worker.exitedAfterDisconnect is set to false after worker.kill(), which doesn't comply with the documentation

[khaled4vokalz]

Version

v18.16.0 (haven't check older version btw)

Platform

Linux 6.5.0-21-generic #21-Ubuntu SMP PREEMPT_DYNAMIC Wed Feb 7 14:17:40 UTC 2024 x86_64 x86_64 x86_64 GNU/Linux

Subsystem

No response

What steps will reproduce the bug?

• Run the below code block using node index.js e.g.

const http = require("node:http");
const process = require("node:process");

if (cluster.isPrimary) {
  // Keep track of http requests
  let numReqs = 0;
  let inter = setInterval(() => {
    console.log(`numReqs = ${numReqs}`);
  }, 1000);

  // Count requests
  function messageHandler(msg) {
    if (msg.cmd && msg.cmd === "notifyRequest") {
      numReqs += 1;
    }
  }

  // Start workers and listen for messages containing notifyRequest
  const numCPUs = require("node:os").availableParallelism();
  for (let i = 0; i < 2; i++) {
    cluster.fork();
  }

  for (const id in cluster.workers) {
    cluster.workers[id].on("message", messageHandler);
  }
  cluster.on("exit", (worker, code, signal) => {
    console.log(
      `--------------EAD---------------${worker.exitedAfterDisconnect}-----`,
    );
    if (worker.exitedAfterDisconnect === true) {
      console.log("Oh, it was just voluntary – no need to worry");
    } else {
      console.log(`-----------EXIT------------------${worker.process.pid}`);
    }
  });
  process.on("SIGINT", async (signal) => {
    console.log("----------------SIGINT----------------------");
    await killChildProcesses(cluster);
    console.log("-----------------DONE------------------");
    process.exit(0);
  });
} else {
  // Worker processes have a http server.
  const server = http
    .Server((req, res) => {
      res.writeHead(200);
      res.end("hello world\n");

      // Notify primary about the request
      process.send({ cmd: "notifyRequest" });
    })
    .listen(4000);
  process.on("SIGTERM", (signal) => {
    console.log("----------------SIGTERM----------------------");
    server.close(() => {
      console.log("----------------------CLOSED---------------------");
      process.exit(0);
    });
  });
}

function killChildProcesses(cluster) {
  return new Promise((resolve) => {
    const aliveWorkers = Object.values(cluster.workers || {}).filter(
      (worker) => {
        return worker && !worker.isDead();
      },
    );

    if (aliveWorkers.length === 0) {
      resolve();
      return;
    }

    let workersAlive = aliveWorkers.length;

    const handleWorkerExit = () => {
      workersAlive--;
      if (workersAlive === 0) {
        resolve();
      }
    };

    aliveWorkers.forEach((worker) => {
      console.log("-----------KILLING-----------");
      worker.on("exit", handleWorkerExit);
      worker.kill();
    });
  });
}

• find out the primary process number for this e.g. ps -ef | grep index.js
• kill the process with SIGINT signal, i.e. kill -SIGINT <pid>
• the resulting console output is as below

----------------SIGINT----------------------
-----------KILLING-----------
-----------KILLING-----------
----------------SIGTERM----------------------
----------------SIGTERM----------------------
----------------------CLOSED---------------------
----------------------CLOSED---------------------
--------------EAD---------------false--0---
-----------EXIT------------------<worker-1-pid>
--------------EAD---------------false--0---
-----------EXIT------------------<worker-2-pid>
-----------------DONE------------------

How often does it reproduce? Is there a required condition?

always

What is the expected behavior? Why is that the expected behavior?

The expected behavior should be

----------------SIGINT----------------------
-----------KILLING-----------
-----------KILLING-----------
----------------SIGTERM----------------------
----------------SIGTERM----------------------
----------------------CLOSED---------------------
----------------------CLOSED---------------------
--------------EAD---------------false--0---
Oh, it was just voluntary – no need to worry
--------------EAD---------------false--0---
Oh, it was just voluntary – no need to worry
-----------------DONE------------------

What do you see instead?

As you can see the worker.exitedAfterDisconnect is not true, but the documentation (https://nodejs.org/api/cluster.html#workerexitedafterdisconnect) states that it should be true if you kill() or disconnect() the worker.

As you can the worker process exited with code 0, I don't understand which other way it can be exited with code 0 other that what the code is doing...

Additional information

No response

[aduh95]

Here's the relevant documentation:

node/doc/api/cluster.md

Lines 447 to 457 in 84716d8

	### `worker.exitedAfterDisconnect`
	<!-- YAML
	added: v6.0.0
	-->
	* {boolean}
	This property is `true` if the worker exited due to `.disconnect()`.
	If the worker exited any other way, it is `false`. If the
	worker has not exited, it is `undefined`.

According to your repro steps, you never call worker.disconnect(), so according to the docs, worker.exitedAfterDisconnect should be false (the worker exited any other way). Am I missing something?

[khaled4vokalz]

okay, I guess it's a types doc issue then? https://github.com/DefinitelyTyped/DefinitelyTyped/blob/master/types/node/v16/cluster.d.ts#L269

Was it previously supported on kill() api as well? or it's just a typo? that's the reason I shared the picture of the typings (I am using types/node version 17)

Also, the DOC says

But from code I don't see something like this... 🤔
https://github.com/nodejs/node/blob/main/lib/internal/cluster/primary.js#L360

Or I am looking into the wrong place? 🤔

[khaled4vokalz]

closing this issue as calling worker#disconnect() works as expected....

[aduh95]

FYI @types/node is not maintained in this repo – but you probably already know that since you linked to the DefinitelyTyped repo.