tencentmusic / cube-studio

cube studio开源云原生一站式机器学习/深度学习/大模型AI平台,支持sso登录,多租户,大数据平台对接,notebook在线开发,拖拉拽任务流pipeline编排,多机多卡分布式训练,超参搜索,推理服务VGPU,边缘计算,serverless,标注平台,自动化标注,数据集管理,大模型微调,vllm大模型推理,llmops,私有知识库,AI模型应用商店,支持模型一键开发/推理/微调,支持国产cpu/gpu/npu芯片,支持RDMA,支持pytorch/tf/mxnet/deepspeed/paddle/colossalai/horovod/spark/ray/volcano分布式
Other
3.45k stars 610 forks source link

私有仓库镜像节点无法运行 failed to look-up entrypoint/cmd for image \"172.40.20.82:8443/aiclube/model_download\" #229

Closed gfoxlin closed 5 months ago

gfoxlin commented 7 months ago

1.私有仓库Harbor v2.8.2-d4c34dc 建立的项目是Public 2.自建测试 Deployment kubectl运行镜像下载成功(排除镜像下载问题 - 本地存在该镜像也有下面的问题) 3.【步骤】将model-download push到私有仓库,》添加本地镜像到系统 》修改‘数据导入导出-model-download’的镜像 》任务流拖出模型导入组件保存 》 运行【必现】

"test-pipeline-misl-ef38-984341727": { "boundaryID": "test-pipeline-misl-ef38", "displayName": "model-download-1707182945889", "finishedAt": "2024-02-06T01:29:12Z", "id": "test-pipeline-misl-ef38-984341727", "message": "task 'test-pipeline-misl-ef38.model-download-1707182945889' errored: failed to look-up entrypoint/cmd for image \"172.40.20.82:8443/aiclube/model_download\", you must either explicitly specify the command, or list the image's command in the index: https://argoproj.github.io/argo-workflows/workflow-executors/#emissary-emissary: Get \"https://172.40.20.82:8443/v2/\": http: server gave HTTP response to HTTPS client", "name": "test-pipeline-misl-ef38.model-download-1707182945889", "phase": "Error", "progress": "0/1", "startedAt": "2024-02-06T01:29:12Z", "templateName": "model-download-1707182945889", "templateScope": "local/test-pipeline-misl-ef38", "type": "Pod" }

data-infra commented 6 months ago

注册模板时,添加上启动命令