OpenRC init script problem with puma/sidekiq cluster daemons

In continuation of #102 (closed) I would like to address here the current problem of the OpenRC init script with puma and sidekiq which both consist of a supervising daemon with some workers.

My impression from the discussion in #102 (closed) is that the OpenRC start-stop-daemon command does not stop all the workers of puma/sidekiq on a rc-service gitlabhq stop.

@asaparov Are you willing to develop an init script using the supervise-daemon you mentioned in #102 (closed)?