This error generally happens when the connector cannot access a Kerberized HDFS cluster. This may happen when:
Visit Prerequisites for more information.
If this happens after a few incremental crawls the incremental data will not be affected, so after this problem is resolved the incremental crawls can be resumed.