AI Fabric tenant could not be reached - on premise

I’ve set up Orchestrator and AIFabric on premise and I’m running into the following issue:

I’ve got the enterprise on premise trial licence.

I’ve found several posts like this in the forum, but people were always using the UiPath cloud which is why their solutions don’t work for me.

Does anyone know how to handle this?

Cheers,
T0Bi

1 Like

Hi @T0Bi
go to your main panel and check you have enabled AI fabric into your relevant tenant
and also go to permission and check your user role/folder permission that granted AI fabric permissions

How do I go into my “main panel” with an on premise installation?
I know how it works with the cloud version, but this does not exist on on-premise:

image

At least I couldn’t find it.

The permissions are all granted.

Hi @T0Bi
Could you please check your permissions on Orchestrator? In particular do you have permissions for ML Packages, ML Skills and ML Logs?

Jeremy

Hi @Jeremy_Tederry

I’ve created an extra role as explained in the troubleshooting section.

This role has all the ML permissions and I gave every user this role.

Anything else I should look out for?

T0Bi

@Jeremy_Tederry

One thing I found was the following warning in the WIndows Event Viewer:

UiPath.IdentityServer.Application.Services.UrlSplitterService Error parsing acr values tenantname:

I use the default tenant when logging in.

Did you complete this step https://docs.uipath.com/ai-fabric/docs/6-verify-installation#provision-an-ai-fabric-tenant (provisioning tenant)? What tenant did you provision? Default?
What Orchestrator version are you using?

1 Like

I feel kind of stupid.

That was the issue…

I’ve had so many problems with the AIFabric installation most of which were only partly or not at all described in the documentation, that I assumed it was again an issue like that.

The other issues were solved by the UiPath support (:

I’ll might do a write-up on all the problems I’ve encountered with the solutions I used, so that others can find it easily. Maybe some things can even be added in the documentation.

Thanks for your quick replies!

  • T0Bi
1 Like

Nice to see that it is working now! It would be awesome if you can send us document with summary of issues you encountered it will help us improving our workflow!

Jeremy

1 Like

This topic was automatically closed 3 days after the last reply. New replies are no longer allowed.