René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"worker.exitedAfterDisconnect is set to `false` after worker.kill(), which doesn't comply with the documentation","articleBody":"### Version\r\n\r\nv18.16.0 (haven't check older version btw)\r\n\r\n### Platform\r\n\r\nLinux \u003cmy-machine-name> 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\r\n\r\n### Subsystem\r\n\r\n_No response_\r\n\r\n### What steps will reproduce the bug?\r\n\r\n- Run the below code block using `node index.js` e.g.\r\n\r\n```const cluster = require(\"node:cluster\");\r\nconst http = require(\"node:http\");\r\nconst process = require(\"node:process\");\r\n\r\nif (cluster.isPrimary) {\r\n  // Keep track of http requests\r\n  let numReqs = 0;\r\n  let inter = setInterval(() => {\r\n    console.log(`numReqs = ${numReqs}`);\r\n  }, 1000);\r\n\r\n  // Count requests\r\n  function messageHandler(msg) {\r\n    if (msg.cmd && msg.cmd === \"notifyRequest\") {\r\n      numReqs += 1;\r\n    }\r\n  }\r\n\r\n  // Start workers and listen for messages containing notifyRequest\r\n  const numCPUs = require(\"node:os\").availableParallelism();\r\n  for (let i = 0; i \u003c 2; i++) {\r\n    cluster.fork();\r\n  }\r\n\r\n  for (const id in cluster.workers) {\r\n    cluster.workers[id].on(\"message\", messageHandler);\r\n  }\r\n  cluster.on(\"exit\", (worker, code, signal) => {\r\n    console.log(\r\n      `--------------EAD---------------${worker.exitedAfterDisconnect}-----`,\r\n    );\r\n    if (worker.exitedAfterDisconnect === true) {\r\n      console.log(\"Oh, it was just voluntary – no need to worry\");\r\n    } else {\r\n      console.log(`-----------EXIT------------------${worker.process.pid}`);\r\n    }\r\n  });\r\n  process.on(\"SIGINT\", async (signal) => {\r\n    console.log(\"----------------SIGINT----------------------\");\r\n    await killChildProcesses(cluster);\r\n    console.log(\"-----------------DONE------------------\");\r\n    process.exit(0);\r\n  });\r\n} else {\r\n  // Worker processes have a http server.\r\n  const server = http\r\n    .Server((req, res) => {\r\n      res.writeHead(200);\r\n      res.end(\"hello world\\n\");\r\n\r\n      // Notify primary about the request\r\n      process.send({ cmd: \"notifyRequest\" });\r\n    })\r\n    .listen(4000);\r\n  process.on(\"SIGTERM\", (signal) => {\r\n    console.log(\"----------------SIGTERM----------------------\");\r\n    server.close(() => {\r\n      console.log(\"----------------------CLOSED---------------------\");\r\n      process.exit(0);\r\n    });\r\n  });\r\n}\r\n\r\nfunction killChildProcesses(cluster) {\r\n  return new Promise((resolve) => {\r\n    const aliveWorkers = Object.values(cluster.workers || {}).filter(\r\n      (worker) => {\r\n        return worker && !worker.isDead();\r\n      },\r\n    );\r\n\r\n    if (aliveWorkers.length === 0) {\r\n      resolve();\r\n      return;\r\n    }\r\n\r\n    let workersAlive = aliveWorkers.length;\r\n\r\n    const handleWorkerExit = () => {\r\n      workersAlive--;\r\n      if (workersAlive === 0) {\r\n        resolve();\r\n      }\r\n    };\r\n\r\n    aliveWorkers.forEach((worker) => {\r\n      console.log(\"-----------KILLING-----------\");\r\n      worker.on(\"exit\", handleWorkerExit);\r\n      worker.kill();\r\n    });\r\n  });\r\n}\r\n\r\n```\r\n- find out the primary process number for this e.g. `ps -ef | grep index.js`\r\n- kill the process with `SIGINT` signal, i.e. `kill -SIGINT \u003cpid>`\r\n- the resulting console output is as below\r\n```\r\n----------------SIGINT----------------------\r\n-----------KILLING-----------\r\n-----------KILLING-----------\r\n----------------SIGTERM----------------------\r\n----------------SIGTERM----------------------\r\n----------------------CLOSED---------------------\r\n----------------------CLOSED---------------------\r\n--------------EAD---------------false--0---\r\n-----------EXIT------------------\u003cworker-1-pid>\r\n--------------EAD---------------false--0---\r\n-----------EXIT------------------\u003cworker-2-pid>\r\n-----------------DONE------------------\r\n\r\n```\r\n### How often does it reproduce? Is there a required condition?\r\n\r\nalways\r\n\r\n### What is the expected behavior? Why is that the expected behavior?\r\n\r\nThe expected behavior should be\r\n```\r\n----------------SIGINT----------------------\r\n-----------KILLING-----------\r\n-----------KILLING-----------\r\n----------------SIGTERM----------------------\r\n----------------SIGTERM----------------------\r\n----------------------CLOSED---------------------\r\n----------------------CLOSED---------------------\r\n--------------EAD---------------false--0---\r\nOh, it was just voluntary – no need to worry\r\n--------------EAD---------------false--0---\r\nOh, it was just voluntary – no need to worry\r\n-----------------DONE------------------\r\n```\r\n\r\n\r\n\r\n### What do you see instead?\r\n\r\nAs 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.\r\n\r\n![image](https://github.com/nodejs/node/assets/14083655/9cea71c5-c608-4260-b654-a60f852e5ec5)\r\n\r\nAs 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...\r\n\r\n### Additional information\r\n\r\n_No response_","author":{"url":"https://github.com/khaled4vokalz","@type":"Person","name":"khaled4vokalz"},"datePublished":"2024-03-01T11:47:07Z","url":"https://github.com/nodejs/node/issues/51934"}