and the user rails has ruby installed via rvm . The user rails can start the command RAILS_ENV=production bundle exec puma -C /var/www/test-optimum/current/config/puma.rb --daemon

What is wrong with this script?

[Unit]
Description=Test Puma HTTP Server
After=network.target

[Service]
Type=simple
User=rails
WorkingDirectory=/var/www/test-optimum/current
ExecStart=/bin/bash -lc 'RAILS_ENV=production bundle exec puma -C /var/www/test-optimum/current/config/puma.rb --daemon'
ExecStop=/home/rails/.rvm/bin/rvm default do bundle exec pumactl -S /var/www/test-optimum/shared/tmp/pids/puma.pid stop

PIDFile=/var/www/test-optimum/shared/tmp/pids/puma.pid
Restart=always

[Install]
WantedBy=multi-user.target

When trying to start it with systemctl start puma I get this errors:

● test-puma.service - test Puma HTTP Server
  Loaded: loaded (/etc/systemd/system/test-puma.service; disabled; vendor preset: enabled)
  Active: failed (Result: exit-code) since Fri 2019-01-04 17:10:43 CET; 6min ago
  Process: 3351 ExecStop=/home/rails/.rvm/bin/rvm default do bundle exec pumactl -S /var/www/test-optimum/shared/tmp/pids/puma.pid stop (code=exited, status=1/FAILURE)
  Process: 3239 ExecStart=/bin/bash -lc RAILS_ENV=production bundle exec puma -C /var/www/test-optimum/current/config/puma.rb --daemon (code=exited, status=0/SUCCESS)
Main PID: 3239 (code=exited, status=0/SUCCESS)
      CPU: 1.179s

Jan 04 17:10:42 test.example.com systemd[1]: test-puma.service: Unit entered failed state.
Jan 04 17:10:42 test.example.com systemd[1]: test-puma.service: Failed with result 'exit-code'.
Jan 04 17:10:43 test.example.com systemd[1]: test-puma.service: Service hold-off time over, scheduling restart.
Jan 04 17:10:43 test.example.com systemd[1]: Stopped test Puma HTTP Server.
Jan 04 17:10:43 test.example.com systemd[1]: test-puma.service: Start request repeated too quickly.
Jan 04 17:10:43 test.example.com systemd[1]: Failed to start test Puma HTTP Server.
Jan 04 17:10:43 test.example.com systemd[1]: test-puma.service: Unit entered failed state.
Jan 04 17:10:43 test.example.com systemd[1]: test-puma.service: Failed with result 'exit-code'.

Answer 1

Here's the solution:

[Unit]
Description=Test Puma HTTP Server
After=network.target

[Service]
Type=simple
User=rails
WorkingDirectory=/var/www/test-optimum/current
ExecStart=/bin/bash -lc 'RAILS_ENV=production bundle exec puma -C /var/www/test-optimum/current/config/puma.rb'
ExecStop=/home/rails/.rvm/bin/rvm default do bundle exec pumactl -S /var/www/test-optimum/shared/tmp/pids/puma.pid stop

PIDFile=/var/www/test-optimum/shared/tmp/pids/puma.pid
Restart=always

[Install]
WantedBy=multi-user.target

Answer 2

What is wrong with this script?

Probably just the type of the service:

Main PID: 3239 (code=exited, status= 0 /SUCCESS)

Seems the main process is forking. Try forking instead of simple : https://www.freedesktop.org/software/systemd/man/systemd.service.html#Options

Answer 3

在我的情况下，我在ExecStart更新到这样的东西时有错误的命令ExecStart=/bin/bash -lc 'RAILS_ENV=production bundle exec puma -C /home/ubuntu/rails_current_server/config/puma.rb'

Answer 4

An improved solution could be:

[Unit]
Description=Test Puma HTTP Server
After=network.target

[Service]
Type=simple
User=rails
WorkingDirectory=/var/www/test-optimum/current
ExecStart=/home/rails/.rvm/bin/rvm default do bundle exec puma -C /var/www/test-optimum/current/config/puma.rb --daemon
ExecStop=/home/rails/.rvm/bin/rvm default do bundle exec pumactl -S /var/www/test-optimum/shared/tmp/pids/puma.pid stop

PIDFile=/var/www/test-optimum/shared/tmp/pids/puma.pid
Restart=always

[Install]
WantedBy=multi-user.target