Running in PM2 cluster mode

[jimboolio]

I'm trying to set up directus to run on PM2 load balanced cluster which would utilize all available cpu cores, but additional instances of Directus exit with App [directus:1] exited with code [1] via signal [SIGINT]

There are no error messages on the console so I guess this could be as designed, if this is the case, it could be beneficial to include this information in the documentation

To Reproduce

My ecosystem.config.js:

module.exports = {
  apps : [{
    name   : "directus",
    instances : "2",
    exec_mode : "cluster",
    script : "npx",
    args : "directus start",
    autorestart : false
  }]
}

Caching is disabled on .env:

CACHE_ENABLED=false
CACHE_SCHEMA=false
CACHE_PERMISSIONS=false
RATE_LIMITER_ENABLED=false

Running pm2 restart ecosystem.config.js --attach outputs the following:

[PM2] Applying action restartProcessId on app [directus](ids: [ 0, 1 ])
[PM2] [directus](0) ✓
[PM2] [directus](1) ✓
┌─────┬─────────────┬─────────────┬─────────┬─────────┬──────────┬────────┬──────┬───────────┬──────────┬──────────┬──────────┬──────────┐
│ id  │ name        │ namespace   │ version │ mode    │ pid      │ uptime │ ↺    │ status    │ cpu      │ mem      │ user     │ watching │
├─────┼─────────────┼─────────────┼─────────┼─────────┼──────────┼────────┼──────┼───────────┼──────────┼──────────┼──────────┼──────────┤
│ 0   │ directus    │ default     │ N/A     │ cluster │ 307      │ 0s     │ 1    │ online    │ 0%       │ 35.2mb   │ adm… │ disabled │
│ 1   │ directus    │ default     │ N/A     │ cluster │ 314      │ 0s     │ 0    │ online    │ 0%       │ 30.0mb   │ adm… │ disabled │
└─────┴─────────────┴─────────────┴─────────┴─────────┴──────────┴────────┴──────┴───────────┴──────────┴──────────┴──────────┴──────────┘
[PM2][INFO] Log streaming apps id: 0 1, exit with Ctrl-C or will exit in 10secs
PM2      | App name:directus id:1 disconnected
PM2      | App [directus:1] exited with code [1] via signal [SIGINT]

And when running pm2 listall but the first directus processes are stopped

Expected behaviour

Directus would run in as many processes as defined in PM2 ecosystem.config.js

Actual behaviour

Only the proces with PM2 id 0 will run, other processes exit

[eikaramba]

maybe try to use a script like this. It helped me to actually get the console output
server.js

const { spawn } = require('child_process');
const ls = spawn('npx',['directus','start']);

ls.stdout.on('data', (data) => {
  console.log(data.toString());
});

ls.stderr.on('data', (data) => {
  console.error(data.toString());
});

ls.on('close', (code) => {
  console.log(`server stopped ${code}`);
});

{
  "apps" : [{
    "name"        : "backend",
    "cwd"         : "/xxxxx",
    "script"      : "./server.js",
    "ignoreWatch" : ["[\\/\\\\]\\./", "node_modules"],
    "watch"       : "false",
    "env": {
        "CONFIG_PATH":"/xxxxx/.env.production",
        "NODE_ENV": "production"
    },
    "exec_mode"  : "cluster_mode",
    "error" : "/xxxxx/logs/backend-err.log",
    "output"   : "/xxxx/logs/backend-out.log",
    "port":1338,
    "time": true
  },

[rijkvanzanten]

@jimboolio is it confirmed that it was Directus that stopped running? The part of the message ... via signal [SIGINT] indicates to me that Directus received the SIGINT command, causing it to stop 🤔

[jimboolio]

This is most likely the issue. With the server.js script from @eikaramba I managed to find out that the other instances get terminated because the directus port is already in use

It looks like that in the pm2 ecosystem.config.js should point to an actual directus startup script/file, not npm or npx - I will update/reply here if I manage to figure that out...

---

Update:

With npx which directus I got the location of the startup script...

So PM2 works as supposed with the following ecosystem.config.js

module.exports = {
  apps : [{
    name   : "directus",
    instances : "max",
    exec_mode : "cluster",
    script : "./node_modules/.bin/directus",
    args : "start",
  }]
}

Thanks for all help!