Skip to content

fix(ci): AUT-2266 retry uv installer download - #3325

Merged
kajalj22 merged 1 commit into
NVIDIA-NeMo:mainfrom
svcnemo-autobot:ci/retry-uv-installer
Sep 11, 2026
Merged

kajalj22 merged 1 commit into
NVIDIA-NeMo:mainfrom
svcnemo-autobot:ci/retry-uv-installer

Conversation

@svcnemo-autobot

Copy link
Copy Markdown
Contributor

Background

Nightly Gym CI run 34561286253 failed before pytest started when the pinned uv installer download was reset by the remote endpoint.

What changed

Add bounded retries and connection, transfer, and overall retry timeouts to the existing curl download.

Details

Buffer the installer before executing it, so a failed or partial download is never piped into sh. The uv version and install destination remain unchanged.

Tested

  • bash -n scripts/ci/setup_dev.sh
  • Behavioral shell harness verified the exact curl arguments and successful pinned uv installation
  • Managed unit:core-unit-tests could not allocate: the proven Gym image was rejected through the permitted pull-through path due invalid/incomplete manifest metadata; GitHub CI remains authoritative

Signed-off-by: svcnemo-autobot <svcnemo-autobot@nvidia.com>
@copy-pr-bot

copy-pr-bot Bot commented Sep 11, 2026

Copy link
Copy Markdown

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

@svcnvidia-nemo-ci svcnvidia-nemo-ci added area:ci CI workflows, test routing, and repository automation Run CICD labels Sep 11, 2026
@svcnvidia-nemo-ci

Copy link
Copy Markdown
Contributor

/ok to test 27bb6f7

@kajalj22
kajalj22 merged commit 8b87da9 into NVIDIA-NeMo:main Sep 11, 2026
41 checks passed
jkyi-nvidia pushed a commit that referenced this pull request Sep 11, 2026
## Background

Nightly Gym CI run 34561286253 failed before pytest started when the
pinned uv installer download was reset by the remote endpoint.

## What changed

Add bounded retries and connection, transfer, and overall retry timeouts
to the existing curl download.

## Details

Buffer the installer before executing it, so a failed or partial
download is never piped into `sh`. The uv version and install
destination remain unchanged.

## Tested

- `bash -n scripts/ci/setup_dev.sh`
- Behavioral shell harness verified the exact curl arguments and
successful pinned uv installation
- Managed `unit:core-unit-tests` could not allocate: the proven Gym
image was rejected through the permitted pull-through path due
invalid/incomplete manifest metadata; GitHub CI remains authoritative

Signed-off-by: svcnemo-autobot <svcnemo-autobot@nvidia.com>
Signed-off-by: Junkeun Yi <jkyi@nvidia.com>

This branch was previously deployed

1 inactive deployment
public β€” 27bb6f7f Deployed Sep 11, 2026 by copy-pr-bot[bot] via release / finalize / notify #2796
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

area:ci CI workflows, test routing, and repository automation Run CICD

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants