There are some cases in which the commit coordinator fails but the removeShard command should not return an error. These cases are when the config server believes the shard to be fully drained and so establishes the commit coordinator but, after blocking migrations and DDLs, the shard is found to not actually be fully drained. In this case, though the coordinator will throw an exception, the config server command will need to handle that error and transform it into the correct status report.

This ticket is to introduce a new error code with extra info containing the draining progress of the shard to facilitate this behavior.