When FileStore.compact() returns the journal.log does not necessarily contain the head created by the compactor. This can lead to problems downstream like e.g. in
OAK-6894 where the compactor tool wrote the wrong (i.e. uncompacted) head to the journal.log.
Proposed fix is to call on of the FileStore.flush() methods after compaction and add a test case that verifies the journal.log contains the correct head state.