HBase Concepts

Question 1

What is a node?

Answer 1

A single computer

Question 2

What is a cluster?

Answer 2

A group of nodes connected and coordinated by certain nodes to perform tasks

Question 3

What is a Master Node?

Answer 3

A node performing coordination tasks

Question 4

What is a Slave Node?

Answer 4

A worker node performing tasks assigned to it

Question 5

What is a Daemon

Answer 5

A process or program that runs in the background

Question 6

Where is table data stored?

Answer 6

In HDFS

Question 7

How is HBase data stored in HDFS?

Answer 7

The data is split into HDFS blocks and stored on multiple nodes in the cluster

Question 8

What is an HBase table split into?

Answer 8

Regions

Question 9

What serves Regions to clients?

Answer 9

Region Servers

Question 10

Can a RegionServer have regions for more than one table?

Answer 10

Yes

Question 11

What is the HBase Master responsible for?

Answer 11

1 - Coordinates which regions are managed by each Region Server
2 - Handles new table creation and other housekeeping operations

Question 12

Can an Hbase cluster have multiple Masters?

Answer 12

Yes, for high availability. But only one can be active at a time.

Question 13

What service handles the coordination of the Masters?

Answer 13

Zookeeper

Question 14

When a cluster has multiple Master, how is the active master determined?

Answer 14

Upon startup all Masters connect to Zookeeper. The first Master to connect, becomes the active master.

Question 15

What happens if the controlling Master fails?

Answer 15

If you have additional master they will compete to run the cluster again.

Question 16

What two servers are typically kept together in the slave nodes?

Answer 16

The data Node and RegionServer

Question 17

List 4 master nodes?

Answer 17

Name Node, Secondary Name Node, Master, Zookeeper

Question 18

What are tables comprised of?

Answer 18

rows, columns and column families

Question 19

How are rows sorted?

Answer 19

They are sorted in rowkey order

Question 20

Can columns in HBase be created on the fly?

Answer 20

Yes

Question 21

If a column for a row does not have a value, does it create a column for that row?

Answer 21

No

Question 22

What is a column family

Answer 22

A collection of columns

Question 23

What is the minimum number of column families that a table must have?

Answer 23

One

Question 24

What delimits the column family from the qualifier

Answer 24

A colon (:)

Question 25

Do all column family members have the same prefix?

Answer 25

Yes. e.g. contactinfo:fname and contactinfo:lname

Question 26

Can you specify the tuning and storage settings at the column family level?

Answer 26

Yes

Question 27

Can you specify the tuning and storage settings at the column level?

Answer 27

No

Question 28

Is there a limit on the number of columns that a column family can have?

Answer 28

No

Question 29

How are columns stored within a column family

Answer 29

Columns within a family are sorted and stored together.

Question 30

Give two examples of when having separte column families are useful?

Answer 30

1 - Data that is not frequently accessed together | 1 - Data that uses different column family options (such as compression)

Question 31

What data type is data in Hbase tables stored as?

Answer 31

A byte array

Question 32

Are empty cells stored?

Answer 32

No

Question 33

How are tables physically stored.

Answer 33

They are stored on a per-column family basis

Question 34

Is there a limit to the data that can be stored in Hbase?

Answer 34

It can store anything that can be serialized into a byte array.

Question 35

In Hbase, what is the equivalent of a primary key?

Answer 35

Rowkey

Question 36

What are the three Hbase Operations?

Answer 36

``` 1 - Get 2 - Scan 3 - Put 4 - Delete 5 - Increment ```

Question 37

What does a Get do?

Answer 37

Retrieves a single row using the row key

Question 38

What does a scan do?

Answer 38

Retrieves all rows

Question 39

How can a scan be constrained?

Answer 39

By specifying a start and end key

Question 40

What does a Put do?

Answer 40

Puts a new row identified by a row key

Question 41

Can you do multiple puts at one time?

Answer 41

Yes

Question 42

What does a delete do?

Answer 42

Marks data as having been deleted. It removes the row identified by the row key.

Question 43

When you do a delete is the data removed from HDFS immediately?

Answer 43

No

Question 44

What is does an increment do?

Answer 44

Allows atomic counters. Allows the value to be initially set or incremented. It can be negatively incremented.

Question 45

What server is responsible for the counters' consistency?

Answer 45

RegionServer

Question 46

How is a increment cell stored?

Answer 46

As a 64-bit integer (a long)

Question 47

What is the default number of versions kept by HBase

Answer 47

3

Question 48

How are versions stored?

Answer 48

They are stored by their timestamp in descending order

Question 49

What is responsible for serving table data

Answer 49

A region server