[https://nvbugs/5450074][fix] Reduce the device memory requirements for testing #6990
Conversation
📝 WalkthroughWalkthroughAdded a null check for ctx_request_id in the context-phase response handling of the disaggregated server, raising ValueError if missing. Updated an integration test config to lower max_batch_size for context and generation servers. Changes
Sequence Diagram(s)sequenceDiagram
participant Client
participant OpenAI_Disagg_Server as Disaggregated Server
participant Context_Server
Client->>OpenAI_Disagg_Server: Context request
OpenAI_Disagg_Server->>Context_Server: Forward request
Context_Server-->>OpenAI_Disagg_Server: Context response
alt Validate response
OpenAI_Disagg_Server->>OpenAI_Disagg_Server: Check choices==1, params!=None, ctx_request_id!=None
opt Invalid
OpenAI_Disagg_Server-->>Client: Raise ValueError
end
opt Valid
OpenAI_Disagg_Server-->>Client: Proceed to next phase
end
end
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Tip 🔌 Remote MCP (Model Context Protocol) integration is now available!Pro plan users can now connect to remote MCP servers from the Integrations page. Connect with popular remote MCPs such as Notion and Linear to add more context to your reviews and chats. 📜 Recent review detailsConfiguration used: .coderabbit.yaml 💡 Knowledge Base configuration:
You can enable these sources in your CodeRabbit configuration. 📒 Files selected for processing (2)
🚧 Files skipped from review as they are similar to previous changes (1)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
🔇 Additional comments (2)
✨ Finishing Touches
🧪 Generate unit tests
Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. 🪧 TipsChatThere are 3 ways to chat with CodeRabbit:
SupportNeed help? Create a ticket on our support page for assistance with any issues or questions. CodeRabbit Commands (Invoked using PR/Issue comments)Type Other keywords and placeholders
Status, Documentation and Community
|
Shixiaowei02
force-pushed
the
xiaoweis/fix-1.0
branch
from
c87de60 to
e555d83
Compare
Shixiaowei02
changed the title
|
/bot run |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 0
🔭 Outside diff range comments (1)
tensorrt_llm/serve/openai_disagg_server.py (1)
202-205: Harden choice count validation (zero choices will currently fall through and crash)If choices is empty, choices[0] will raise. Align with your other check that enforces “exactly one choice.”
Apply this diff:
- choices = ctx_response.choices - if len(choices) > 1: - raise ValueError("Disagg server returned more than one choice. This is currently not supported in disaggregated server.") + choices = ctx_response.choices + if len(choices) != 1: + raise ValueError("Context server did not return a single choice. This is not expected")
🧹 Nitpick comments (6)
tensorrt_llm/serve/openai_disagg_server.py (5)
207-208: Make the error message actionable by including ctx_server and the offending fieldIncluding the server URL and field name significantly speeds up triage when a backend misbehaves.
Apply this diff:
- raise ValueError("Invalid disaggregated params in context phase response.") + raise ValueError(f"Invalid disaggregated params in context phase response from {ctx_server}: ctx_request_id is None")
147-147: Fix error messages that intend to interpolate type namesThe braces won’t interpolate without f-strings, making the messages confusing when debugging.
Apply this diff:
- raise TypeError("Invalid request type: {type(gen_req).__name__}") + raise TypeError(f"Invalid request type: {type(gen_req).__name__}") @@ - raise ValueError("Invalid request type: {type(req).__name__}") + raise ValueError(f"Invalid request type: {type(req).__name__}")Also applies to: 259-259
106-111: create_error_response ignores the caller-provided status and always raises 500This will misclassify client vs server errors. If this helper is used, it should honor the status code and not mask it as 500.
Apply this diff:
- def create_error_response( - message: str, - status_code: HTTPStatus = HTTPStatus.BAD_REQUEST) -> ErrorResponse: - raise HTTPException(status_code=500, detail=f"Internal server error {message}") + def create_error_response( + message: str, + status_code: HTTPStatus = HTTPStatus.BAD_REQUEST) -> None: + # Always raise; callers rely on exception flow. Preserve the provided status. + raise HTTPException(status_code=int(status_code), detail=str(message))Additionally, add this import near the other typing imports:
from typing import NoReturn # if you prefer `-> NoReturn` as return type
34-34: Typo prevents re-enabling yapf formatting“enale” should be “enable”; otherwise yapf may remain disabled for the rest of the file.
Apply this diff:
-# yapf: enale +# yapf: enable
385-394: Avoid shadowing builtins (“iter”) in readiness loopMinor readability nit: use a different name to avoid shadowing the built-in iter().
Apply this diff:
- iter = 0 + attempt = 0 @@ - logger.info( - f"[{iter}] Servers are not ready. Waiting for {unready_servers}..." - ) + logger.info(f"[{attempt}] Servers are not ready. Waiting for {unready_servers}...") @@ - iter += 1 + attempt += 1tests/integration/defs/disaggregated/test_configs/disagg_config_diff_max_tokens.yaml (1)
1-24: Batch/Token Limits in Disaggregated Test ConfigsI scanned all
tests/integration/defs/disaggregatedYAMLs and found that while you’ve added lower values in each file, most still include a second “default” section with high limits (e.g.max_batch_size: 256,max_num_tokens: 4096). To keep test memory usage consistently low, consider aligning or removing those high defaults.• Representative files with high defaults:
– test_configs/disagg_config_cache_aware_balance.yaml
– test_configs/disagg_config_ctxpp2_genpp2.yaml
– test_configs/disagg_config_ctxtp2_genpp2.yaml
– test_configs/disagg_config_ctxtp2pp2_gentp2pp2.yaml
– …and several others that still list256/4096.Recommendation: Reduce or remove the second block of limits in each config so tests only run with the smaller batch/token settings you intend.
📜 Review details
Configuration used: .coderabbit.yaml
Review profile: CHILL
Plan: Pro
💡 Knowledge Base configuration:
- MCP integration is disabled by default for public repositories
- Jira integration is disabled by default for public repositories
- Linear integration is disabled by default for public repositories
You can enable these sources in your CodeRabbit configuration.
📒 Files selected for processing (2)
tensorrt_llm/serve/openai_disagg_server.py(1 hunks)tests/integration/defs/disaggregated/test_configs/disagg_config_diff_max_tokens.yaml(1 hunks)
🧰 Additional context used
📓 Path-based instructions (2)
**/*.py
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
**/*.py: Python code must target Python 3.8+
Python indentation: 4 spaces, no tabs
Maintain module namespace in imports (from package.subpackage import foo; then use foo.SomeClass())
Python file names use snake_case
Python class names use PascalCase
Python functions/methods and local variables use snake_case; variables starting with a number get k_ prefix (e.g., k_99th_percentile)
Global variables use G_ prefixed UPPER_SNAKE_CASE (e.g., G_MY_GLOBAL)
Constants use UPPER_SNAKE_CASE in Python
Avoid shadowing variables from outer scopes in Python
Initialize all externally visible members of a Python class in init
Prefer docstrings for interfaces used outside a file; comments for local code
Use Google-style docstrings for classes and functions (Sphinx-parsable)
Document attributes/variables inline with short docstrings
Avoid reflection when simple alternatives exist (e.g., prefer explicit parameters over dict(**locals()))
In try/except, catch the narrowest exceptions possible
For duck-typing with try/except, keep try body minimal and put logic in else
Files:
tensorrt_llm/serve/openai_disagg_server.py
**/*.{cpp,cxx,cc,cu,h,hpp,hxx,hh,cuh,py}
📄 CodeRabbit Inference Engine (CODING_GUIDELINES.md)
Prepend NVIDIA copyright header (current year) to all source files
Files:
tensorrt_llm/serve/openai_disagg_server.py
🔇 Additional comments (3)
tensorrt_llm/serve/openai_disagg_server.py (1)
207-208: Good defensive check for missing ctx_request_idThe explicit None-check on ctx_request_id before proceeding to generation helps catch malformed upstream responses early.
tests/integration/defs/disaggregated/test_configs/disagg_config_diff_max_tokens.yaml (2)
11-11: Lowering context max_batch_size to 64 aligns with the PR goal to reduce test-time VRAM footprintChange looks good and should help tests pass on smaller GPUs.
19-19: Lowering generation max_batch_size to 32 further reduces device memory usageLGTM for memory-constrained CI/test environments.
|
/bot run |
Shixiaowei02
force-pushed
the
xiaoweis/fix-1.0
branch
from
e555d83 to
0946f6e
Compare
|
/bot run |
|
PR_Github #15685 [ run ] triggered by Bot |
|
PR_Github #15685 [ run ] completed with state |
Signed-off-by: Shixiaowei02 <[email protected]>
Shixiaowei02
force-pushed
the
xiaoweis/fix-1.0
branch
from
0946f6e to
2cd62b9
Compare
|
/bot skip --comment "It has passed in 15685" |
|
PR_Github #16014 [ skip ] triggered by Bot |
|
PR_Github #16014 [ skip ] completed with state |
…or testing (#6990) Signed-off-by: Shixiaowei02 <[email protected]>
…or testing (NVIDIA#6990) Signed-off-by: Shixiaowei02 <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…or testing (NVIDIA#6990) Signed-off-by: Shixiaowei02 <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…or testing (NVIDIA#6990) Signed-off-by: Shixiaowei02 <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…or testing (NVIDIA#6990) Signed-off-by: Shixiaowei02 <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
…or testing (NVIDIA#6990) Signed-off-by: Shixiaowei02 <[email protected]> Signed-off-by: Wangshanshan <[email protected]>
5 participants
Summary by CodeRabbit
Description
Test Coverage
GitHub Bot Help
/bot [-h] ['run', 'kill', 'skip', 'reuse-pipeline'] ...Provide a user friendly way for developers to interact with a Jenkins server.
Run
/bot [-h|--help]to print this help message.See details below for each supported subcommand.
Details
run [--reuse-test (optional)pipeline-id --disable-fail-fast --skip-test --stage-list "A10-PyTorch-1, xxx" --gpu-type "A30, H100_PCIe" --test-backend "pytorch, cpp" --add-multi-gpu-test --only-multi-gpu-test --disable-multi-gpu-test --post-merge --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx" --detailed-log --debug(experimental)]Launch build/test pipelines. All previously running jobs will be killed.
--reuse-test (optional)pipeline-id(OPTIONAL) : Allow the new pipeline to reuse build artifacts and skip successful test stages from a specified pipeline or the last pipeline if no pipeline-id is indicated. If the Git commit ID has changed, this option will be always ignored. The DEFAULT behavior of the bot is to reuse build artifacts and successful test results from the last pipeline.--disable-reuse-test(OPTIONAL) : Explicitly prevent the pipeline from reusing build artifacts and skipping successful test stages from a previous pipeline. Ensure that all builds and tests are run regardless of previous successes.--disable-fail-fast(OPTIONAL) : Disable fail fast on build/tests/infra failures.--skip-test(OPTIONAL) : Skip all test stages, but still run build stages, package stages and sanity check stages. Note: Does NOT update GitHub check status.--stage-list "A10-PyTorch-1, xxx"(OPTIONAL) : Only run the specified test stages. Examples: "A10-PyTorch-1, xxx". Note: Does NOT update GitHub check status.--gpu-type "A30, H100_PCIe"(OPTIONAL) : Only run the test stages on the specified GPU types. Examples: "A30, H100_PCIe". Note: Does NOT update GitHub check status.--test-backend "pytorch, cpp"(OPTIONAL) : Skip test stages which don't match the specified backends. Only support [pytorch, cpp, tensorrt, triton]. Examples: "pytorch, cpp" (does not run test stages with tensorrt or triton backend). Note: Does NOT update GitHub pipeline status.--only-multi-gpu-test(OPTIONAL) : Only run the multi-GPU tests. Note: Does NOT update GitHub check status.--disable-multi-gpu-test(OPTIONAL) : Disable the multi-GPU tests. Note: Does NOT update GitHub check status.--add-multi-gpu-test(OPTIONAL) : Force run the multi-GPU tests in addition to running L0 pre-merge pipeline.--post-merge(OPTIONAL) : Run the L0 post-merge pipeline instead of the ordinary L0 pre-merge pipeline.--extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx"(OPTIONAL) : Run the ordinary L0 pre-merge pipeline and specified test stages. Examples: --extra-stage "H100_PCIe-TensorRT-Post-Merge-1, xxx".--detailed-log(OPTIONAL) : Enable flushing out all logs to the Jenkins console. This will significantly increase the log volume and may slow down the job.--debug(OPTIONAL) : Experimental feature. Enable access to the CI container for debugging purpose. Note: Specify exactly one stage in thestage-listparameter to access the appropriate container environment. Note: Does NOT update GitHub check status.For guidance on mapping tests to stage names, see
docs/source/reference/ci-overview.mdand the
scripts/test_to_stage_mapping.pyhelper.kill
killKill all running builds associated with pull request.
skip
skip --comment COMMENTSkip testing for latest commit on pull request.
--comment "Reason for skipping build/test"is required. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.reuse-pipeline
reuse-pipelineReuse a previous pipeline to validate current commit. This action will also kill all currently running builds associated with the pull request. IMPORTANT NOTE: This is dangerous since lack of user care and validation can cause top of tree to break.