@actions/artifact: a non-200 blob response is not destroyed before the retry, so the process waits for the 3-min socket timeout

Open Beginner friendly
#2,575 0 comments 0 reactions 0 assignees View on GitHub

Maintainers usually reply within 1 day

Nobody has claimed this yet.

Assessment

Difficulty
2/5
Estimated time
1-3 hours
Newbie friendliness
78/100
Issue type
Bug
Clarity
Clearly specified
Activity status
Active
Tech stack
github, node.js, typescript
Domain
backend, devtools

Research direction

Start in streamExtractExternal in the @actions/artifact package, where the issue says a non-200 response is left open before the retry. Read the response handling and its tests, then reproduce the 404-then-200 case with the local mock described in the issue. Done when the retry succeeds and the process exits promptly instead of waiting for the socket timeout.

Written by the indexing model from the issue text.

Description

Package: @actions/artifact, as bundled in actions/download-artifact@v8.

What happens.

  1. The first GET of the blob URL gets a non-200 answer (seen with 404 and 503).
  2. streamExtractExternal throws without consuming or destroying the response.
  3. The download retries after 5 s, the retry succeeds, and the step logs success.
  4. The first response's socket stays open. @actions/http-client's default request timeout (req.setTimeout(3 * 60000)) then keeps node alive until it fires, so the step ends about 180 s after the blob request instead of at once.

Reproducer. We used a local mock of the artifact API whose blob URL answers 404 (or 503) once, then 200. We ran download-artifact@v8's dist/index.js with GITHUB_API_URL pointed at the mock:

  • With the one error, it logs Failed to download artifact after 1 retries due to Unexpected HTTP response from blob storage: 404 Not Found, then success, and exits at 180.3 s.
  • With a clean 200, it exits in 0.2 s.

No proxy is involved.

In real runs (self-hosted Linux runners, 181 download steps over 40 workflow runs): the step ran past 60 s 5 times. Each time it ended 181.4 s after its "Redirecting to blob download url" line, and the download itself took about 7.7 s (one failed GET, the 5 s wait, a 1.6 s retry).

Suggested fix. In streamExtractExternal, destroy response.message (or read it to the end) before throwing on a non-200 status, so the failed attempt's socket is released.

Workaround on our side: a 1-min step timeout on the download, then a second try.

Dominant language
TypeScript
Stars
5.9k
Forks
1.8k
Avg merge
15m
Merged PRs (30d)
1

Getting set up

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

More from actions/toolkit

All issues in actions/toolkit

Similar issues

More TypeScript issues

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.