[Bug]: CUDA illegal memory access error when `enable_prefix_caching=True`

[mpoemsl]

Your current environment

The output of `python collect_env.py`
PyTorch version: 2.3.0+cu121
Is debug build: False
CUDA used to build PyTorch: 12.1
ROCM used to build PyTorch: N/A

OS: Ubuntu 20.04.6 LTS (x86_64)
GCC version: (Ubuntu 9.4.0-1ubuntu1~20.04.2) 9.4.0
Clang version: Could not collect
CMake version: version 3.29.3
Libc version: glibc-2.31

Python version: 3.11.9 (main, Apr  6 2024, 17:59:24) [GCC 9.4.0] (64-bit runtime)
Python platform: Linux-5.15.0-107-generic-x86_64-with-glibc2.31
Is CUDA available: True
CUDA runtime version: Could not collect
CUDA_MODULE_LOADING set to: LAZY
GPU models and configuration:
GPU 0: NVIDIA RTX A6000
GPU 1: NVIDIA RTX A6000
GPU 2: NVIDIA RTX A6000
GPU 3: NVIDIA RTX A6000
GPU 4: NVIDIA RTX A6000
GPU 5: NVIDIA RTX A6000
GPU 6: NVIDIA RTX A6000
GPU 7: NVIDIA RTX A6000

Nvidia driver version: 535.161.08
cuDNN version: Could not collect
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
Byte Order:                         Little Endian
Address sizes:                      43 bits physical, 48 bits virtual
CPU(s):                             256
On-line CPU(s) list:                0-254
Off-line CPU(s) list:               255
Thread(s) per core:                 1
Core(s) per socket:                 64
Socket(s):                          2
NUMA node(s):                       2
Vendor ID:                          AuthenticAMD
CPU family:                         23
Model:                              49
Model name:                         AMD EPYC 7662 64-Core Processor
Stepping:                           0
Frequency boost:                    enabled
CPU MHz:                            1499.325
CPU max MHz:                        2154.2959
CPU min MHz:                        1500.0000
BogoMIPS:                           4000.28
Virtualization:                     AMD-V
L1d cache:                          2 MiB
L1i cache:                          2 MiB
L2 cache:                           32 MiB
L3 cache:                           256 MiB
NUMA node0 CPU(s):                  0-63,128-191
NUMA node1 CPU(s):                  64-127,192-254
Vulnerability Gather data sampling: Not affected
Vulnerability Itlb multihit:        Not affected
Vulnerability L1tf:                 Not affected
Vulnerability Mds:                  Not affected
Vulnerability Meltdown:             Not affected
Vulnerability Mmio stale data:      Not affected
Vulnerability Retbleed:             Mitigation; untrained return thunk; SMT enabled with STIBP protection
Vulnerability Spec rstack overflow: Mitigation; safe RET
Vulnerability Spec store bypass:    Mitigation; Speculative Store Bypass disabled via prctl and seccomp
Vulnerability Spectre v1:           Mitigation; usercopy/swapgs barriers and __user pointer sanitization
Vulnerability Spectre v2:           Mitigation; Retpolines; IBPB conditional; STIBP always-on; RSB filling; PBRSB-eIBRS Not affected; BHI Not affected
Vulnerability Srbds:                Not affected
Vulnerability Tsx async abort:      Not affected
Flags:                              fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd amd_ppin arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip rdpid overflow_recov succor smca sme sev sev_es

Versions of relevant libraries:
[pip3] mypy-extensions==1.0.0
[pip3] numpy==1.26.4
[pip3] nvidia-nccl-cu12==2.20.5
[pip3] torch==2.3.0
[pip3] transformers==4.41.1
[pip3] triton==2.3.0
[conda] Could not collect
ROCM Version: Could not collect
Neuron SDK Version: N/A
vLLM Version: 0.5.0.post1
vLLM Build Flags:
CUDA Archs: Not Set; ROCm: Disabled; Neuron: Disabled
GPU Topology:
GPU0	GPU1	GPU2	GPU3	GPU4	GPU5	GPU6	GPU7	CPU Affinity	NUMA Affinity	GPU NUMA ID
GPU0	 X 	NV4	NODE	NODE	SYS	SYS	SYS	SYS	0-63,128-191	0		N/A
GPU1	NV4	 X 	NODE	NODE	SYS	SYS	SYS	SYS	0-63,128-191	0		N/A
GPU2	NODE	NODE	 X 	NV4	SYS	SYS	SYS	SYS	0-63,128-191	0		N/A
GPU3	NODE	NODE	NV4	 X 	SYS	SYS	SYS	SYS	0-63,128-191	0		N/A
GPU4	SYS	SYS	SYS	SYS	 X 	NV4	NODE	NODE	64-127,192-254	1		N/A
GPU5	SYS	SYS	SYS	SYS	NV4	 X 	NODE	NODE	64-127,192-254	1		N/A
GPU6	SYS	SYS	SYS	SYS	NODE	NODE	 X 	NV4	64-127,192-254	1		N/A
GPU7	SYS	SYS	SYS	SYS	NODE	NODE	NV4	 X 	64-127,192-254	1		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

While testing the new version, I ran into this CUDA error (not immediately, after a few successful iterations).

llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [rank1]:[E ProcessGroupNCCL.cpp:1414] [PG 2 Rank 1] Process group watchdog thread terminated with exception: CUDA error: an illegal me
mory access was encountered
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m Exception raised from c10_cuda_check_implementation at ../c10/cuda/CUDAException.cpp:43 (most recent call first):
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fec74098897 in /usr/local/lib/python3.11/dist-packages/torch/
lib/libc10.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::string const&) + 0x64 (0x7fec74048b25 in /usr/local
/lib/python3.11/dist-packages/torch/lib/libc10.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #2: c10::cuda::c10_cuda_check_implementation(int, char const*, char const*, int, bool) + 0x118 (0x7fec8c150718 in /usr/local/lib
/python3.11/dist-packages/torch/lib/libc10_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #3: c10d::ProcessGroupNCCL::WorkNCCL::finishedGPUExecutionInternal() const + 0x56 (0x7fbcd5978e36 in /usr/local/lib/python3.11/d
ist-packages/torch/lib/libtorch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #4: c10d::ProcessGroupNCCL::WorkNCCL::isCompleted() + 0x58 (0x7fbcd597cf38 in /usr/local/lib/python3.11/dist-packages/torch/lib/
libtorch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #5: c10d::ProcessGroupNCCL::watchdogHandler() + 0x77c (0x7fbcd59825ac in /usr/local/lib/python3.11/dist-packages/torch/lib/libto
rch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #6: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7fbcd598331c in /usr/local/lib/python3.11/dist-packages/torch/lib/libt
orch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #7: <unknown function> + 0xd6df4 (0x7fec90c95df4 in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #8: <unknown function> + 0x8609 (0x7fec92bb5609 in /usr/lib/x86_64-linux-gnu/libpthread.so.0)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #9: clone + 0x43 (0x7fec92cef353 in /usr/lib/x86_64-linux-gnu/libc.so.6)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [2024-06-14 10:35:20,297 E 14390 15027] logging.cc:101: Unhandled exception: N3c1016DistBackendErrorE. what(): [PG 2 Rank 1] Process g
roup watchdog thread terminated with exception: CUDA error: an illegal memory access was encountered
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m CUDA kernel errors might be asynchronously reported at some other API call, so the stacktrace below might be incorrect.
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m For debugging consider passing CUDA_LAUNCH_BLOCKING=1.
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m Compile with `TORCH_USE_CUDA_DSA` to enable device-side assertions.
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m Exception raised from c10_cuda_check_implementation at ../c10/cuda/CUDAException.cpp:43 (most recent call first):
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fec74098897 in /usr/local/lib/python3.11/dist-packages/torch/
lib/libc10.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::string const&) + 0x64 (0x7fec74048b25 in /usr/local
/lib/python3.11/dist-packages/torch/lib/libc10.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #2: c10::cuda::c10_cuda_check_implementation(int, char const*, char const*, int, bool) + 0x118 (0x7fec8c150718 in /usr/local/lib
/python3.11/dist-packages/torch/lib/libc10_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #3: c10d::ProcessGroupNCCL::WorkNCCL::finishedGPUExecutionInternal() const + 0x56 (0x7fbcd5978e36 in /usr/local/lib/python3.11/d
ist-packages/torch/lib/libtorch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #4: c10d::ProcessGroupNCCL::WorkNCCL::isCompleted() + 0x58 (0x7fbcd597cf38 in /usr/local/lib/python3.11/dist-packages/torch/lib/
libtorch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #5: c10d::ProcessGroupNCCL::watchdogHandler() + 0x77c (0x7fbcd59825ac in /usr/local/lib/python3.11/dist-packages/torch/lib/libto
rch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #6: c10d::ProcessGroupNCCL::ncclCommWatchdog() + 0x10c (0x7fbcd598331c in /usr/local/lib/python3.11/dist-packages/torch/lib/libt
orch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #7: <unknown function> + 0xd6df4 (0x7fec90c95df4 in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #8: <unknown function> + 0x8609 (0x7fec92bb5609 in /usr/lib/x86_64-linux-gnu/libpthread.so.0)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #9: clone + 0x43 (0x7fec92cef353 in /usr/lib/x86_64-linux-gnu/libc.so.6)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m Exception raised from ncclCommWatchdog at ../torch/csrc/distributed/c10d/ProcessGroupNCCL.cpp:1418 (most recent call first):
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #0: c10::Error::Error(c10::SourceLocation, std::string) + 0x57 (0x7fec74098897 in /usr/local/lib/python3.11/dist-packages/torch/
lib/libc10.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #1: <unknown function> + 0xe32e33 (0x7fbcd5605e33 in /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #2: <unknown function> + 0xd6df4 (0x7fec90c95df4 in /usr/lib/x86_64-linux-gnu/libstdc++.so.6)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #3: <unknown function> + 0x8609 (0x7fec92bb5609 in /usr/lib/x86_64-linux-gnu/libpthread.so.0)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m frame #4: clone + 0x43 (0x7fec92cef353 in /usr/lib/x86_64-linux-gnu/libc.so.6)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [2024-06-14 10:35:20,329 E 14390 15027] logging.cc:108: Stack trace:
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m  /usr/local/lib/python3.11/dist-packages/ray/_raylet.so(+0x101867a) [0x7fec91dc367a] ray::operator<<()
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/local/lib/python3.11/dist-packages/ray/_raylet.so(+0x101b138) [0x7fec91dc6138] ray::TerminateHandler()
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0xaa37c) [0x7fec90c6937c]
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0xaa3e7) [0x7fec90c693e7]
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0xaa36f) [0x7fec90c6936f]
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/local/lib/python3.11/dist-packages/torch/lib/libtorch_cuda.so(+0xe32ee4) [0x7fbcd5605ee4] c10d::ProcessGroupNCCL::ncclCommWatchdo
g()
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/lib/x86_64-linux-gnu/libstdc++.so.6(+0xd6df4) [0x7fec90c95df4]
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/lib/x86_64-linux-gnu/libpthread.so.0(+0x8609) [0x7fec92bb5609] start_thread
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m /usr/lib/x86_64-linux-gnu/libc.so.6(clone+0x43) [0x7fec92cef353] __clone
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m *** SIGABRT received at time=1718361320 on cpu 156 ***
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m PC: @     0x7fec92c1300b  (unknown)  raise
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m     @     0x7fec92c13090       2192  (unknown)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m     @     0x7fec90c6937c  (unknown)  (unknown)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m     @     0x7fec90c69090  (unknown)  (unknown)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [2024-06-14 10:35:20,331 E 14390 15027] logging.cc:365: *** SIGABRT received at time=1718361320 on cpu 156 ***
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [2024-06-14 10:35:20,331 E 14390 15027] logging.cc:365: PC: @     0x7fec92c1300b  (unknown)  raise
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [2024-06-14 10:35:20,331 E 14390 15027] logging.cc:365:     @     0x7fec92c13090       2192  (unknown)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [2024-06-14 10:35:20,331 E 14390 15027] logging.cc:365:     @     0x7fec90c6937c  (unknown)  (unknown)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m [2024-06-14 10:35:20,332 E 14390 15027] logging.cc:365:     @     0x7fec90c69090  (unknown)  (unknown)
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m Fatal Python error: Aborted
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14390)ESC[0m Extension modules: msgpack._cmsgpack, google._upb._message, psutil._psutil_linux, psutil._psutil_posix, setproctitle, yaml._yaml, char
set_normalizer.md, uvloop.loop, ray._raylet, numpy.core._multiarray_umath, numpy.core._multiarray_tests, numpy.linalg._umath_linalg, numpy.fft._pocketfft_internal, numpy.random._common, numpy.random.bit_generator,
numpy.random._bounded_integers, numpy.random._mt19937, numpy.random.mtrand, numpy.random._philox, numpy.random._pcg64, numpy.random._sfc64, numpy.random._generator, torch._C, torch._C._fft, torch._C._linalg, torch.
_C._nested, torch._C._nn, torch._C._sparse, torch._C._special, sentencepiece._sentencepiece, pyarrow.lib, pyarrow._json, PIL._imaging, cuda_utils, __triton_launcher (total: 35)
llm-inference-mixtral-20240612-2-1  | ESC[33m(raylet)ESC[0m A worker died or was killed while executing a task by an unexpected system error. To troubleshoot the problem, check the logs for the dead worker. RayTask
 ID: ffffffffffffffff268bbf7e251c226eef173be001000000 Worker ID: 0a04e75572848f8edda22427a258283dfdee451cc8858a13c0c1bd7e Node ID: 6461451955b118f9eefdf9dfc3b248c3ce82bd7e1809df449be56a98 Worker IP address: 172.21.
0.18 Worker port: 40797 Worker PID: 14390 Worker exit type: SYSTEM_ERROR Worker exit detail: Worker unexpectedly exits with a connection error code 2. End of file. There are some potential root causes. (1) The proc
ess is killed by SIGKILL by OOM killer due to high memory usage. (2) ray stop --force is called. (3) The worker is crashed unexpectedly due to SIGSEGV or other unexpected errors.
llm-inference-mixtral-20240612-2-1  | ESC[36m(RayWorkerWrapper pid=14862)ESC[0m INFO 06-14 10:33:31 model_runner.py:965] Graph capturing finished in 13 secs.ESC[32m [repeated 2x across cluster]ESC[0m
llm-inference-mixtral-20240612-2-1  | ERROR 06-14 10:36:20 async_llm_engine.py:535] Engine iteration timed out. This should never happen!
llm-inference-mixtral-20240612-2-1  | ERROR 06-14 10:36:20 async_llm_engine.py:52] Engine background task failed

The only change compared to the previously working setup was the VLLM version upgrade and enabling prefix caching. The model served is Mixtral 8x7B (unquantized).

Note that CUDA_VISIBLE_DEVICES was set to 0,1,2,3 with tensor_parallel=4 to only use the first four GPUs (the machine has eight in total).

Update: no issues with v0.4.3