What happend while formatting hadoop namenode

Question

I have setup a single node hadoop in my VM. While doing so, before starting all the daemons, I had run the command to format the namenode:

  bin/hadoop namenode -format

Now, I want to know what exactly happens in hadoop namenode formatting. Basically, I would like to know why this step is requirement during hadoop installation.

Answer 1

Basically, NameNode is the master daemon that manages all the hdfs data and maintains meta data w.r.t. the data stored in hdfs. Meta data are nothing but infrormation about the hdfs data blocks i.e. data block location, the hdfs directory structure, permissions, etc. Now, when you format the NameNode using the command mentioned, you are instructing NameNode to format the meta data info w.r.t. all the data nodes. This allow you to use all the hdfs space as the info regarding the old hdfs data has been deleted.