Managing the Multi-Cloud Object Gateway

You can use NooBaa to manage your s3 compatible object storage resources, like IBM Object Storage, AWS s3, or Azure Blob storage. With NooBaa, you can also create and manage your object storage consistently across clusters and zones.

Setting up backing stores by using the NooBaa CLI

Backing stores are NooBaa resources for managing your s3 compatible services and buckets. When you create a backing store, you provide your s3 service credentials like your access key ID, secret access key, and endpoint for your object storage service. After you add backing stores to your cluster, you can create bucket classes which allow you to configure data federation policies for your s3 compatible storage services.

After you deploy ODF, you can configure more backing stores in your storage cluster. You can create a backing store by using any s3 compatible object store such as AWS or IBM Cloud Object Storage.

You can also create and manage your backing stores in the Red Hat OpenShift web console.

To add a backing store to your storage cluster by using the NooBaa CLI:

Access your Red Hat OpenShift cluster.

Install the NooBaa Operator CLI.

Run the noobaa backingstore create command to see a list of supported backing store types.

noobaa backingstore create

Example output

Create backing store

Available Commands:
aws-s3               Create aws-s3 backing store
azure-blob           Create azure-blob backing store
google-cloud-storage Create google-cloud-storage backing store
ibm-cos              Create ibm-cos backing store
pv-pool              Create pv-pool backing store
s3-compatible        Create s3-compatible backing store

Get the details of the service that you want to use. If you want to set up an IBM Cloud Object Storage, get your HMAC credentials. For more information, see Using HMAC credentials. The following example command shows the configuration parameters that are required to create a backing store by using an IBM Cloud Object Storage service instance.
```
noobaa backingstore create ibm-cos <backing-store-name> -n openshift-storage --access-key=<access-key> --endpoint=<endpoint> --secret-key=<secret-key> --target-bucket<target-bucket>
```

Get the details of your backing store.

noobaa backingstore status <backing-store-name> -n openshift-storage

Creating a bucket class that uses two backing stores with a mirror placement policy

After you have installed ODF and configured at least two backing stores, you can create a bucket class that uses multiple backing stores. Bucket classes allow you to configure data federation policies across your backing stores. Then, when your apps write objects to your bucket, you can mirror or spread the objects across two or more backing stores.

Before you begin Install the NooBaa CLI)

Get the status of your NooBaa instance. For clusters in IBM Cloud, target the openshift-storage namespace.
```
noobaa status -n openshift-storage
```

Verify you have at least two backing stores.

noobaa backingstore list -n openshift-storage

Create a bucket class that uses two of your backing stores and has a Mirror policy to mirror objects across two or more backing stores.

noobaa bucketclass create <bucketclass-name> --backingstores=[<backingstore_one>,<backingstore_two>] --placement='<placement>' -n <namespace>

Optional You can also create a bucket class by using a YAML configuration file.

apiVersion: noobaa.io/v1alpha1
kind: BucketClass
metadata:
labels:
    app: noobaa
name: mirror-bucket-class
namespace: openshift-storage
spec:
placementPolicy:
    tiers:
    - backingStores:
    - backing-store-one
    - backing-store-two
    placement: Mirror

Creating a storage class

After you create backing stores and a bucket class, create a storage class to make your bucket class resources available across the namespaces in your cluster.

Create a storage class YAML file that contains a name for your storage class in the format <name>-noobaa.noobaa.io, the name of the bucket class that you want to use, and the provisioner set to openshift-storage.noobaa.io/obc.
```
metadata:
    name: <name>-noobaa.noobaa.io
  parameters:
bucketclass: <bucket-class-name>
provisioner: openshift-storage.noobaa.io/obc
reclaimPolicy: Delete
volumeBindingMode: Immediate
```
Create the storage class in your cluster.
```
oc create -f sc.yaml
```
Verify the storage class is created.
```
oc get sc
```

Creating an Object Bucket Claim

After you create backing stores and a bucket class, you can create an object bucket claim that you can use to claim storage resources for your s3 apps.

Create an object bucket claim YAML file that references the bucket class that you created earlier.

apiVersion: objectbucket.io/v1alpha1
kind: ObjectBucketClaim
metadata:
    name: cos-obc
  spec:
storageClassName: cos-noobaa.noobaa.io
bucketName: cos-bucket

Create the object bucket claim in your cluster.
```
oc apply -f obc.yaml
```
Verify the OBC is created.
```
noobaa obc list
```
List the secrets and config maps in your cluster and verify that the corresponding secret and config map for your OBC are created.
```
oc get secrets | grep <obc-name> && oc get cm | grep <obc-name>
```

Deploying an s3 app in the Multicloud Object Gateway

Create an app YAML file that references the secret and config map from your OBC. The following example app uses the banst/awscli image and configures the AWS CLI.

apiVersion: v1
kind: Pod
metadata:
    name: app
labels:
  app: app
spec:
    containers:
    - name: app
    envFrom:
    - secretRef:
        name: <secret-name>
    - configMapRef:
        name: <config-map-name>
    image: banst/awscli
    command:
        - sh
        - "-c"
        - |
            echo "----> Configuring S3 endpoint ...";
            pip install awscli-plugin-endpoint;
            aws configure set plugins.endpoint awscli_plugin_endpoint;
            aws configure set s3.endpoint_url https://s3.openshift-storage.svc;
            echo "----> Configuring certificates ...";
            aws configure set ca_bundle /var/run/secrets/kubernetes.io/serviceaccount/service-ca.crt;
            echo "----> Copying files ...";
            aws s3 cp --recursive /etc s3://$BUCKET_NAME;
            echo "----> List files ...";
            aws s3 ls $BUCKET_NAME;
            echo "----> Done.";

Create the app in your cluster.
```
oc create -f app.yaml
```
Verify your app is running.
```
oc get pods
```