The ShardingDDLCoordinatorService blocks the acquisition of DDL locks on step-down to ensure the serialization of DDL operations and other operations that also acquire DDL locks, even with step-downs.

There are a pair of issues according to this process that should be addressed under this ticket:

• During the step-down, the ShardingDDLCoordinatorService is blocking the acquisition of DDL locks after interrupting the DDL operations. So, there is a small window of time where another operation could kick in and get a DDL lock when it shouldn't (because another DDL operation may be in progress).

• During the shutdown process is happening something similar but we are even not calling ShardingDDLCoordinatorService::_afterStepDown(), which is the method that blocks the DDL lock acquisition.