Nuxt memory leak universal mode

[WilliamDASILVA]

Versions

• nuxt: 2.14.1
• node: 10.16.0

Reproduction

Additional Details

nuxt.config

module.exports = {
  mode: 'universal',
  srcDir: './src',
  server,
  head: {
    title: pkg.name,
    noscript: [{ innerHTML: 'This website requires JavaScript to run properly.' }],
    htmlAttrs: {
      dir: 'ltr'
    },
    meta: [
      { charset: 'utf-8' },
      { name: 'viewport', content: 'width=device-width, initial-scale=1' },
      { hid: 'description', name: 'description', content: '' }
    ],
    link: [
      { rel: 'icon', type: 'image/x-icon', href: '/favicon.ico' }
    ]
  },
  css: [
    { src: '@/assets/scss/style.scss', lang: 'scss' }
  ],
  plugins: [
    '~/plugins/vee-validate.js'
  ],
  modules: [
    'nuxt-stripe-module',
    '@nuxtjs/style-resources',
    '@nuxtjs/axios',
    '@nuxtjs/sitemap',
    '@nuxtjs/sentry',
    '@nuxtjs/robots',
    ['vue-wait/nuxt', { useVuex: true }],
    ['@nuxtjs/google-analytics', {
      id: 'UA-TEST'
    }]
  ],
  stripe: {
    publishableKey: process.env.STRIPE_PUBLISHABLE_KEY || ''
  },
  sentry: {
    dsn: process.env.SENTRY_DSN || ''
  },
  sitemap: {
    path: '/sitemap.xml'
  },
  styleResources: {
    scss: [
      'assets/scss/_variables.scss',
      'assets/scss/_mixins.scss'
    ]
  },
  build: {
    transpile: [
      'vee-validate/dist/rules'
    ]
  },
  env: {
    appUrl: process.env.VUE_APP_APP_URL
  },
  axios: {
    baseURL: process.env.API_BASE_URL,
    browserBaseURL: process.env.API_BROWSER_BASE_URL
  },
  tailwindcss: {
    cssPath: '~/assets/css/tailwind.css'
  },
  buildModules: [
    '@nuxtjs/pwa',
    '@nuxtjs/tailwindcss'
  ]
}

store.js

import Vue from 'vue'
import Vuex from 'vuex'

Vue.use(Vuex)

const store = new Vuex.Store({
  state: () => ({}),
  actions: {},
  mutations: {},
  getters: {}
})

export default () => store

Steps to reproduce

I've build my Nuxt app that uses Vuex, then I sent a certain amount of requests to it, the memory usage keeps growing without coming down. At some point (approx. 4-6 hours in a normal usage, less than a minute during a stress-test), the server crashes because it runs out of memory.

In this screenshot, you can see 2 snapshots. The first one is just after the server started. The second is after a 1 minute stress-test right after

By removing the lines in the .nuxt/index and .nuxt/dist/server/server.js:

  const registerModule = store.registerModule
  store.registerModule = (path, rawModule, options) => registerModule.call(store, path, rawModule, Object.assign({ preserveState: process.client }, options))

Then restarting the server and doing the same stress-test, I'm not getting the huge memory leak spike.

What is Expected?

I expect the server to live longer than some hours in production, in a real usage. I don't want to restart it everytime it crashes with a balancer to have a good uptime.

What is actually happening?

I'm getting a huge memory usage and the server keeps crashing.

[galvez]

What is the value of NODE_ENV on your production environment?

[WilliamDASILVA]

@galvez My NODE_ENV is production. Even if I change the NODE_ENV to something else than production, I'm getting the same results.

"start:inspect": "NODE_ENV=production node --inspect=0.0.0.0 ./node_modules/nuxt-start/bin/nuxt-start.js",

My quick-dirty-bandaid fix/workaround was to remove the usage of the registerModule by adding a new plugin that would delete that function, since I'm quite sure I'm not using it anywhere.

/**
 * Bandaid to resolve the issue with store.registerModule
 * since we're not using registerModule anywhere.
 */
export default ({ store }) => {
  delete store.registerModule
}

The same test with my workaround:

[galvez]

@WilliamDASILVA should be fixed in latest Nuxt release.