Removing a container should also remove its cidfile.

[jperville]

Running a named container fails when its cidfile already exists (eg. left over from a previous deployment).

How to reproduce:

vagrant@default-ubuntu-1404:~$ sudo docker run  --cidfile="/var/run/my-sample-app-0.cid" --detach=true --env-file="/etc/sample-docker-app.conf" --name="sample-docker-app-0" --publish="7022:22" --publish="7080:80" --volume="/var/log/nginx:/var/log/nginx/sample-docker-app" docker.si.perfect-memory.com/sample-docker-app:latest /sbin/my_init --enable-insecure-key
a1628f0d268404532cd19b0c8eadd2c01f8bb4558b174ebe8993ea946386e00d
vagrant@default-ubuntu-1404:~$ sudo docker rm -f sample-docker-app-0
sample-docker-app-0
vagrant@default-ubuntu-1404:~$ sudo docker run  --cidfile="/var/run/my-sample-app-0.cid" --detach=true --env-file="/etc/sample-docker-app.conf" --name="sample-docker-app-0" --publish="7022:22" --publish="7080:80" --volume="/var/log/nginx:/var/log/nginx/sample-docker-app" docker.si.perfect-memory.com/sample-docker-app:latest /sbin/my_init --enable-insecure-key
2014/06/11 15:17:32 Container ID file found, make sure the other container isn't running or delete /var/run/my-sample-app-0.cid

This PR removes the cidfile every time the container is removed (during redeploy) so that the container can be recreated safely, as suggested by the docker run error message in the above example.

[djdefi]

Seem to be running into this just using the redeploy action as described in the README.md example.

Container is running, make a change to the image, push to registry, and then when run chef-client again it fails to run the new docker container because the cidfile already exists. Also it seems to not be stopping the old running container.

[bflad]

This seems like a necessary fix. Thanks for the help!

[jperville]

Thank you very much @bflad