MB-52931 Drop index (not just instance) after async recovery is done

If an index drop is received while async recovery is in
progress, indexer will process the drop only after async
recovery is done. At this time, delete the index as there
is no concept of replica drop in serverless mode.

Otherwise, the index defnID will remain in meta and this
can lead to index creation failures when another index with
same name on same keyspace is created.

Change-Id: I743ebe269766fb8e13835a3fa5fbcc61bc74ef34

secondary/manager/lifecycle.go

@@ -1030,7 +1030,7 @@ func (m *LifecycleMgr) handleInstAsyncRecoveryDone(content []byte) error {
 			"while async recovery was in progress", inst.InstId)
 
 		// Instance is dropped while async recovery is in progress. Drop the index
-		return m.DeleteIndexInstance(inst.Defn.DefnId, inst.InstId, true, false, true, common.NewShardRebalanceRequestContext())
+		return m.DeleteIndexInstance(inst.Defn.DefnId, inst.InstId, true, false, false, common.NewShardRebalanceRequestContext())
 	}
 
 	return nil