Horizontal scaling mental model in MongoDB

Introduction

Horizontal scaling means adding more machines to handle more data or users. It helps systems grow smoothly without slowing down.

When your database grows too big for one server to handle.

When many users access your app at the same time and one server gets overloaded.

When you want to avoid downtime by spreading data across multiple servers.

When you want to improve read and write speeds by dividing the work.

When you want to easily add more servers as your app becomes popular.

Syntax

MongoDB

sh.enableSharding("databaseName")
sh.shardCollection("databaseName.collectionName", { shardKey: 1 })

Use sh.enableSharding to turn on sharding for a database.

Use sh.shardCollection to split a collection using a shard key.

Examples

This enables sharding on the shopDB database and shards the orders collection by orderId.

MongoDB

sh.enableSharding("shopDB")
sh.shardCollection("shopDB.orders", { orderId: 1 })

This sets up horizontal scaling for user profiles by splitting data based on userId.

MongoDB

sh.enableSharding("userDB")
sh.shardCollection("userDB.profiles", { userId: 1 })

Sample Program

This example shows how to set up horizontal scaling in MongoDB by adding servers as shards, enabling sharding on a database, choosing a shard key, and letting MongoDB distribute data automatically. Queries are routed to the correct server based on the shard key.

MongoDB

/* Step 1: Add shards (servers) to the cluster */
sh.addShard("server1:27017")
sh.addShard("server2:27017")

/* Step 2: Enable sharding on the database */
sh.enableSharding("ecommerceDB")

/* Step 3: Choose a shard key and shard the collection */
sh.shardCollection("ecommerceDB.products", { productId: 1 })

/* Step 4: MongoDB automatically distributes data based on productId */

/* Step 5: When a query comes in, MongoDB routes it to the right shard */

Important Notes

Choosing the right shard key is very important for balanced data distribution.

Too many writes to one shard can cause a hotspot and slow down the system.

MongoDB manages routing and balancing automatically once sharding is set up.

Summary

Horizontal scaling means adding more servers to share the load.

MongoDB uses sharding to split data across servers based on a shard key.

This helps handle more data and users without slowing down.

Practice

1. What does horizontal scaling in MongoDB primarily involve?

easy

A. Adding more servers to share the data load

B. Increasing the CPU power of a single server

C. Using faster disks on the existing server

D. Compressing data to save space

Solution

1. Step 1: Understand horizontal scaling concept

   Horizontal scaling means adding more machines to handle more data or users.

2. Step 2: Apply to MongoDB context

   MongoDB uses sharding to distribute data across multiple servers, which is horizontal scaling.

3. Final Answer:

   Adding more servers to share the data load -> Option A

4. Quick Check:

   Horizontal scaling = adding servers [OK]

Hint: Horizontal scaling means more servers, not bigger servers [OK]

Common Mistakes:

• Confusing horizontal scaling with vertical scaling
• Thinking faster disks alone is horizontal scaling
• Mixing data compression with scaling

2. Which MongoDB feature is used to implement horizontal scaling?

easy

A. Sharding

B. Indexing

C. Replication

D. Aggregation

Solution

1. Step 1: Recall MongoDB scaling features

   Replication copies data for backup and availability, not scaling data horizontally.

2. Step 2: Identify horizontal scaling feature

   Sharding splits data across servers to scale horizontally.

3. Final Answer:

   Sharding -> Option A

4. Quick Check:

   Horizontal scaling = Sharding [OK]

Hint: Sharding splits data across servers horizontally [OK]

Common Mistakes:

• Confusing replication with sharding
• Thinking indexing scales data horizontally
• Mixing aggregation with scaling

3. Given a MongoDB cluster with 3 shards, each holding 1 million documents, what happens when you add a 4th shard?

medium

A. The cluster stops working until manually rebalanced

B. New data only goes to the 4th shard, old data stays on first 3

C. All data duplicates on the 4th shard

D. Data automatically redistributes to balance across all 4 shards

Solution

1. Step 1: Understand shard addition effect

   Adding a shard triggers MongoDB to redistribute chunks to balance data load.

2. Step 2: Analyze options

   Data automatically redistributes to balance across all 4 shards matches MongoDB's automatic balancing; others describe incorrect behaviors.

3. Final Answer:

   Data automatically redistributes to balance across all 4 shards -> Option D

4. Quick Check:

   Shard addition triggers auto-balancing [OK]

Hint: Adding shards triggers automatic data balancing [OK]

Common Mistakes:

• Assuming new shard only gets new data
• Thinking data duplicates on new shard
• Believing manual rebalance is always required

4. You configured a MongoDB sharded cluster but notice all queries hit only one shard. What is the likely cause?

medium

A. Replication is not enabled

B. Shard key is not properly chosen or missing

C. Indexes are missing on all shards

D. The cluster has too many shards

Solution

1. Step 1: Understand shard key role

   The shard key determines how data is distributed and which shard handles queries.

2. Step 2: Diagnose query routing issue

   If all queries hit one shard, shard key is likely missing or poorly chosen, causing no data distribution.

3. Final Answer:

   Shard key is not properly chosen or missing -> Option B

4. Quick Check:

   Shard key controls query distribution [OK]

Hint: Proper shard key ensures queries spread across shards [OK]

Common Mistakes:

• Blaming replication for query routing
• Assuming indexes affect shard routing
• Thinking too many shards cause single shard hits

5. You have a MongoDB cluster with 5 shards. Your application experiences uneven load because one shard stores much more data than others. What is the best solution to fix this?

hard

A. Increase RAM on the overloaded shard server

B. Add more replicas to the overloaded shard

C. Choose a better shard key that evenly distributes data

D. Disable sharding and use a single server

Solution

1. Step 1: Identify cause of uneven load

   Uneven data distribution usually means the shard key does not distribute data evenly.

2. Step 2: Select best fix

   Changing to a better shard key balances data and load across shards, improving performance.

3. Final Answer:

   Choose a better shard key that evenly distributes data -> Option C

4. Quick Check:

   Balanced shard key = even load [OK]

Hint: Good shard key balances data evenly across shards [OK]

Common Mistakes:

• Adding replicas does not fix data imbalance
• Increasing RAM only helps performance, not distribution
• Disabling sharding removes scaling benefits