Docker and Node: FATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memory

Question

I have a node NextJS app that I am containerizing with Docker and then deploying to GCP Cloud Run. While Docker runs without issue locally, on deployment to Cloud Run it runs into an error. When I look at the error logs it looks like it goes through various steps like installing the packages and running a migration without a problem, but then fails unexpectedly with the error message:

FATAL ERROR: Reached heap limit Allocation failed - JavaScript heap out of memory 
Uncaught signal: 6, pid=13, tid=13, fault_addr=0.
Uncaught signal: 6, pid=1, tid=1, fault_addr=0.
Container terminated on signal 6.

I find it a bit bizarre to run into that with such a simple application and am wondering what it means and options to prevent this from occurring (Smaller docker size?) or resolve (Expand the node memory?)

Not sure what additional information is needed, but maybe relevant are

Package.json (Scripts):

"scripts": {
    "dev": "nodemon server.js",
    "build": "next build",
    "start": "NODE_ENV=production npm run db:migrate && node server.js",
    "test": "jest --watch",
    "test:ci": "jest --ci",
    "db:migrate": "node_modules/.bin/sequelize db:migrate",
    "db:migrate:undo": "node_modules/.bin/sequelize db:migrate:undo",
    "db:migrate:undo:all": "node_modules/.bin/sequelize db:migrate:undo:all",
    "db:seed": "node_modules/.bin/sequelize db:seed:all",
    "db:seed:undo:all": "node_modules/.bin/sequelize db:seed:undo:all"
  },

Dockerfile:

version: "3.9"
services:
    redis:
        image: redis:alpine
    database:
        image: postgres
        ports:
            - "5432:5432"
        environment: 
            POSTGRES_USER: 'postgres'
            POSTGRES_PASSWORD: ''
            POSTGRES_DB: ${DB_DATABASE}
        volumes:
            - nextjs_auth_boilerplate:/var/lib/postgresql/data/ # persist data even if container shuts down
    app:
        image: nextjs-auth-boilerplate
        build: .
        depends_on:
            - redis
            - database
        command: ["./wait-for-it.sh", "database:5432", "--", "npm", "start"]
        ports:
            - "3000:3000"
        environment: 
            - REDIS_HOST=redis
            - DB_HOSTNAME=database
volumes:
    nextjs_auth_boilerplate:

Answer 1

It's hard to tell what went wrong without knowing the commands you used to deploy your code to Cloud Run, but the file you posted is not a Dockerfile . It's a docker compose file, and you cannot use docker-compose.yml in Google Cloud Run . This means that your Cloud Run service will have neither Redis nor PostgreSQL.

My wild guess is that you deployed your NextJS app, and then the db:migrate command of the Sequelize CLI didn't find the database and ended up emitting SIGABRT (signal 6).

You could try executing node server.js using a bigger V8 memory heap (see Node.js heap out of memory ), but I doubt it would solve your problem.

Possibly related:

• When does a process get SIGABRT (signal 6)?
• abrtd: Node Process was killed by signal 6 (SIGABRT)