[https://nvbugs/6487038][test] Unwaive GB300 Kimi gen-only

[chienchunhung]

Summary

• remove the waiver for the exact GB300 Kimi-K2.5 Thinking FP4 concurrency-4096 disaggregated gen_only perf-sanity test tracked by NVBUG 6487038
• keep every other waiver from the bulk NVBUG 6561566 bucket unchanged
• preserve the Python transceiver's default kv_transfer_timeout_ms=60000; this PR has no runtime or configuration changes

Why

PR #16832 fixed the Python-transceiver/V1 fabric-memory allocation path, and PR #16717 originally unwaived this test. PR #17272 later re-waived it in a bulk post-merge waiver after synthetic Test terminated unexpectedly records caused by the GEN sentinel/Slurm teardown path.

The relevant harness issues are now fixed on main by:

• [None][fix] Align perf launcher with pytest shard #17134, which aligns the multi-node launcher with the pytest shard
• [None][fix] Bound GEN log sentinel wait #17140, which bounds the GEN log-sentinel wait and prevents a completed workload from becoming a Slurm teardown cancellation

Verification

Exact test:

perf/test_perf_sanity.py::test_e2e[disagg_upload-gen_only-gb300_kimi-k25-thinking-fp4_8k1k_con4096_ctx1_dep4_gen1_dep16_eplb0_mtp0_ccb-NIXL]

Run	Result	Duration	Notes
PR #16920 pipeline #51223 (report, completion)	PASSED	878,816 ms	Verified with the 60-second KV-transfer timeout
Main Post-Merge #2886 (report)	PASSED	806,478 ms	Exact main post-merge result
PR #17121 pipeline #52153 (report, completion)	PASSED	849,092 ms	Post-#17140 targeted pass; tested commit contains #17140's merge commit

Scope

This PR removes one waiver line only. It is intentionally narrower than the bulk unwaive in #17350 and relies only on evidence for this exact test.

Local validation

• pre-commit run --files tests/integration/test_lists/waives.txt — all applicable hooks passed

Dev Engineer Review

• Removed only the exact GB300 Kimi-K2.5 Thinking FP4, concurrency-4096, disaggregated gen_only waiver.
• Preserved all other NVBUG 6561566 waivers.
• The change is correctly scoped to tests/integration/test_lists/waives.txt.
• No runtime, configuration, API, or performance changes were introduced.
• The waiver file format remains valid.
• No duplicate or unrelated waiver changes were added.
• The targeted test passed in three reported runs.
• Pre-commit validation passed for the modified file.

QA Engineer Review

• No test-db/ or qa/ files were modified.
• One entry was removed from tests/integration/test_lists/waives.txt.
• The removed entry covered the GB300 Kimi-K2.5 Thinking FP4 disaggregated gen_only performance-sanity test.
• Verdict: sufficient.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Enterprise

Run ID: c45ff2e8-0efe-4a1b-8087-825c02b58644

📥 Commits

Reviewing files that changed from the base of the PR and between 5a47974 and f655225.

📒 Files selected for processing (1)

• tests/integration/test_lists/waives.txt

💤 Files with no reviewable changes (1)

• tests/integration/test_lists/waives.txt

---

Walkthrough

The change removes the skip entry for the GB300 Kimi-K2 Thinking disaggregated performance test associated with nvbugs/6561566.

Changes

Performance waiver

Layer / File(s)	Summary
Remove obsolete test waiver tests/integration/test_lists/waives.txt	Removes the skip entry for the GB300 Kimi-K2 Thinking disaggregated performance test.

Estimated code review effort: 1 (Trivial) | ~2 minutes

Possibly related PRs

• NVIDIA/TensorRT-LLM#17385: Modifies the same waiver list for a different test configuration.

Suggested reviewers: shixiaowei02, chenfeiz0326

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly identifies the NVBugs ticket, test-related change, and removal of the GB300 Kimi gen-only waiver.
Description check	✅ Passed	The description explains the change, rationale, scope, exact test coverage, verification results, and local validation.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[chienchunhung]

/bot skip --comment "Un-waiving the tests that have been validated by CI run."

[tensorrt-cicd]

PR_Github #64426 [ skip ] triggered by Bot. Commit: f655225 Link to invocation

[tensorrt-cicd]

PR_Github #64426 [ skip ] completed with state SUCCESS. Commit: f655225
Skipping testing for commit f655225

Link to invocation