Hi all, I have a small cluster with one namenode1 and one datanode.
I want to configure another namenode2 to replace the namenode1 by only replicating the files in namenode directory of namenode1 to namenode2 and changing IP of namenode2 to namenode1's. I tried this, the replacing namenode2 can work with datanode, like start/stop hdfs, create directory, delete directory, a new mapreduce job.... But when I want to check the directories and files created by namenode1, nothing found. So i suspect that the blocking-mapping information is not included in namenode directory of namenode1. Am I right? Does anyone know how the namenode manages the mapping block information. Please give me some ideas. If I am wrong. Please correct me. Thanks a looooooot. Shengdi