How to manually checkpoint Secondary Name Node
There are scenarios when checkpointing fails ( due to various reasons, for example network issues ). In those cases we can manually run checkpointing to allow NameNode and Secondary NameNode to sync in the metadata. Manual steps to checkpoint is as below: 1. On primary NameNode, we will try to save latest metadata to the fsimage as : hdfs dfsadmin -safemode enter hdfs dfsadmin -saveNamespace hdfs dfsadmin -safemode leave 2. we will now run manual checkpointing: Stop secondary NN or HDFS service or Cluster . Now execute following command on the Secondary NameNode: mv /data/dfs/snn /data/dfs/snn.old hdfs secondarynamenode -checkpoint force Start the SNN or HDFS service or Cluster .