From http://docs.opennebula.org/5.2/operation/vm_management/vm_templates.html, a vm can have three types of disk:
- Use a persistent Image: changes to the disk image will persist after the VM is terminated.
- Use a non-persistent Image: a copy of the source Image is used, changes made to the VM disk will be lost.
- Volatile: disks are created on the fly on the target host. After the VM is terminated the disk is disposed.
As I understand the VM use non-persistent disk/image only lost data when VM is terminated. Is that correct? I am very appreciated for anyone help me explain more about non-persistent image/disk.
I had trouble with this concept when I started using OpenNebula also. It made more sense to me when I started thinking about it in terms of what the public cloud providers do. If you spin up a machine at Linode for example, the data on your disks will only persist until you destroy that machine. You can power-off and reboot the machine as much as you like, but the disk will stick around until you delete the VM.
ONE works the same way–most of our usage (99+%) is with non-persistent disks. ONE has some extra features that most cloud providers don’t though, you can ‘undeploy’ a VM and it will shut down and be essentially removed from the hypervisor until you ‘deploy’ it again, but the disk won’t be destroyed even though it’s non-persistent.
If you need to save the data that’s on a non-persistent disk you can do a ‘save as’ from SunStone or use the CLI to turn an existing non-persistent disk into a persistent one.
Does that help?
Thank you for your very informative and valuable reply to my question. It helped me to understand better some Open Nebula concepts. Once again, thank you for your help.