docker client not passing signals to dockerd #3793
Description
We'd like a signal to the docker client to trigger a stop, so if the docker client is killed, it cleans up the containers.
Filing bug per @crosbymichael request for how to reproduce.
$ docker run -sig-proxy=true busybox sleep 3000
(other terminal)
$ docker ps
e8767eed1283 busybox:latest sleep 3000 28 seconds ago Up 28 seconds insane_pasteur
$ ps aux|grep 'busybox sleep'
782 docker docker run -sig-proxy=true busybox sleep 3000
$ kill 782
$ docker ps
e8767eed1283 busybox:latest sleep 3000 45 seconds ago Up 45 seconds insane_pasteur
... still running ...
$ kill -9 782
$ ps aux | grep 782
$
... docker run (client) has been killed ... but the container is still running.
$ docker ps
e8767eed1283 busybox:latest sleep 3000 3 minutes ago Up 3 minutes insane_pasteur
#1249 appears to have the functionality needed to fix this, but that appears to have been removed. Is there any other way to get this functionality?