All 15 Clustering & Worker Threads questions
More in Clustering & Worker Threads
What is the difference between the 'sticky' and default load-balancing strategies when using Node's cluster module, and when does it matter? What is the difference between the number of CPU cores and the optimal number of cluster workers for a given Node.js application? What is the difference between spawning a Worker Thread and spawning a Child Process for running a CPU-intensive JavaScript function? How do you gracefully terminate a worker thread, and what's the difference between worker.terminate() and letting it exit naturally? How do you use SharedArrayBuffer to share memory directly between Node.js worker threads, and what are the safety considerations? What is a worker thread pool, and why would you build one rather than creating a new Worker for every task? How does zero-downtime restart work with Node's cluster module or PM2, and why is it important for production deployments? How do you handle an uncaught exception inside a specific cluster worker without taking down the entire application? What is the WorkerData option when creating a Worker, and how does it differ from communicating via postMessage() after creation? What is the difference in behavior between the 'primary' and 'worker' role in Node's cluster module regarding which code paths execute?
Advertisement
Node.js › Clustering & Worker Threads

What happens to in-memory state (like an in-memory cache) when using the Node.js cluster module, and how do you handle this correctly?

Advanced

Answer

Each cluster worker is a separate OS process with its own memory -- an in-memory cache exists separately and inconsistently in each worker. The fix is moving shared state to an external store like Redis, accessible consistently regardless of which worker handles a given request.
// Broken with clustering: each worker has its own separate cache
const cache = new Map();

// Correct: external shared store
await redis.set(key, value);
Real-world example A team clustering across four cores noticed roughly 1 in 4 requests seemed to miss recently cached data, traced to an in-memory Map existing separately in each worker; moving to Redis resolved it.

Common follow-ups: What's the correct way to implement rate limiting correctly across cluster workers given this same issue?;How does this apply to WebSocket connections handled by a specific worker?

Caching with Redis;WebSockets & Real-Time Communication
All 15 Clustering & Worker Threads questions