pumactl restart or SIGUSR2 actually kills puma server instead of restarting

[vizcay]

Environment

Puma: 3.9.1
Ruby: 2.3.3p222

I've just configured puma in a new rails 5.0 application, config/puma/production.rb:

threads 4, 8
workers 2
bind "unix://tmp/sockets/puma.sock"
daemonize true
preload_app!
activate_control_app "unix://tmp/sockets/pumactl.sock", { auth_token: '12345' }
pidfile "tmp/pids/puma.pid"
state_path "tmp/pids/puma.state"
plugin :tmp_restart

on_worker_boot do
  ActiveRecord::Base.establish_connection if defined?(ActiveRecord)
end

Problem

Puma runs great by bundle exec puma -e production. I can stop and restart the server by bundle exec pumactl -P tmp/pids/puma.pid start and bundle exec pumactl -P tmp/pids/puma.pid stop.

But when I try to restart bundle exec pumactl -P tmp/pids/puma.pid restart the puma server gets killed. Also, I've tried to send the signal directly by sudo kill -SIGUSR2 #pid# to be sure pumactl is not responsible.

Any ideas?

[prashantvithani]

The issue might be similar to #1308

[vizcay]

@prashantvithani yes it's looks like an exact duplicate, sorry.