hadoop 2.4.1 2 namenodes + 11 datanodes A disk failure occurs, but did not occur under this block. So I tested the removing disk. Change the config and restart the datanode. before <name>dfs.datanode.data.dir</name> <value>file:/disk01/data/dfs/data,file:/disk02/data/dfs/data,file:/disk03/data/dfs/data</value> after <name>dfs.datanode.data.dir</name> <value>file:/disk01/data/dfs/data,file:/disk02/data/dfs/data</value> I tested all of the data nodes. Occurs under-blocks in some cases and replication blocks. Strangely this did not occur under the block in some cases. And the error occurred. 2014-10-01 17:21:43,841 ERROR org.apache.hadoop.hdfs.server.datanode.DataNode: datanode-192.168.10.18:90010:DataXceiver error processing READ_BLOCK operation src: /192.168.10.18:38954 dst: /192.168.10.18:90010 org.apache.hadoop.hdfs.server.datanode.ReplicaNotFoundException: Replica not found for BP-665073150-127.0.0.1-1393573366664:blk_1073756063_15239
