Improve robustness of scheduler re-indexing

Wrap the scheduler re-indexing code in a try/catch block so it won't die if there's an invalid event (bad data in mh_scheduled_event table).