Uncaught AssertionError thrown due to a possible race condition?
Nobody has claimed this yet.
- Dominant language
- JavaScript
- Stars
- 7.7k
- Forks
- 880
- Avg merge
- 2d 16h
- Merged PRs (30d)
- 68
Description
Bug Description
I run an application which makes a lot of requests per second using HTTP/2 over a fixed pool of about 500 Dispatchers.
Everything works great but sometimes when the upstream tends to have a higher error rate, these errors come in the form of a weird assertion error within our logs as uncaught promise rejections.
Reproducible By
I am not sure how I could provide a reproducible example for this as it does happen randomly and is certainly tied to upstream request failures.
Expected Behavior
The request failure should be properly thrown and caught by the request itself instead of the global unhandled rejection handler within our application.
Logs & Screenshots
AssertionError [ERR_ASSERTION]: The expression evaluated to a falsy value:
assert(this.callback)
at RequestHandler.onConnect (/root/cluster-worker/node_modules/undici/lib/api/api-request.js:82:5)
at Request.onConnect (/root/cluster-worker/node_modules/undici/lib/core/request.js:228:29)
at writeH2 (/root/cluster-worker/node_modules/undici/lib/dispatcher/client-h2.js:345:13)
at Object.write (/root/cluster-worker/node_modules/undici/lib/dispatcher/client-h2.js:119:14)
at _resume (/root/cluster-worker/node_modules/undici/lib/dispatcher/client.js:601:50)
at resume (/root/cluster-worker/node_modules/undici/lib/dispatcher/client.js:523:3)
at Client.<computed> (/root/cluster-worker/node_modules/undici/lib/dispatcher/client.js:250:31)
at ClientHttp2Stream.<anonymous> (/root/cluster-worker/node_modules/undici/lib/dispatcher/client-h2.js:531:22)
at Object.onceWrapper (node:events:632:28)
at ClientHttp2Stream.emit (node:events:530:35)
Environment
Distributor ID: Ubuntu
Description: Ubuntu 24.04.1 LTS
Release: 24.04
Codename: noble
Node: v20.18.2
Additional context
I have looked through the source code and may have a clue as to what is potentially causing this.
The error is thrown in this snippet: https://github.com/nodejs/undici/blob/c7f3d77011234fe243c317ada1398044032342cc/lib/api/api-request.js#L76-L86
As we can observe, the culprit is that the assert(this.callback) fails thus something is nullifying the callback before the onConnect is fired.
There are only two places in the rest of this file where this is done which is within the onHeader() and onError() methods. The onHeader() is likely not the culprit as well the onConnect will always fire before the onHeader due to the request lifecycle.
The onError() thus remains to be the only culprit which can be causing this where the onError is for some reason firing before the onConnect().
In fact, there seems to be a TODO: already there which potentially questions this problematic this.callback = null operation on line 146: https://github.com/nodejs/undici/blob/c7f3d77011234fe243c317ada1398044032342cc/lib/api/api-request.js#L141-L150
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start with lib/api/api-request.js around onConnect(), onHeader(), and onError(), then trace the request lifecycle through lib/core/request.js and lib/dispatcher/client-h2.js. Investigate how an upstream failure can clear the callback before onConnect() and whether the existing TODO explains the ordering. Done should include a regression case showing request failures are caught without an uncaught AssertionError.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- javascript, nodejs
- Domain
- networking
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 38/100