Hadoop reads the file

HDFS - Read the file

1. The client sends a read request to NameNode
2. NameNode returns all the locks of the files and the DataNodes in which they are located (including replication nodes)
3. The client reads the data directly from DataNode, and if the DataNode read fails (DataNode fails or the check code is not correct), it reads from the replication node (if the read data is on the machine, read directly, otherwise read over the network)