[BUG] Existing ext4 xattr VM scenario exceeds its QEMU deadline #90

Open
opened 2026-09-25 02:25:26 +02:00 by erikinkinen · 0 comments
Owner

Summary

The existing ext4 xattr positive VM scenario exceeds its declared 120-second QEMU deadline in original Integration CI1784. The complete catalog therefore fails even though Rust and Markdown jobs succeed. A deterministic isolated reproduction and underlying cause are not yet established.

Reproduction

Observed once in original run1784 at Integration 9c5f766369d3ee9e3fdc8e59f16571dd4837f997, worker 4, scenario subsystem-e2fs-fat-ext4-xattr-positive. The bootloader VM stage starts QEMU at 21:12:49 UTC on 24 September 2026 and reports qemu timed out after 120s at 21:14:50 UTC. Preserve the same catalog selection, concurrency and deadline when investigating. No unchanged job has been rerun.

Expected and actual behavior

Expected: the existing positive scenario satisfies its success and cleanup checks within its declared deadline.

Actual: the overall QEMU timeout expires. The final catalog reports 489 passed and two failed out of 491; the other failure is the historically affected realm-startup scenario. Neither a shared cause nor a relation to the quota timeout is established. This observation is an overall deadline failure, not a reported serial-progress deadline failure.

Environment and identity

  • Integration: 9c5f766369d3ee9e3fdc8e59f16571dd4837f997, original PR run1784, Integration job, worker 4.
  • Scenario: subsystem-e2fs-fat-ext4-xattr-positive, subsystem-positive, vm-full-system.
  • Runner: twelve catalog workers; the original declared QEMU limit is 120 seconds.
  • Image digest, exact hardware pressure and last known passing identical artifact: not established by this log excerpt; retain the original catalog/source selection when reproducing.
  • Authority: existing VM fixture and filesystem routes only; no extra grants or wider storage access are authorized by this report.

Evidence and prior checks

The complete original Integration log and both companion job logs have been retained and verified. Rust and Markdown pass; the VM catalog fails. The final summary and original timeout remain evidence even when a later run passes. No deadline was increased, assertion removed or original job retried.

Tracking

Track correction and unchanged-limit VM proof under Phase 6 completion and Integration PR12. The quota timeout is a separate existing observation; do not assume the same cause. The historical realm preparation failure has later native correction evidence and does not explain this timeout.

## Summary The existing ext4 xattr positive VM scenario exceeds its declared 120-second QEMU deadline in original Integration CI1784. The complete catalog therefore fails even though Rust and Markdown jobs succeed. A deterministic isolated reproduction and underlying cause are not yet established. ## Reproduction Observed once in [original run1784](https://git.erikinkinen.fi/erix/integration/actions/runs/1784) at Integration `9c5f766369d3ee9e3fdc8e59f16571dd4837f997`, worker 4, scenario `subsystem-e2fs-fat-ext4-xattr-positive`. The bootloader VM stage starts QEMU at 21:12:49 UTC on 24 September 2026 and reports `qemu timed out after 120s` at 21:14:50 UTC. Preserve the same catalog selection, concurrency and deadline when investigating. No unchanged job has been rerun. ## Expected and actual behavior Expected: the existing positive scenario satisfies its success and cleanup checks within its declared deadline. Actual: the overall QEMU timeout expires. The final catalog reports 489 passed and two failed out of 491; the other failure is the historically affected realm-startup scenario. Neither a shared cause nor a relation to the quota timeout is established. This observation is an overall deadline failure, not a reported serial-progress deadline failure. ## Environment and identity - Integration: `9c5f766369d3ee9e3fdc8e59f16571dd4837f997`, original PR run1784, Integration job, worker 4. - Scenario: `subsystem-e2fs-fat-ext4-xattr-positive`, subsystem-positive, vm-full-system. - Runner: twelve catalog workers; the original declared QEMU limit is 120 seconds. - Image digest, exact hardware pressure and last known passing identical artifact: not established by this log excerpt; retain the original catalog/source selection when reproducing. - Authority: existing VM fixture and filesystem routes only; no extra grants or wider storage access are authorized by this report. ## Evidence and prior checks The complete original Integration log and both companion job logs have been retained and verified. Rust and Markdown pass; the VM catalog fails. The final summary and original timeout remain evidence even when a later run passes. No deadline was increased, assertion removed or original job retried. ## Tracking Track correction and unchanged-limit VM proof under [Phase 6 completion](https://git.erikinkinen.fi/erix/integration/issues/65) and [Integration PR12](https://git.erikinkinen.fi/erix/integration/pulls/12). The [quota timeout](https://git.erikinkinen.fi/erix/integration/issues/18) is a separate existing observation; do not assume the same cause. The historical [realm preparation failure](https://git.erikinkinen.fi/erix/integration/issues/87) has later native correction evidence and does not explain this timeout.
erikinkinen added this to the Alpha 1 milestone 2026-10-30 09:39:19 +01:00
Sign in to join this conversation.
No milestone
No project
No assignees
1 participant
Notifications
Due date
The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Dependencies

No dependencies set

Reference
erix/integration#90
No description provided.