I am very curious to to know how HDFS reserves and allocates storage space on a Data node.
Say if I have 500GB hard disk in one of the Data node, out of it 400GB is allocated for /data partition and hadoop was said to sit on it.
Here how much space out of 400GB will be taken by hadoop daemons for the storage of the input splits(blocks).
Does it reserve all the storage space priorly or does it allocate on demand?.
and also wanted to know whether formatting Name node while setting up the cluster has anything to do with this.
Many Thanks...