[FEATURE] Introduce a corrupted index state

[noCharger]

Is your feature request related to a problem?

Flint index can be deleted via OpenSearch Delete index API and flint index will stay in stale state since the metadata log entry is dangling. Plus, the next transaction call to clean up the metadata log entry doesn't execute for some of the APIs like refreshIndex and recoverIndex

What solution would you like?

Introduce a corrupted index state when OpenSearch index is deleted. The benefit is that people will notice the index has stopped operating instead of being confused why they cannot delete the index (comparing the alternatives)

What alternatives have you considered?

Recreate an empty OpenSearch index.