Install Milvus Standalone

You can install Milvus standalone with Docker Compose or on Kubernetes.

You can also build Milvus from source code.

Installing Milvus with Docker Compose can only be used for testing and cannot be used in production.

1. Start a local Kubernetes cluster:

$ minikube start

2. Start Milvus

Helm package manager for Kubernetes can be used to simplify this process.

Add the Milvus chart repository:

$ helm repo add milvus

Update your Milvus chart to the latest version:

$ helm repo update

Install Milvus Helm chart:

Provide a release name for identifying your Milvus deployment.

This tutorial uses my-release as the release name. To use a different release name, adjust the release name in subsequent commands.

Install Milvus standalone:

$ helm install my-release milvus/milvus
For more details, see Milvus Helm chart.

If Milvus boots successfully, each Milvus’ pod shows 1/1 under READY:

$ kubectl get pods
NAME                                               READY   STATUS      RESTARTS   AGE
my-release-etcd-0                                  1/1     Running     0          30s
my-release-milvus-standalone-54c4f88cb9-f84pf      1/1     Running     0          30s
my-release-minio-5564fbbddc-mz7f5                  1/1     Running     0          30s

3. Connect to Milvus

In a new terminal, port-forward your Milvus service to your local machine:

$ kubectl port-forward service/my-release-milvus 19530
Forwarding from -> 19530

4. Uninstall Milvus

$ helm uninstall my-release

5. Stop the cluster

To shut down the minikube virtual machine but preserve all the resources you’ve created, run the following command to stop the cluster:

$ minikube stop
Run minikube start to restart the cluster.

6. Delete the cluster

If you will not need to restart the cluster, run the following command to delete the minikube virtual machine as well as the persistent volume and all the created resources:

minikube delete
To retain logs, copy them from each pod's stderr before deleting the cluster and all its resources. To access a pod's standard error stream, run kubectl logs (podname).
Is this page helpful?
Scored Successfully!