HDFS

Question 1

How are block replicas are assigned to node

Answer 1

file is split up into 128 blocks. Replicate each block 3 times and each replica is sent to different machines.

Question 2

Calculate how much network overhead is incurred when reading/writing files in HDFS.

Question 3

which information is stored on the name node and data nodes.

Answer 3

name node: the file and directory names, replication factor, list of all data nodes
data node: contents of the file

Question 4

What does HDFS do to account for hardware failures

Answer 4

Replicate the data to multiple nodes.

Question 5

What is write-once-read-many access model

Answer 5

HDFS can write a file only once but you can read it many times.

Question 6

What is the most common replication factor in HDFS

Answer 6

3

Question 7

Which of the following are restrictions imposed by HDFS

Answer 7

Once a file is closed, its contents can no longer be modified