Does the App Router support PM2 cluster mode?

[developer-choi]

https://github.com/developer-choi/test-pm2

I installed Next.js using npx [email protected]
and then added PM2 globally for testing. npm i -g pm2

However, I encountered the following error in cluster mode.

That every time I run That every time I run pm2 reload ecosystem.config.js, the service crashes, the service crashes,
and when the service is down, running pm2 reload ecosystem.config.js brings it back to life.

From my experience, out of 10 times I run pm2 reload ecosystem.config.js, the service crashes about 8 times. It’s not 100%, though.

As you can see, both pm2 and nextjs code were basic examples, but I had problems.

So I looked at the error log file. pm2 logs

So I looked at pm2 and nextjs official documents, but I couldn't find the answer I wanted.

It doesn't have to be pm2. I want to do uninterrupted distribution. What should I do?

[neo87cs]

Hi @developer-choi ,
We are using it in staging/production yet (next 14), i had to do some test to obtain the working version but looks like it is working with this ecosystem (in our case the script, with pnpm, has to be bypassed by pointing directly to next installed dist):

module.exports = {
  apps: [
    {
      args: 'start',
      cwd: process.env.PWD,
      env_development: {
        NODE_ENV: 'development',
      },
      env_production: {
        NODE_ENV: 'production',
        PORT: 3000,
      },
      env_staging: {
        NODE_ENV: 'production',
        PORT: 3000,
      },
      exec_mode: 'cluster',
      instances: 2,
      kill_timeout: 120000,
      listen_timeout: 10000,
      name: 'app',
      script: './node_modules/next/dist/bin/next',
    },
  ],
};

be aware that the default mode of pm2 ecosystem is fork, it means you will get 2 different servers running, if you want to achieve hot reload the correct exec mode is cluster

[quyphan97]

Do you have any other solution? I updated it like that and it worked but the memory increased 56mb -> >500mb and continues to increase