HBase architecture (RegionServer, HMaster) in Hadoop - Cheat Sheet & Quick Revision

Recall & Review

beginner

What is the role of the HMaster in HBase architecture?

The HMaster manages the cluster, handles metadata operations, assigns regions to RegionServers, and monitors their health. It acts like a manager coordinating the work of RegionServers.

beginner

What is a RegionServer in HBase?

A RegionServer hosts and manages regions, which are subsets of tables. It handles read and write requests for the data in those regions, similar to a worker serving data to users.

intermediate

How does HBase split data across RegionServers?

HBase splits tables into regions based on row keys. Each RegionServer manages one or more regions, distributing data and load evenly across the cluster.

intermediate

What happens if a RegionServer fails in HBase?

The HMaster detects the failure and reassigns the regions managed by the failed RegionServer to other active RegionServers to keep the system running smoothly.

advanced

Why is HMaster considered a single point of control but not a single point of failure?

HMaster controls the cluster but does not serve data directly. If it fails, a backup HMaster or a standby HMaster takes over, so data serving continues without interruption.

What component in HBase assigns regions to RegionServers?

A ZooKeeper

B HMaster

C RegionServer

D HDFS

The HMaster is responsible for assigning regions to RegionServers.

Which component handles read and write requests in HBase?

A HMaster

B DataNode

C NameNode

D RegionServer

RegionServers handle the actual read and write requests for data.

If a RegionServer fails, what does HBase do?

A Reassigns regions to other RegionServers

B Stops all operations

C Deletes the data

D Restarts the HMaster

The HMaster reassigns the regions from the failed RegionServer to other active RegionServers.

What is a region in HBase?

A A client application

B A backup server

C A subset of a table managed by a RegionServer

D A metadata file

A region is a part of a table that a RegionServer manages.

Which system helps HBase maintain high availability of the HMaster?

A Backup HMaster

B Multiple RegionServers

C HDFS replication

D ZooKeeper

A backup HMaster can take over if the active HMaster fails, ensuring high availability.

Explain the roles of HMaster and RegionServer in HBase architecture.

Think of HMaster as the manager and RegionServers as workers.

Describe what happens when a RegionServer fails in an HBase cluster.

Focus on fault tolerance and recovery.

Practice

1. What is the primary role of the HMaster in HBase architecture?

easy

A. To backup data to external storage

B. To store the actual data in tables

C. To serve client read and write requests directly

D. To manage the cluster and assign regions to RegionServers

Solution

1. Step 1: Understand HMaster's function

   The HMaster is responsible for managing the overall HBase cluster, including assigning regions to RegionServers and monitoring their health.

2. Step 2: Differentiate from RegionServer role

   RegionServers handle storing data and serving client requests, not the HMaster.

3. Final Answer:

   To manage the cluster and assign regions to RegionServers -> Option D

4. Quick Check:

   HMaster = Cluster manager [OK]

Hint: HMaster manages, RegionServers serve data [OK]

Common Mistakes:

• Confusing HMaster with RegionServer
• Thinking HMaster stores data
• Assuming HMaster handles client requests

2. Which of the following is the correct statement about RegionServers in HBase?

easy

A. RegionServers manage the cluster and assign regions

B. RegionServers store data and handle client requests

C. RegionServers only backup data to HMaster

D. RegionServers are used to run MapReduce jobs

Solution

1. Step 1: Identify RegionServer responsibilities

   RegionServers are responsible for storing data in regions and serving client read/write requests.

2. Step 2: Exclude incorrect roles

   Cluster management is done by HMaster, not RegionServers. Backup and MapReduce are separate concerns.

3. Final Answer:

   RegionServers store data and handle client requests -> Option B

4. Quick Check:

   RegionServers = Data storage and serving [OK]

Hint: RegionServers serve data, not manage cluster [OK]

Common Mistakes:

• Mixing up RegionServer and HMaster roles
• Thinking RegionServers manage cluster
• Confusing RegionServers with MapReduce nodes

3. Consider the following scenario: A client sends a read request to HBase. Which component handles this request and what happens next?

1. Client sends read request
2. ?

medium

A. RegionServer receives the request and serves the data

B. HMaster receives the request and returns data directly

C. Zookeeper receives the request and forwards to RegionServer

D. HDFS handles the request and sends data to client

Solution

1. Step 1: Identify request handling component

   RegionServers handle client read and write requests by serving data stored in regions.

2. Step 2: Understand role of other components

   HMaster manages cluster but does not serve data directly. Zookeeper coordinates but does not serve data. HDFS stores files but does not handle client requests directly.

3. Final Answer:

   RegionServer receives the request and serves the data -> Option A

4. Quick Check:

   Client read -> RegionServer serves data [OK]

Hint: Client requests go to RegionServer, not HMaster [OK]

Common Mistakes:

• Assuming HMaster serves client data
• Confusing Zookeeper with data serving
• Thinking HDFS directly serves client requests

4. You notice that your HBase cluster is not assigning regions to RegionServers properly. Which of the following is the most likely cause?

medium

A. RegionServers are down or unreachable

B. Zookeeper quorum is not available

C. HMaster is down or not running

D. HDFS storage is full

Solution

1. Step 1: Understand region assignment process

   HMaster is responsible for assigning regions to RegionServers and managing cluster state.

2. Step 2: Analyze failure causes

   If regions are not assigned, it usually means HMaster is down or not running. RegionServers being down affects serving data but not assignment logic. Zookeeper issues cause coordination problems but not direct assignment failure. HDFS full affects storage but not assignment.

3. Final Answer:

   HMaster is down or not running -> Option C

4. Quick Check:

   Region assignment needs HMaster [OK]

Hint: Region assignment depends on HMaster running [OK]

Common Mistakes:

• Blaming RegionServers for assignment issues
• Ignoring HMaster's cluster management role
• Confusing Zookeeper issues with region assignment

5. In a large HBase cluster, you want to improve scalability and fault tolerance. Which architectural approach involving HMaster and RegionServers is best suited for this goal?

hard

A. Use multiple RegionServers but only one HMaster with failover standby

B. Use multiple HMasters in active-active mode and many RegionServers

C. Use a single HMaster and increase RegionServers count

D. Use multiple HMasters with no RegionServers

Solution

1. Step 1: Understand HMaster and RegionServer roles in scalability

   RegionServers store data and serve requests, so increasing their number improves scalability.

2. Step 2: Understand HMaster fault tolerance

   HBase supports one active HMaster with standby HMasters for failover, not active-active multiple HMasters.

3. Step 3: Evaluate options

   Use multiple HMasters in active-active mode and many RegionServers is incorrect because active-active HMasters are not supported. Use a single HMaster and increase RegionServers count lacks fault tolerance for HMaster. Use multiple HMasters with no RegionServers is invalid as RegionServers are needed to store data.

4. Final Answer:

   Use multiple RegionServers but only one HMaster with failover standby -> Option A

5. Quick Check:

   One active HMaster + standby + many RegionServers = scalable & fault tolerant [OK]

Hint: One active HMaster with standby, many RegionServers [OK]

Common Mistakes:

• Thinking multiple active HMasters improve fault tolerance
• Ignoring standby HMaster role
• Assuming RegionServers can replace HMaster