Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

The following scenarios of deploying the DPOD Cloud Agent on k8s cloud providers are provided only as examples.
The details may change between different cloud providers, k8s versions and k8s cluster custom configuration.
You are expected to modify and adjust the commands according to your specific case.

The following variables are variable is used across this document:

Code Block
DPOD_CLOUD_AGENT_VERSION="1.0.21.0"
DPOD_CLOUD_AGENT_OPERATOR_VERSION="1.1.0"

Configure Amazon EKS for Cloud Agent Deployment

Configure cluster nodes for registry mirror (containerd)

  • Create an EC2 Launch Template for the cluster nodes.

  • In the Launch Template Advanced details -> User data add the following configuration:
    Change the configuration according to your environment (at least the host attribute).

    Code Block
    MIME-Version: 1.0
    Content-Type: multipart/mixed; boundary="//"
    
    --//
    Content-Type: text/x-shellscript; charset="us-ascii"
    #!/bin/bash
    
    mkdir -p /etc/containerd/certs.d/icr.io
    printf 'server = "https://icr.io"\n[host."https://icr.io/dpod"]\n    capabilities = ["pull", "resolve"]\n    skip_verify = false\n    override_path = true\n' | tee /etc/containerd/certs.d/icr.io/hosts.toml
    --//--
  • Make sure to create the cluster with Node Group that uses the new Launch Template.

Configure Google GKE for Cloud Agent Deployment

Configure cluster nodes for registry mirror (containerd)

  • A DaemonSet can be used to configure the GKE cluster nodes containerd runtime.

  • K8s DaemonSet ensures that all (or some) nodes run a copy of a pod, which executes logic which in this case can be used to configure the registry mirror as described in Cloud Agent Mirroring - K8S (Optional).

  • Further information about k8s DaemonSet can be found in the k8s documentation.

  • A generic example of “startup-script” DeamonSet can be found here.

...

2.0"

Deploy DPOD Cloud Agent on k8s cluster using OLM

...