Cloud-Init is the defacto multi-distribution package that handles early initialization of a virtual machine instance. Using Cloud-Init, one can configure network devices and ssh keys on the hypervisor side. When the VM starts the first time, the Cloud-Init software inside the VM applies those settings.
Many Linux distributions provides ready-to-use Cloud-Init images, mostly designed for 'OpenStack'. Those images also works with {pve}. While it may be convenient to use such read-to-use images, we usually recommend to prepare those images by yourself. That way you know exactly what is installed, and you can easily customize the image for your needs.
Once you created such image, it is best practice to convert it into a VM template. It is really fast to create linked clones of VM templates, so this is a very fast way to roll out new VM instances. You just need to configure the network (any maybe ssh keys) before you start the new VM.
We recommend the use of SSH key-based authentication to login to VMs provisioned by Cloud-Init. It is also possible to set a password, but {pve} needs to store an encrypted version of that password inside the Cloud-Init data. So this is not as safe as using SSH key-based authentication.
{pve} generates an ISO image to pass the Cloud-Init data to the VM. So all Cloud-Init VMs needs to have an assigned CDROM drive for that purpose. Also, many Cloud-Init Images assumes to have a serial console, so it is best to add a serial console and use that as display for those VMs.
The first step is to prepare your VM. You can basically use any VM, and simply install the Cloud-Init packages inside the VM you want to prepare. On Debian/Ubuntu based systems this is as simple as:
apt-get install cloud-init
Many distributions provides ready-to-use Cloud-Init images (provided
as .qcow2 files), so as alternative you can simply download and
import such image. For the following example, we will use the cloud
images provided by Ubuntu at https://cloud-images.ubuntu.com.
# download the image wget https://cloud-images.ubuntu.com/bionic/current/bionic-server-cloudimg-amd64.img # create a new VM qm create 9000 --memory 2048 --net0 virtio,bridge=vmbr0 # import the downloaded disk to local-lvm storage qm importdisk 9000 bionic-server-cloudimg-amd64.img local-lvm # finally attach the new disk to the VM as scsi drive qm set 9000 --scsihw virtio-scsi-pci --scsi0 local-lvm:vm-9000-disk-1
|
Note
|
Ubuntu Cloud-Init images requires the virtio-scsi-pci
controller type for SCSI drives.
|
The next step is to configure a CDROM drive, used to pass the Cloud-Init data to the VM.
qm set 9000 --ide2 local-lvm:cloudinit
We want to boot directly from the Cloud-Init image, so we set the
bootdisk parameter to scsi0 and restrict BIOS to boot from disk
only. This simply speeds up booting, because VM BIOS skips testing for
a bootable CDROM.
qm set 9000 --boot c --bootdisk scsi0
We also want to configure a serial console and use that as display. Many Cloud-Init images rely on that, because it is an requirement for OpenStack images.
qm set 9000 --serial0 socket --vga serial0
Finally, it is usually a good idea to transform such VM into a template. You can create linked clones with them, so deployment from VM templates is much faster than creating a full clone (copy).
qm template 9000
You can easily deploy such template by cloning:
qm clone 9000 123 --name ubuntu2
Then configure the SSH public key used for authentication, and the IP setup
qm set 123 --sshkey ~/.ssh/id_rsa.pub qm set 123 --ipconfig0 ip=10.0.10.123/24,gw=10.0.10.1
You can configure all Cloud-Init options using a single command. I just split above example to separate commands to reduce the line length. Also make sure you adopt the IP setup for your environment.