[Bug] Segmentation fault occurs and the machine with openEuler os was automatically reboots

[jiajie-yang]

Checklist

• 1. I have searched related issues but cannot get the expected help.
• 2. The bug has not been fixed in the latest version.

Describe the bug

In my A40/T4 card environment, when the service of lmdeploy is run, after the service address is printed out, a Segmentation fault occurs and exits without any other error message. The most important point is that about a minute after the above error occurred, my machine was automatically restarted very quickly, which is very confusing!

Reproduction

lmdeploy serve api_server internlm2-chat-7b --tp 2 --log-level DEBUG

Environment

sys.platform: linux
Python: 3.10.14 (main, May  6 2024, 19:42:50) [GCC 11.2.0]
CUDA available: True
MUSA available: False
numpy_random_seed: 2147483648
GPU 0,1,2,3,4,5,6,7: NVIDIA A40
CUDA_HOME: None
GCC: gcc (GCC) 10.3.1
PyTorch: 2.2.2+cu121
PyTorch compiling details: PyTorch built with:
  - GCC 9.3
  - C++ Version: 201703
  - Intel(R) oneAPI Math Kernel Library Version 2022.2-Product Build 20220804 for Intel(R) 64 architecture applications
  - Intel(R) MKL-DNN v3.3.2 (Git Hash 2dc95a2ad0841e29db8b22fbccaf3e5da7992b01)
  - OpenMP 201511 (a.k.a. OpenMP 4.5)
  - LAPACK is enabled (usually provided by MKL)
  - NNPACK is enabled
  - CPU capability usage: AVX512
  - CUDA Runtime 12.1
  - NVCC architecture flags: -gencode;arch=compute_50,code=sm_50;-gencode;arch=compute_60,code=sm_60;-gencode;arch=compute_70,code=sm_70;-gencode;arch=compute_75,code=sm_75;-gencode;arch=compute_80,code=sm_80;-gencode;arch=compute_86,code=sm_86;-gencode;arch=compute_90,code=sm_90
  - CuDNN 8.9.2
  - Magma 2.6.1
  - Build settings: BLAS_INFO=mkl, BUILD_TYPE=Release, CUDA_VERSION=12.1, CUDNN_VERSION=8.9.2, CXX_COMPILER=/opt/rh/devtoolset-9/root/usr/bin/c++, CXX_FLAGS= -D_GLIBCXX_USE_CXX11_ABI=0 -fabi-version=11 -fvisibility-inlines-hidden -DUSE_PTHREADPOOL -DNDEBUG -DUSE_KINETO -DLIBKINETO_NOROCTRACER -DUSE_FBGEMM -DUSE_QNNPACK -DUSE_PYTORCH_QNNPACK -DUSE_XNNPACK -DSYMBOLICATE_MOBILE_DEBUG_HANDLE -O2 -fPIC -Wall -Wextra -Werror=return-type -Werror=non-virtual-dtor -Werror=bool-operation -Wnarrowing -Wno-missing-field-initializers -Wno-type-limits -Wno-array-bounds -Wno-unknown-pragmas -Wno-unused-parameter -Wno-unused-function -Wno-unused-result -Wno-strict-overflow -Wno-strict-aliasing -Wno-stringop-overflow -Wsuggest-override -Wno-psabi -Wno-error=pedantic -Wno-error=old-style-cast -Wno-missing-braces -fdiagnostics-color=always -faligned-new -Wno-unused-but-set-variable -Wno-maybe-uninitialized -fno-math-errno -fno-trapping-math -Werror=format -Wno-stringop-overflow, LAPACK_INFO=mkl, PERF_WITH_AVX=1, PERF_WITH_AVX2=1, PERF_WITH_AVX512=1, TORCH_VERSION=2.2.2, USE_CUDA=ON, USE_CUDNN=ON, USE_EXCEPTION_PTR=1, USE_GFLAGS=OFF, USE_GLOG=OFF, USE_MKL=ON, USE_MKLDNN=ON, USE_MPI=OFF, USE_NCCL=1, USE_NNPACK=ON, USE_OPENMP=ON, USE_ROCM=OFF, USE_ROCM_KERNEL_ASSERT=OFF, 

TorchVision: 0.17.2+cu121
LMDeploy: 0.5.0+
transformers: 4.41.2
gradio: Not Found
fastapi: 0.111.0
pydantic: 2.8.0
triton: 2.2.0

+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.86.10              Driver Version: 535.86.10    CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA A40                     On  | 00000000:0B:00.0 Off |                    0 |
|  0%   52C    P0              84W / 300W |      4MiB / 46068MiB |      0%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
|   1  NVIDIA A40                     On  | 00000000:0C:00.0 Off |                    0 |
|  0%   53C    P0              85W / 300W |      4MiB / 46068MiB |      0%      Default |
|                                         |                      |                  N/A |
+-----------------------------------------+----------------------+----------------------+
...

Error traceback

HINT:    Please open http://0.0.0.0:23333 in a browser for detailed api usage!!!
HINT:    Please open http://0.0.0.0:23333 in a browser for detailed api usage!!!
HINT:    Please open http://0.0.0.0:23333 in a browser for detailed api usage!!!
Segmentation fault

[jiajie-yang]

My lmdeploy is installed by pip install lmdeploy. The system and kernel information is as follows:

$ cat /etc/os-release 
NAME="openEuler"
VERSION="22.03 LTS"
ID="openEuler"
VERSION_ID="22.03"
PRETTY_NAME="openEuler 22.03 LTS"
ANSI_COLOR="0;31"

$ uname -a
Linux standalone-amd64-a40-26-163 5.10.0-60.94.0.118.oe2203.x86_64 #1 SMP Mon Sep 4 12:14:17 CST 2023 x86_64 x86_64 x86_64 GNU/Linux

[jiajie-yang]

Is it because the system is openEuler? There doesn't seem to be this problem on centos and ubuntu systems.

[zhyncs]

The Python package is compiled with manylinux2014_x86_64 and I only have the server with CentOS 7 or Ubuntu 22.04. Maybe it's a corner case and hard for me to reproduce because I don't have the environment.

[lvhan028]

Is it because the system is openEuler? There doesn't seem to be this problem on centos and ubuntu systems.

Can you try to build the project from source? Here is the guide: https://lmdeploy.readthedocs.io/en/latest/build.html

[jiajie-yang]

The Python package is compiled with manylinux2014_x86_64 and I only have the server with CentOS 7 or Ubuntu 22.04. Maybe it's a corner case and hard for me to reproduce because I don't have the environment.

It is almost certain that the kernel version (5.10.0-60.94.0.118.oe2203.x86_64) is the reason. Because it seems that upgrading or downgrading the kernel version will solve the problem.

[jiajie-yang]

Is it because the system is openEuler? There doesn't seem to be this problem on centos and ubuntu systems.

Can you try to build the project from source? Here is the guide: https://lmdeploy.readthedocs.io/en/latest/build.html

That doesn't seem to be necessary, as it seems that upgrading or downgrading the kernel version will solve the problem.