Ideal replication factor in Hadoop is 3. However the replication factor can be changed in hdfs-site configuration file.

Reason for 3 being the ideal replication factor: Multiple copies of data blocks or input splits are stored in different nodes of the Hadoop cluster. Because of the Hadoop HDFS’s rack awareness, copies of data are stored in Datanodes of different racks as well. So, even if the whole rack goes down, data blocks can be accessed from the nodes on different rack. At any given point of time, if one Datanode goes down and one Datanode is not available, there will always be a third Datanode on which the required data block is stored.

Even i also need same help with big data hadoop  

if one Datanode goes down and one Datanode is not available

good information

 The default replication factor is 3 which can be configured as per the requirement; it can be changed to 2 (less than 3) or can be increased (more than 3.).

Increasing/Decreasing the replication factor in HDFS has a impact on Hadoop cluster performance.

Hi, while this has been answered, I wanted to add a bit more detail explanation. 

HDFS has a default replication factor of 3. 

Files are broken down into blocks and each block is replicated (3)times and stored in various points on the cluster.  This is to make the data more resilient.  You can change the replication factor to whatever you want.  If you lower the replication factor you increase the risk of losing data.  If you increase the number of replicas you decrease the risk of losing data, however, you are using more storage and it takes longer to write the files.  

While you can increase the replication factor for 'golden data', its also possible to make copies of the files and to also copy them off cluster to another cluster or cold storage if the data is valuable.