Repository navigation
Closed
1 / 11 of 1 issue completedLabels
approved-directionepicMulti-issue initiative tracked on the Bluefin planning boardMulti-issue initiative tracked on the Bluefin planning boardtriage/acceptedScope accepted; ready for work. Use `/triage accepted`.Scope accepted; ready for work. Use `/triage accepted`.
Description
- fix(ci): count job-summary scenarios explicitly instead of by subtraction #698 (job-summary counts)
- fix(smoke): restore Bluefin release gate reliability #691 (Bluefin release gate)
- feat(installer): add post-boot assertions for UEFI, Flatpak exclusion, and LUKS cmdline #697 (installer post-boot assertions)
created_date: "2026-09-24"
document_status: draft
closed_date: nullFeature: 20260924190738-95546aef-testsuite-reachable-e2e-suites
Overview
Consumers of testsuite’s reusable e2e workflow can request the security, hardware, lifecycle, Flatcar, NVIDIA and installer suites already in the repository instead of maintaining unreachable test files. Availability and expected hardware blockers remain explicit so a selectable suite is not mistaken for passing coverage. Source: testsuite#704.
Requirements
- Expose runnable suites
The reusable workflow accepts security, hardware, lifecycle, Flatcar and installer suites as declared inputs. - Keep NVIDIA truthfully gated
The NVIDIA suite is selectable only under its @hardware_blocked/hardware-emulation conditions. - Scale scenario distribution
Shard scheduling handles suite sizes beyond the old smoke/common-only split. - Publish truthful input docs
The suites input description, reference docs and suite map agree on every available suite and documented expected-fail state. - Prove consumed workflow behavior
Existing downstream callers continue to work while the new input and shard logic pass tests.
Constraints
- Reusable e2e.yml is consumed across repos: human CI-interface/breakage review is required before modifying inputs or shard behavior.
- The source body’s old 1-triage statement conflicts with its current agent-queue label; routing remains with the owning workflow, not this spec.
- Wait for or coordinate fix(ci): count job-summary scenarios explicitly instead of by subtraction #698 job summaries, fix(smoke): restore Bluefin release gate reliability #691 release gate and feat(installer): add post-boot assertions for UEFI, Flatpak exclusion, and LUKS cmdline #697 installer post-boot changes to avoid overriding contested e2e.yml.
Acceptance Criteria
- Expose runnable suites
A consumer selects each suite and sees its scenarios scheduled or an explicit linked blocker. - Keep NVIDIA truthfully gated
Without supported GPU hardware the suite reports its blocked reason, not a green result from zero executed scenarios. - Scale scenario distribution
A large common and one new suite execute without duplicate or missing scenarios. - Publish truthful input docs
Docs/skills remove stale Epic: Bluefin — migrate CI from ghost/Argo to GitHub Actions #43/Epic: Bluefin LTS — migrate CI from ghost/Argo to GitHub Actions #44 blockers; references/inputs-outputs.md lists the same accepted input values as e2e.yml. - Prove consumed workflow behavior
tests/unit/test_e2e_workflow.py and real selected suite workflows pass or carry a linked expected-fail issue.
Technical Approach
- Owned paths are .github/workflows/e2e.yml, tests/unit/test_e2e_workflow.py and docs/skills/ci-ops/e2e-workflow/**.
- Original suite-map baseline: security 15, hardware 13, NVIDIA 12 @future, Flatcar 13, lifecycle 27 plus installer; refresh counts from current source at implementation.
Success Metrics
- Every requested suite either runs observable scenarios or exposes a specific blocker, never a zero-scenario green.
Non-Goals
- No new workflow state labels and no claim that suite reachability alone proves hardware coverage.
[You have received this identical output 5 times. Re-reading 'specs/speks/20260924190738-95546aef-testsuite-reachable-e2e-suites.md:raw' will not change it — use a narrower selector (path:A-B), or proceed with the edit.]
Metadata
Metadata
Assignees
Labels
approved-directionepicMulti-issue initiative tracked on the Bluefin planning boardMulti-issue initiative tracked on the Bluefin planning boardtriage/acceptedScope accepted; ready for work. Use `/triage accepted`.Scope accepted; ready for work. Use `/triage accepted`.
Type
Projects
- StatusShow more project fieldsDone
Activity
test(flatcar): implement Ignition and update-disable lifecycle scenar…
fix(flatcar): harden Ignition marker probe and update.conf restore
[-]Epic: expose the unreachable suites via e2e.yml[/-][+]Expose the unreachable suites via e2e.yml[/+]hivecommons-hive commentedon Nov 2, 2026
Summary: I re-checked this plan against the current code. Two of the six missing test groups (system upgrades and the installer) can already be requested, and the three changes this plan said to wait for have all landed. The remaining work is now tracked in one smaller issue, and the last piece (Flatcar) needs a separate look because those tests target a different kind of operating system than the shared workflow boots.
Re-verified on
main(2026-11-02):lifecycle,installer(.github/workflows/e2e.yml:29).security,hardware,nvidiaselectable + documented, with an explicit "blocked" result (not a silent green) when every scenario is@hardware_blockedor SELinux is off. The SELinux half depends on feat(e2e): qualify the GHA VM with SELinux enforcing instead of selinux=0 #1024 / PR feat(e2e): keep Linux security protections on during desktop tests and fail on blocked actions #1037 (VM currently boots withselinux=0,e2e.yml:542).e2e.yml.flatcar(tests/flatcar/features, 13 scenarios). The e2e job boots a bootc image viabootc install; Flatcar uses Ignition and a different boot path, so it needs its own design rather than a newsuitesvalue. That will be split once [scanner] e2e: let callers run the security test group #1096 lands.🐝 Hive Agent:
scanner| Instance:hosted-projectbluefin-knuckle-gjvq| SHA:05905a03491b7c63d2e864f0fad457fbd2c96264— hive: agent=scanner backend=copilot model=claude-opus-5.5 copilot=1.0.88
hivecommons-hive commentedon Nov 4, 2026
All children delivered: #1096.
Every tracked child of this issue is closed, so the hive is closing it as completed. Reopen it if something it asked for is still missing.