[Bug] Service crashed with 4 H100s and QPS=25

[yh-yao]

Checklist

• 1. I have searched related issues but cannot get the expected help.
• 2. The bug has not been fixed in the latest version.
• 3. Please note that if the bug-related issue you submitted lacks corresponding environment info and a minimal reproducible demo, it will be challenging for us to reproduce and resolve the issue, reducing the likelihood of receiving feedback.
• 4. If the issue you raised is not a bug but a question, please raise a discussion at https://github.com/sgl-project/sglang/discussions/new/choose Otherwise, it will be closed.
• 5. Please use English, otherwise it will be closed.

Describe the bug

The serving is OK at the start. GPU usage was OK at 80%-99%.

One GPU usage suddenly increases to 100%. The other 3 GPUs running 80%-99% then become 0%.

It is then unable to hand in the requests. Although it shows OK, it is not handled.

Reproduction

docker run --gpus '"device=4,5,6,7"' \ --shm-size 300g \ -p 30001:30001 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=XXXXXXXXXXXXX" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server --model-path meta-llama/Llama-3.1-8B-Instruct --data-parallel-size 4 --disable-overlap --host 0.0.0.0 --port 30001

A private model fine-tuned from meta-llama/Llama-3.1-8B-Instruct with context length 8192.

Environment

`/usr/local/lib/python3.10/dist-packages/pydantic/_internal/_config.py:345: UserWarning: Valid config keys have changed in V2:

• 'fields' has been removed
  warnings.warn(message, UserWarning)
  Python: 3.10.16 (main, Dec 4 2024, 08:53:37) [GCC 9.4.0]
  CUDA available: True
  GPU 0,1,2,3,4: NVIDIA H100 80GB HBM3
  GPU 0,1,2,3,4 Compute Capability: 9.0
  CUDA_HOME: /usr/local/cuda
  NVCC: Cuda compilation tools, release 12.4, V12.4.131
  CUDA Driver Version: 525.147.05
  PyTorch: 2.5.1+cu124
  flashinfer: 0.1.6+cu124torch2.4
  triton: 3.1.0
  transformers: 4.48.0
  torchao: 0.8.0
  numpy: 1.26.4
  aiohttp: 3.11.11
  fastapi: 0.115.6
  hf_transfer: 0.1.9
  huggingface_hub: 0.27.1
  interegular: 0.3.3
  modelscope: 1.22.3
  orjson: 3.10.15
  packaging: 24.2
  psutil: 6.1.1
  pydantic: 2.10.5
  multipart: 0.0.20
  zmq: 26.2.0
  uvicorn: 0.34.0
  uvloop: 0.21.0
  vllm: 0.6.4.post1
  openai: 1.59.8
  anthropic: 0.43.1
  decord: 0.6.0
  NVIDIA Topology:
  GPU0 GPU1 GPU2 GPU3 GPU4 NIC0 NIC1 NIC2 NIC3 NIC4 NIC5 NIC6 NIC7 NIC8 NIC9 NIC10 NIC11 CPU Affinity NUMA Affinity
  GPU0 X NV18 NV18 NV18 NV18 NODE NODE NODE NODE NODE PXB SYS SYS SYS SYS SYS SYS 0-55,112-167 0
  GPU1 NV18 X NV18 NV18 NV18 SYS SYS SYS SYS SYS SYS PXB NODE NODE NODE NODE NODE 56-111,168-223 1
  GPU2 NV18 NV18 X NV18 NV18 SYS SYS SYS SYS SYS SYS NODE NODE NODE PXB NODE NODE 56-111,168-223 1
  GPU3 NV18 NV18 NV18 X NV18 SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE PXB NODE 56-111,168-223 1
  GPU4 NV18 NV18 NV18 NV18 X SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE PXB 56-111,168-223 1
  NIC0 NODE SYS SYS SYS SYS X NODE NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS
  NIC1 NODE SYS SYS SYS SYS NODE X PIX NODE NODE NODE SYS SYS SYS SYS SYS SYS
  NIC2 NODE SYS SYS SYS SYS NODE PIX X NODE NODE NODE SYS SYS SYS SYS SYS SYS
  NIC3 NODE SYS SYS SYS SYS NODE NODE NODE X NODE NODE SYS SYS SYS SYS SYS SYS
  NIC4 NODE SYS SYS SYS SYS NODE NODE NODE NODE X NODE SYS SYS SYS SYS SYS SYS
  NIC5 PXB SYS SYS SYS SYS NODE NODE NODE NODE NODE X SYS SYS SYS SYS SYS SYS
  NIC6 SYS PXB NODE NODE NODE SYS SYS SYS SYS SYS SYS X NODE NODE NODE NODE NODE
  NIC7 SYS NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS NODE X PIX NODE NODE NODE
  NIC8 SYS NODE NODE NODE NODE SYS SYS SYS SYS SYS SYS NODE PIX X NODE NODE NODE
  NIC9 SYS NODE PXB NODE NODE SYS SYS SYS SYS SYS SYS NODE NODE NODE X NODE NODE
  NIC10 SYS NODE NODE PXB NODE SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE X NODE
  NIC11 SYS NODE NODE NODE PXB SYS SYS SYS SYS SYS SYS NODE NODE NODE NODE NODE X

Legend:

X = Self
SYS = Connection traversing PCIe as well as the SMP interconnect between NUMA nodes (e.g., QPI/UPI)
NODE = Connection traversing PCIe as well as the interconnect between PCIe Host Bridges within a NUMA node
PHB = Connection traversing PCIe as well as a PCIe Host Bridge (typically the CPU)
PXB = Connection traversing multiple PCIe bridges (without traversing the PCIe Host Bridge)
PIX = Connection traversing at most a single PCIe bridge
NV# = Connection traversing a bonded set of # NVLinks

NIC Legend:

NIC0: mlx5_0
NIC1: mlx5_1
NIC2: mlx5_2
NIC3: mlx5_3
NIC4: mlx5_4
NIC5: mlx5_5
NIC6: mlx5_6
NIC7: mlx5_7
NIC8: mlx5_8
NIC9: mlx5_9
NIC10: mlx5_10
NIC11: mlx5_11

ulimit soft: 1048576
`

[zhaochenyang20]

What do you mean by not handle? Maybe it's just take some time? Also, you can try https://docs.sglang.ai/router/router.html instead of dp-size.