Pods always complete successfully after restart

[meurig]

If I remove write permission to the git repo then on the first run the pod fails and spits out an error.

However, kubernetes is then restarting the pod and on the second run it passes (without attempting the git push).

I've added a pull request which I think addresses the underlying issue (that we're not starting with a fresh copy of the repo on the second attempt). But I'm aware that I might be missing something obvious, so any thoughts would be appreciated.