Welcome to the Linux Foundation Forum!

"Saving progress" with labs

I'm up to Lab 5.1 and noticing the labs seem to build off all the progress that has been made in previous labs. For example, 5.1 requires working "simpleapp" deployments from lab 3. I saved the YAML files I used for lab 3 but I'm also having to go through all the Dockerfile creation, setting up a registry, etc, because the simpleapp image expects a registry to pull from. Not to mention the kubeadm + calico install + config from lab 1.

I'm getting through 1-2 labs a week and deleting the GCE instances I'm using for the labs in between, because I don't want to be paying for them to run when I'm not using them. But that means I have a lot of re-work to do with every lab.

Is there a way to save the entire lab environment including all the kubeadm setup, Docker registry and whatever else so I can quickly spin up a new instance with everything I need? I'm not super familiar with GCE yet.


  • I see that I can create disk snapshots and images https://cloud.google.com/compute/docs/disks/create-snapshots , but I'm not sure what exactly will be left in working order if I take a snapshot, delete the instances, and create new instances from the snapshots. Will I need to configure kubeadm or the private docker registry again?

  • Or maybe I'm better to just stop and restart the instances rather than deleting them. I'll try that for a bit and see how much they charge me.

  • chrispokorni
    chrispokorni Posts: 2,028

    Hi @ronniegane,
    I stop my instances when not in use, and I don't think I am being charged much - if at all, to have my VMs stopped overnight, or even over a few days. Upon restart, my cluster is up and running and I can just continue my work, without any data/config loss.
    As far as snapshots go, I am not sure what are the charges to create, store and then use them.

  • Thanks @chrispokorni , stopping and starting instances seems to work pretty well - all the pods I had running restarted just fine, I had to re-push the simpleapp image to the docker registry but that makes sense.


Upcoming Training