redmine, puma, nginx and crontab - bash

Question

Wanted to make puma start automatically when system reboots.
This is my crontab line (standard user):
# m h dom mon dow command
SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/local/bin:/usr/bin
HOME=/home/seven/
#reboot /home/seven/mars/start.sh
the file start.sh
#!/bin/bash
/usr/bin/screen -dmS mars /home/seven/mars/puma -e production
Is I run the file by myself:
./start.sh
All works fine, screen starts, puma starts. Brill
But if I reboot the machine, nothing happens, screen and puma are not loaded.
What am I doing wrong?
Many thanks

Answer 1

One possible way to run puma, is with systemd as proposed by Puma's team at
https://github.com/puma/puma/blob/master/docs/systemd.md
This way would also let you
restart the service by simply typing
systemctl restart redmine
and you can get status by
systemctl status redmine
so the output looks like:
● redmine.service - Puma HTTP Server
Loaded: loaded (/etc/systemd/system/redmine.service; enabled; vendor preset: enabled)
Active: active (running) since Sun 2021-01-03 06:25:16 CET; 2 days ago
Main PID: 29598 (ruby)
Tasks: 6 (limit: 4915)
CGroup: /system.slice/redmine.service
└─29598 puma 4.2.1 (tcp://0.0.0.0:3000) [redmine]
Jan 03 06:25:17 srv-redmine puma[29598]: Puma starting in single mode...
Jan 03 06:25:17 srv-redmine puma[29598]: * Version 4.2.1 (ruby 2.6.3-p62), codename: Distant Airhorns
Jan 03 06:25:17 srv-redmine puma[29598]: * Min threads: 0, max threads: 16
Jan 03 06:25:17 srv-redmine puma[29598]: * Environment: development
Jan 03 06:25:19 srv-redmine puma[29598]: * Listening on tcp://0.0.0.0:3000
Jan 03 06:25:19 srv-redmine puma[29598]: Use Ctrl-C to stop
Place the following code into: /etc/systemd/system/redmine.service
[Unit]
Description=Puma HTTP Server
After=network.target
[Service]
Type=simple
# Preferably configure a non-privileged user
User=testuser
# Specify the path to your puma application root
WorkingDirectory=/home/testuser/redmine
# Helpful for debugging socket activation, etc.
# Environment=PUMA_DEBUG=1
# Setting secret_key_base for rails production environment. We can set other Environment variables the same way, for example PRODUCTION_DATABASE_PASSWORD
#Environment=SECRET_KEY_BASE=b7fbccc14d4018631dd739e8777a3bef95ee8b3c9d8d51f14f1e63e613b17b92d2f4e726ccbd0d388555991c9e90d3924b8aa0f89e43eff800774ba29
# The command to start Puma, use 'which puma' to get puma's bin path, specify your config/puma.rb file
ExecStart=/home/testuser/.rvm/wrappers/my_app/puma -C /home/testuser/redmine/config/puma.rb
Restart=always
KillMode=process
#RemainAfterExit=yes
#KillMode=none
[Install]
WantedBy=multi-user.target
Make sure to adjust user and path in above code to fit your system, by replacing testuser with your real Redmine user.
puma.rb file can look something like this:
port ENV['PORT'] || 3000
stdout_redirect '/home/testuser/redmine/log/puma.stderr.log', '/home/testtser/redmine/log/puma.stdout.log'
#daemonize true
#workers 3
#threads 5,5
on_worker_boot do
ActiveSupport.on_load(:active_record) do
ActiveRecord::Base.establish_connection
end
end
For more info about puma confg take a look at:
https://github.com/puma/puma

Answer 2

Thanks to Aleksandar Pavić for point out that I can use systemd!!!
I am the type that if I want something to work in one way, I ignore everything else because I want thinks to work the way I intended in first place. So in short, me been stupid.
Systemd was not that simple in the end, but reading the puma documentation helped a lot.
My redmine.service file:
[Unit]
Description=Puma HTTP Server
After=network.target
# Uncomment for socket activation (see below)
#Requires=puma.socket
[Service]
Type=simple
#WatchdogSec=10
# Preferably configure a non-privileged user
User=seven
# Specify the path to your puma application root
WorkingDirectory=/home/seven/mars/
# The command to start Puma, use 'which puma' to get puma's bin path, specify your config/puma.rb file
ExecStart=/bin/bash -lc '/home/seven/.rvm/gems/ruby-2.7.2/bin/puma -C /home/seven/mars/config/puma.rb'
Restart=always
KillMode=process
[Install]
WantedBy=multi-user.target
and my puma.rb
# config/puma.rb
workers Integer(ENV['PUMA_WORKERS'] || 2)
threads Integer(ENV['MIN_THREADS'] || 0), Integer(ENV['MAX_THREADS'] || 16)
rackup DefaultRackup
port ENV['PORT'] || 9292
environment ENV['RACK_ENV'] || 'production'
lowlevel_error_handler do |e|
Rollbar.critical(e)
[500, {}, ["An error has occurred, and engineers have been informed. Please reload the page. If you continue to have problems, contact hq#starfleet-command.co\n"]]
end
After this systemd started to work with one exception. I was still getting an error when I wanted to enable the process:
sudo systemctl enable redmine.service
Synchronizing state of redmine.service with SysV service script with /lib/systemd/systemd-sysv-install.
Executing: /lib/systemd/systemd-sysv-install enable redmine
update-rc.d: error: cannot find a LSB script for redmine
I needed to remove the redmine file in /etc/init.d/ and enable redmine.service again to re-create the start script
Now all works as it should :) Many thanks to all!!!

Answer 3

Your details are not sufficient to answer but in your case i think that you created crontab file manually. if so why you just don't use that following command to add a cron job:
crontab -e
If you want to add it manually, make sure that your file has execute mode, add it to /etc/cron.d and reload your cron service with sudo service cron reload.
But if it doesn't help you, try to see the logs with the following command:
journalctl -xe
This may help you to debug your job when your system reboots.

Answer 4

so, this is the start.sh script:
#!/bin/bash
DISPLAY=:0; export DISPLAY
SHELL=/bin/bash
PWD=/home/seven/mars
LOGNAME=seven
XDG_SESSION_TYPE=tty
MOTD_SHOWN=pam
HOME=/home/seven
LANG=C.UTF-8
USER=seven
# main script to start screen and after puma.sh
cd /home/seven/mars/ && /usr/bin/screen -dmS mars /home/seven/puma.sh
what I did not is, I am trying to run a second script form the one that cron is starting. The puma.sh
Content:
#!/bin/bash
DISPLAY=:0; export DISPLAY
SHELL=/bin/bash
PWD=/home/seven/mars
LOGNAME=seven
XDG_SESSION_TYPE=tty
MOTD_SHOWN=pam
HOME=/home/seven
LANG=C.UTF-8
USER=seven
# main script to start puma
cd /home/seven/mars/ && puma -e production -w 2 -t 0:16
when I run the start script manually: ./start.sh in Cli all works fine.
When I reboot the machine (ubuntu 20.04 LTS) the script is been triggered, screen starts but the second script is not starting.
Now, I didi try to put the puma -e production ... after the screen -dmS mars ... (without the second script line), but that ended up in screen not starting after reboot.
Again, the start.sh work perfectly when triggered manually under Cli, just not when triggered by Cron, well not fully.