• Search for:
    11 Oct, 2021
    Kustomize Transformer Configurations for Cluster API v1beta1
    Cloud

    The topic of combining kustomize with Cluster API (CAPI) is a topic I’ve touched on several times over the last 18-24 months. I first touched on this topic in November 2019 with a post on using kustomize with CAPI manifests. A short while later, I discovered a way to change the configurations for the kustomize transformers to make it easier to use it with CAPI. That resulted in two posts on changing the kustomize transformers: one for v1alpha2 and one for v1alpha3 (since there were changes to the API between versions). In this post, I’ll revisit kustomize transformer configurations again, this time for CAPI v1beta1 (the API version corresponding to the CAPI 1.0 release).

    In the v1alpha2 post (the first post on modifying kustomize transformer configurations), I mentioned that changes were needed to the NameReference and CommonLabel transformers. In the v1alpha3 post, I mentioned that the changes to the CommonLabel transformer became largely optional; if you are planning on adding additional labels to MachineDeployments, then the change to the CommonLabels transformer is required, but otherwise you could probably get by without it.

    For v1beta1, the necessary changes are very similar to v1alpha3, and (for the most part) are focused on the NameReference transformer. The NameReference transformer tracks references between objects, so that if the name of an object changes—perhaps due to use of the namePrefix or nameSuffix directives in the kustomization.yaml file—references to that object are also appropriately renamed.

    Here are the CAPI-related changes needed for the NameReference transformer:

    - kind: Cluster group: cluster.x-k8s.io version: v1beta1 fieldSpecs: - path: spec/clusterName kind: MachineDeployment - path: spec/template/spec/clusterName kind: MachineDeployment - kind: AWSCluster group: infrastructure.cluster.x-k8s.io version: v1beta1 fieldSpecs: - path: spec/infrastructureRef/name kind: Cluster - kind: KubeadmControlPlane group: controlplane.cluster.x-k8s.io version: v1beta1 fieldSpecs: - path: spec/controlPlaneRef/name kind: Cluster - kind: AWSMachine group: infrastructure.cluster.x-k8s.io version: v1beta1 fieldSpecs: - path: spec/infrastructureRef/name kind: Machine - kind: KubeadmConfig group: bootstrap.cluster.x-k8s.io version: v1beta1 fieldSpecs: - path: spec/bootstrap/configRef/name kind: Machine - kind: AWSMachineTemplate group: infrastructure.cluster.x-k8s.io version: v1beta1 fieldSpecs: - path: spec/template/spec/infrastructureRef/name kind: MachineDeployment - path: spec/machineTemplate/infrastructureRef/name kind: KubeadmControlPlane - kind: KubeadmConfigTemplate group: bootstrap.cluster.x-k8s.io version: v1beta1 fieldSpecs: - path: spec/template/spec/bootstrap/configRef/name kind: MachineDeployment

    Generally, you’d append this content to the default NameReference transformer configuration, which you’d obtain using kustomize config save. However, somewhere in the Kustomize 3.8.4 release timeframe, the kustomize config save command for extracting the default transformer configurations was removed, and I have yet to figure out another way of getting this information. In theory, when using kustomize with CAPI manifests, you wouldn’t need any of the default NameReference transformer configurations, but I haven’t conducted any thorough testing of that theory (yet).

    Aside from replacing all instances of v1alpha3 with v1beta1, the only other difference in the YAML shown above compared to YAML in the the v1alpha3 post is a change to the fieldSpecs list for AWSMachineTemplate. Previously, the KubeadmControlPlane referenced an underlying AWSMachineTemplate at the path spec/infrastructureTemplate/name. In v1beta1, the KubeadmControlPlane object now references an AWSMachineTemplate at the path spec/machineTemplate/infrastructureRef/name.

    As mentioned in both of the previous posts, you’ll need to put this content in a file (I use namereference.yaml) and then specify the path to this configuration in kustomization.yaml, like this:

    configurations: - /path/to/customized/namereference.yaml

    I hope this information is useful to readers. Feel free to find me on the Kubernetes Slack instance if you have questions, and I’ll do my best to help answer them. You’re also welcome to contact me on Twitter (DMs are open). Thanks!

    …

    Read MoreKustomize Transformer Configurations for Cluster API v1beta1

    8 Oct, 2021
    Technology Short Take 146
    Cloud

    Welcome to Technology Short Take #146! Over the last couple of weeks, I’ve gathered a few technology-related links for you all. There’s some networking stuff, a few security links, and even a hardware-related article. But enough with the introduction—let’s get into the content!

    Networking

    Servers/Hardware

    • Chris Mellor speculates that Cisco UCS may be on the way out; Kevin Houston responds with a “I don’t think so.” Who will be correct? I guess we will just have to wait and see.

    Security

    Cloud Computing/Cloud Management

    Operating Systems/Applications

    Storage

    • Cloudflare recently introduced its own object storage offering, announced in this blog post. Cloudflare’s offering, called R2, offers an S3-compatible API and no egress fees, among other features.

    Virtualization

    Although this Tech Short Take is a tad shorter than usual, I still hope that you found something useful in here. Feel free to hit me up on Twitter if you have any feedback. Enjoy!

    …

    Read MoreTechnology Short Take 146

    7 Oct, 2021
    Installing Cilium via a ClusterResourceSet
    Cloud

    In this post, I’m going to walk you through how to install Cilium onto a Cluster API-managed workload cluster using a ClusterResourceSet. It’s reasonable to consider this post a follow-up to my earlier post that walked you through using a ClusterResourceSet to install Calico. There’s no need to read the earlier post, though, as this post includes all the information (or links to the information) you need. Ready? Let’s jump in!

    Prerequisites

    If you aren’t already familiar with Cluster API—hereafter just referred to as CAPI—then I would encourage you to read my introduction to Cluster API post. Although it is a bit dated (it was written in the very early days of the project, which recently released version 1.0). Some of the commands referenced in that post have changed, but the underlying concepts remain valid. If you’re not familiar with Cilium, check out their introduction to the project for more information. Finally, if you’re not familiar at all with the idea of ClusterResourceSets, you can read my earlier post or check out the ClusterResourceSet CAEP document.

    Installing Cilium via a ClusterResourceSet

    If you want to install Cilium via a ClusterResourceSet, the process looks something like this:

    1. Create a ConfigMap with the instructions for installing Cilium.
    2. Create a ClusterResourceSet that references the ConfigMap.
    3. Profit (when you deploy matching workload clusters).

    Let’s look at these steps in a bit more detail.

    Creating the Installation ConfigMap

    The Cilium docs generally recommend the use of the cilium CLI tool to install Cilium. The reasoning behind this is, as I understand it, that the cilium CLI tool can interrogate the Kubernetes cluster to gather information and then attempt to pick the best configuration options for you. Using helm is also another option recommended by the docs. For our purposes, however, neither of those approaches will work—using a ClusterResourceSet means you need to be able to supply YAML manifests.

    Fortunately, the fact that Cilium supports Helm gives us a path forward via the use of helm template to render the templates locally. As per the docs on helm template, there are some caveats/considerations, but this was the only way I found to create YAML manifests for installing Cilium.

    So, the first step to creating the ConfigMap you need is to set up the Helm repository:

    helm repo add cilium https://helm.cilium.io

    Then render the templates locally:

    helm template cilium cilium/cilium --version 1.10.4 \
--namespace kube-system > cilium-1.10.4.yaml

    You may need to specify additional options/values as needed in the above command in order to accommodate your specific environment or requirements, of course.

    Once you have the templates rendered, then create the ConfigMap that the ClusterResourceSet needs:

    kubectl create configmap cilium-crs-cm --from-file=cilium-1.10.4.yaml

    This ConfigMap should be created on the appropriate CAPI management cluster, so ensure your Kubernetes context is set correctly.

    Creating the ClusterResourceSet

    Now you’re ready to create the ClusterResourceSet. Here’s an example you could use as a starting point:

    ---
apiVersion: addons.cluster.x-k8s.io/v1alpha4
kind: ClusterResourceSet
metadata: name: cilium-crs namespace: default
spec: clusterSelector: matchLabels: cni: cilium  resources: - kind: ConfigMap name: cilium-crs-cm

    You can see that the ClusterResourceSet references the ConfigMap, which in turn contains the YAML to install Cilium (and that’s the YAML applied against matching workload clusters).

    Deploy a Matching Workload Cluster

    What determines a matching workload cluster? The clusterSelector portion of the ClusterResourceSet. In the example above, the ClusterResourceSet’s clusterSelector specifies that workload clusters should have the cni: cilium label attached.

    The label should be part of CAPI’s Cluster object, like this:

    ---
apiVersion: cluster.x-k8s.io/v1alpha4
kind: Cluster
metadata: name: cilium-cluster namespace: cilium-test labels: cni: cilium

    When CAPI creates a workload cluster with that label and value, as shown in the example CAPI manifest above, then the ClusterResourceSet will automatically apply the contents of the ConfigMap against the cluster after it has been fully provisioned. The result: Cilium gets installed automatically on the new workload cluster!

    I hope this post is useful. If you have any questions or any feedback, I’d love to hear from you! Feel free to find me on Twitter, or connect with me on the Kubernetes Slack instance. For more Cilium information and assistance, I’d encourage you to check out the Cilium Slack community.

    …

    Read MoreInstalling Cilium via a ClusterResourceSet