Clean External Build On Failure, master branch (2022.04.22.)

After some pondering I came up with this proposal for handling the type of issue described by @aundrus in ATLINFR-4497.

As a reminder, we've recently seen a number of cases where (due to network and/or GitLab issues) a given external, like ROOT would fail to build in the CI, and the build directory got "so broken" that we had to schedule a cleanup of the affected build node(s).

@fwinkl, what do you think? I only wanted to remove anything in case of a failure when we are in "CI mode" in the build script. Since if something goes wrong in a nightly, keeping all artifacts should help with diagnosing what may have gone wrong. At the same time I don't just clean out the build directory even in the CI, I only remove certain directories. This way most of the logs would still remain on the CI node, allowing for at least some amount of debugging.

Still, I'm not convinced that this is the best way of doing it... but I think it may be a little better than not doing anything at all. 🤔