Is there a windows executable for kubectl-tg? If I download the kubectl plugin and add it to the path on my windows system I’ll receive the following message when running kubectl plugin list
kubectl-tg identified as a kubectl plugin, but it is not executable
@nick_h This might help, it is a guide that walks you through the installation process. Let me know if you’re still running into any issues with the install!
This page describes the steps to install and uninstall the TigerGraph Kubernetes Operator on your Kubernetes cluster.
Kubernetes Operator support is currently a Preview Feature. Preview Features give users an early look at future production-level features. Preview Features should not be used for production deployments.
The Operator only needs to be installed once for each cluster. If you run into errors saying that Custom Resource Definition (CRD) already exists, it is likely that someone else who has access to your cluster has already installed the operator. Therefore, you can safely ignore the error and skip installation.
The following software must be installed on your computer:
You have configured the kubectl command-line tool for access to your running Kubernetes cluster.
kubectl pluginRun the command below to install the TigerGraph plugin for the kubectl command-line tool:
curl https://dl.tigergraph.com/k8s/0.0.4/kubectl-tg -o kubectl-tg
sudo install kubectl-tg /usr/local/bin/
The plugin is installed on your local machine. You do not need to install the plugin for each cluster you operate on.
Run the command below to initialize the TigerGraph Kubernetes Operator.
$ kubectl tg init
Add the -n flag with your namespace to use a custom namespace. The full options for the init command are listed here:
kubectl tg init --help
Install the operator
Examples:
# install the operator in the current namespace
kubectl tg init
# install the operator in the specified namespace
kubectl tg init --namespace tg-tenant1
# install the operator in the specified namespace, with specified helm repo and image pull secret
kubectl tg init --namespace tg-tenant1 --helm-repo https://yourhelprepo.com --image-pull-secret yoursecret
# install the operator in the specified namespace, with specified operator version, watch name namespace, cpu and memory
kubectl tg init --cluster-scope false --version OPERATOR_VERSION --operator-size 3 --operator-watch-namespace tigergraph --operator-cpu 1000m --operator-memory 1024Mi --namespace tigergraph
Options:
-n, --namespace : set namespace to deploy TG cluster, default as default
--helm-repo : set the specified helm repo to install operator, default as https://dl.tigergraph.com/charts
--docker-registry : set docker registry to download tigergraph image, default as docker.io
--docker-image-repo : set docker image repo for image name, default as tigergraph.
-p, --image-pull-secret : set imagePullSecret of docker registry, default as tigergraph-operator-image-pull-secrets-default
--image-pull-policy: set pull policy of image, available policy: IfNotPresent, Always, and Never, default is IfNotPresent
--operator-version: set TG K8S operator version
--operator-size : set the replicas of operator's deployment for high availability, default is 3
--operator-cpu : set request cpu of operator, default as 1000m
--operator-memory : set request memory of operator, default as 1024Mi
--operator-watch-namespace : set watch namespaces of operator, blank string as default indicate all namespace, multiple namespaces are separated by commas, as ns1\,ns2
--cluster-scope : set true to deploy operator with ClusterRole, set false to deploy with Role,
so that you can deploy mutiple operators in one cluster, default as true
The --cluster-scope option during creation allows you to specify whether to manage resources across the whole cluster (cluster scoped, the default option) or across only a namespace (namespace scoped).
In this example, resources will only be managed in the tigergraph namespace.
kubectl tg init --cluster-scope false --namespace tigergraph
Run the following command to uninstall the Operator:
$ kubectl tg uninstall
This command only uninstalls the Operator. The clusters created by the Operator and the Custom Resource Definitions (CRD) defined when you first installed the Operator are not deleted by this command.
Thanks Jon. I have read through this documentation. Windows support doesn’t seem to be mentioned.