VScode web cannot load for AML compute instance

microsoft / vscode-tools-for-ai

Azure Machine Learning for Visual Studio Code, previously called Visual Studio Code Tools for AI, is an extension to easily build, train, and deploy machine learning models to the cloud or the edge with Azure Machine Learning service.

Other

321 stars 91 forks source link

VScode web cannot load for AML compute instance #2207

Closed jrd232f closed 8 months ago

jrd232f commented 8 months ago

Does this occur consistently%3F Repro steps:

spin up a compute instance
click on the VSCode Web application

Error message: Failed to authenticate. Error type: AuthError

Version: 0.18.0 Browser Name: Chrome Browser Version: 118.0.0.0 OS: Windows 10 64-bit Product: Visual Studio Code Product Version: 1.83.1 Language: en

Call Stack

``` kt.getSession webOpener.js:2:706485 https:/webOpener.js:2:785512 Ar.measure webOpener.js:2:792628 Tr.getSession webOpener.js:2:785470 Tr.create webOpener.js:2:785206 $n.resolve webOpener.js:2:742078 https:/webOpener.js:2:772343 Ar.measure webOpener.js:2:792628 Lr.doRoute webOpener.js:2:772300 Module.Ir webOpener.js:2:793231 ```

jrd232f commented 8 months ago

sevillal commented 8 months ago

@jrd232f I apologize for the issue you are facing. Could you please send us the logs you get by following these instructions?

jrd232f commented 8 months ago

I guess the failed connection won't allow me to view any logs on the VSCode interface. Here's a screenshot of the VSCode interface after I clicked on "Cancel" from the error window uploaded previously. The "Reload" button only regenerates the error window itself, and the "Report an Issue" button leads to the Github Issue page.

sevillal commented 8 months ago

@jrd232f thanks for looking into that. I think you should be able to get the logs by clicking on the top left menu -> View -> Output. Then you can look for the right outputs using the dropdown.

jrd232f commented 8 months ago

@jrd232f thanks for looking into that. I think you should be able to get the logs by clicking on the top left menu -> View -> Output. Then you can look for the right outputs using the dropdown.

This again opens the empty output panel at the bottom of the screen. It's the same with the screenshot:

sevillal commented 8 months ago

@jrd232f in order to see the logs you need to select the right output channel, you can do that by selection vscode.dev and Azure ML Remote Web options from the following dropdown:

jrd232f commented 8 months ago

Thanks for pointing that out. Here are the logs:

sevillal commented 8 months ago

@jrd232f could you please also send the log from the vscode.dev output channel?

sevillal commented 8 months ago

@jrd232f following up regarding this issue. Is there anything I can do to help? If you are open we can have a triage call next week, please send me your availability and I can schedule some time. Please reach out to me to sevillal@microsoft.com with you availability and email so I can setup the call.

jrd232f commented 8 months ago

Hi there, I found another solution for me which is to install the jupyter-git extension inside the jupyterlab app of the compute. For now I don't need to use vscode. Thanks