Implement a strategy for stale containers

[agis]

Right now, we remove containers in StartContainer() after they run successfully. However, it might happen, for example in case of a mistry failure while a container was running (eg. SIGKILL), that containers are left around with nobody to remove them.

We should decide on how to treat such cases.

[agis]

This also manifests itself in tests. If we stop a running test suite (Ctrl-C), it might happen that there are still containers left around, because they were created but never started. So when the tests run the next time, ContainerCreate() will fail since the container already exists.

A solution might be the following:

• if ContainerCreate() returns an error, check if the error is because a container with the same name exists. If so, then just use it. If not, return the error. I believe we can do this since container names depend on the project files.
• on boot, remove existing mistry's containers (namespaced)

I'm not quite sure about the 2nd point.

[agis]

After some discussion, perhaps the simplest solution would be to remove the container if it exists.