Opennebula 6.10 Fireedge 400 Bad Request

When upgrading opennebula running on Ubuntu 22.04 from 6.8 to 6.10 everything was smooth and the upgrade was successful. Tested by instantiating a desktop vm, connected to the vm via guacamole and all was good.
At this point I decided to do one last apt update && apt -y upgrade, the upgrade did not have anything to upgrade but I saw the remove unused packages prompt and without thinking I ran the command which promptly removed netplan, guacd and many other packages.

After searching the internet I found an article where someone had the same problem. Following the article I downloaded the deb files and reinstalled what was missing. That reinstated netplan but guacd was still missing. I copied the /usr/share/one/guacd folder from a backup as well as the openebula-guacd.service file. Started everything up and I was back in business. The only issue now is that guacamole still is not working.

Guacamole throws a 400 Bad Request


Versions of the related components and OS (frontend, hypervisors, VMs):
O/S Ubuntu 22.04
frontend: opennebula 6.10
hypervisor : kvm
vm’s : ubuntu 22.04/almalinux 8

Steps to reproduce:
Run multiple sudo apt update && sudo apt -y upgrade until you see the remove unused packages command shown and running command.
Current results:
Running remove unused packages command, removes opennebula required packages.
Expected results:
Opennebula required packages not flagged to be removed.

It doesn’t fail for that request. That request is for validation of opennebula support, you must have another one that starts with /vm/guacamole

Hi Jorge,
Thanks for your reply and help.
Ive not been able to find any errors in the logs and not able to find any /vm/guacamole in the developer tools debugger.
All I can find that looks like an error is the following:

If there is any guidance you can give, I’m all ears.

It would be great if we could find and fix this issue, especially as the upgrade from 6.8 to 6.10 went without a hitch. It was just me trying to cleanup the O/S after the upgrade. (Lesson Learnt)
Hopefully, I won’t have to backup my network templates, vm images and vm templates, destroy opennebula 6.10 then re-install and re-instate.

Hi, Just in case it helps, this was the procedure I followed to recover netplan after running apt autoremove.

All features are working except guacamole was not available so I copied the /usr/share/one/guacd folder from a backup and created openebula-guacd.service, guacd started ok and looks good but still not guac vnc :anguished:

Check the browser logs. you should see a request like this: http://localhost:2616/fireedge/api/vm/<VM_ID>/guacamole/<TYPE> check that where it says <VM_ID> is the id of the VM and that in the <TYPE> is the valid type (vnc/ssh/rdp) this request should deliver a json that in its data position are encrypted connection data!

Although I think that in your case what may be happening is that the server that opennebula-guacd should not be running. it should be running on port 4822.

Hi, Thanks for getting back to me, as you can see from the attached screenshots everything looks ok in the browser logs.



And guacd is running on port 4822.

Looking at the status of guacd, I see a msg that vnc is not installed.
Could this be the problem?

Hi Support,

Regarding the issue I’m having described above with Guacamole VNC sessions not connecting.
I’m assuming vnc was also removed when I ran apt autoremove

How would I go about installing vnc on the controller node?
Thanks in advance.

I have the same 400 error code like this on Ubuntu 22.04 with OpenNebula 6.10. I did no apt autoremove, just follow the doc.

Hi, The apt autoremove was a mistake on my part after everything was working fine. I’m trying to not do a re-install.
My opennebula 6.10 is working fine except I am not able to vnc to any vm’s because the guac service is saying vnc is not installed and I’m not sure of the best way to install vnc on opennebula 6.10

The apt autoremove seems to have caused havoc for another part of my 6.10 system.
I’m seeing lots of gem errors when I try to add marketplace app to my minio s3 marketplace.
I have a mini 6.10 server using the same s3 minio marketplace without errors.

Just decided, backing up templates and images then trashing and rebuilding 6.10.
Instead of installing manually I’ll probably try one of the following:
https://docs.opennebula.io/6.10/provision_clusters/edge_clusters/onprem_cluster.html

There are things in your installation that I don’t know why they fail, but if you want you can not use the opennebula-guacd package and use the guacamole package

Thanks for your help and suggestions everyone.
After spending a lot of time backing up images to trash and re-install I decided to :

  • sudo apt -y install guacd
  • sudo apt-get install opennebula opennebula-gate opennebula-flow opennebula-provision opennebula-fireedge python3-pyone
  • sudo apt update && sudo apt -y upgrade
  • sudo reboot.

Guacomle and vnc is now working again.
:rocket: :rocket: :rocket: