Welcome to the Linux Foundation Forum!

Using GCP / AWS / Azure Linux VMS

I am planning on using GCPs GCE to provision and manage my Ubuntu VMs alongside the GCP Cloud Shell. Later on (since I already have working experience with GKE) I plan on using K8s on GCP.

Are there going to be things that I am not going to be able to do? I would rather spend the time to do things as much as Cloud Native as possible and add Terraform for the stuff I don't want to click through every time.

Comments

  • juliovnjuliovn Posts: 5

    From what I can tell you can use any of the big names cloud providers however those do come at a cost which might not be worth it just for studying purposes. And at least from GCP when I tried it wouldn't let me set up k8s with the free trial credit they give you. In the end I opted for setting up a local cluster with vagrant which I think works really well in simulating all that.

  • DRpandaMDDRpandaMD Posts: 17

    @juliovn
    So I've had this newish GCP account while studying for my Associate GCP Cloud Engineer cert with Acloudguru.

    And I recently went through Nigel Poulton's Kubernetes DeepDive course on the same account.

    As of recent, I think earlier this year or late last year. GCP gives you 1 management node per region for free. You just have to manged the worker nodes. I used the N1-standard machines in a cluster of 3 and my auto scaler set to 3min, 3 req 5 max.
    Anything smaller and the cluster yo-yos a lot as all those operational pods are fighting for resources and get preempted over and over.

    When you are done you go back over to the cluster management on the GCP side remove the auto scaler and set the node count to 0 and save. That way you won't get charged AND you don't have to reconfigure a K8s cluster over and over.

    Late last night, I found out that the GCP cloud shell env has Terraform preloaded. I went to work creating a small terraform file to give me a VM. I got it working and that is how I will quickly manage spinning up new VMs as clicking over and over gets boring and repetitive.

    Oh and as bonus using gcloud compute ssh {instance name} it manages your SSH key FOR you. So I can work contained without accidental leakage.

    For anyone else that is interested and is somewhat familiar with Terraform let me know and I'll share that file with you.

  • Hi @DRpandaMD ,

    At least for DevOps Fundamentals (LFS261) we'll use GCP (Google Cloud Platform). It gives $300 free, so we'll be using that free credit. For K8 I'll let Tim Serewicz to answer.

    Many regards and welcome to the bootcamp!

    Luis.

  • DRpandaMDDRpandaMD Posts: 17

    @luisviveropena great then I am already in the correct mental frame. GCP is <3

  • juliovnjuliovn Posts: 5

    Hi @DRpandaMD,

    Well that is good to know, thanks for the tips!

  • cognextechcognextech Posts: 1
    edited July 2020

    Meanwhile if you are learner, and looking for AWS training then visit Cognex technology

  • serewiczserewicz Posts: 899

    Hello,
    Our Kubernetes courses labs are tested on GCP, but are written to be able to run on many different platforms such as AWS, Digital Ocean, VirtualBox and bare metal. Each will have some considerations, perhaps a couple of extra steps. In the long run we basically want Ubuntu 18.04 instances that don't have a firewall between them.

Sign In or Register to comment.