Investigate possibility of race conditions in ReactiveDeleteCoordinator

[Sanne]

As I was working on something unrelated, I spotted a suspicious pattern within ReactiveDeleteCoordinator : it declares a field of type CompletableFuture.

This looks suspiciously similar to the pattern we have recently fixed for updates in #1687 - I am not sure if this could manifest with a similar problem, but opening this to not forget to investigate.

[vineetgit]

@Sanne Good day, is this missed out? Or expected ? or not needed. A project I am doing seem to be hit by this, where we are using (orphanRemoval) for a one-to-many relation. Is there a work-around for it?

[DavideD]

Hello,
@Sanne identified a problematic pattern, but we don't have a test at the moment and nobody else mentioned any problem related to this area. So we haven't investigated it yet.

If you could provide a testcase, or at least give us more details, we will check what's going on sooner.