Error deploying 6.8 and importing host and VMs that ran on 5.12

I already have an On Premise cloud implemented with on in version 5.12. However, recently everything stopped working due to a power outage and we lost access to the SQLite database. We did another implementation of Sunstone and other server packages on another server in version 6.8.0. As a result, I’m having trouble connecting to hosts that were running on version 5.12. Also, on other hosts that we recently deployed, which were not on cloud 5.12, I can: Connect to the host, Create VM, and see the resources. However, after the VM is created, I cannot access it. The VM is born, but becomes inaccessible.

What should I do?

What exactly is the VM being inaccessible ? You have several ways to interact with the VM like VNC and SSH access.

In general

  • Confirm that the VM really exists on the hypervisor node by running virsh list
  • If the VM has VNC check if you can see the login prompt
  • If the VM has network, check if you can reach the VM IP address
    • You can also add a public SSH key to the VM and try accessing it with the corresponding private key
    • Alternatively you can set the password on the CONTEXT section and use it on the VNC login prompt or SSH

The issue is that, although OpenNebula indicates that the VM has an IP, I can’t even ping that IP. I have questions about how to configure the Virtual Network in the best way for the VM to be available. I tried several combinations, but still the VM is not accessible by IP or VNC.

root@torvalds:~# onevm list
60 oneadmin oneadmin 14020 runn 1 2G desktopg04 1d 19h06

Note that the IP is the IP that got assigned to the VM from the address space. For the VM to actually have that IP configured inside the Guest OS you need to make sure the VM as a working contextualization.