Why Sentinel provides high availability in Redis - Visual Breakdown

Concept Flow - Why Sentinel provides high availability

Start: Redis Master and Slaves Running

↓

Sentinel Monitors Master

↓

Detects Master Failure?

No→Continue Monitoring

Yes↓

Sentinel Elects New Master from Slaves

↓

Update Clients to Use New Master

↓

Resume Normal Operations with New Master

Sentinel watches the Redis master. If it fails, Sentinel picks a new master from slaves and tells clients to use it, keeping the system available.

Execution Sample

Redis

1. Sentinel monitors master
2. Detects failure
3. Elects new master
4. Updates clients
5. System continues

This sequence shows how Sentinel detects failure and switches to a new master to keep Redis available.

Execution Table

Step	Sentinel Action	Condition	Result	System State
1	Monitor master	Master is healthy	No fail detected	Master active, slaves syncing
2	Monitor master	Master not responding	Fail detected	Master down
3	Start failover	Fail confirmed by quorum	Elect new master	One slave promoted
4	Update clients	New master elected	Clients redirected	Clients connected to new master
5	Resume monitoring	New master active	Normal operation	High availability restored

💡 Sentinel stops failover after new master is elected and clients are updated

Variable Tracker

Variable	Start	After Step 2	After Step 3	After Step 4	Final
Master Status	Healthy	Down	Down	Down	New Master Active
Failover State	None	Fail Detected	Failover In Progress	Failover Done	None
Client Connection	To Master	To Master (Failing)	Switching	To New Master	To New Master

Key Moments - 2 Insights

Why does Sentinel wait for quorum before electing a new master?

Sentinel waits for quorum to avoid false failover; this ensures multiple Sentinels agree the master is down (see execution_table step 3).

What happens if clients are not updated after failover?

Clients would keep trying the old master and fail; updating clients (step 4) is crucial to maintain availability.

Visual Quiz - 3 Questions

Look at the execution_table at which step does Sentinel detect the master failure?

A Step 1

B Step 3

C Step 2

D Step 4

💡 Hint

Check the 'Condition' and 'Result' columns in execution_table row for Step 2

According to variable_tracker, what is the client connection state after step 4?

A Connected to new master

B Switching connection

C Connected to old master

D Disconnected

💡 Hint

Look at 'Client Connection' row under 'After Step 4' in variable_tracker

If Sentinel did not update clients after failover, what would happen?

A Clients connect to new master automatically

B Clients fail to connect and lose availability

C Clients continue normal operation

D Sentinel restarts the master

💡 Hint

Refer to key_moments explanation about client updates after failover

Concept Snapshot

Sentinel monitors Redis master continuously.
If master fails, Sentinel confirms failure with quorum.
Sentinel elects a slave as new master.
Clients are updated to use new master.
This automatic failover keeps Redis highly available.

Full Transcript

Redis Sentinel provides high availability by monitoring the master server. It continuously checks if the master is working. When Sentinel detects the master is down, it waits for agreement from other Sentinels (quorum) to avoid mistakes. Then, Sentinel chooses one of the slaves to become the new master. After that, Sentinel updates all clients to connect to this new master. This process ensures Redis keeps working without manual intervention, maintaining high availability.

Practice

1. What is the main reason Redis Sentinel provides high availability?

easy

A. It compresses data to save storage space.

B. It increases the speed of Redis queries.

C. It encrypts Redis data for security.

D. It monitors Redis servers and promotes replicas automatically on failure.

Solution

1. Step 1: Understand Sentinel's role in monitoring

   Sentinel continuously watches Redis servers to detect if the master fails.

2. Step 2: Recognize automatic failover

   When the master fails, Sentinel promotes a replica to master automatically to keep service running.

3. Final Answer:

   It monitors Redis servers and promotes replicas automatically on failure. -> Option D

4. Quick Check:

   Sentinel = Automatic failover [OK]

Hint: Sentinel watches and switches masters automatically [OK]

Common Mistakes:

• Confusing Sentinel with data compression tools
• Thinking Sentinel speeds up queries
• Assuming Sentinel only encrypts data

2. Which of the following is the correct way to configure Redis Sentinel to monitor a master named mymaster on port 6379?

easy

A. sentinel monitor mymaster 127.0.0.1 6379 2

B. sentinel watch mymaster 127.0.0.1 6379 2

C. sentinel track mymaster 127.0.0.1 6379 2

D. sentinel check mymaster 127.0.0.1 6379 2

Solution

1. Step 1: Recall Sentinel configuration command

   The correct command to start monitoring a master is sentinel monitor.

2. Step 2: Verify command syntax

   The syntax is sentinel monitor <name> <ip> <port> <quorum>, matching sentinel monitor mymaster 127.0.0.1 6379 2.

3. Final Answer:

   sentinel monitor mymaster 127.0.0.1 6379 2 -> Option A

4. Quick Check:

   Monitor command = sentinel monitor [OK]

Hint: Use 'sentinel monitor' to watch a master server [OK]

Common Mistakes:

• Using 'watch' instead of 'monitor'
• Confusing 'track' or 'check' commands
• Wrong command syntax order

3. Given a Redis Sentinel setup monitoring a master and two replicas, what will happen if the master server stops responding?

medium

A. Sentinel will promote one replica to master automatically.

B. Sentinel will do nothing and wait for manual restart.

C. Sentinel will shut down all Redis servers.

D. Sentinel will delete all data to prevent corruption.

Solution

1. Step 1: Understand Sentinel's failure detection

   Sentinel detects when the master is down by monitoring its health.

2. Step 2: Recognize Sentinel's failover action

   Upon detecting failure, Sentinel promotes a replica to master automatically to maintain availability.

3. Final Answer:

   Sentinel will promote one replica to master automatically. -> Option A

4. Quick Check:

   Failover = Replica promoted [OK]

Hint: Master down? Sentinel promotes replica automatically [OK]

Common Mistakes:

• Thinking Sentinel waits for manual restart
• Assuming Sentinel shuts down servers
• Believing Sentinel deletes data

4. You configured Redis Sentinel but it never promotes a replica after the master fails. What is the most likely cause?

medium

A. The replicas have more memory than the master.

B. The Redis version is too new for Sentinel to work.

C. Sentinel quorum is not reached due to insufficient Sentinel nodes.

D. Sentinel does not support automatic failover.

Solution

1. Step 1: Understand Sentinel quorum requirement

   Sentinel requires a quorum (minimum number of Sentinels agreeing) to trigger failover.

2. Step 2: Identify cause of no failover

   If quorum is not reached, Sentinel will not promote replicas automatically.

3. Final Answer:

   Sentinel quorum is not reached due to insufficient Sentinel nodes. -> Option C

4. Quick Check:

   Failover needs quorum [OK]

Hint: Check Sentinel quorum count if failover doesn't happen [OK]

Common Mistakes:

• Blaming replica memory size
• Assuming Redis version blocks Sentinel
• Thinking Sentinel lacks failover feature

5. In a Redis Sentinel cluster with 5 Sentinel nodes monitoring 3 replicas and 1 master, what happens if 2 Sentinels lose network connectivity but the master fails?

hard

A. All replicas are shut down to avoid split-brain.

B. Failover proceeds because quorum of Sentinels still agree on master failure.

C. Sentinel automatically adds new nodes to replace disconnected ones.

D. Failover is blocked because any Sentinel failure stops the process.

Solution

1. Step 1: Calculate quorum in a 5-node Sentinel cluster

   Quorum is usually a majority; with 5 nodes, quorum is 3.

2. Step 2: Assess failover with 2 disconnected Sentinels

   3 Sentinels remain connected, so quorum is met and failover proceeds.

3. Step 3: Understand Sentinel behavior on partial network loss

   Sentinel does not block failover if quorum is reached despite some nodes being disconnected.

4. Final Answer:

   Failover proceeds because quorum of Sentinels still agree on master failure. -> Option B

5. Quick Check:

   Quorum met = failover proceeds [OK]

Hint: Failover needs majority Sentinels, not all connected [OK]

Common Mistakes:

• Thinking any Sentinel loss blocks failover
• Assuming Sentinel auto-adds nodes
• Believing replicas shut down on master failure