Creating 16.04 images-things that changed

Just an FYI until official support for Ubuntu 16.04.
I had to change a few things to get cloudmin to create a “usable” image from an “empty system” install of Ubuntu.

  1. I created and empty system from 16.04x64 server iso and installed a base install with ssh server. (2 issues, must manually assign IP info, and must not include a swap partition. Can only have 1 partition for cloudmin to be able to re-size the image.)
  2. SSH root access must be enabled in /etc/ssh/sshd.conf. Add ‘PermitRootLogin yes’ and comment out ‘PermitRootLogin prohibit-password’ reboot or restart ssh. After that, cloudmin can set root to use sshkey or password.
  3. create a cloudmin image from the now modified “empty system”.
  4. then create a cloudmin instance using the new 16.04 cloudmin image.
  5. Boot the new system, It will look dead/locked because network interface names have changed. Ubuntu 16.04 no longer uses eth0 names but cloudmin puts eth0 in the /etc/network/interfaces file.
    A. use root or the sudo user initially created during “empty system” install to login to the graphical remote console.
    B. edit /etc/network/interfaces and change eth0 in 2 places to ens3. (yours maybe different if you dont use virtIO net adapter in your VM’s) You can check the original /etc/network/interfaces file in the empty system to see what your adapters should be called. Then sudo /etc/init.d/networking restart will get your network up. It should now be pingable, ssh’able and cloumin’able. Only issue is cloudmin still shows there’s an eth0 interface. But it starts, reboots, shutsdown, all fine.
  6. Lastly delete the sudo user that was initially created.

You now have a 16.04 virtual machine.
Only issue is you’ll have to edit the interfaces file and remove sudo user every time you create an image from this cloudmin image.
Or you could create another cloudmin image from this image after deleting sudo user. But will still have to edit the interfaces file every time.

VirutalminGPL also installed via the ./install.sh with no issues other than the path to webalizer.conf has changed to /etc/webalizer/webalizer.conf. Changed it in the module confg for webalizer.

Nevermind.
While this works to host an empty system or image created system, in the long run I was getting hundreds of buffer I/O errors on loopback device. Only stopped after deleting the empty system install. Just shutting it down didnt help. The VM had to be deleted before I/O errors would stop.