Unable to restart Rasa X installed with REI

[Deprecated] Rasa X Community Edition
1

Hello,

Having followed the instructions given here: Installing and Deploying Rasa X (Using Rasa Ephemeral Installer) - YouTube

I have been able to successfully deploy a Rasa X instance on a GCP VM with Ubuntu 20.04 without specifying a project.

(base) rasa-8gb:~$ rasactl status
Name:                           physical-exam
Status:                         Running                   
URL:                            http://X.X.X.X:30040
Version:                        1.0.1                     
Enterprise:                     inactive                  
Rasa production version:        2.8.15                    
Rasa worker version:            2.8.15                    
Project path:                   not defined

If I restart the VM according to the following procedure:

  1. Stop the deployment using rasactl
  2. Shut down the VM
  3. Restart the VM
  4. Restart the deployment using rasactl

Everything seems to work fine, i.e. the 4th command completes successfully and I have a status similar to the above.

However, if I instead create the exact same deployment, but this time specifying a project directory, it will initially work fine:

(base) rasa-8gb:~/bot-pipeline/physical_exam$ rasactl status
Name:                           physical-exam                                    
Status:                         Running                                          
URL:                            http://X.X.X.X:30036                       
Version:                        1.0.1                                            
Enterprise:                     inactive                                         
Rasa production version:        2.8.15                                           
Rasa worker version:            2.8.15                                           
Project path:                   /home/bdi_idb_nrc_cnrc/bot-pipeline/physical_exam

But, if I reboot the VM, step 4 above fails with the error:

...
2022-03-04T02:28:43.285Z        INFO    Deployment is not ready: physical-exam/rasa-x-app. 0 out of 1 expected pods are ready
2022-03-04T02:28:46.085Z        INFO    Deployment is not ready: physical-exam/rasa-x-app. 0 out of 1 expected pods are ready
2022-03-04T02:28:48.884Z        INFO    Deployment is not ready: physical-exam/rasa-x-app. 0 out of 1 expected pods are ready
2022-03-04T02:28:49.290Z        INFO    warning: Upgrade "rasa-x" failed: timed out waiting for the condition
Error: timed out waiting for the condition

The behavior above has been observed even for quite large settings for the --wait-timeout flag in rasactl. While rasactrl is running I see very little usage of the processor and kubectl reports no resources in the namespace.

Does anybody have some experience with this problem? Is there another method that I should be following in rebooting the VM to avoid this?

Thanks!

2

Hey, @jlh

Were you able to resolve this issue? I ran into issues with Rasa X on GCP as well

3

Hi @GiedreM ,

Unfortunately, no, I was able to reproduce the problem on all sorts of installations, but couldn’t get it to restart nicely. So I just went back to using the docker-compose installation.