[None][feat] Assert attention DP disabled when KV connector is in use

[jthomson04]

Summary

• KV connector initialization currently has no compatibility check against attention data parallelism. The combination has not been validated and there is no established plumbing for the connector scheduler/worker to reason about per-DP-rank KV state.
• Adds a NotImplementedError in py_executor_creator._create_py_executor that fires when mapping.enable_attention_dp is True and a kv_connector_config is provided, matching the style of the existing scheduler-policy and VSWA guards.

Test plan

• Launch with kv_connector_config + enable_attention_dp=True and confirm the new NotImplementedError is raised before any connector worker is instantiated.
• Launch with kv_connector_config + enable_attention_dp=False and confirm startup is unchanged.

Summary by CodeRabbit

• Bug Fixes
  • Enhanced configuration validation to prevent incompatible feature combinations. The system now explicitly rejects the concurrent use of KV connector with attention data parallelism, providing a clear error message. This improvement prevents unexpected behavior and configuration errors by blocking unsupported feature interactions at initialization time.

[jthomson04]

/bot run --disable-fail-fast

[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: 12d3bbf9-1eef-4b37-88bb-07471a695968

📥 Commits

Reviewing files that changed from the base of the PR and between 55e2727 and afffd53.

📒 Files selected for processing (1)

• tensorrt_llm/_torch/pyexecutor/py_executor_creator.py

---

📝 Walkthrough

Walkthrough

A validation check is added to create_py_executor that explicitly rejects configurations combining KV connector with attention data parallelism by checking mapping.enable_attention_dp and raising NotImplementedError if both are enabled.

Changes

Cohort / File(s)	Summary
Validation Check tensorrt_llm/_torch/pyexecutor/py_executor_creator.py	Added validation to reject unsupported configuration combining KV connector with attention data parallelism; check occurs after scheduler/VSWA compatibility validations.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~4 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly and specifically summarizes the main change: adding an assertion that attention DP is disabled when KV connector is in use.
Description check	✅ Passed	The PR description includes a clear summary of the issue and the solution, and explicitly documents the test plan with checkmarks indicating tests have been performed.
Docstring Coverage	✅ Passed	Docstring coverage is 100.00% which is sufficient. The required threshold is 80.00%.
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.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

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

[tensorrt-cicd]

PR_Github #45439 [ run ] triggered by Bot. Commit: afffd53 Link to invocation

[tensorrt-cicd]

PR_Github #45439 [ run ] completed with state FAILURE. Commit: afffd53
/LLM/main/L0_MergeRequest_PR pipeline #35671 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

• Please check the failed tests and fix your PR
• If you cannot view the failures, ask the CI triggerer to share details
• Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

[jthomson04]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #45508 [ run ] triggered by Bot. Commit: 4828836 Link to invocation

[tensorrt-cicd]

PR_Github #45508 [ run ] completed with state SUCCESS. Commit: 4828836
/LLM/main/L0_MergeRequest_PR pipeline #35732 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

• Please check the failed tests and fix your PR
• If you cannot view the failures, ask the CI triggerer to share details
• Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

[jthomson04]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #45517 [ run ] triggered by Bot. Commit: 61d7d3f Link to invocation

[tensorrt-cicd]

PR_Github #45517 [ run ] completed with state SUCCESS. Commit: 61d7d3f
/LLM/main/L0_MergeRequest_PR pipeline #35741 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

• Please check the failed tests and fix your PR
• If you cannot view the failures, ask the CI triggerer to share details
• Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

[pcastonguay]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #45736 [ run ] triggered by Bot. Commit: 61d7d3f Link to invocation

[tensorrt-cicd]

PR_Github #45736 [ run ] completed with state FAILURE. Commit: 61d7d3f
/LLM/main/L0_MergeRequest_PR pipeline #35932 completed with status: 'FAILURE'

CI Report

⚠️ Action Required:

• Please check the failed tests and fix your PR
• If you cannot view the failures, ask the CI triggerer to share details
• Once fixed, request an NVIDIA team member to trigger CI again

Link to invocation

[pcastonguay]

/bot run --disable-fail-fast

[tensorrt-cicd]

PR_Github #45781 [ run ] triggered by Bot. Commit: 61d7d3f Link to invocation

[tensorrt-cicd]

PR_Github #45781 [ run ] completed with state SUCCESS. Commit: 61d7d3f
/LLM/main/L0_MergeRequest_PR pipeline #35971 completed with status: 'SUCCESS'

CI Report

Link to invocation