SigOpt: Bring Your Own Cluster
You can connect to SigOpt and run experiments using any Kubernetes cluster! If you are using Google Cloud Platform, Microsoft Azure, or bringing your own cluster on AWS, use this section to learn how to connect your existing Kubernetes cluster to SigOpt.
Configuration Back to Top
SigOpt uses a kubeconfig file to connect to your cluster.
If you already have your kubeconfig file, you can use it to connect to SigOpt with the
$ sigopt cluster connect \ --cluster-name my-cluster \ --provider custom \ --registry my-registry \ # example: gcr.io/[project-id] --kubeconfig $HOME/.kube/config
For example, if you have a Google Kubernetes Engine (GKE) cluster up and running, and have the
gcloud CLI installed, you can run the following command to generate your kubeconfig:
$ gcloud container clusters get-credentials <CLUSTER-NAME> [--region <GCP-REGION>]
Initial Setup Back to Top
SigOpt needs to create some extra resources before running on an existing cluster. Please run the following command after connecting to your cluster for the first time:
$ sigopt cluster install-plugins
This only needs to be done one time for each custom cluster you use.
Custom Image Registries Back to Top
Clusters with the provider
custom will use Docker Hub as their default container registry.
To override the default options with a custom image registry, provide the
--registry argument when you connect to your cluster:
$ sigopt cluster connect \ --cluster-name tiny-cluster \ --provider custom \ --kubeconfig /path/to/kubeconfig \ --registry myregistrydomain:port
Kubeconfig File Not Found Back to Top
For this section, you will need to install kubectl, a tool for communicating with Kubernetes clusters.
Test whether or not your development environment is properly setup to talk to your Kubernetes cluster:
$ kubectl get svc NAME TYPE CLUSTER-IP EXTERNAL-IP PORT(S) AGE svc/kubernetes ClusterIP 10.100.0.1 <none> 443/TCP 1m
If your output looks like the one above, you are ready to proceed to the next step. If your output looks significantly different, or contains an error, please contact your Kubernetes administrator to help you configure access to your cluster.
Connect to SigOpt
kubectl to copy the kubeconfig to a local file, and then connect to SigOpt.
$ kubectl config view >> kubeconfig $ sigopt cluster connect \ --cluster-name my-cluster \ --provider custom \ --registry my-registry \ # example: gcr.io/[project-id] --kubeconfig kubeconfig
AWS cluster create and manageDockerfile: define model training environment