Automation Suite on Linux installation guide

Last updated Apr 29, 2025

Failed ML skill deployment due to token expiry

Description

Due to a bug in containerd, the deployment of ML skills can fail as a token expires before larger image downloads can be completed, especially under conditions of slower network connectivity.

Solution

To address this issue, take the following steps:

Cordon one of the nodes.
Run /opt/node-drain.sh.
Add the mirror of the registry containing the image to the registries.yaml file. Replace <localregistryurl> with your local registry URL:
```
registry.uipath.com:
  endpoint: 
    "<localregistryurl>"registry.uipath.com:
  endpoint: 
    "<localregistryurl>"
```
Restart the RKE2 server.
Uncordon the node.
Test the deployment with one of the ML skills.