[Bug]: Unable to serve Llama3 using vLLM Docker container

[vecorro]

Your current environment

Collecting environment information...
PyTorch version: 2.0.1+cu117
Is debug build: False
CUDA used to build PyTorch: 11.7
ROCM used to build PyTorch: N/A

OS: Ubuntu 22.04.4 LTS (x86_64)
GCC version: (Ubuntu 12.3.0-1ubuntu1~22.04) 12.3.0
Clang version: Could not collect
CMake version: version 3.24.4
Libc version: glibc-2.35

Python version: 3.10.13 (main, Sep 11 2023, 13:44:35) [GCC 11.2.0] (64-bit runtime)
Python platform: Linux-6.5.0-28-generic-x86_64-with-glibc2.35
Is CUDA available: True
CUDA runtime version: 12.1.66
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration: GPU 0: GRID A100-40C
Nvidia driver version: 535.154.05
cuDNN version: Probably one of the following:
/usr/lib/x86_64-linux-gnu/libcudnn.so.8.9.7
/usr/lib/x86_64-linux-gnu/libcudnn_adv_infer.so.8.9.7
/usr/lib/x86_64-linux-gnu/libcudnn_adv_train.so.8.9.7
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_infer.so.8.9.7
/usr/lib/x86_64-linux-gnu/libcudnn_cnn_train.so.8.9.7
/usr/lib/x86_64-linux-gnu/libcudnn_ops_infer.so.8.9.7
/usr/lib/x86_64-linux-gnu/libcudnn_ops_train.so.8.9.7
HIP runtime version: N/A
MIOpen runtime version: N/A
Is XNNPACK available: True

CPU:
Architecture:                       x86_64
CPU op-mode(s):                     32-bit, 64-bit
Address sizes:                      45 bits physical, 48 bits virtual
Byte Order:                         Little Endian
CPU(s):                             32
On-line CPU(s) list:                0-31
Vendor ID:                          GenuineIntel
Model name:                         Intel(R) Xeon(R) Gold 6330 CPU @ 2.00GHz
CPU family:                         6
Model:                              106
Thread(s) per core:                 1
Core(s) per socket:                 16
Socket(s):                          2
Stepping:                           6
BogoMIPS:                           3990.62
Flags:                              fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ss ht syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon rep_good nopl xtopology tsc_reliable nonstop_tsc cpuid tsc_known_freq pni pclmulqdq ssse3 fma cx16 pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand hypervisor lahf_lm abm 3dnowprefetch invpcid_single ssbd ibrs ibpb stibp ibrs_enhanced fsgsbase tsc_adjust bmi1 avx2 smep bmi2 erms invpcid avx512f avx512dq rdseed adx smap avx512ifma clflushopt clwb avx512cd sha_ni avx512bw avx512vl xsaveopt xsavec xgetbv1 xsaves wbnoinvd arat avx512vbmi umip pku ospke avx512_vbmi2 gfni vaes vpclmulqdq avx512_vnni avx512_bitalg avx512_vpopcntdq rdpid fsrm md_clear flush_l1d arch_capabilities
Hypervisor vendor:                  VMware
Virtualization type:                full
L1d cache:                          1.5 MiB (32 instances)
L1i cache:                          1 MiB (32 instances)
L2 cache:                           40 MiB (32 instances)
L3 cache:                           84 MiB (2 instances)
NUMA node(s):                       2
NUMA node0 CPU(s):                  0-15
NUMA node1 CPU(s):                  16-31
Vulnerability Gather data sampling: Unknown: Dependent on hypervisor status
Vulnerability Itlb multihit:        KVM: Mitigation: VMX unsupported
Vulnerability L1tf:                 Not affected
Vulnerability Mds:                  Not affected
Vulnerability Meltdown:             Not affected
Vulnerability Mmio stale data:      Vulnerable: Clear CPU buffers attempted, no microcode; SMT Host state unknown
Vulnerability Retbleed:             Not affected
Vulnerability Spec rstack overflow: Not affected
Vulnerability Spec store bypass:    Mitigation; Speculative Store Bypass disabled via prctl
Vulnerability Spectre v1:           Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:           Mitigation; Enhanced / Automatic IBRS, IBPB conditional, RSB filling, PBRSB-eIBRS SW sequence
Vulnerability Srbds:                Not affected
Vulnerability Tsx async abort:      Not affected

Versions of relevant libraries:
[pip3] flake8==3.8.4
[pip3] mypy-extensions==1.0.0
[pip3] numpy==1.25.0
[pip3] nvidia-nccl-cu11==2.14.3
[pip3] onnx==1.14.1
[pip3] onnxruntime-gpu==1.16.0
[pip3] torch==2.0.1
[pip3] triton==2.0.0
[conda] numpy                     1.25.0                   pypi_0    pypi
[conda] nvidia-nccl-cu11          2.14.3                   pypi_0    pypi
[conda] torch                     2.0.1                    pypi_0    pypi
[conda] triton                    2.0.0                    pypi_0    pypiROCM Version: Could not collect
Neuron SDK Version: N/A
vLLM Version: N/A
vLLM Build Flags:
CUDA Archs: Not Set; ROCm: Disabled; Neuron: Disabled
GPU Topology:
GPU0	CPU Affinity	NUMA Affinity	GPU NUMA ID
GPU0	 X 	0-15	0		N/A

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

🐛 Describe the bug

I'm trying to run Llama3 using Docker this way:

sudo docker run --runtime nvidia --gpus all \
	 -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HUGGING_FACE_HUB_TOKEN=TOK" \
    -p 8000:8000 \
    --ipc=host \
    --env CUDA_VISIBLE_DEVICES=0 \
    vllm/vllm-openai:latest \
    --model meta-llama/Meta-Llama-3-8B-Instruct \
    --enforce-eager

My GPU is properly configured:

Thu May  9 18:48:13 2024       
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.154.05             Driver Version: 535.154.05   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  GRID A100-40C                  On  | 00000000:03:00.0 Off |                    0 |
| N/A   N/A    P0              N/A /  N/A |     35MiB / 40960MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+
                                                                                         
+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
|    0   N/A  N/A      2664      G   /usr/lib/xorg/Xorg                           35MiB |
+---------------------------------------------------------------------------------------+

but I get the following error:

/usr/local/lib/python3.10/dist-packages/huggingface_hub/file_download.py:1132: FutureWarning: `resume_download` is deprecated and will be removed in version 1.0.0. Downloads always resume when possible. If you want to force a new download, use `force_download=True`.
  warnings.warn(
INFO 05-10 00:32:00 llm_engine.py:100] Initializing an LLM engine (v0.4.2) with config: model='meta-llama/Meta-Llama-3-8B-Instruct', speculative_config=None, tokenizer='meta-llama/Meta-Llama-3-8B-Instruct', skip_tokenizer_init=False, tokenizer_mode=auto, revision=None, tokenizer_revision=None, trust_remote_code=False, dtype=torch.bfloat16, max_seq_len=8192, download_dir=None, load_format=LoadFormat.AUTO, tensor_parallel_size=1, disable_custom_all_reduce=False, quantization=None, enforce_eager=True, kv_cache_dtype=auto, quantization_param_path=None, device_config=cuda, decoding_config=DecodingConfig(guided_decoding_backend='outlines'), seed=0, served_model_name=meta-llama/Meta-Llama-3-8B-Instruct)
Special tokens have been added in the vocabulary, make sure the associated word embeddings are fine-tuned or trained.
INFO 05-10 00:32:01 utils.py:660] Found nccl from library /root/.config/vllm/nccl/cu12/libnccl.so.2.18.1
INFO 05-10 00:32:02 selector.py:27] Using FlashAttention-2 backend.
[rank0]: Traceback (most recent call last):
[rank0]:   File "/usr/lib/python3.10/runpy.py", line 196, in _run_module_as_main
[rank0]:     return _run_code(code, main_globals, None,
[rank0]:   File "/usr/lib/python3.10/runpy.py", line 86, in _run_code
[rank0]:     exec(code, run_globals)
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/entrypoints/openai/api_server.py", line 168, in <module>
[rank0]:     engine = AsyncLLMEngine.from_engine_args(
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/engine/async_llm_engine.py", line 366, in from_engine_args
[rank0]:     engine = cls(
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/engine/async_llm_engine.py", line 324, in __init__
[rank0]:     self.engine = self._init_engine(*args, **kwargs)
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/engine/async_llm_engine.py", line 442, in _init_engine
[rank0]:     return engine_class(*args, **kwargs)
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/engine/llm_engine.py", line 160, in __init__
[rank0]:     self.model_executor = executor_class(
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/executor/executor_base.py", line 41, in __init__
[rank0]:     self._init_executor()
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/executor/gpu_executor.py", line 23, in _init_executor
[rank0]:     self._init_non_spec_worker()
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/executor/gpu_executor.py", line 68, in _init_non_spec_worker
[rank0]:     self.driver_worker.init_device()
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/worker/worker.py", line 111, in init_device
[rank0]:     init_worker_distributed_environment(self.parallel_config, self.rank,
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/vllm/worker/worker.py", line 313, in init_worker_distributed_environment
[rank0]:     torch.distributed.all_reduce(torch.zeros(1).cuda())
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/torch/distributed/c10d_logger.py", line 75, in wrapper
[rank0]:     return func(*args, **kwargs)
[rank0]:   File "/usr/local/lib/python3.10/dist-packages/torch/distributed/distributed_c10d.py", line 2219, in all_reduce
[rank0]:     work = group.allreduce([tensor], opts)
[rank0]: torch.distributed.DistBackendError: NCCL error in: ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1970, unhandled system error (run with NCCL_DEBUG=INFO for details), NCCL version 2.20.5
[rank0]: ncclSystemError: System call (e.g. socket, malloc) or external library call failed or device error. 
[rank0]: Last error:
[rank0]: nvmlDeviceGetP2PStatus(0,0,NVML_P2P_CAPS_INDEX_READ) failed: Invalid Argument

[vecorro]

It works with the container tag [v0.3.3]. It does not work with either v0.4.2 or v0.4.1

[yspaik]

We have the same symptom as well. we have conducted important function tests on a single A100 using vllm 0.3.3, but after upgrading the version to vllm 0.4 (and vllm 0.4.2 as well), it won't run at all with the same error message.

[ywang96]

@youkaichao Do you have any insights on this?

FWIW, 0.4.1 works for me without the custom all-reduce operation but 0.4.2 does expose some issues as well

[youkaichao]

The error trace points to pytorch distributed, so that's not what I know. It's inside pytorch i think.

The problem looks strange, because you only have 1 GPUs while pytorch tries to read the p2p status between 0 and 0 (essentially the GPU itself).

One educated guess, maybe you can try to upgrade the driver version, I remember several issues can be solved by upgrading to driver 540. 535 seems to be buggy.

[vecorro]

Thank you, @youkaichao. You're right. After upgrading the NVIDIA driver to v550.x vLLM 0.4.2 worked properly