[Alex-Welsh]

Mirrorlists seem to pull packages late. Pull direct from source instead

[coderabbitai]

📝 Summary

Updated EPEL and Rocky Linux repository URLs to use direct upstream paths instead of mirrorlists. This covers EPEL 8 and Rocky Linux 8.10 and 9.8 repositories, including architecture and source variants.

The RPM repository definitions now use direct Fedora and Rocky Linux repository paths instead of country-specific mirror lists. The changes cover EPEL 8, 9 and 10, and Rocky Linux 8.10, 9.8 and 10.2.

Changes: RPM repository URLs

Layer / File(s)	Summary
EPEL repository paths ansible/inventory/group_vars/all/rpm-package-repos	EPEL 8, 9 and 10 URLs now point to direct Fedora paths. EPEL 10 URLs use the 10.2 upstream path while retaining the epel/10 base paths and existing sync settings.
Rocky Linux repository paths ansible/inventory/group_vars/all/rpm-package-repos	Rocky Linux 8.10, 9.8 and 10.2 URLs now point to direct Rocky Linux paths. The changes cover binary, aarch64 and source repositories.

Priority: ➖ Normal

Estimated code review effort: 2 (Simple) | ~10 minutes

Merge Risk: 🔵 Low · up to 89080

This change moves the RPM repository definitions from mirror lists to direct upstream URLs. If Fedora serves a bot-protection challenge to the sync host, EPEL syncing could fail. Fetch repomd.xml and one RPM from the sync host to confirm before relying on the change; the risk is unconfirmed and easy to check.

Pre-merge checks: 3

Check name	Status	Explanation
Docstring Coverage		No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0…
Linked Issues check		Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check		Check skipped because no linked issues were found for this pull request.

Full details: Docstring Coverage

Explanation

No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff. Analyzed 0 functions across 0 files. (1 skipped: 1 unsupported.)

• Autofix · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

---

ℹ️ Review info

⚙️ Run configuration

• Source mode: Auto
• Repository: stackhpc/stackhpc-release-train
• Branch: rpm-repos-drop-mirrorlists
• File: .coderabbit.yaml
• Commit: 890801c
• Layers: Repository YAML
• Review profile: ASSERTIVE
• Plan: Advanced
• Run ID: ae2fe216-b54f-46ac-ab02-6f20457ad199

📥 Commits

Reviewing files that changed from the base of the PR and between 23b6321 and 890801c.

📒 Files selected for processing: 1

• ansible/inventory/group_vars/all/rpm-package-repos

Included review availability: This review used your included allowance. Your plan provides up to 1 included review per hour; 0 remain after this review.

[Alex-Welsh]

Test sync running here for a few random repos. It'll probably time out if I ran it without a limit https://github.com/stackhpc/stackhpc-release-train/actions/runs/39949786576/job/120401166586

[jackhodgkiss]

Test sync running here for a few random repos. It'll probably time out if I ran it without a limit https://github.com/stackhpc/stackhpc-release-train/actions/runs/39949786576/job/120401166586

Should we increase the time out on the job?

[Alex-Welsh]

Test sync running here for a few random repos. It'll probably time out if I ran it without a limit https://github.com/stackhpc/stackhpc-release-train/actions/runs/39949786576/job/120401166586

Should we increase the time out on the job?

For now, it's not really causing us any issues. I think the timeouts are quite high already, and it's only a problem if we run the manual workflow (usually that's limited to a smaller set of repos). Nightly syncs are parallel so rarely hit the limit