When starting SIFS (Soft Index File Store) with a corrupt index it must rebuild it. When it does it will run the compactor a file that should have been compacted, unfortunately the Compactor has not started at this point and gets a NPE. This in turn puts the Segment in an invalid state.

This then prevents the server from shutting down cleanly meaning it gets stuck in this state.

The big problem is this error encountered is never even logged. We should log the error (so that if a future issue happens we can more easily debug) and prevent this specific case from happening.