Help automating environment config in Chapter 3

andreizimin

Is there a way to auto-configure the master and worker VMs as per Chapter 3 requirements?
If I got it right, we'll need a configured master and worker VMs with Docker/kubeadm/kubectl/calico/etc. installed and configured on them for each lab going forward. I killed these instances after I was done with Chapter 3 to avoid charges of having resources on standby. Thank you for your help!

chrispokorni

Hi @andreizimin,

It is possible to automate the provisioning of your VM instances, together with the installation of software components and the cluster bootstrapping. The automation can be achieved with config management tools, and/or startup scripts on each VM.

However, I would take a closer look at the resources billing policy of the cloud provider, because there may be a way to reduce and/or possibly eliminate resources cost by simply "stopping" VMs (not "standby", "pause", or "suspend") instead of deleting/terminating them. "Stopping" VMs preserves all your work, and the next time you "start" your VMs you get to continue your work without any re-installs.

Regards,
-Chris

andreizimin

@chrispokorni said:
Hi @andreizimin,

It is possible to automate the provisioning of your VM instances, together with the installation of software components and the cluster bootstrapping. The automation can be achieved with config management tools, and/or startup scripts on each VM.

However, I would take a closer look at the resources billing policy of the cloud provider, because there may be a way to reduce and/or possibly eliminate resources cost by simply "stopping" VMs (not "standby", "pause", or "suspend") instead of deleting/terminating them. "Stopping" VMs preserves all your work, and the next time you "start" your VMs you get to continue your work without any re-installs.

Regards,
-Chris

Thank you Chris! I explored stopping the VM instead of killing it - you are still being charged for the disks allocated to it - will look into the pricing, maybe it's pennies.

If the cost of standby is substantial, what is the route to automating the provisioning?

chrispokorni

Hi @andreizimin,

Costs associated with disk space should not be significant.

However, for instance provisioning automation and cluster install/bootstrapping there are various tools that may work, depending on the infrastructure - cloud provider, local hypervisor, etc... You may be able to use Terraform, CloudFormation, Vagrant, and possibly others. In the past learners shared in this forum github repositories with such automation procedures. Using the search feature of this forum should help to find them.

Regards,
-Chris