So I think I figured cloud-init!

Previously I wrote about cloud-init as part of something a cloud service provider might offer. (Or perhaps I simply called it user-data; I don't accurately recall) Recently I have learned that it is something provisioned by the operating system itself. (Especially linux based servers). I do not know the history of the os feature as such, but if you are someone who plays around with Oracle Virtualbox for provisioning VMs then you need to know how to benefit from cloud-init.

I had to go through several webpages and perform several experiments to get it right. So in this post I will collate the steps for you to get easily started.

1. (Step 1) TLDR: Grab the vmdk file. Create a blank VM (no disk) and attach the disk you have downloaded.
   
   You need to grab a cloud image. Every distro has it. Its a vmdk file, that is it is a virtual box disk file. (Not the traditional iso). If you created a blank VM and attached this disk, and, powered up the VM, the os will boot and display the login prompt. But since you don't know the credentials, you cannot proceed to work with the vm instance any further. 
   
   
2. Compose the cloud-init disk. Multiple steps required here. Hence find the dedicated section below
   
   
3. Attach the cloud-init disk to the above, and, boot.
   

After the VM boots, you can login with the user profile you specified in step #2. It will have all the necessary artifacts you have specified in the same step.

The benefit is that now you have a VM that is configured with all the necessary software you want to further your exploration. 

Goodbye secure practices👋🫡

Most cloud-init tutorials will talk about creating different users, public key based ssh authentication, configuring sudo to not prompt for password, etc. I am assuming you are novice to the whole concept of cloud-init, and, you are working in some kind of personal self-exploratory capacity. 

Bottom line is that those secure practices are meant for professionals or experts. I am assuming most readers reading this post are trying to become one, and, they understand that they should not perform these steps for their professional work.

Configuring the cloud-init disk

Most of the information here is obtained from this thread: https://superuser.com/questions/827977/use-cloud-init-with-virtualbox/853957#853957

I will reproduce the commands in a slightly different way below. Now is probably a good time to check out some cloud-init docs and tutorial videos. It should give you a precursor to the stuff that I write, for e.g. in the user-data or meta-data file below. The tutorials you find online, are vastly different from what you are going to go through out here.

0. What am I doing here?

I am creating an instance with nodejs pre-installed. I have started off with an ubuntu cloud image. So when you log-in you should theoretically work with node in a out-of-box style. You log-in to the os with username/password: ubuntu/ubuntu. The hostname of the provisioned machine is osbox03. All this is done by cloud-init. The process would download nodejs and make its binaries globally available. However, for cloud-init to work in this manner, we need to create a disk with a certain label, copy some files over to that disk; files which have the necessary cloud-init configuration. This is outlined in one or more of the steps below. In the end you will also find the link to a gist which has all the data and commands you need to type.

1. create a user-data file:

#cloud-config
users:
  - default

ssh_pwauth: true
chpasswd: { expire: false }
preserve_hostname: False
hostname: osbox03
runcmd:
  - [ ls, -l, / ]
  - [ sh, -xc, "echo $(date) ': hello world!'" ]
  - [ sh, -c, echo "=========hello world=========" ]
  - [ mkdir, "/home/ubuntu/nodejs" ]
  - [ wget, https://nodejs.org/dist/v20.11.1/node-v20.11.1-linux-x64.tar.xz, -O, /home/ubuntu/nodejs/node-v20.11.1-linux-x64.tar.xz ]
  - [ tar, xvf, /home/ubuntu/nodejs/node-v20.11.1-linux-x64.tar.xz, -C, /home/ubuntu/nodejs/ ]
  - [ ln, -s, /home/ubuntu/nodejs/node-v20.11.1-linux-x64/bin/node, /bin/node ]
  - [ ln, -s, /home/ubuntu/nodejs/node-v20.11.1-linux-x64/bin/npx, /bin/npx ]
  - [ ln, -s, /home/ubuntu/nodejs/node-v20.11.1-linux-x64/bin/npm, /bin/npm ]
  - [ rm, /home/ubuntu/nodejs/node-v20.11.1-linux-x64.tar.xz ]

system_info:
  default_user:
    name: ubuntu
    plain_text_passwd: 'ubuntu'
    shell: /bin/bash
    lock_passwd: false
    gecos: ubuntu user

2. Create meta-data file:

instance-id: my-instance-1

3. Create the cloud-init disk:

Follow these steps:

# Create empty virtual hard drive file
dd if=/dev/zero of=config.img bs=1 count=0 seek=2M

# put correct filesystem and disk label on
mkfs.vfat -n cidata config.img

# mount it somewhere so you can put the config data on
sudo mount config.img /mnt

Copy the user-data and meta-data files to /mnt, and, then unmount:

sudo cp user-data meta-data /mnt

sudo umount /mnt

config.img is hydrated with the cloud-init setup. We need to convert the file with img extension to vmdk extension.

  
sudo apt-get install qemu-kvm
qemu-img convert -O vmdk  config.img config.vmdk

Now attach config.vmdk to the VM created in step #1, and power it up.

Now after you have powered up your VM, you can physically log-in to the terminal. You can quickly inspect the /home/ubuntu/nodejs folder. If contents don't exist, you may have to wait a while for cloud-init to conclude its work. You can run the following commands to inspect the cloud-init output:

cat /var/log/cloud-init-output.log

If anything fails you will learn about it through the above output.And if everything works out you can type the following and self-confirm everything works:

node  --version && npm --version

An alternative command you can run to assess the status of cloud-init is below:

cloud-init status

Thaks all folks! 

Gist: https://gist.github.com/deostroll/bcb18a5d25f533b4aad3f27566219bf9