15. Sharding Architecture: mongos Router, Config Servers & Shard Key Selection • MongoDB & NoSQL Architecture: CRUD to Sharding & WiredTiger • Rishabh Ki Kaksha

MongoDB & NoSQL Architecture: CRUD to Sharding & WiredTiger › 15. Sharding Architecture: mongos Router, Config Servers & Shard Key Selection

50 mins

← Prev Next →

Course Syllabus

Module 1: MongoDB Foundations, BSON & CRUD Mechanics

01. NoSQL vs SQL: CAP Theorem & Document-Oriented Architecture 02. BSON Data Types & Document Schema Design (Embedding vs Referencing) 03. High-Performance CRUD Operations, BulkWrite & Write Concerns 04. Schema Validation with JSON Schema ($jsonSchema)

Module 2: Advanced Aggregation Framework & Data Pipelines

05. Aggregation Pipeline Core: $match, $project, $group, $sort & $limit 06. Complex Transformations: $unwind, $lookup (Joins) & $facet (Multi-Facet Analytics) 07. Window Operators ($setWindowFields) & Accumulator Functions

Module 3: Indexing Strategies & Query Optimization

08. B-Tree Indexes in MongoDB: Single Field, Compound & Partial Indexes 09. Special Indexes: TTL (Time-To-Live), Geospatial (2dsphere) & Text Indexes 10. Query Profiling, explain("executionStats") & Index Covering Analysis

Module 4: Storage Engine Internals (WiredTiger) & ACID Transactions

11. WiredTiger Storage Engine Architecture: Cache, Compression & Checkpoints 12. Read Concerns (local, majority, linearizable) & Write Concerns (w:majority, j:true) 13. Multi-Document Distributed ACID Transactions with Sessions

Module 5: High Availability, Replica Sets, Sharding & Backups

14. Replica Sets Architecture: Primary Election, Oplog & Automatic Failover 15. Sharding Architecture: mongos Router, Config Servers & Shard Key Selection 16. Range vs Hash Sharding, Chunk Balancing & Jumbo Chunks Mitigation 17. Change Streams: Building Real-Time Event-Driven Pipelines with MongoDB 18. Production Backup Strategies: mongodump/mongorestore, Point-in-Time & Atlas CLI

Interactive Lesson Studio

# 15. Sharding Architecture: mongos Router, Config Servers & Shard Key Selection

> MongoDB Notes by Rishabh - Topic 15

---

### 1. Horizontal Scaling via Sharding (Petabyte Scale 🌐)

Jab data 1 single server ki disk/RAM limit (e.g. > 4TB) cross kar jaye:
*   **Sharding** data ko multiple independent replica sets (**Shards**) me divide kar deta hai.

```
                    [ Client Applications ]
                               ⬇
                     [ mongos Query Router ]
                     (Stateless routing proxy)
                     /          |          \
                    ⬇           ⬇           ⬇
          [ Config Server ]  [ Shard 1 ]  [ Shard 2 ]
          (Metadata Catalog) (Data 0-50M) (Data 50M-100M)
```

---

### 2. The 3 Core Components of a Sharded Cluster:

1. **`mongos` (Query Router):** Stateless routing engine jo client queries ko right shard par dispatch karta hai.
2. **Config Servers (3-Node Replica Set):** Cluster ka master map store karta hai (Kaunsa chunk kaunse shard par hai).
3. **Shards:** Actual data store karne wale independent Replica Sets.

---

### 3. Shard Key Selection: The #1 Architectural Decision 👑

*   **Targeted Query (Good):** Query me Shard Key present hai (`db.users.find({ country: "IN" })`) ➔ `mongos` direct Shard 1 ko request bhejta hai.
*   **Scatter-Gather Query (Bad):** Query me Shard Key missing hai (`db.users.find({ age: 25 })`) ➔ `mongos` saare shards par broadcast bhejta hai aur results merge karta hai (Slow!).

---

### 4. Master Comparison Matrix

---

### 5. Interview Ke 3 Sawal Is Topic Se

1. **Auto-incrementing timestamp ya `_id` par range sharding kyu fail ho jaati hai?** -> Monotonic Hotspotting! Har naya record sabse highest range me aata hai, jisse saare writes 1 hi shard par girte hain aur baki shards idle rehte hain.
2. **MongoDB 5.0+ me Shard Key ko re-shard ya change kiya ja sakta hai kya?** -> Haan! MongoDB 5.0 me `reshardCollection` command live zero-downtime re-sharding support karti hai.
3. **`mongos` instances kitne deploy karne chahiye?** -> Har application server par ya load balancer ke piche multiple stateless `mongos` instances chalaye jaate hain taaki router single point of failure na bane.

---
**Next Topic:** 16. Range vs Hash Sharding, Chunk Balancing & Jumbo Chunks Mitigation

Loading lesson content...

Python 3.12

Full IDE

Live Execution Terminal

Click "Execute Live" to run code.
                

Lesson Mastery AI Quiz

AI Doubt Solver

CERTIFICATE OF COMPLETION

Rishabh Gehlot

PHP 8+ & Laravel 11 Master Track