无法设置Web服务环境，因为其配置状态为“失败".设置状态成功. Azure机器学习. [英] Web-service environment cannot be set, as its provisioning state is Failed. Provisioning state succeeded is required. Azure Machine Learning.

问题描述

几个月前，我创建了一个环境"通过模型管理，以便通过此环境部署Web服务.我计划将这种环境用于集群部署.由于我尚未准备好部署任何模型， 我决定降级VM，关闭它们并减少节点数，以避免产生任何进一步的费用.

2个月后，我已经准备好将模型作为Web服务部署到此环境.

我已打开虚拟机，并将其(1个主代理和2个代理)升级到D3_v2规范.

问题:

环境"资源(类型为Microsoft.MachineLearningCompute/operationalizationClusters)不再出现​​在资源组中(在门户上)，现在它是隐藏的".尝试在此环境中发布Web服务时， 我收到此错误:无法设置"组*******和名称*********的资源，因为其配置状态为失败.设置状态必须成功."

在列出我的可用环境时，CLI指出此特定环境存在配置错误".代码为" InternalError "的代码；和消息"无法部署评分前端".

为什么会这样?我是否让虚拟机关闭时间过长?我仍然有一个包含kubernetes集群和相关资源的资源组.因此，我不想简单地创建另一个环境，因为这会 进而生成另一个资源组.

是否可以某种方式修复环境的Provisioning状态?还是至少创建一个新环境并将其分配给现有集群?

任何帮助将不胜感激.

此致

Andreas

解决方案

很抱歉听到您遇到此问题.请参考以下博客以获取解决方案:

https://support.kemptechnologies.com/hc/zh-cn/articles/210752506-Azure-VLM-Provisioning-State-results-In-Failed-Message

请告诉我您是否还有进一步的挑战.

此致

雨桐

A few months ago I created an "environment" through Model Management in order to deploy web services via this environment.   I planned on using this environment for cluster deployments.  As i didn't have any models ready to deploy, I decided to downgrade the VMs, switch them off and decrease the node count to avoid incurring any further costs.

2 months later I have model ready to deploy as a web-service to this environment.

I have switched on the VMs and upgraded (1 master and 2 agents) to D3_v2 specification. 

The problem:

The "environment" resource (of type Microsoft.MachineLearningCompute/operationalizationClusters) no longer appears in the resource group (on the portal), it is now "hidden".  When trying to publish a web-service to this environment, I receive this error: "Resource with group ******* and name ********* cannot be set, as its provisioning state is Failed. Provisioning state succeeded is required."

When listing my available environments, the CLI states this particular environment has "Provisioning Errors" with code "InternalError" and message "Failed to deploy scoring front-end".

Why has this happened?  Have I left the VMs switched off for too long?  I still have a resource group containing the kubernetes cluster and associated resources.  For that reason I do not want to simply create another environment as this will in turn generate another resource group.  

Is it possible to fix the Provisioning state of the environment somehow? Or at least create a new environment and assign it to the existing cluster?

Any help would be much appreciated.

Regards,

Andreas

解决方案

Hi,

Sorry to hear you are suffering from this issue. Please refer to following blog for solution:

https://support.kemptechnologies.com/hc/en-us/articles/210752506-Azure-VLM-Provisioning-State-results-In-Failed-Message

Please tell me if you have further challenge.

Regards,

Yutong

这篇关于无法设置Web服务环境，因为其配置状态为“失败".设置状态成功. Azure机器学习.的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！