惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

D
Docker
爱范儿
爱范儿
人人都是产品经理
人人都是产品经理
博客园 - 司徒正美
cs.AI updates on arXiv.org
cs.AI updates on arXiv.org
量子位
罗磊的独立博客
cs.CL updates on arXiv.org
cs.CL updates on arXiv.org
小众软件
小众软件
C
Cybersecurity and Infrastructure Security Agency CISA
Cyberwarzone
Cyberwarzone
大猫的无限游戏
大猫的无限游戏
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
雷峰网
雷峰网
Simon Willison's Weblog
Simon Willison's Weblog
The Cloudflare Blog
博客园 - 三生石上(FineUI控件)
D
Darknet – Hacking Tools, Hacker News & Cyber Security
C
Cyber Attacks, Cyber Crime and Cyber Security
博客园_首页
博客园 - 叶小钗
V
Vulnerabilities – Threatpost
T
The Exploit Database - CXSecurity.com
T
Tailwind CSS Blog
IT之家
IT之家
博客园 - 聂微东
Spread Privacy
Spread Privacy
V2EX - 技术
V2EX - 技术
S
Security Affairs
宝玉的分享
宝玉的分享
V
V2EX
C
Cisco Blogs
博客园 - Franky
美团技术团队
酷 壳 – CoolShell
酷 壳 – CoolShell
月光博客
月光博客
S
Securelist
J
Java Code Geeks
Webroot Blog
Webroot Blog
让小产品的独立变现更简单 - ezindie.com
让小产品的独立变现更简单 - ezindie.com
P
Proofpoint News Feed
Last Week in AI
Last Week in AI
L
LINUX DO - 热门话题
NISL@THU
NISL@THU
WordPress大学
WordPress大学
W
WeLiveSecurity
T
Threatpost
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
腾讯CDC
阮一峰的网络日志
阮一峰的网络日志

DEV Community

Authentication Security Deep Dive: From Brute Force to Salted Hashing (With Java Examples) Why AI Systems Don’t Fail — They Drift Spilling beans for how i learn for exam😁"Reinforcement Learning Cheat Sheet" I Replaced Chrome with Safari for AI Browser Automation. Here's What Broke (and What Finally Worked) How Python Borrows Other People's Work The $40 Architecture: Processing 1 Billion API Requests with 99.99% Uptime Vibe Coding: A Workflow Guide (From Zero to SaaS) Most webhook security guides protect the wrong side. The scary part is delivery. Headless CMS for TanStack Start: Build a Blog with Cosmic EU Age Verification App "Hacked in 2 Minutes" — What Actually Happened Comfy Cloud’s delete function does not actually remove files Running AI Models on GPU Cloud Servers: A Beginner Guide Event-driven media intelligence with AWS Step Functions and Bedrock I scored 500 AI prompts across 8 quality dimensions — here's what broke How to Call Google Gemini API from Next.js (Free Tier, No Backend Needed) The Portal Protocol: Reclaiming Human Connection in the Age of AI How to Fix Your Team's Scattered Knowledge Problem With a Self-Hosted Forum Intro to tc Cloud Functors: A Graph-First Mental Model for the Modern Cloud Designing Multi-Tenant Backends With Both Ownership and Team Access I Built a Neumorphic CSS Library with 77+ Components — Here's What I Learned PostgreSQL Performance Optimization: Why Connection Pooling Is Critical at Scale Cómo construí un SaaS multi-rubro para gestionar expensas en Argentina con FastAPI + Vue 3 🚀 I Built an Ethical Hacking Scanner Tool – Open Source Project I Replaced /usage and /context in Claude Code With a Single Statusline A Pythonic Way to Handle Emails (IMAP/SMTP) with Auto-Discovery and AI-Ready Design I Collected 8.9 Million Polymarket Price Points — Here's What I Found About How Markets Really Move EcoTrack AI — Carbon Footprint Tracker & Dashboard Everyone's Using AI. No One Agrees How. 5 self-hosted ebook managers worth trying in 2026 Building Your First AI Agent with LangChain: From Chatbot to Autonomous Assistant Common SOC 2 Failures (Real World) Stop Vibe-Checking Your AI App: A Practical Guide to Evals How to Use SonarQube and SonarScanner Locally to Level Up Your Code Quality Your Next To-Do App Is Dead — I Replaced Mine with an OpenClaw AI Sign a Nostr event in 60 lines of Python using coincurve — no nostr-sdk, no nbxplorer, no rust toolchain ITGC Audit Explained Like You’re in Big 4 Patch Tuesday abril 2026: Microsoft parcha 163 vulnerabilidades y un zero-day en SharePoint Stop scraping everything: a better way to track competitor price changes Listing on MCPize + the Official MCP Registry while routing payments OUTSIDE the marketplace — how I kept 100% of my x402 revenue Building an AI-Powered Risk Intelligence System Using Serverless Architecture Why We Ripped Function Overloading Out of Our AI Toolchain Testing AI-Generated Code: How to Actually Know If It Works SaaS Churn Is Killing Your Business. Here Is What to Do About It (Without a Support Team) The Speed of AI Is No Longer Linear - And Self-Improving Models Are Why How to Implement RBAC for MCP Tools: A Practical Guide for Engineering Teams From Standard Quote to Persuasive Proposal: AI Automation for Arborists I built a CLI that scaffolds complete multi-tenant SaaS apps Axios CVE-2025–62718: The Silent SSRF Bug That Could Be Hiding in Your Node.js App Right Now The dashboard that ended our friendship Data Pipelines Explained Simply (and How to Build Them with Python) The Hidden Cost of AI Systems Nobody Talks About. undefined vs undeclared, and how typeof behaves Switching from file-based jobs to NATS/Kafka in Rust without changing code io_uring Adventures: Rust Servers That Love Syscalls Why Agentic AI is Killing the Traditional Database The POUR principles of web accessibility for developers and designers Quantum Neural Network 3D — A Deep Dive into Interactive WebGL Visualization How To Install Caveman In Codex On macOS And Windows Automation Pipeline Reliability: Why Your Workflow Breaks When Nobody Is Watching I Built an 'Open World' AI Coding Agent — It Works From ANY Folder From Freelancing to Product: A Tech Service Company's SaaS Transformation China's AI Giants: Adding Tencent Hunyuan & ByteDance Doubao to AI University (74 Providers) On the Vibe Coders and Their Lies clerk: Auto-Summarize Your Claude Code Sessions AI Weekly — 2026/04/10–04/17 | The Model Lockdown Is Here, but the Toolchain Is the Real Battleground AI 週報 — 2026/04/10–2026/04/17 模型封鎖潮來了,但工具鏈才是真戰場 Maybe this is how Open-Source apps are born... 🚀 Fine-Tune LLMs with LoRA and QLoRA: 2026 Guide tRPC v11 + Next.js App Router: End-to-End Type Safety Without the Boilerplate ShadCN UI in 2026: Why I Stopped Installing Component Libraries and Started Owning My Components SaaS Billing in React Server Components: Stripe + Supabase Without a Single `useEffect` Join our DEV Weekend Challenge — $1,000 in Prizes Across TEN winners! Submissions Due April 20 at 6:59 AM UTC. Implementing FSRS Spaced Repetition in Flutter + Supabase — Adding Memory Science to an AI Learning App "I Texted My Localhost From the Train — Claude Code Fixed the Bug Before I Got Home" I Built a Sales Prep AI and It Went Deeper Than Expected Design to Code #2: One JSON, Eleven Outputs Solving the 100M-Row Problem: A Summary Table Pattern for High-Volume Push Notification Logs Flutter Web With Wasm: What Actually Changes For Developers I Built 50 Royalty-Free Soundtracks for My Side Project in a Weekend Using AI Music Generation The Vibe Coding Security Checklist: 7 Things to Check Before You Ship Stop Letting Googlebot Guess Fix Your React App's SEO Right Desconstruindo o Streaming do LinkedIn: Como Criar um Engine de Extração de Vídeo de Alta Performance com HLS e FFmpeg (EDA Part-1) EDA (Exploratory Data Analysis) Explained With Real Life — Why Looking at Your Data Is the Most Important Step in Machine Learning Brand Relationship Management at Scale: Our 4-Touch Outreach System for 200+ Brands Why String.fromEnvironment() Might Return an Empty String in Dart JGuardrails 1.0.0 — Hardening Java LLM Apps Against Jailbreaks, Toxicity, and Prompt Injection Plan and Schedule a Full Week of Threads Content From One Claude Conversation Coding Cat Oran Ep3, Five Tables Changed Everything Updated: BFF Pattern I'm done watching freelancers get buried by 200 proposals. So I'm building the alternative. This is my first post BFS Algorithm in Java Step by Step Tutorial with Examples Tracking LLM Pricing Monthly: An Open Dataset for 22 AI Models How We Measure Content ROI on a Comparison Site: Revenue Attribution Without Perfect Data Introducing Nova AI Ops: The AI-Native Operating System for SRE Teams I built a free desktop video downloader for Windows — Grabbit How Talkie OCR Helps Vision-Impaired & Dyslexic Users Read the World Around Them VRCFaceTracking安装和iPhone面捕配置教程,有bug Even CrowdStrike Can't See Your Agents The Automation Gold Rush: What n8n Workflows and Claude Are Opening Up for Developers Right Now
The SSR State Management Trap: defineStore vs defineContextStore
reactuse.com · 2026-05-06 · via DEV Community

You ship a Next.js-style SSR app with a global user store. A month later, a user files a bug: "I logged in and saw someone else's profile." You can't reproduce it locally. Production logs are useless. The session cookie looks fine.

What happened: your global store is a singleton on the server. Under concurrent requests, one request's setUser() wrote to the same object another request was reading. The second user hit a warm server, got a hydration payload containing the first user's data, and saw it for a split second before React reconciled.

This is the SSR state management trap. It's one of the oldest footguns in server rendering, and almost every framework makes it possible. In this post: how it happens at the module level, how Pareto's defineStore and defineContextStore address it, and the decision rule for which one to use.

Why global state leaks across requests on the server

In the browser, your app is one process per user. A module-level variable is a per-user variable — there's no one else sharing it.

On the server, one Node process serves thousands of users. A module-level variable is shared across every request. If request A mutates it during its render, and request B reads it during its render, B sees A's state.

// state.ts (imported by server and client)
let currentUser = null

export function setCurrentUser(user) { currentUser = user }
export function getCurrentUser() { return currentUser }

Enter fullscreen mode Exit fullscreen mode

Works in the browser. On the server, two requests racing through:

t=0ms:  Request A starts, calls setCurrentUser({ id: 'alice' })
t=1ms:  Request B starts, calls setCurrentUser({ id: 'bob' })
t=2ms:  Request A reads getCurrentUser() → { id: 'bob' }  ← leaked

Enter fullscreen mode Exit fullscreen mode

Alice's rendered HTML contains Bob's name. Alice's hydration payload contains Bob's data. Whatever the server sends to Alice's browser now references Bob.

This doesn't require a state library. A global Map, a module.cache, a let declaration — all vulnerable. State libraries just dress up the same pattern.

What Pareto's two store types do

Pareto has two APIs that look almost identical but have opposite SSR behavior.

defineStore — one instance per process

import { defineStore } from '@paretojs/core/store'

const themeStore = defineStore((set) => ({
  mode: 'light' as 'light' | 'dark',
  toggle: () => set((d) => {
    d.mode = d.mode === 'light' ? 'dark' : 'light'
  }),
}))

Enter fullscreen mode Exit fullscreen mode

On the server, there's one themeStore per Node process. Every SSR request reads and writes the same state. This is the dangerous primitive.

On the client, there's one themeStore per browser tab — which is what you actually want for client-side state.

defineContextStore — one instance per React render

import { defineContextStore } from '@paretojs/core/store'

const { Provider, useStore } = defineContextStore((initialUser) => (set) => ({
  user: initialUser,
  setUser: (user) => set((d) => { d.user = user }),
}))

Enter fullscreen mode Exit fullscreen mode

Context stores are scoped to a <Provider>. Each SSR request renders its own provider tree, gets its own store instance, and can't see any other request's state.

function App({ user }) {
  return (
    <Provider initialData={user}>
      <Dashboard />
    </Provider>
  )
}

Enter fullscreen mode Exit fullscreen mode

Two concurrent requests create two isolated stores. Alice's render never touches Bob's store. The pattern is guaranteed safe.

The decision rule

Ask: "Does the initial state of this store differ per user or per request?"

  • Yes — use defineContextStore. User identity, auth tokens, per-tenant config, feature flags that depend on the viewer, shopping cart tied to a session.
  • NodefineStore is fine. UI theme, sidebar open/closed, recently-viewed modals, client-side caches keyed on data that's already safe to share.

For the "no" case, the reasoning is: if every SSR request reads the same initial state, there's nothing for concurrent requests to leak to each other. The theme is 'light' for every new request until JavaScript hydrates and the client reads localStorage. No per-request state, no leak surface.

If you're ever unsure, pick defineContextStore. It's always safe. You can downgrade to defineStore later once you've confirmed the state is truly request-independent.

A leaky example and its fix

Here's a realistic pattern that leaks:

// ❌ Bad: user identity in a global store
import { defineStore } from '@paretojs/core/store'

export const userStore = defineStore((set) => ({
  user: null as User | null,
  setUser: (user: User) => set((d) => { d.user = user }),
  isAdmin: false,
}))

Enter fullscreen mode Exit fullscreen mode

Then in a loader or layout:

// ❌ Mutating a global store during SSR
export async function loader(ctx: LoaderContext) {
  const user = await getUserFromSession(ctx.request)
  userStore.setState((d) => {
    d.user = user
    d.isAdmin = user.role === 'admin'
  })
  return { user }
}

Enter fullscreen mode Exit fullscreen mode

This leaks in both directions. Concurrent requests overwrite d.user. Requests that arrive after a previous user's session read stale data. The bug looks like random user-mixup reports.

The fix:

// ✅ Good: context store, scoped to the render
import { defineContextStore } from '@paretojs/core/store'

export const { Provider: UserProvider, useStore: useUser } =
  defineContextStore((initial: { user: User; isAdmin: boolean }) => (set) => ({
    user: initial.user,
    isAdmin: initial.isAdmin,
    setUser: (user: User) => set((d) => { d.user = user }),
  }))

Enter fullscreen mode Exit fullscreen mode

Wrap the app in a provider in the root layout:

// app/layout.tsx
import { UserProvider } from './stores/user'

export default function RootLayout({ children, loaderData }) {
  const { user, isAdmin } = loaderData as LoaderData
  return (
    <UserProvider initialData={{ user, isAdmin }}>
      {children}
    </UserProvider>
  )
}

Enter fullscreen mode Exit fullscreen mode

Now every request gets its own store instance, initialized from its own loader data. No shared object. No leak path.

When defineStore is the right call

Not every store is dangerous. Plenty of client-state is genuinely the same on every SSR request and only diverges after hydration.

UI preferences

export const uiStore = defineStore((set) => ({
  sidebarOpen: true,
  commandPaletteOpen: false,
  toggleSidebar: () => set((d) => { d.sidebarOpen = !d.sidebarOpen }),
}))

Enter fullscreen mode Exit fullscreen mode

On the server, every request starts with sidebarOpen: true. No user data, no request-dependent initial state. Safe.

Client-only feature toggles

export const devStore = defineStore((set) => ({
  showDebugOverlay: false,
  toggleDebugOverlay: () => set((d) => { d.showDebugOverlay = !d.showDebugOverlay }),
}))

Enter fullscreen mode Exit fullscreen mode

Debug UI that only the developer uses. Initial state is identical for every request; all mutations happen on the client.

The "I'll just reset it at the start of each request" trick

A pattern people try when they realize globals leak: reset the store at the start of every SSR request.

// ❌ Doesn't work
export async function loader(ctx) {
  userStore.setState(() => ({ user: null, isAdmin: false }))
  const user = await getUserFromSession(ctx.request)
  userStore.setState((d) => { d.user = user })
  return { user }
}

Enter fullscreen mode Exit fullscreen mode

This has two problems:

  1. Reset and populate are not atomic. Another request can interleave between them and see the reset state — or the first request's populated state before its own reset.
  2. Multiple renders share one store. Pareto's streaming SSR holds a response open while deferred promises resolve. During that window, another request can run through the same loader and overwrite the store.

The "reset per request" pattern is what AsyncLocalStorage exists to solve, and what context stores are in Pareto. Don't try to work around it with manual resets.

What about Zustand / Jotai / Redux?

Same principle. If the library's default API is a module-level store, it's unsafe on the server. Most of these libraries document an "SSR mode" that requires you to instantiate a per-request store and pass it through context.

Pareto's defineContextStore is that pattern as a first-class API. If you're coming from Zustand, defineContextStore is the shape you already know from Zustand's createStore + React context wrapper — just built in.

Pareto's defineStore is the global-singleton shortcut for cases where you've verified the shortcut is safe.

The bigger pattern

SSR adds a dimension most state management libraries weren't originally designed for. The browser gave you "one app, one state tree, forever." The server gives you "one process, many concurrent state trees, for a few milliseconds each." Any API that pretends these are the same will eventually leak.

Pareto's answer: two APIs with different defaults.

  • defineStore for state where "one process, one state" is genuinely what you want
  • defineContextStore for state where "each render tree, its own state" is what you need

Pick the one that matches the shape of the state. If you can't decide, defineContextStore is never wrong.

npx create-pareto@latest my-app
cd my-app && npm install && npm run dev

Enter fullscreen mode Exit fullscreen mode


Pareto is a lightweight, streaming-first React SSR framework built on Vite. Documentation