AI Center Install Freezes During GPU Installation After Restarting Docker

How to handle when AI Center Install freezes during GPU installation after restarting docker ?

Issue Description : AI Center Install freezes during GPU installation after restarting docker. Script will just show:


################################# restarting docker #####################################

This issue could also be encountered when trying to add the GPU.

Diagnosing Steps:

  1. If this issue occurs, open a new console.
  2. Revert the changes made by the installer by executing the following command
    • sudo bash -c 'echo \ "{ \"exec-opts\": [\"native.cgroupdriver=systemd\"]}" > /etc/docker/daemon.json'
  3. Restart the machine.
  4. Log back into the machine
  5. Wait for the cluster to come up. (just run kubectl get pods, if it returns the cluster is up)
  6. Run the command: kubectl get nodes
    • Remember the node name that is returned.
  7. Once the cluster is up, shutdown the cluster.
    • sudo /opt/ekco/shutdown.sh
  8. After this is done, try adding the GPU manually. See more on AI Center Infrastructure Installer - Adding GPU Manually
  9. Docker should be able to restart. If the installer hangs again, open a support ticket with uipath.
  10. After the install, run the following command:
    • kubectl cordon <node name>
  11. After this step, validate access to AI Center.