[Chapel GA Tests] 2026-09-16

[chapelhpcBot]

The following GitHub Actions jobs failed:

• full-checks / Chapel paratest on darwin / Test test/classes/hannah (darwin), step: "upload test results"
• full-checks / Chapel paratest on darwin / Test test/classes/forwarding (darwin), step: "download Chapel build"
• full-checks / Chapel paratest on darwin / Aggregate results (darwin), step: "aggregate and report results"

Ran 746 total jobs at commit b193726903a9af810a7132a2f61f346ba517e82f "Fix typo in suppressif name (#29415)".

GitHub Actions run info:

• SHA: b193726903a9af810a7132a2f61f346ba517e82f
• Run ID: 35069513823
• Run attempt: 1
• Run number: 26
• Triggering event name: schedule

[annarift]

One job failed in uploading an artifact, another in downloading, both getting a 403 (Forbidden). Timestamps show the failures were hit in the same second.

I wonder if we might be hitting a rate-limit in having so many paratest child jobs uploading and downloading artifacts in a short window of time. We could resolve that by limiting the number of concurrent child jobs, but that would significantly increase the runtime when really all we need is the upload and download steps to be staggered. Maybe a random 0-60 second delay at the start of each child job?

[annarift]

PR'd Add random delay at start of paratest child jobs by arifthpe · Pull Request #29418 · chapel-lang/chapel · GitHub for random delay.