Hadoop Secondary NameNode

Due to its unfortunate naming, the Secondary NameNode  (SNN) is sometimes 

confused with a failover backup for NameNode. It most certainly is not. The SNN 

only serves to periodically clean up and tighten the fi lesystem’s state information in 

NameNode, helping NameNode become more effi cient. NameNode manages the 

fi  lesystem’s state information using two fi les,  FsImage  and EditLog . The fi  le FsImage is 

a snapshot of the fi  lesystem at some checkpoint, and EditLog records each incremental 

change ( delta ) to the fi lesystem after that checkpoint. These two fi les can completely 

determine the current state of the fi  lesystem. When you initialize NameNode, it merges 

these two fi les to create a new snapshot. At the end of NameNode’s initialization, 

FsImage will contain the new snapshot and EditLog will be empty. Afterward any 

operation that changes the state of HDFS is appended to EditLog, whereas FsImage will 

remain unchanged. When you shut down NameNode and restart it, the consolidation 

will take place again and make a new snapshot. Note that the two fi les are only for 

retaining the fi  lesystem’s state information while NameNode is not running (either 

intentionally shut down or due to system malfunction ). NameNode keeps in memory 

a constantly maintained copy of the fi lesystem’s state information  to quickly answer 

queries about the fi lesystem 

Usage：hadoop secondarynamenode [-checkpoint [force]] | [-geteditsize]