Wait to wait for processes to stop

[ralfthewise]

Couldn't find anything in the docs or a quick scour of the code. Ideally I would like something like:

bundle exec eye stop resque --wait

which wouldn't return until the signaled process(s) stopped. This is because ultimately I want to be able to do:

bundle exec eye stop all --wait
bundle exec eye quit

[kostya]

Right now this is not possible. I am not a fan of wait at all, bluepill has some form of wait and mary bugs related with it. All commands is a signals to process, which scheduled in some queue.
Also, i not see many reasons for that, in my every day deploy i never needed wait, and even quit. If it was, i just do it manually. What the use case, if you need stop all and quit?

[kostya]

about this https://github.com/kostya/eye/wiki/Run-ruby-apps-that-use-a-separate-bundle-than-eye-itself, i guess this is can be easier:

Eye.application :some do
   clear_bundler_env
   env "BUNDLE_GEMFILE" => "some_file"
end

[ralfthewise]

Ah, excellent, I hadn't seen that method in the DSL, will update the wiki. Thanks!

[ralfthewise]

As for the wait issue, it's not a big deal so if you're strongly opposed to it I can certainly work around it. Our situation is such that we have a cluster of EC2 servers running resque jobs that grows and shrinks based on demand. Some of the resque jobs can be fairly long running (15-20 minutes), and when we want to shrink the cluster we have to make sure all resque jobs are finished on it before shutting the server down.

In general though it would be nice for some way to tell if all of the processes eye manages are stopped. Obviously we can just parse the output of "eye oinfo" or some such, and that's probably what we'll end up doing.

[kostya]

in 0.6.rc added options to eye quit --stop_all and --timeout 600
eye q -s to stop all processes and quit.

[ralfthewise]

Awesome, thanks @kostya will test this out next week