今天刚上电算力服务器：

配置：

1. openEuler release 22.03 (LTS-SP1)， aarch64，300I-Duo(310P3)npu卡，npu驱动已经安装

2. 在社区通过镜像下载了swr.cn-south-1.myhuaweicloud.com/ascendhub/mindie:1.0.0-300I-Duo-py311-openeuler24.03-lts

3. 模型使用：DeepSeek-R1-Distill-Qwen-14B（https://gitee.com/ascend/ModelZoo-PyTorch/tree/master/MindIE/LLM/DeepSeek/DeepSeek-R1-Distill-Qwen-14B）

4. 启动，然后进入docker修改相关参数

docker run -it -d --shm-size=4g \ 

    --privileged \ 

    --name deepseek14b_2 \ 

    --device=/dev/davinci_manager \ 

    --device=/dev/hisi_hdc \ 

    --device=/dev/devmm_svm \ 

    -p 22222:1025 \ 

    -v /usr/local/Ascend/driver:/usr/local/Ascend/driver:ro \ 

    -v /usr/local/sbin:/usr/local/sbin:ro \ 

    -v /home/model/deepseek-r1-qwen-14b:/home/weight:rw \ 

    swr.cn-south-1.myhuaweicloud.com/ascendhub/mindie:1.0.0-300I-Duo-py311-openeuler24.03-lts bash 

5. 在docker内启动： ./mindieservice_daemon

[2025-08-26 11:41:31.710053+08:00] [114] [116] [server] [ERROR] [llm_infer_model_instance.cpp:234] : [LLMInferModelInstance::Init] llmManager_ init fail! 

[2025-08-26 11:41:31.710112+08:00] [114] [116] [server] [ERROR] [llm_infer_model.cpp:28] : ModelInstance init failed 

[2025-08-26 11:41:31.710175+08:00] [114] [116] [server] [ERROR] [llm_infer_engine.cpp:48] : [MIE05E040001] [llm_backend] LLMInferEngine initialize LLMInferModels failed. please check the error stacks for more details 

[2025-08-26 11:41:31.710243+08:00] [114] [116] [server] [ERROR] [infer_backend_manager.cpp:272] : [MIE04E060108] [backend_manager] Failed to init engine: mindieservice_llm_engine 

[2025-08-26 11:41:31.710276+08:00] [114] [116] [server] [ERROR] [EndPoint endpoint.cpp:50] ERR: Failed to init backendManager! Please check in the mindservice.log, pythonlog.log(default path: mindie-llm/logs) or console output.