What to do if you repeatedly format a cluster

note: if you just want to solve this problem you can skip 1,2 and go straight to the 3 and 4 solution steps

  1. one-click start cluster view where the datanode’s log is
    sh start-all.sh

enter the log view
use shift+g to enter the last row mode, in the upward turn, see the first INFO, there is WARN below, there is a prompt message, about the datenode clusterID and namenode clusterID is not consistent.

  1. enter CD/export/servers/hadoop – server – cdh5.14.0/hadoopDatas/datanodeDatas/current/
    see cat VERSION

    this is consistent with the namenode old ID
    3. delete current

    for each node in the cluster

    1. just restart the cluster, and check whether all the starts have been successful
      sh start-all-sh

Read More: