Fixes #600 - Add init: true to scheduler and railsserver to prevent zombie process accumulation

[mgruner]

No description provided.

[mgruner]

@derpixler @fliebe92 what do you think about this? Since this behaviour affects all containers running the Rails stack, I would suggest to add init: true to the shared service definition, and disabling it only for zammad-nginx, which does not run the rails stack.

If you agree, I would think that a small UAT might be good to ensure there is no negative behaviour change. @ralf401 could you help here?

[fliebe92]

@derpixler @fliebe92 what do you think about this? Since this behaviour affects all containers running the Rails stack, I would suggest to add init: true to the shared service definition, and disabling it only for zammad-nginx, which does not run the rails stack.

If you agree, I would think that a small UAT might be good to ensure there is no negative behaviour change. @ralf401 could you help here?

Sounds good to me.

[derpixler]

@mg it is smarter as my suggestion, go ahead.

[ralf401]

@mgruner I played a bit around after deploying a stack from this PR and setting up a dummy email server. Didn't notice any unexpected behavior. However, not sure it I tested every important aspect. I just triggered some email notifications and checked the processes on the railsserver. Result: PID was 2403 in my case (docker-init got 1) and I could not see any duplicated processes.

Let me know if you think I should test additional stuff and/or wait longer to confirm no zombie process exists.

[mgruner]

Awesome, thank you @ralf401