On step up complete, the range deleter service is asynchronously spawning the recovery of existing range deletions before setting its state as "up".

Since the scheduling of new range deletions is disallowed while the service is not up, this may lead to errors in case a very fast migration commits (or simply when an already committed migration gets recovered).

A possible solution to overcome this problem would be to spawn range deletion recovery when step-up begins and waiting for the component to be up before transitioning to primary.