Jobs with artifacts lead to infinite "still running..." when using podman

[ebousse]

Minimal .gitlab-ci.yml illustrating the issue

stages:
  - stage1
  - stage2

task1:
  image: docker.io/library/alpine:latest
  stage: stage1
  script:
    - mkdir -p output
    - echo "hello" > output/artifact
  artifacts:
    paths:
    - output

task2:
  stage: stage2
  image: docker.io/library/alpine:latest
  script:
    - cat output/artifact

Note that this problem only happens when artifacts are declared.

Expected behavior

When running:

gitlab-ci-local --container-executable podman

I would expect the build to be swift and successful.

However, for an uncanny reason, the build never finishes, and this happen on my system:

$ gitlab-ci-local --container-executable podman
Using fallback git commit data
Unable to retrieve default remote branch, falling back to `main`.
Using fallback git remote data
parsing and downloads finished in 48 ms.
json schema validated in 174 ms
task1 starting docker.io/library/alpine:latest (stage1)
task1 copied to podman volumes in 698 ms
task1 $ mkdir -p output
task1 $ echo "hello" > output/artifact
task1 finished in 1.24 s
task1 exported artifacts in 587 ms
task1 copied artifacts to cwd in 50 ms
task2 starting docker.io/library/alpine:latest (stage2)
task2 copied to podman volumes in 567 ms
task1 > still running...
task2 > still running...
task1 > still running...
task2 > still running...
task1 > still running...
task2 > still running...
task1 > still running...
task2 > still running...
task1 > still running...
task2 > still running...
task1 > still running...
task2 > still running...
task1 > still running...
task2 > still running..
…

• The problem does not occur with Docker.
• I tried reseting and/or upgrading Podman, but without any change.
• This started happening I don't know exactly when, but gitlab-ci-local worked perfectly fine with Podman for years before this started happening.

Host information
Fedora 42

problem observed with:

• gitlab-ci-local 4.58.0
• gitlab-ci-local commit 530c186

Containerd binary
podman version 5.4.1

Additional context

• I investigated my podman configuration, but it seems I am using the recommended configuration (especially for the storage driver)
• I'm using Podman outside gitlab-ci-local without any issues. I even tried reproducing what gitlab-ci-local should do, and I could not find the problem.

[ebousse]

Additional information, this is the error trace when I use Ctrl+c to kill the stuck job:

Error: Command failed with exit code 1: podman cp /home/bousse-e/Téléchargements/buggitlabci/.gitlab-ci-local/artifacts/task1/. 62d6589c0fc0e6e8e5fcf9067be4dbf4fc728e9e5cbbd22dc514531a38bfbd83:/builds/fallback.group/fallback.project
    at makeError (/snapshot/gitlab-ci-local/node_modules/execa/lib/error.js:60:11)
    at <anonymous> (/snapshot/gitlab-ci-local/node_modules/execa/index.js:118:26)
    at process.processTicksAndRejections (node:internal/process/task_queues:105:5)
    at async Promise.all (index 0)
    at async Zt.copyArtifactsIn (/snapshot/gitlab-ci-local/src/job.ts:1155:9)
    at async Zt.execScripts (/snapshot/gitlab-ci-local/src/job.ts:940:9)
    at async Zt.execPreScripts (/snapshot/gitlab-ci-local/src/job.ts:763:36)
    at async Zt.start (/snapshot/gitlab-ci-local/src/job.ts:655:9)
    at async <anonymous> (/snapshot/gitlab-ci-local/node_modules/p-map/index.js:57:22)

[aasseman]

It seems that it's due to a podman update. I was (and still am) running gitlab-ci-local 4.57.0 for the past few weeks. Worked fine until I tried using it today.