AI Center standalone troubleshooting

This section provides troubleshooting information for AI Center in standalone environment.

The sections below are specific for AI Center.

Important: The AI Center standalone installation uses the same installer as Automation Suite. Some pages from the AI Center standalone troubleshooting section can lead to the corresponding page in the Automation Suite. In this situation, the steps are the same in both situations and there are no particularities in the specific procedure for AI Center.

Make sure that you follow the procedure suited to your needs.

Expired Identity Token

Description

In some cases, when the AI Center installation takes more than an hour (usually during airgapped installations), the initial Orchestrator token supplied in the input.json file will expire and the AI Center registration with Identity Server fails. Follow the steps below to recover it.

Recovery steps

1. Log in to https://alm.<LB DNS> using the admin username. To obtain the password, run the following command:

   kubectl -n argocd get secret argocd-admin-password -o jsonpath={.data.password} | base64 -d

2. Go to ArgoCD and click on the aicenter tile.
   
   

   

   
   
3. Click APP DETAILS and go to the Manifest tab.
4. In the Manifest tab, click Edit.
5. Obtain the new identity token by updating the accessToken field in the Manifest tab and click Save.
   
   

   

   
   

Synchronization starts automatically and is completed.

Message: Curl: (92) HTTP/2 Stream 0 Was Not Closed Cleanly: HTTP_1_1_REQUIRED (err 13)

Description

The following error message can occur while installing AI Center standalone: curl: (92) HTTP/2 stream 0 was not closed cleanly: HTTP_1_1_REQUIRED (err 13).

Solution

Make sure that you are using TLS 1.2 and HTTP/2 before proceeding with the installation.

How to recreate databases

If there is an issue with your databases, you can recreate them from scratch directly post-installation.

You can do this by running an SQL command to drop all the DBs and recreate them as follows:

USE [master]
ALTER DATABASE [AutomationSuite_AICenter] SET SINGLE_USER WITH ROLLBACK IMMEDIATE
DROP DATABASE [AutomationSuite_AICenter]
CREATE DATABASE [AutomationSuite_AICenter]
GO

Installer cannot connect to ArgoCD to check if password was reset

Description

This issue might occur during fabric installation. The installer might fail with below similar error.

appproject.argoproj.io/fabric created
configmap/argocd-cm configured
[INFO] [2021-09-02T09:21:15+0000]: Checking if ArgoCD password was reset, looking for secrets/argocd-admin-password.
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:21:16+0000]: Secret not found, trying to log in with initial password...1/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:21:36+0000]: Secret not found, trying to log in with initial password...2/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:21:56+0000]: Secret not found, trying to log in with initial password...3/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:22:16+0000]: Secret not found, trying to log in with initial password...4/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:22:36+0000]: Secret not found, trying to log in with initial password...5/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:22:56+0000]: Secret not found, trying to log in with initial password...6/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:23:17+0000]: Secret not found, trying to log in with initial password...7/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:23:37+0000]: Secret not found, trying to log in with initial password...8/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:23:57+0000]: Secret not found, trying to log in with initial password...9/10
FATA[0000] dial tcp: lookup remusr-sf on 168.63.129.16:53: no such host
[INFO] [2021-09-02T09:24:17+0000]: Secret not found, trying to log in with initial password...10/10
[ERROR][2021-09-02T09:24:37+0000]: Failed to log in

Solution 1

Check all the required subdomains and ensure they are configured correctly and are routable as follows:

getent ahosts automationsuite.mycompany.com | awk '{print $1}' | sort | uniq
getent ahosts alm.automationsuite.mycompany.com | awk '{print $1}' | sort | uniq
getent ahosts registry.automationsuite.mycompany.com | awk '{print $1}' | sort | uniq
getent ahosts monitoring.automationsuite.mycompany.com | awk '{print $1}' | sort | uniq
getent ahosts objectstore.automationsuite.mycompany.com | awk '{print $1}' | sort | uniq

Important: Replace automationsuite.mycompany.com with you cluster FQDN.

If the above commands/lines do not return a routable IP address, then the subdomain required for AI Center is not configured properly.

Note:

This error is encountered when the DNS is not public.

You need to add the Private DNS Zone (for Azure) or route 53 (for AWS).

If above commands return the proper IP address, then follow the below steps.

Solution 2

1. Delete the ArgoCD namespace by executing following command:

   export KUBECONFIG=/etc/rancher/rke2/rke2.yaml
   export PATH=$PATH:/var/lib/rancher/rke2/bin
   kubectl delete namespace argocd

2. Run the following command to verify:

   kubectl get namespace

There should not be ArgoCD namespace in the output of this command.

Note: Once the ArgoCD namespace is deleted, resume with the installation.

Issues while accessing AI Center

For issues related to accessing AI Center, make sure to follow the steps from the following sections:

• Enabling kubectl
• Managing certificates
• Allowing Robots to communicate with AI Center

Note: In case you are using self sign certificate you will also need to visit https://objectstore.${CONFIG_CLUSTER_FQDN} url once with every browser that you want to use in order to be able to interact with storage.

Enabling AI Center on the restored cluster

After restoring a cluster, take the following steps.

1. Connect to the restored cluster.
2. Delete the existing secrets.
   1. List the secrets:

      kubectl -n uipath get secrets | grep aicenter

   2. Delete the following secrets:

      Kubectl -n uipath delete secrets aicenter-cluster-credentials aicenter-deploy-token-XXXXX aicenter-jobs-token-XXXXX

      Replace XXXXX with the characters listed at step 2b.
   3. Get the default token name and delete it:

      Kubectl -n uipath get secrets | grep default-token
      Kubectl -n uipath delete secrets default-token-XXXXX

3. Go to ArgoCD and sync the AI Center application.
4. Restart all AI Center-related services:

   kubectl -n uipath rollout restart deployment ai-app-deployment ai-helper-deployment ai-pkgmanager-deployment ai-deployer-deployment ai-trainer-deployment ai-appmanager-deployment

Enabling AI Center on the restored cluster

After restoring a cluster, take the following steps.

1. Connect to the restored cluster.
2. Delete the existing secrets.
   1. List the secrets:

      kubectl -n uipath get secrets | grep aicenter

   2. Delete the following secrets:

      Kubectl -n uipath delete secrets aicenter-cluster-credentials aicenter-deploy-token-XXXXX aicenter-jobs-token-XXXXX

      Replace XXXXX with the characters listed at step 2b.
   3. Get the default token name and delete it:

      Kubectl -n uipath get secrets | grep default-token
      Kubectl -n uipath delete secrets default-token-XXXXX

3. Go to ArgoCD and sync the AI Center application.
4. Restart all AI Center-related services:

   kubectl -n uipath rollout restart deployment ai-app-deployment ai-helper-deployment ai-pkgmanager-deployment ai-deployer-deployment ai-trainer-deployment ai-appmanager-deployment