mindie 2.0 版本部署qwen 2.5vl 72B，并发数为40时出现如下错误Segmentation fault (core dumped)，导致服务直接挂死_MindIE_昇腾论坛

# mindie 2.0 版本 部署qwen 2.5vl 72B，并发数为40时 出现如下错误Segmentation fault (core dumped)，导致服务直接挂死_MindIE_昇腾论坛

## 概述

本文档基于昇腾社区论坛帖子生成的技术教程。

**原始链接**: https://www.hiascend.com/forum/thread-0267186315092089043-1-1.html

**生成时间**: 2025-08-27 10:34:00

---

## 问题描述

SystemInfo 硬件 910BC3 mindie 2.0.RC2 config.json { "Version": "1.0.0", "ServerConfig": { "ipAddress": "172.17.0.10", "managementIpAddress": "127.0.0.2", "port": 8000, "managementPort": 1026, "metricsPort": 1027, "allowAllZeroIpListening": false, "maxLinkNum": 1000, "httpsEnabled": false, "fullTextEnabled": false, "tlsCaPath": "security/ca/", "tlsCaFile": [ "ca.pem" ], "tlsCert": "security/certs/server.pem", "tlsPk": "security/keys/server.key.pem", "tlsPkPwd": "security/pass/key_pwd.txt", "tlsCrlPath": "security/certs/", "tlsCrlFiles": [ "server_crl.pem" ], "managementTlsCaFile": [ "management_ca.pem" ], "managementTlsCert": "security/certs/management/server.pem", "managementTlsPk": "security/keys/management/server.key.pem", "managementTlsPkPwd": "security/pass/management/key_pwd.txt", "managementTlsCrlPath": "security/management/certs/", "managementTlsCrlFiles": [ "server_crl.pem" ], "kmcKsfMaster": "tools/pmt/master/ksfa", "kmcKsfStandby": "tools/pmt/standby/ksfb", "inferMode": "standa...

## 相关代码

### 代码示例 1

```
{
"Version": "1.0.0",
"ServerConfig": {
"ipAddress": "172.17.0.10",
"managementIpAddress": "127.0.0.2",
"port": 8000,
"managementPort": 1026,
"metricsPort": 1027,
"allowAllZeroIpListening": false,
"maxLinkNum": 1000,
"httpsEnabled": false,
"fullTextEnabled": false,
"tlsCaPath": "security/ca/",
"tlsCaFile": [
"ca.pem"
],
"tlsCert": "security/certs/server.pem",
"tlsPk": "security/keys/server.key.pem",
"tlsPkPwd": "security/pass/key_pwd.txt",
"tlsCrlPath": "security/certs/",
"tlsCrlFiles": [
"server_crl.pem"
],
"managementTlsCaFile": [
"management_ca.pem"
],
"managementTlsCert": "security/certs/management/server.pem",
"managementTlsPk": "security/keys/management/server.key.pem",
"managementTlsPkPwd": "security/pass/management/key_pwd.txt",
"managementTlsCrlPath": "security/management/certs/",
"managementTlsCrlFiles": [
"server_crl.pem"
],
"kmcKsfMaster": "tools/pmt/master/ksfa",
"kmcKsfStandby": "tools/pmt/standby/ksfb",
"inferMode": "standard",
"interCommTLSEnabled": true,
"interCommPort": 1121,
"interCommTlsCaPath": "security/grpc/ca/",
"interCommTlsCaFiles": [
"ca.pem"
],
"interCommTlsCert": "security/grpc/certs/server.pem",
"interCommPk": "security/grpc/keys/server.key.pem",
"interCommPkPwd": "security/grpc/pass/key_pwd.txt",
"interCommTlsCrlPath": "security/grpc/certs/",
"interCommTlsCrlFiles": [
"server_crl.pem"
],
"openAiSupport": "vllm",
"tokenTimeout": 600,
"e2eTimeout": 600,
"distDPServerEnabled": false
},
"BackendConfig": {
"backendName": "mindieservice_llm_engine",
"modelInstanceNumber": 1,
"npuDeviceIds": [
[
0,
1,
2,
3
]
],
"tokenizerProcessNumber": 8,
"multiNodesInferEnabled": false,
"multiNodesInferPort": 1120,
"interNodeTLSEnabled": true,
"interNodeTlsCaPath": "security/grpc/ca/",
"interNodeTlsCaFiles": [
"ca.pem"
],
"interNodeTlsCert": "security/grpc/certs/server.pem",
"interNodeTlsPk": "security/grpc/keys/server.key.pem",
"interNodeTlsPkPwd": "security/grpc/pass/mindie_server_key_pwd.txt",
"interNodeTlsCrlPath": "security/grpc/certs/",
"interNodeTlsCrlFiles": [
"server_crl.pem"
],
"interNodeKmcKsfMaster": "tools/pmt/master/ksfa",
"interNodeKmcKsfStandby": "tools/pmt/standby/ksfb",
"ModelDeployConfig": {
"maxSeqLen": 26000,
"maxInputTokenLen": 20000,
"truncation": false,
"ModelConfig": [
{
"modelInstanceType": "Standard",
"modelName": "qwen2_5_vl_72B",
"modelWeightPath": "/invoker-deploy/Qwen_local_57_b1275e8d/in/Qwen2___5-VL-72B-Instruct",
"worldSize": 4,
"cpuMemSize": 5,
"npuMemSize": -1,
"backendType": "atb",
"trustRemoteCode": false
}
]
},
"ScheduleConfig": {
"templateType": "Standard",
"templateName": "Standard_LLM",
"cacheBlockSize": 128,
"maxPrefillBatchSize": 50,
"maxPrefillTokens": 30000,
"prefillTimeMsPerReq": 150,
"prefillPolicyType": 0,
"decodeTimeMsPerReq": 50,
"decodePolicyType": 0,
"maxBatchSize": 200,
"maxIterTimes": 6000,
"maxPreemptCount": 0,
"supportSelectBatch": false,
"maxQueueDelayMicroseconds": 5000
}
}
}
复制
```

### 代码示例 2

```
{
"Version": "1.0.0",
"ServerConfig": {
"ipAddress": "172.17.0.10",
"managementIpAddress": "127.0.0.2",
"port": 8000,
"managementPort": 1026,
"metricsPort": 1027,
"allowAllZeroIpListening": false,
"maxLinkNum": 1000,
"httpsEnabled": false,
"fullTextEnabled": false,
"tlsCaPath": "security/ca/",
"tlsCaFile": [
"ca.pem"
],
"tlsCert": "security/certs/server.pem",
"tlsPk": "security/keys/server.key.pem",
"tlsPkPwd": "security/pass/key_pwd.txt",
"tlsCrlPath": "security/certs/",
"tlsCrlFiles": [
"server_crl.pem"
],
"managementTlsCaFile": [
"management_ca.pem"
],
"managementTlsCert": "security/certs/management/server.pem",
"managementTlsPk": "security/keys/management/server.key.pem",
"managementTlsPkPwd": "security/pass/management/key_pwd.txt",
"managementTlsCrlPath": "security/management/certs/",
"managementTlsCrlFiles": [
"server_crl.pem"
],
"kmcKsfMaster": "tools/pmt/master/ksfa",
"kmcKsfStandby": "tools/pmt/standby/ksfb",
"inferMode": "standard",
"interCommTLSEnabled": true,
"interCommPort": 1121,
"interCommTlsCaPath": "security/grpc/ca/",
"interCommTlsCaFiles": [
"ca.pem"
],
"interCommTlsCert": "security/grpc/certs/server.pem",
"interCommPk": "security/grpc/keys/server.key.pem",
"interCommPkPwd": "security/grpc/pass/key_pwd.txt",
"interCommTlsCrlPath": "security/grpc/certs/",
"interCommTlsCrlFiles": [
"server_crl.pem"
],
"openAiSupport": "vllm",
"tokenTimeout": 600,
"e2eTimeout": 600,
"distDPServerEnabled": false
},
"BackendConfig": {
"backendName": "mindieservice_llm_engine",
"modelInstanceNumber": 1,
"npuDeviceIds": [
[
0,
1,
2,
3
]
],
"tokenizerProcessNumber": 8,
"multiNodesInferEnabled": false,
"multiNodesInferPort": 1120,
"interNodeTLSEnabled": true,
"interNodeTlsCaPath": "security/grpc/ca/",
"interNodeTlsCaFiles": [
"ca.pem"
],
"interNodeTlsCert": "security/grpc/certs/server.pem",
"interNodeTlsPk": "security/grpc/keys/server.key.pem",
"interNodeTlsPkPwd": "security/grpc/pass/mindie_server_key_pwd.txt",
"interNodeTlsCrlPath": "security/grpc/certs/",
"interNodeTlsCrlFiles": [
"server_crl.pem"
],
"interNodeKmcKsfMaster": "tools/pmt/master/ksfa",
"interNodeKmcKsfStandby": "tools/pmt/standby/ksfb",
"ModelDeployConfig": {
"maxSeqLen": 26000,
"maxInputTokenLen": 20000,
"truncation": false,
"ModelConfig": [
{
"modelInstanceType": "Standard",
"modelName": "qwen2_5_vl_72B",
"modelWeightPath": "/invoker-deploy/Qwen_local_57_b1275e8d/in/Qwen2___5-VL-72B-Instruct",
"worldSize": 4,
"cpuMemSize": 5,
"npuMemSize": -1,
"backendType": "atb",
"trustRemoteCode": false
}
]
},
"ScheduleConfig": {
"templateType": "Standard",
"templateName": "Standard_LLM",
"cacheBlockSize": 128,
"maxPrefillBatchSize": 50,
"maxPrefillTokens": 30000,
"prefillTimeMsPerReq": 150,
"prefillPolicyType": 0,
"decodeTimeMsPerReq": 50,
"decodePolicyType": 0,
"maxBatchSize": 200,
"maxIterTimes": 6000,
"maxPreemptCount": 0,
"supportSelectBatch": false,
"maxQueueDelayMicroseconds": 5000
}
}
}
```

### 代码示例 3

```
============ Serving Benchmark Result ============

Successful requests:                     20

Benchmark duration (s):                  55.80

Total input tokens:                      89051

Total generated tokens:                  2741

Request throughput (req/s):              0.36

Output token throughput (tok/s):         49.12

Total Token throughput (tok/s):          1645.07

---------------Time to First Token----------------

Mean TTFT (ms):                          33009.60

Median TTFT (ms):                        33009.08

P99 TTFT (ms):                           33034.09

-----Time per Output Token (excl. 1st token)------

Mean TPOT (ms):                          136.80

Median TPOT (ms):                        137.92

P99 TPOT (ms):                           138.75

---------------Inter-token Latency----------------

Mean ITL (ms):                           135.52

Median ITL (ms):                         138.23

P99 ITL (ms):                            145.32

---------------------Accuracy---------------------

Accuracy Rate (%):                       65.00

==================================================

root@5fd5f429c6cb:/home/project# ^C

root@5fd5f429c6cb:/home/project#  cd /home/project ; /usr/bin/env /usr/local/bin/python /root/.vscode-server/extensions/ms-python.debugpy-2025.8.0-linux-x64/bundled/libs/debugpy/adapter/../../debugpy/launcher 60067 -- /home/project/learning_project/benchmarks/benchmark_serving.py --backend openai-chat --model Qwen/Qwen2_5-VL-72B-Instruct --endpoint /v1/chat/completions --dataset-name phonetest --dataset-path /home/project/dataset/phonetest/web_nj_action_0426_grpo.json --num-prompts 30 --served_model_name qwen2_5_vl_72B --save_result --host 10.20.42.105 --tokenizer /home/temp/llm_tokenizer/Qwen/Qwen2_5-VL-72B-Instruct --port 32781

INFO 06-26 08:59:02 [__init__.py:244] Automatically detected platform cuda.

Namespace(backend='openai-chat', base_url=None, host='10.20.42.105', port=32781, endpoint='/v1/chat/completions', dataset_name='phonetest', dataset_path='/home/project/dataset/phonetest/web_nj_action_0426_grpo.json', max_concurrency=None, model='Qwen/Qwen2_5-VL-72B-Instruct', tokenizer='/home/temp/llm_tokenizer/Qwen/Qwen2_5-VL-72B-Instruct', use_beam_search=False, num_prompts=30, logprobs=None, request_rate=inf, burstiness=1.0, seed=0, trust_remote_code=False, ascend=False, disable_tqdm=False, profile=False, save_result=True, save_detailed=False, append_result=False, metadata=None, result_dir=None, result_filename=None, ignore_eos=False, percentile_metrics='ttft,tpot,itl,accuracy_rate', metric_percentiles='99', goodput=None, custom_output_len=256, custom_skip_chat_template=False, phonetest_output_len=1024, sonnet_input_len=550, sonnet_output_len=150, sonnet_prefix_len=200, sharegpt_output_len=None, random_input_len=1024, random_output_len=128, random_range_ratio=0.0, random_prefix_len=0, hf_subset=None, hf_split=None, hf_output_len=None, top_p=None, top_k=None, min_p=None, temperature=None, tokenizer_mode='auto', served_model_name='qwen2_5_vl_72B', lora_modules=None)

Starting initial single prompt test run...

Initial test run completed. Starting main benchmark run...

Traffic request rate: inf

Burstiness factor: 1.0 (Poisson process)

Maximum request concurrency: None

100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 30/30 [01:25<00:00,  2.86s/it]

============ Serving Benchmark Result ============

Successful requests:                     30

Benchmark duration (s):                  85.87

Total input tokens:                      133938

Total generated tokens:                  4137

Request throughput (req/s):              0.35

Output token throughput (tok/s):         48.18

Total Token throughput (tok/s):          1607.91

---------------Time to First Token----------------

Mean TTFT (ms):                          41686.26

Median TTFT (ms):                        33241.36

P99 TTFT (ms):                           66323.77

-----Time per Output Token (excl. 1st token)------

Mean TPOT (ms):                          199.20

Median TPOT (ms):                        206.81

P99 TPOT (ms):                           250.82

---------------Inter-token Latency----------------

Mean ITL (ms):                           200.35

Median ITL (ms):                         137.54

P99 ITL (ms):                            1651.45

---------------------Accuracy---------------------

Accuracy Rate (%):                       66.67

==================================================

============ Serving Benchmark Result ============

Successful requests:                     1

Benchmark duration (s):                  62.06

Total input tokens:                      4187

Total generated tokens:                  102

Request throughput (req/s):              0.02

Output token throughput (tok/s):         1.64

Total Token throughput (tok/s):          69.11

---------------Time to First Token----------------

Mean TTFT (ms):                          34565.55

Median TTFT (ms):                        34565.55

P99 TTFT (ms):                           34565.55

-----Time per Output Token (excl. 1st token)------

Mean TPOT (ms):                          143.41

Median TPOT (ms):                        143.41

P99 TPOT (ms):                           143.41

---------------Inter-token Latency----------------

Mean ITL (ms):                           142.01

Median ITL (ms):                         131.28

P99 ITL (ms):                            163.19

---------------------Accuracy---------------------

Accuracy Rate (%):                       100.00

==================================================
复制
```

### 代码示例 4

```
============ Serving Benchmark Result ============

Successful requests:                     20

Benchmark duration (s):                  55.80

Total input tokens:                      89051

Total generated tokens:                  2741

Request throughput (req/s):              0.36

Output token throughput (tok/s):         49.12

Total Token throughput (tok/s):          1645.07

---------------Time to First Token----------------

Mean TTFT (ms):                          33009.60

Median TTFT (ms):                        33009.08

P99 TTFT (ms):                           33034.09

-----Time per Output Token (excl. 1st token)------

Mean TPOT (ms):                          136.80

Median TPOT (ms):                        137.92

P99 TPOT (ms):                           138.75

---------------Inter-token Latency----------------

Mean ITL (ms):                           135.52

Median ITL (ms):                         138.23

P99 ITL (ms):                            145.32

---------------------Accuracy---------------------

Accuracy Rate (%):                       65.00

==================================================

root@5fd5f429c6cb:/home/project# ^C

root@5fd5f429c6cb:/home/project#  cd /home/project ; /usr/bin/env /usr/local/bin/python /root/.vscode-server/extensions/ms-python.debugpy-2025.8.0-linux-x64/bundled/libs/debugpy/adapter/../../debugpy/launcher 60067 -- /home/project/learning_project/benchmarks/benchmark_serving.py --backend openai-chat --model Qwen/Qwen2_5-VL-72B-Instruct --endpoint /v1/chat/completions --dataset-name phonetest --dataset-path /home/project/dataset/phonetest/web_nj_action_0426_grpo.json --num-prompts 30 --served_model_name qwen2_5_vl_72B --save_result --host 10.20.42.105 --tokenizer /home/temp/llm_tokenizer/Qwen/Qwen2_5-VL-72B-Instruct --port 32781

INFO 06-26 08:59:02 [__init__.py:244] Automatically detected platform cuda.

Namespace(backend='openai-chat', base_url=None, host='10.20.42.105', port=32781, endpoint='/v1/chat/completions', dataset_name='phonetest', dataset_path='/home/project/dataset/phonetest/web_nj_action_0426_grpo.json', max_concurrency=None, model='Qwen/Qwen2_5-VL-72B-Instruct', tokenizer='/home/temp/llm_tokenizer/Qwen/Qwen2_5-VL-72B-Instruct', use_beam_search=False, num_prompts=30, logprobs=None, request_rate=inf, burstiness=1.0, seed=0, trust_remote_code=False, ascend=False, disable_tqdm=False, profile=False, save_result=True, save_detailed=False, append_result=False, metadata=None, result_dir=None, result_filename=None, ignore_eos=False, percentile_metrics='ttft,tpot,itl,accuracy_rate', metric_percentiles='99', goodput=None, custom_output_len=256, custom_skip_chat_template=False, phonetest_output_len=1024, sonnet_input_len=550, sonnet_output_len=150, sonnet_prefix_len=200, sharegpt_output_len=None, random_input_len=1024, random_output_len=128, random_range_ratio=0.0, random_prefix_len=0, hf_subset=None, hf_split=None, hf_output_len=None, top_p=None, top_k=None, min_p=None, temperature=None, tokenizer_mode='auto', served_model_name='qwen2_5_vl_72B', lora_modules=None)

Starting initial single prompt test run...

Initial test run completed. Starting main benchmark run...

Traffic request rate: inf

Burstiness factor: 1.0 (Poisson process)

Maximum request concurrency: None

100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 30/30 [01:25<00:00,  2.86s/it]

============ Serving Benchmark Result ============

Successful requests:                     30

Benchmark duration (s):                  85.87

Total input tokens:                      133938

Total generated tokens:                  4137

Request throughput (req/s):              0.35

Output token throughput (tok/s):         48.18

Total Token throughput (tok/s):          1607.91

---------------Time to First Token----------------

Mean TTFT (ms):                          41686.26

Median TTFT (ms):                        33241.36

P99 TTFT (ms):                           66323.77

-----Time per Output Token (excl. 1st token)------

Mean TPOT (ms):                          199.20

Median TPOT (ms):                        206.81

P99 TPOT (ms):                           250.82

---------------Inter-token Latency----------------

Mean ITL (ms):                           200.35

Median ITL (ms):                         137.54

P99 ITL (ms):                            1651.45

---------------------Accuracy---------------------

Accuracy Rate (%):                       66.67

==================================================

============ Serving Benchmark Result ============

Successful requests:                     1

Benchmark duration (s):                  62.06

Total input tokens:                      4187

Total generated tokens:                  102

Request throughput (req/s):              0.02

Output token throughput (tok/s):         1.64

Total Token throughput (tok/s):          69.11

---------------Time to First Token----------------

Mean TTFT (ms):                          34565.55

Median TTFT (ms):                        34565.55

P99 TTFT (ms):                           34565.55

-----Time per Output Token (excl. 1st token)------

Mean TPOT (ms):                          143.41

Median TPOT (ms):                        143.41

P99 TPOT (ms):                           143.41

---------------Inter-token Latency----------------

Mean ITL (ms):                           142.01

Median ITL (ms):                         131.28

P99 ITL (ms):                            163.19

---------------------Accuracy---------------------

Accuracy Rate (%):                       100.00

==================================================
```

### 代码示例 5

```
Traceback (most recent call last):
File "/home/project/learning_project/benchmarks/benchmark_serving.py", line 1311, in <module>
main(args)
File "/home/project/learning_project/benchmarks/benchmark_serving.py", line 840, in main
benchmark_result = asyncio.run(
File "/usr/local/lib/python3.10/asyncio/runners.py", line 44, in run
return loop.run_until_complete(main)
File "/usr/local/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete
return future.result()
File "/home/project/learning_project/benchmarks/benchmark_serving.py", line 329, in benchmark
raise ValueError(
ValueError: Initial test run failed - Please make sure benchmark arguments are correctly specified. Error: Traceback (most recent call last):
File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1115, in _wrap_create_connection
sock = await aiohappyeyeballs.start_connection(
File "/usr/local/lib/python3.10/site-packages/aiohappyeyeballs/impl.py", line 122, in start_connection
raise first_exception
File "/usr/local/lib/python3.10/site-packages/aiohappyeyeballs/impl.py", line 73, in start_connection
sock = await _connect_sock(
File "/usr/local/lib/python3.10/site-packages/aiohappyeyeballs/impl.py", line 208, in _connect_sock
await loop.sock_connect(sock, address)
File "/usr/local/lib/python3.10/asyncio/selector_events.py", line 501, in sock_connect
return await fut
File "/usr/local/lib/python3.10/asyncio/selector_events.py", line 541, in _sock_connect_cb
raise OSError(err, f'Connect call failed {address}')
ConnectionRefusedError: [Errno 111] Connect call failed ('10.20.42.105', 32781)

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "/home/project/learning_project/benchmarks/backend_request_func.py", line 406, in async_request_openai_chat_completions
async with session.post(
File "/usr/local/lib/python3.10/site-packages/aiohttp/client.py", line 1425, in __aenter__
self._resp: _RetType = await self._coro
File "/usr/local/lib/python3.10/site-packages/aiohttp/client.py", line 703, in _request
conn = await self._connector.connect(
File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 548, in connect
proto = await self._create_connection(req, traces, timeout)
File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1056, in _create_connection
_, proto = await self._create_direct_connection(req, traces, timeout)
File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1400, in _create_direct_connection
raise last_exc
File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1369, in _create_direct_connection
transp, proto = await self._wrap_create_connection(
File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1130, in _wrap_create_connection
raise client_error(req.connection_key, exc) from exc
aiohttp.client_exceptions.ClientConnectorError: Cannot connect to host 10.20.42.105:32781 ssl:default [Connect call failed ('10.20.42.105', 32781)]
复制
```

## 完整内容

SystemInfo 硬件 910BC3 mindie 2.0.RC2 config.json { "Version": "1.0.0", "ServerConfig": { "ipAddress": "172.17.0.10", "managementIpAddress": "127.0.0.2", "port": 8000, "managementPort": 1026, "metricsPort": 1027, "allowAllZeroIpListening": false, "maxLinkNum": 1000, "httpsEnabled": false, "fullTextEnabled": false, "tlsCaPath": "security/ca/", "tlsCaFile": [ "ca.pem" ], "tlsCert": "security/certs/server.pem", "tlsPk": "security/keys/server.key.pem", "tlsPkPwd": "security/pass/key_pwd.txt", "tlsCrlPath": "security/certs/", "tlsCrlFiles": [ "server_crl.pem" ], "managementTlsCaFile": [ "management_ca.pem" ], "managementTlsCert": "security/certs/management/server.pem", "managementTlsPk": "security/keys/management/server.key.pem", "managementTlsPkPwd": "security/pass/management/key_pwd.txt", "managementTlsCrlPath": "security/management/certs/", "managementTlsCrlFiles": [ "server_crl.pem" ], "kmcKsfMaster": "tools/pmt/master/ksfa", "kmcKsfStandby": "tools/pmt/standby/ksfb", "inferMode": "standard", "interCommTLSEnabled": true, "interCommPort": 1121, "interCommTlsCaPath": "security/grpc/ca/", "interCommTlsCaFiles": [ "ca.pem" ], "interCommTlsCert": "security/grpc/certs/server.pem", "interCommPk": "security/grpc/keys/server.key.pem", "interCommPkPwd": "security/grpc/pass/key_pwd.txt", "interCommTlsCrlPath": "security/grpc/certs/", "interCommTlsCrlFiles": [ "server_crl.pem" ], "openAiSupport": "vllm", "tokenTimeout": 600, "e2eTimeout": 600, "distDPServerEnabled": false }, "BackendConfig": { "backendName": "mindieservice_llm_engine", "modelInstanceNumber": 1, "npuDeviceIds": [ [ 0, 1, 2, 3 ] ], "tokenizerProcessNumber": 8, "multiNodesInferEnabled": false, "multiNodesInferPort": 1120, "interNodeTLSEnabled": true, "interNodeTlsCaPath": "security/grpc/ca/", "interNodeTlsCaFiles": [ "ca.pem" ], "interNodeTlsCert": "security/grpc/certs/server.pem", "interNodeTlsPk": "security/grpc/keys/server.key.pem", "interNodeTlsPkPwd": "security/grpc/pass/mindie_server_key_pwd.txt", "interNodeTlsCrlPath": "security/grpc/certs/", "interNodeTlsCrlFiles": [ "server_crl.pem" ], "interNodeKmcKsfMaster": "tools/pmt/master/ksfa", "interNodeKmcKsfStandby": "tools/pmt/standby/ksfb", "ModelDeployConfig": { "maxSeqLen": 26000, "maxInputTokenLen": 20000, "truncation": false, "ModelConfig": [ { "modelInstanceType": "Standard", "modelName": "qwen2_5_vl_72B", "modelWeightPath": "/invoker-deploy/Qwen_local_57_b1275e8d/in/Qwen2___5-VL-72B-Instruct", "worldSize": 4, "cpuMemSize": 5, "npuMemSize": -1, "backendType": "atb", "trustRemoteCode": false } ] }, "ScheduleConfig": { "templateType": "Standard", "templateName": "Standard_LLM", "cacheBlockSize": 128, "maxPrefillBatchSize": 50, "maxPrefillTokens": 30000, "prefillTimeMsPerReq": 150, "prefillPolicyType": 0, "decodeTimeMsPerReq": 50, "decodePolicyType": 0, "maxBatchSize": 200, "maxIterTimes": 6000, "maxPreemptCount": 0, "supportSelectBatch": false, "maxQueueDelayMicroseconds": 5000 } } } 复制 测试结果 ============ Serving Benchmark Result ============ Successful requests: 20 Benchmark duration (s): 55.80 Total input tokens: 89051 Total generated tokens: 2741 Request throughput (req/s): 0.36 Output token throughput (tok/s): 49.12 Total Token throughput (tok/s): 1645.07 ---------------Time to First Token---------------- Mean TTFT (ms): 33009.60 Median TTFT (ms): 33009.08 P99 TTFT (ms): 33034.09 -----Time per Output Token (excl. 1st token)------ Mean TPOT (ms): 136.80 Median TPOT (ms): 137.92 P99 TPOT (ms): 138.75 ---------------Inter-token Latency---------------- Mean ITL (ms): 135.52 Median ITL (ms): 138.23 P99 ITL (ms): 145.32 ---------------------Accuracy--------------------- Accuracy Rate (%): 65.00 ================================================== root@5fd5f429c6cb:/home/project# C root@5fd5f429c6cb:/home/project# cd /home/project ; /usr/bin/env /usr/local/bin/python /root/.vscode-server/extensions/ms-python.debugpy-2025.8.0-linux-x64/bundled/libs/debugpy/adapter/../../debugpy/launcher 60067 -- /home/project/learning_project/benchmarks/benchmark_serving.py --backend openai-chat --model Qwen/Qwen2_5-VL-72B-Instruct --endpoint /v1/chat/completions --dataset-name phonetest --dataset-path /home/project/dataset/phonetest/web_nj_action_0426_grpo.json --num-prompts 30 --served_model_name qwen2_5_vl_72B --save_result --host 10.20.42.105 --tokenizer /home/temp/llm_tokenizer/Qwen/Qwen2_5-VL-72B-Instruct --port 32781 INFO 06-26 08:59:02 [__init__.py:244] Automatically detected platform cuda. Namespace(backend='openai-chat', base_url=None, host='10.20.42.105', port=32781, endpoint='/v1/chat/completions', dataset_name='phonetest', dataset_path='/home/project/dataset/phonetest/web_nj_action_0426_grpo.json', max_concurrency=None, model='Qwen/Qwen2_5-VL-72B-Instruct', tokenizer='/home/temp/llm_tokenizer/Qwen/Qwen2_5-VL-72B-Instruct', use_beam_search=False, num_prompts=30, logprobs=None, request_rate=inf, burstiness=1.0, seed=0, trust_remote_code=False, ascend=False, disable_tqdm=False, profile=False, save_result=True, save_detailed=False, append_result=False, metadata=None, result_dir=None, result_filename=None, ignore_eos=False, percentile_metrics='ttft,tpot,itl,accuracy_rate', metric_percentiles='99', goodput=None, custom_output_len=256, custom_skip_chat_template=False, phonetest_output_len=1024, sonnet_input_len=550, sonnet_output_len=150, sonnet_prefix_len=200, sharegpt_output_len=None, random_input_len=1024, random_output_len=128, random_range_ratio=0.0, random_prefix_len=0, hf_subset=None, hf_split=None, hf_output_len=None, top_p=None, top_k=None, min_p=None, temperature=None, tokenizer_mode='auto', served_model_name='qwen2_5_vl_72B', lora_modules=None) Starting initial single prompt test run... Initial test run completed. Starting main benchmark run... Traffic request rate: inf Burstiness factor: 1.0 (Poisson process) Maximum request concurrency: None 100%|| 30/30 [01:25<00:00, 2.86s/it] ============ Serving Benchmark Result ============ Successful requests: 30 Benchmark duration (s): 85.87 Total input tokens: 133938 Total generated tokens: 4137 Request throughput (req/s): 0.35 Output token throughput (tok/s): 48.18 Total Token throughput (tok/s): 1607.91 ---------------Time to First Token---------------- Mean TTFT (ms): 41686.26 Median TTFT (ms): 33241.36 P99 TTFT (ms): 66323.77 -----Time per Output Token (excl. 1st token)------ Mean TPOT (ms): 199.20 Median TPOT (ms): 206.81 P99 TPOT (ms): 250.82 ---------------Inter-token Latency---------------- Mean ITL (ms): 200.35 Median ITL (ms): 137.54 P99 ITL (ms): 1651.45 ---------------------Accuracy--------------------- Accuracy Rate (%): 66.67 ================================================== ============ Serving Benchmark Result ============ Successful requests: 1 Benchmark duration (s): 62.06 Total input tokens: 4187 Total generated tokens: 102 Request throughput (req/s): 0.02 Output token throughput (tok/s): 1.64 Total Token throughput (tok/s): 69.11 ---------------Time to First Token---------------- Mean TTFT (ms): 34565.55 Median TTFT (ms): 34565.55 P99 TTFT (ms): 34565.55 -----Time per Output Token (excl. 1st token)------ Mean TPOT (ms): 143.41 Median TPOT (ms): 143.41 P99 TPOT (ms): 143.41 ---------------Inter-token Latency---------------- Mean ITL (ms): 142.01 Median ITL (ms): 131.28 P99 ITL (ms): 163.19 ---------------------Accuracy--------------------- Accuracy Rate (%): 100.00 ================================================== 复制 再次发送请求服务挂死 Traceback (most recent call last): File "/home/project/learning_project/benchmarks/benchmark_serving.py", line 1311, in <module> main(args) File "/home/project/learning_project/benchmarks/benchmark_serving.py", line 840, in main benchmark_result = asyncio.run( File "/usr/local/lib/python3.10/asyncio/runners.py", line 44, in run return loop.run_until_complete(main) File "/usr/local/lib/python3.10/asyncio/base_events.py", line 649, in run_until_complete return future.result() File "/home/project/learning_project/benchmarks/benchmark_serving.py", line 329, in benchmark raise ValueError( ValueError: Initial test run failed - Please make sure benchmark arguments are correctly specified. Error: Traceback (most recent call last): File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1115, in _wrap_create_connection sock = await aiohappyeyeballs.start_connection( File "/usr/local/lib/python3.10/site-packages/aiohappyeyeballs/impl.py", line 122, in start_connection raise first_exception File "/usr/local/lib/python3.10/site-packages/aiohappyeyeballs/impl.py", line 73, in start_connection sock = await _connect_sock( File "/usr/local/lib/python3.10/site-packages/aiohappyeyeballs/impl.py", line 208, in _connect_sock await loop.sock_connect(sock, address) File "/usr/local/lib/python3.10/asyncio/selector_events.py", line 501, in sock_connect return await fut File "/usr/local/lib/python3.10/asyncio/selector_events.py", line 541, in _sock_connect_cb raise OSError(err, f'Connect call failed {address}') ConnectionRefusedError: [Errno 111] Connect call failed ('10.20.42.105', 32781) The above exception was the direct cause of the following exception: Traceback (most recent call last): File "/home/project/learning_project/benchmarks/backend_request_func.py", line 406, in async_request_openai_chat_completions async with session.post( File "/usr/local/lib/python3.10/site-packages/aiohttp/client.py", line 1425, in __aenter__ self._resp: _RetType = await self._coro File "/usr/local/lib/python3.10/site-packages/aiohttp/client.py", line 703, in _request conn = await self._connector.connect( File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 548, in connect proto = await self._create_connection(req, traces, timeout) File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1056, in _create_connection _, proto = await self._create_direct_connection(req, traces, timeout) File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1400, in _create_direct_connection raise last_exc File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1369, in _create_direct_connection transp, proto = await self._wrap_create_connection( File "/usr/local/lib/python3.10/site-packages/aiohttp/connector.py", line 1130, in _wrap_create_connection raise client_error(req.connection_key, exc) from exc aiohttp.client_exceptions.ClientConnectorError: Cannot connect to host 10.20.42.105:32781 ssl:default [Connect call failed ('10.20.42.105', 32781)] 复制 日志信息 /root/mindie/log/debug/mindie-server_19285_202506261647866.log [2025-06-26 17:07:08.881+08:00] [19285] [22974] [server] [INFO] [http_server.cpp:65] : [endpoint] Http receive response from 172.17.0.10:8000, status code is 200, body len 0 [2025-06-26 17:07:08.881+08:00] [19285] [22911] [server] [INFO] [llm_infer_model_instance.cpp:463] : [llm_backend] LLM backend enqueue request finished, request id: endpoint_common_118 [2025-06-26 17:07:08.882+08:00] [19285] [22911] [server] [INFO] [http_server.cpp:65] : [endpoint] Http receive response from 172.17.0.10:8000, status code is 200, body len 0 [2025-06-26 17:07:08.884+08:00] [19285] [22894] [server] [INFO] [llm_infer_model_instance.cpp:463] : [llm_backend] LLM backend enqueue request finished, request id: endpoint_common_110 [2025-06-26 17:07:08.884+08:00] [19285] [22894] [server] [INFO] [http_server.cpp:65] : [endpoint] Http receive response from 172.17.0.10:8000, status code is 200, body len 0 [2025-06-26 17:07:08.921+08:00] [19285] [22928] [server] [INFO] [llm_infer_model_instance.cpp:463] : [llm_backend] LLM backend enqueue request finished, request id: endpoint_common_111 [2025-06-26 17:07:08.921+08:00] [19285] [22928] [server] [INFO] [http_server.cpp:65] : [endpoint] Http receive response from 172.17.0.10:8000, status code is 200, body len 0 [2025-06-26 17:07:09.001+08:00] [19285] [22947] [server] [INFO] [llm_infer_model_instance.cpp:463] : [llm_backend] LLM backend enqueue request finished, request id: endpoint_common_104 [2025-06-26 17:07:09.001+08:00] [19285] [22947] [server] [INFO] [http_server.cpp:65] : [endpoint] Http receive response from 172.17.0.10:8000, status code is 200, body len 0 [2025-06-26 17:07:56.495+08:00] [19285] [22917] [server] [INFO] [common_wrapper.cpp:455] : [endpoint] ResponseCallback begin to send last response. RequestId is endpoint_common_95 [2025-06-26 17:07:56.516+08:00] [19285] [22551] [server] [ERROR] [base_wrapper.cpp:44] : [MIE04E020414] [endpoint] Failed to get ibis seqs id by output 'IBIS_SEQS_ID' not found in response [2025-06-26 17:07:56.516+08:00] [19285] [22551] [server] [ERROR] [base_wrapper.cpp:312] : [MIE04E020414] [endpoint] Failed to generate result for parse seq id error. [2025-06-26 17:07:56.516+08:00] [19285] [22868] [server] [ERROR] [base_wrapper.cpp:44] : [MIE04E020414] [endpoint] Failed to get ibis seqs id by output 'IBIS_SEQS_ID' not found in response [2025-06-26 17:07:56.516+08:00] [19285] [22868] [server] [ERROR] [base_wrapper.cpp:312] : [MIE04E020414] [endpoint] Failed to generate result for parse seq id error. [2025-06-26 17:07:58.884+08:00] [19285] [22553] [server] [WARN] [llm_daemon.cpp:74] : [MIE04W01011A] [daemon] Received exit signal[17] 复制 /root/mindie/log/debug/mindie-batchscheduler_19492_20250626164427423.log [2025-06-26 17:07:55.931+08:00] [19492] [19516] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:523] : [model_backend] slave receive msg len: 46340 [2025-06-26 17:07:56.071+08:00] [19492] [19516] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:523] : [model_backend] slave receive msg len: 46340 [2025-06-26 17:07:56.212+08:00] [19492] [19516] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:523] : [model_backend] slave receive msg len: 46340 [2025-06-26 17:07:56.354+08:00] [19492] [19516] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:523] : [model_backend] slave receive msg len: 46340 [2025-06-26 17:07:56.026+0800] [19492] [281473516958048] [batchscheduler] [ERROR] [standard_model.py:188] : [Model] >>> global rank-2 Execute type:1, Exception:Get data from share memory error: [Errno 2] No such file or directory: '/psm_236f54d9' Traceback (most recent call last): File "/usr/local/Ascend/atb-models/atb_llm/utils/shm_utils.py", line 249, in get_data_from_shm shm = shared_memory.SharedMemory(name=shm_name)  File "/usr/lib64/python3.11/multiprocessing/shared_memory.py", line 104, in __init__ self._fd = _posixshmem.shm_open(  FileNotFoundError: [Errno 2] No such file or directory: '/psm_236f54d9' The above exception was the direct cause of the following exception: Traceback (most recent call last): File "/usr/local/lib/python3.11/site-packages/model_wrapper/standard_model.py", line 176, in execute return self._prefill(requests)  File "/usr/local/lib/python3.11/site-packages/model_wrapper/standard_model.py", line 403, in _prefill return self.__generate(requests, is_prefill=True, is_mix=False)  File "/usr/local/lib/python3.11/site-packages/model_wrapper/standard_model.py", line 555, in __generate generate_output = self.__handle_requests(requests, is_prefill, is_mix)  File "/usr/local/lib/python3.11/site-packages/model_wrapper/standard_model.py", line 910, in __handle_requests generate_output = self.generator.generate_token(metadata)  File "/usr/local/lib/python3.11/site-packages/mindie_llm/text_generator/generator.py", line 327, in generate_token raise e File "/usr/local/lib/python3.11/site-packages/mindie_llm/text_generator/generator.py", line 308, in generate_token generation_output = self.plugin.generate_token(input_metadata)  File "/usr/local/lib/python3.11/site-packages/mindie_llm/utils/decorators/time_decorator.py", line 91, in wrapper return func(*args, **kwargs)  File "/usr/local/lib/python3.11/site-packages/mindie_llm/text_generator/plugins/plugin_manager.py", line 102, in generate_token result = self.generator_backend.forward(model_inputs, q_lens=self.plugin_data_param.q_len,  File "/usr/local/lib/python3.11/site-packages/mindie_llm/utils/decorators/time_decorator.py", line 69, in wrapper return func(*args, **kwargs)  File "/usr/local/lib/python3.11/site-packages/mindie_llm/text_generator/adapter/generator_torch.py", line 209, in forward logits = self._forward(model_inputs, **kwargs)  File "/usr/local/lib/python3.11/site-packages/mindie_llm/text_generator/adapter/generator_torch.py", line 554, in _forward logits = self.model_wrapper.forward(model_inputs, self.cache_pool.npu_cache, **kwargs)  File "/usr/local/lib/python3.11/site-packages/mindie_llm/modeling/model_wrapper/atb/atb_model_wrapper.py", line 166, in forward result = self.forward_tensor(  File "/usr/local/lib/python3.11/site-packages/mindie_llm/modeling/model_wrapper/atb/atb_model_wrapper.py", line 206, in forward_tensor result = self.model_runner.forward(  File "/usr/local/Ascend/atb-models/atb_llm/runner/model_runner.py", line 297, in forward res = self.model.forward(**kwargs)  File "/usr/local/Ascend/atb-models/atb_llm/models/qwen2_vl/flash_causal_qwen2_vl.py", line 118, in forward inputs_embeds, image_grid_thw, video_grid_thw, second_per_grid_ts = self.prepare_prefill_token_service(  File "/usr/local/Ascend/atb-models/atb_llm/models/qwen2_vl/flash_causal_qwen2_vl.py", line 169, in prepare_prefill_token_service shared_array = get_data_from_shm(shm_value, shape_value, np.uint8, self.weights.device)  File "/usr/local/Ascend/atb-models/atb_llm/utils/shm_utils.py", line 253, in get_data_from_shm raise ValueError(f"Get data from share memory error: {e}") from e ValueError: Get data from share memory error: [Errno 2] No such file or directory: '/psm_236f54d9' [2025-06-26 17:07:56.494+08:00] [19492] [19516] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:523] : [model_backend] slave receive msg len: 49631 [2025-06-26 17:07:56.575+08:00] [19492] [19516] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:523] : [model_backend] slave receive msg len: 47247 [2025-06-26 17:07:58.858+08:00] [19492] [19516] [batchscheduler] [INFO] [slave_IPC_communicator.cpp:523] : [model_backend] slave receive msg len: 46340 复制

---

## 技术要点总结

基于以上内容，主要技术要点包括：

1. **问题类型**: 错误处理
2. **涉及技术**: HTTPS, SSL, NPU, MindIE, 昇腾, CANN, AI
3. **解决方案**: 请参考完整内容中的解决方案

## 相关资源

- 昇腾社区: https://www.hiascend.com/
- 昇腾论坛: https://www.hiascend.com/forum/

---

*本文档由AI自动生成，仅供参考。如有疑问，请参考原始帖子。*

mindie 2.0 版本 部署qwen 2.5vl 72B，并发数为40时 出现如下错误Segmentation fault (core dumped)，导致服务直接挂死_MindIE_昇腾论坛

mindie 2.0 版本部署qwen 2.5vl 72B，并发数为40时出现如下错误Segmentation fault (core dumped)，导致服务直接挂死_MindIE_昇腾论坛