Version 5.2.0. Multiple system failures resulted in multiple non-recoverable disk errors - fsck fails. Apparently, these have corrupted the only copies of files within the volume. We have accepted the data loss, but have readable data directories within the volume. When attempting to delete some of the corrupt files, we get this error:
thealy@t2:~$ hadoop fs -rm -r /pcaps/5/6
17/06/14 13:01:59 INFO Configuration.deprecation: io.bytes.per.checksum is deprecated. Instead, use dfs.bytes-per-checksum
17/06/14 13:01:59 INFO fs.TrashPolicyDefault: Namenode trash configuration: Deletion interval = 0 minutes, Emptier interval = 0 minutes.
2017-06-14 13:02:00,0295 ERROR Client fs/client/fileclient/cc/client.cc:2551 Thread: 30050 Rmdirs failed for dir 6,Readdirplus rpc error No data available(61) fid 670053.10059.613190
2017-06-14 13:02:00,0295 ERROR JniCommon fs/client/fileclient/cc/jni_MapRClient.cc:963 Thread: 30050 remove: File /pcaps/5/6, rpc error, No data available(61)
17/06/14 13:02:00 ERROR fs.MapRFileSystem: Failed to delete path maprfs:///pcaps/5/6, error: No data available (61)
rm: `/pcaps/5/6': Input/output error
How can we "force" the delete to get rid of the files/directories that are corrupt?