[trtllm-agent]

Description

Temporarily waive 6 failed case(s) on CPU. Opened automatically by auto-waive (draft; a human reviews and merges).

The native main-break detector (native-p3/k3-w12h-fresh12h-nopass-hostclock/1.0.1) reports unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[False-already-complete] broken on main for CPU: failure signature attributeerror: types.simplenamespace object has no attribute guided_failures, onset 2026-11-10T00:39:54Z, latest failure 2026-11-10T01:34:17Z. Policy native_p3@1.0.0 selected it; https://nvbugs/7071872 tracks the fix. Scope CPU is the stage prefix that ran the failing test.

• Bug: https://nvbugs/7071872 (details, logs and 7-day test history are in the bug)
• Restore coverage when: Remove this line when https://nvbugs/7071872 is fixed or closed and the test passes on CPU post-merge; un-waive follows the bug, not the PR.
• Scope reviewed by: auto-waive at 2026-11-10T03:35:48.586002+00:00

Tests in this break event:

• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[False-already-complete] on CPU
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[False-cancelled] on CPU
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[True-already-complete] on CPU
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[True-cancelled] on CPU
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_deferred_lora_generation_resumes_from_harvested_host_token on CPU
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_overlap_marker_changes_only_after_previous_sample_is_harvested[empty] on CPU

PRs whose pre-merge runs hit this break (rebase or re-run after this merges):

• [None][perf] Fuse residual add, RMSNorm and static FP8 quantization for the FP8 consumers of Qwen3-Next norms #21209
• [None][feat] BREAKING: Add native KV cache V2 sleep/wake #21221
• [TRTLLM-16303][refact] Keep hybrid speculative replay state in KV cache regions #21324
• [TRTLLM-16731][perf] Read Kimi-Linear expert shards with pread into pinned host buffers #21588

full:CPU/unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[False-already-complete] SKIP (https://nvbugs/7071872)
full:CPU/unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[False-cancelled] SKIP (https://nvbugs/7071872)
full:CPU/unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[True-already-complete] SKIP (https://nvbugs/7071872)
full:CPU/unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[True-cancelled] SKIP (https://nvbugs/7071872)
full:CPU/unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_deferred_lora_generation_resumes_from_harvested_host_token SKIP (https://nvbugs/7071872)
full:CPU/unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_overlap_marker_changes_only_after_previous_sample_is_harvested[empty] SKIP (https://nvbugs/7071872)

What to check before merging:

• The bug's test-history links show the cases failing on main, not only on the trigger PR.
• The scope CPU matches the failing stage and no narrower waiver already exists.
• Nothing else is needed from you after merging: the bug owner removes these lines with the fix.

Provenance

• Detector native-p3/k3-w12h-fresh12h-nopass-hostclock/1.0.1, policy native_p3@1.0.0
• Run 7a88ffc3-7e37-499b-8be3-ed7dab742609; snapshot SHA-256 6a32e90c3f2a91dcda75de4d36c5f4904e43746883653d10d6429611c58a1952; config SHA-256 80248ef10d45ba9db91b7f0a8258df726f95897e1dc5b17a275cf58bc305d1f5
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[False-already-complete]: incident a64e525ed1ef76607f24c830e39dc0b119d339be728f423a352ad5d7e8fed36f, observation 27b96953e60bf9d8fdda22dbcbda9ad00a94284a1ef8938372baa5fbeb921e01
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[False-cancelled]: incident fff18a462f4baef568efef2d478c29ccf0c47541dc31d0480d5b71384398e799, observation efb3930c55702e2fff7f921d3904640f27629088d5a257396ae6356666371564
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[True-already-complete]: incident ce6303b46dab46b23d3d31d74357f2073b2207cbf1804f867130336b18b2f073, observation f6d8f7a379b2468880e22f552d119545105edd15e5ff98a75189a2922f3af48b
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary[True-cancelled]: incident d79ec313419532cfb2cfd0da284018579036085883c07f3a1ba054644fd946e1, observation 03537f08fd7c66decb3c6fd92c1e351db3403ad46f3ce6b8e5cb60d949f7ecc0
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_deferred_lora_generation_resumes_from_harvested_host_token: incident cdc48b0a8d26e5609734ba025d8832c3b10ac4698e3a6d02aae95faef19d7b26, observation 3c5c7e9c1f33dde6f9bc6328fbeaadcd35a18df27a0f3fda890c1c5d11237f09
• unittest/_torch/disaggregation/test_disagg_idle_cancellation.py::test_overlap_marker_changes_only_after_previous_sample_is_harvested[empty]: incident d81f34d092fede8a59bcdb504a931ec94d02d08eb303e154e4d1827d60032350, observation 394899c09e78f3bdd550334a778873c5f3e4314fbd6f6e74abca18bae7458a54

Test Coverage

Review the incident evidence and selector scope before merging. This proposal does not
establish that main is fixed. Confirm restoration separately after the corrective change.

PR Checklist

• Please check this after reviewing the above items as appropriate for this PR.

Dev Engineer Review

The change adds six CPU waivers to tests/integration/test_lists/waives.txt. Each waiver names bug 7071872. No source code changes are included.

QA Engineer Review

The six entries cover four parameterized cases of test_cancelled_overlap_request_retires_at_the_existing_post_sampler_boundary, plus test_deferred_lora_generation_resumes_from_harvested_host_token and the [empty] case of test_overlap_marker_changes_only_after_previous_sample_is_harvested. The waivers prevent these cases from running in the full CPU list. The supplied evidence does not establish whether the bug is fixed or whether the waivers are listed in other CI or manual-QA lists. Coverage verdict: needs follow-up.

Per-File QA Perspective

• tests/integration/test_lists/waives.txt: Adds six full:CPU skip entries for disaggregation idle-cancellation tests. Confirm the waivers are removed when the bug is fixed and the tests pass on CPU.

[github-actions]

Automatically added "ci: full pre-merge approved" because this PR has satisfied the required GitHub review approvals. Unresolved review conversations and other required checks remain independent merge requirements.

[mzweilz]

/bot skip --comment "waive only"

[coderabbitai]

No actionable comments were generated in the recent review.

Walkthrough

Six CPU disaggregation tests were added to the waived-test list. Each entry marks a test as SKIP and references NVIDIA bug 7071872.

Changes

Test waivers

Layer / File(s)	Summary
Add CPU disaggregation test waivers tests/integration/test_lists/waives.txt	Adds six full:CPU test entries: four overlap-request cancellation cases, one deferred LoRA generation case, and one overlap-marker case. Each entry references bug 7071872.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Change: Other

Possibly related issues

• Issue 21932 — The added entries use full:CPU platform-scoped waivers, which relate to the issue’s subject of exact waive-entry matching and platform scope.

Merge Risk: Minimal · up to 2b9f9

Six CPU cases are temporarily skipped under bug 7071872. Their reported main-branch failures remain unverified.

Architecture risk: Low · up to 2b9f9

The change affects 1 system.

Changed systems: tests

Architecture concerns
No architecture-level concerns identified.

Review details

Systems and components

• observed — tests (service) was modified; 1 changed file maps to changed impact.

Before / after behavior

• observed — Modified behavior in tests/integration/test_lists/waives.txt: Added six full:CPU disaggregation tests to the skip list: four parameterized overlap-request cancellation cases, one deferred LoRA generation case, and one overlap-marker case. All reference bug 7071872.

Pre-merge checks: 5

Check name	Status	Explanation
Title check		The title follows the repository format and clearly identifies the CPU test waiver linked to NVBugs 7071872.
Description check		The description explains the failure, the six waived cases, the CPU scope, the bug link, restoration criteria, and review checks. It includes the required Description, Test Coverage, and PR Checklist …
Docstring Coverage		No eligible functions found in the changed files. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentat…
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: Description check

Explanation

The description explains the failure, the six waived cases, the CPU scope, the bug link, restoration criteria, and review checks. It includes the required Description, Test Coverage, and PR Checklist sections.

Full details: Docstring Coverage

Explanation

No eligible functions found in the changed files. Skipping docstring coverage check. Docstring coverage is scoped to functions touched by this diff with cyclomatic complexity of at least 2. Documentation-only edits and deletions within surviving functions are included; unchanged diff context and fully deleted functions are excluded. Analyzed 0 functions across 0 files. (1 skipped: 1 unsupported.)

Finishing Touches

Generate unit tests (beta)

• Create a new PR

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

Recent review info

Run configuration

• Source mode: Auto
• Repository: NVIDIA/TensorRT-LLM
• Branch: main
• File: .coderabbit.yaml
• Commit: 615831c
• Layers: Repository YAML
• Review profile: CHILL
• Plan: Enterprise
• Run ID: e2487608-4959-4126-a117-d2f6448e05d8

Commits

Reviewing files that changed from the base of the PR and between 615831c and 2b9f9f4.

Files selected for processing: 1

• tests/integration/test_lists/waives.txt

Included review availability: This review used your included allowance. Your plan provides up to 12 included reviews per hour; 11 remain after this review.

[tensorrt-cicd]

PR_Github #83656 [ skip ] triggered by Bot. Commit: 2b9f9f4 Link to invocation

[tensorrt-cicd]

PR_Github #83656 [ skip ] completed with state SUCCESS. Commit: 2b9f9f4
Skipping testing for commit 2b9f9f4

Link to invocation

[sunnyqgg]

Waiver review: 0 Blocker, 0 Possibly blocking, 1 Required, 1 Nice to have, both inline. The six lines are correct and complete. Since #20903, _process_previous_batch reads previous_batch.guided_failures, and the tests added in #21508 pass a SimpleNamespace double without that field. That makes exactly these six cases fail deterministically on main; [adp-deferred] and the rest of the file never reach that read. The ids match the parametrization, scripts/check_test_list.py --validate passes, the full:CPU/ scope matches where these cpu_only tests run, and https://nvbugs/7071872 is the bug filed for this break. The one required item is coordinating with #21963, which waives the same six ids for the same bug.

[sunnyqgg]

[Required] This duplicates #21963. That PR came from the CI Report bot at HuiGao-NV's request and is already approved and /bot skipped. It waives exactly these six test ids for the same https://nvbugs/7071872, unprefixed there and under full:CPU/ here. The two hunks are far apart, so both PRs merge cleanly. The duplicate-waive pre-commit check keys on the raw line (scripts/check_test_list.py:1085-1091), so it won't flag full:CPU/unittest/... next to unittest/.... Both forms take effect on the CPU stages, so whichever lands second only adds six redundant waivers. The fix for the bug then has to remove twelve lines, and removing one set leaves the tests skipped. Please link the two PRs and land only one: if #21963 merges first, close this PR instead of merging it; otherwise close #21963.

[sunnyqgg]

[Nice to have] For nvbugs/7071872, which has no engineer assigned yet: the root cause is a stale test double, and the product code is fine. Since #20903, _process_previous_batch reads previous_batch.guided_failures and self.capturable_guided_decoder through _take_capturable_guided_failures (py_executor.py:7606-7608, 11870-11871). The tests added in #21508 build previous_batch as a SimpleNamespace without those fields (test_disagg_idle_cancellation.py:616, :681), and the fixture never sets capturable_guided_decoder. Three edits should make all six pass again, as #20903 already did in test_disagg_loop_transcript.py:141 and :335-340: add guided_batch_id=None, guided_failures=None to both namespaces and engine.capturable_guided_decoder = None to the fixture. The fix PR can then remove these waivers, and #21963's too if it lands.

[yuxianq]

I have a PR #21994 to fix this bug, maybe we can wait for it instead.