Deploy a MinIO Tenant

This procedure documents deploying a MinIO Tenant onto a stock Kubernetes cluster using either Kustomize or MinIO’s Helm Charts.

Deploying Single-Node topologies requires additional configurations not covered in this documentation. You can alternatively use a simple Kubernetes YAML object to describe a Single-Node topology for local testing and evaluation as necessary. MinIO does not recommend nor support single-node deployment topologies for production environments.

This documentation assumes familiarity with all referenced Kubernetes concepts, utilities, and procedures. While this documentation may provide guidance for configuring or deploying Kubernetes-related resources on a best-effort basis, it is not a replacement for the official Kubernetes Documentation.

Prerequisites

MinIO Kubernetes Operator

The procedures on this page requires a valid installation of the MinIO Kubernetes Operator and assumes the local host has a matching installation of the MinIO Kubernetes Operator. This procedure assumes the latest stable Operator, version 6.0.4.

See Deploy the MinIO Operator for complete documentation on deploying the MinIO Operator.

EKS Cluster with EBS-Optimized EC2 Nodes

This procedure assumes an existing EKS cluster with at least four EC2 nodes. The EC2 nodes should have matching machine types and configurations to ensure predictable performance with MinIO.

MinIO provides hardware guidelines for selecting the appropriate EC2 instance class and size. MinIO strongly recommends selecting EBS-optimized instances with at least 25Gbps Network bandwidth as a baseline for performance.

For more complete information on the available EC2 and EBS resources, see EC2 Instance Types and EBS Volume Types. MinIO SUBNET customers should reach out to MinIO engineering as part of architecture planning for assistance in selecting the optimal instance and volume types for the target workload and performance goals.

Persistent Volumes

Exclusive access to drives

MinIO requires exclusive access to the drives or volumes provided for object storage. No other processes, software, scripts, or persons should perform any actions directly on the drives or volumes provided to MinIO or the objects or files MinIO places on them.

Unless directed by MinIO Engineering, do not use scripts or tools to directly modify, delete, or move any of the data shards, parity shards, or metadata files on the provided drives, including from one drive or node to another. Such operations are very likely to result in widespread corruption and data loss beyond MinIO’s ability to heal.

MinIO Tenants on EKS must use the EBS CSI Driver to provision the necessary underlying persistent volumes. MinIO strongly recommends using SSD-backed EBS volumes for best performance. MinIO strongly recommends deploying EBS-based PVs with the XFS filesystem. Create a StorageClass for the MinIO EBS PVs and set the csi.storage.k8s.io/fstype parameter to xfs . For more information on EBS resources, see EBS Volume Types. For more information on StorageClass Parameters, see StorageClass Parameters.

Deploy a MinIO Tenant using Kustomize

The following procedure uses kubectl -k to deploy a MinIO Tenant using the base Kustomization template in the MinIO Operator Github repository.

You can select a different base or pre-built template from the repository as your starting point, or build your own Kustomization resources using the MinIO Custom Resource Documentation.

Important

If you use Kustomize to deploy a MinIO Tenant, you must use Kustomize to manage or upgrade that deployment. Do not use kubectl krew, a Helm Chart, or similar methods to manage or upgrade the MinIO Tenant.

This procedure is not exhaustive of all possible configuration options available in the Tenant CRD. It provides a baseline from which you can modify and tailor the Tenant to your requirements.

Create a YAML object for the Tenant

Use the kubectl kustomize command to produce a YAML file containing all Kubernetes resources necessary to deploy the base Tenant:
```
kubectl kustomize https://github.com/minio/operator/examples/kustomization/base/ > tenant-base.yaml
```
The command creates a single YAML file with multiple objects separated by the --- line. Open the file in your preferred editor.

The following steps reference each object based on it’s kind and metadata.name fields:

Configure the Tenant topology

The kind: Tenant object describes the MinIO Tenant.

The following fields share the spec.pools[0] prefix and control the number of servers, volumes per server, and storage class of all pods deployed in the Tenant:

Field	Description
`servers`	The number of MinIO pods to deploy in the Server Pool.
`volumesPerServer`	The number of persistent volumes to attach to each MinIO pod (`servers`). The Operator generates `volumesPerServer x servers` Persistant Volume Claims for the Tenant.
`volumeClaimTemplate.spec.storageClassName`	The Kubernetes storage class to associate with the generated Persistent Volume Claims. If no storage class exists matching the specified value or if the specified storage class cannot meet the requested number of PVCs or storage capacity, the Tenant may fail to start.
`volumeClaimTemplate.spec.resources.requests.storage`	The amount of storage to request for each generated PVC.

Configure Tenant Affinity or Anti-Affinity

The MinIO Operator supports the following Kubernetes Affinity and Anti-Affinity configurations:
- Node Affinity (spec.pools[n].nodeAffinity)
- Pod Affinity (spec.pools[n].podAffinity)
- Pod Anti-Affinity (spec.pools[n].podAntiAffinity)
MinIO recommends configuring Tenants with Pod Anti-Affinity to ensure that the Kubernetes schedule does not schedule multiple pods on the same worker node.

If you have specific worker nodes on which you want to deploy the tenant, pass those node labels or filters to the nodeAffinity field to constrain the scheduler to place pods on those nodes.

Configure Network Encryption

The MinIO Tenant CRD provides the following fields from which you can configure tenant TLS network encryption:

Field	Description
`tenant.certificate.requestAutoCert`	Enable or disable MinIO automatic TLS certificate generation Defaults to `true` or enabled if omitted.
`tenant.certificate.certConfig`	Customize the behavior of automatic TLS, if enabled.
`tenant.certificate.externalCertSecret`	Enable TLS for multiple hostnames via Server Name Indication (SNI) Specify one or more Kubernetes secrets of type `kubernetes.io/tls` or `cert-manager`.
`tenant.certificate.externalCACertSecret`	Enable validation of client TLS certificates signed by unknown, third-party, or internal Certificate Authorities (CA). Specify one or more Kubernetes secrets of type `kubernetes.io/tls` containing the full chain of CA certificates for a given authority.

Configure MinIO Environment Variables

You can set MinIO Server environment variables using the tenant.configuration field.

Field	Description
`tenant.configuration`	Specify a Kubernetes opaque secret whose data payload `config.env` contains each MinIO environment variable you want to set. The `config.env` data payload must be a base64-encoded string. You can create a local file, set your environment variables, and then use `cat LOCALFILE \| base64` to create the payload.

Field

Description

tenant.configuration

Specify a Kubernetes opaque secret whose data payload config.env contains each MinIO environment variable you want to set.

The config.env data payload must be a base64-encoded string. You can create a local file, set your environment variables, and then use cat LOCALFILE | base64 to create the payload.

The YAML includes an object kind: Secret with metadata.name: storage-configuration that sets the root username, password, erasure parity settings, and enables Tenant Console.

Modify this as needed to reflect your Tenant requirements.

Review the Namespace

The YAML object kind: Namespace sets the default namespace for the Tenant to minio-tenant.

You can change this value to create a different namespace for the Tenant. You must change all metadata.namespace values in the YAML file to match the Namespace.
Deploy the Tenant

Use the kubectl apply -f command to deploy the Tenant.
```
kubectl apply -f tenant-base.yaml
```
The command creates each of the resources specified in the YAML object at the configured namespace.

You can monitor the progress using the following command:
```
watch kubectl get all -n minio-tenant
```
Expose the Tenant MinIO S3 API port

To test the MinIO Client mc from your local machine, forward the MinIO port and create an alias.
- Forward the Tenant’s MinIO port:
```
kubectl port-forward svc/MINIO_TENANT_NAME-hl 9000 -n MINIO_TENANT_NAMESPACE
```
- Create an alias for the Tenant service:
```
mc alias set myminio https://localhost:9000 minio minio123 --insecure
```
You can use mc mb to create a bucket on the Tenant:
```
mc mb myminio/mybucket --insecure
```
If you deployed your MinIO Tenant using TLS certificates minted by a trusted Certificate Authority (CA) you can omit the --insecure flag.

See Connect to the Tenant for specific instructions.

Connect to the Tenant

The MinIO Operator creates services for the MinIO Tenant.

Use the kubectl get svc -n NAMESPACE command to review the deployed services:

kubectl get svc -n TENANT-NAMESPACE

NAME                               TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)          AGE
minio                              LoadBalancer   10.97.114.60     <pending>     443:30979/TCP    2d3h
TENANT-NAMESPACE-console             LoadBalancer   10.106.103.247   <pending>     9443:32095/TCP   2d3h
TENANT-NAMESPACE-hl                  ClusterIP      None             <none>        9000/TCP         2d3h

The minio service corresponds to the MinIO Tenant service. Applications should use this service for performing operations against the MinIO Tenant.
The *-console service corresponds to the MinIO Console. Administrators should use this service for accessing the MinIO Console and performing administrative operations on the MinIO Tenant.

The remaining services support Tenant operations and are not intended for consumption by users or administrators.

By default each service is visible only within the Kubernetes cluster. Applications deployed inside the cluster can access the services using the CLUSTER-IP.

Applications external to the Kubernetes cluster can access the services using the EXTERNAL-IP. This value is only populated for Kubernetes clusters configured for Ingress or a similar network access service. Kubernetes provides multiple options for configuring external access to services.

See the Kubernetes documentation on Publishing Services (ServiceTypes) and Ingress for more complete information on configuring external access to services.