I ran into a problem where the container does not start after 10 startups.
The reason for this is `/tmp/.X10-lock` file are not deleted when running `docker stop`.
On next `docker-start`, `/tmp/.X11-lock` is allocated instead.
This keeps going until `/tmp/.X19-lock` and from that it gives up.
Another workaround is to use `--tmpfs` in `docker run`.
But not all users always use it.
