Bugfix: Shutdown cleanly when SIGTERM is received

Prior, if restic received SIGTERM it'd just immediately terminate skipping
cleanup- resulting in potential issues like stale locks being left behind.

This primarily effected containerized restic invocations- they use SIGTERM-
but this could be triggered via a simple `killall restic` in addition.

https://github.com/restic/restic/pull/4703