Welcome to the Linux Foundation Forum!

Lab 3.2 -- Failing to Create PVs

Hi all,

I'm getting stuck on Lab 3.2 with the initial service creation. I'm running Ubuntu 18.04 on AWS. With the course tarbell I got, there's no "easyregistry.yaml", only a "edited-localregistry.yaml", so I'm working off that.

When I try to run "kubectl create -f /edited-localregistry.yaml", all the services/deployments/pods are created but get stuck at pending.

I tracked it down to a FailedBinding with the PVCs:

Is this a firewall issue? I haven't worked on AWS before and couldn't find related issues online.

Thanks!
Thomas

Comments

  • To clarify, I saw similar posts where the issue was deploying multiple times. I deleted all services/deployments/pods/persistentvolumeclaims in between each create.

  • thorhammer18
    thorhammer18 Posts: 3
    edited May 2021

    Update:
    AWS doesn't install a default StorageClass in EC2 instances, which can be done by following this guide: https://docs.aws.amazon.com/eks/latest/userguide/storage-classes.html

    However, that doesn't fix the PVC issue. My error is :
    Type Reason Age From Message
    ---- ------ ---- ---- -------
    Warning ProvisioningFailed 70s (x17 over 22m) persistentvolume-controller Failed to provision volume with StorageClass "gp2": Failed to get AWS Cloud Provider. GetCloudProvider returned instead

    Previous issues with this error recommended setting the --cloud-provider and --cloud-config flags with kubelet. Doing so didn't resolve my issues as I was erroring out with AWS credentials.

    I moved back to my local shell and ran the commands with the AWS CLI (I have my credentials set up in my local config) but had difficulty actually sending the command. I'll update once I figure it out but wanted to post to see if this provided additional insight. Definitely feel like I've missed something.

  • chrispokorni
    chrispokorni Posts: 2,346

    Hi @thorhammer18,

    The guide you are referencing is for Amazon EKS, and we are not using the Amazon managed Kubernetes service, instead we are using the upstream Kubernetes distribution. Features and plugins will work differently between the two Kubernetes distributions.

    Regards,
    -Chris

Categories

Upcoming Training