Storage

In this guide, we provide an example of how to install SeaweedFS and integrating it into Mender.

export STORAGE_CLASS="default"
export STORAGE_BUCKET="replace-with-your-bucket-name"

cat >seaweedfs.yml <<EOF
filer:
  s3:
    enabled: true
    enableAuth: true
    createBuckets:
      - name: "${STORAGE_BUCKET}"
  storageClass: ${STORAGE_CLASS}

s3:
  enabled: true
  enableAuth: true
EOF

helm repo add seaweedfs https://seaweedfs.github.io/seaweedfs/helm
helm repo update
helm install seaweedfs --wait -f seaweedfs.yml  seaweedfs/seaweedfs

Finally, export the following environment variables, needed for installing Mender in the later steps:

export AWS_ACCESS_KEY_ID=$(kubectl get secret seaweedfs-s3-secret -o jsonpath='{.data.admin_access_key_id}' |base64 -d)
export AWS_SECRET_ACCESS_KEY=$(kubectl  get secret seaweedfs-s3-secret -o jsonpath='{.data.admin_secret_access_key}' |base64 -d)
export AWS_REGION="us-east-1"
export STORAGE_ENDPOINT="http://seaweedfs-s3:8333"

kubectl port-forward svc/seaweedfs-s3 8333:8333 &
aws s3 ls --endpoint-url="http://localhost:8333"

Please refer to the SeaweedFS official documentation for further information about setting up a SeaweedFS deployment.

Migration from MinIO

If you are currently using MinIO as the storage backend for Mender, you can migrate to SeaweedFS by following these steps:

# export data from MinIO
export AWS_ACCESS_KEY_ID="<your minio access key>"
export AWS_SECRET_ACCESS_KEY="<your minio secret key>"

mkdir mender-artifact-storage
# Tip: make sure you have enough disk space to store the data

aws s3 cp --recursive --endpoint-url=<your minio endpoint> s3://mender-artifact-storage/ ./mender-artifact-storage/

# import data to SeaweedFS
export AWS_ACCESS_KEY_ID=$(kubectl get secret seaweedfs-s3-secret -o jsonpath='{.data.admin_access_key_id}' |base64 -d)
export AWS_SECRET_ACCESS_KEY=$(kubectl  get secret seaweedfs-s3-secret -o jsonpath='{.data.admin_secret_access_key}' |base64 -d)

kubectl port-forward svc/seaweedfs-s3 8333:8333
aws s3 cp --recursive ./mender-artifact-storage/ --endpoint-url=http://localhost:8333 s3://mender-artifact-storage/

In this guide, we provide an example of how to use Amazon S3 as a Mender Artifact Storage.

Create a new bucket in Amazon S3, then a IAM user and its access key with the proper permissions to access the bucket.

You can find the required permissions in the Requirements section

Then, export the following environment variables:

export AWS_ACCESS_KEY_ID="replace-with-your-access-key-id"
export AWS_SECRET_ACCESS_KEY="replace-with-your-secret-access-key"
export AWS_REGION="replace-with-your-aws-region"
export STORAGE_BUCKET="replace-with-your-bucket-name"
export STORAGE_ENDPOINT="https://s3.${AWS_REGION}.amazonaws.com"

Alternatively, you can configure and use a Service Account to access the bucket.

export AWS_ACCESS_KEY_ID=""
export AWS_SECRET_ACCESS_KEY=""
export AWS_SERVICE_ACCOUNT_NAME="replace-with-your-service-account-name"
export AWS_REGION="replace-with-your-aws-region"
export STORAGE_BUCKET="replace-with-your-bucket-name"
export STORAGE_ENDPOINT="https://s3.${AWS_REGION}.amazonaws.com"

As a reference, here's your Helm Chart customization to make Mender use Amazon S3 as the storage backend:

global:
  s3:
    AWS_URI: "https://s3.<your-aws-region>.amazonaws.com"
    AWS_BUCKET: "<name-of-your-bucket>"
    AWS_REGION: "<your-aws-region>"
    AWS_ACCESS_KEY_ID: "<your-access-key-id>"
    AWS_SECRET_ACCESS_KEY: "<your-secret-access-key>"
    AWS_FORCE_PATH_STYLE: "false"

With a Service Account:

global:
  s3:
    AWS_SERVICE_ACCOUNT_NAME: "<service-account-name>"
    AWS_URI: "https://s3.<your-aws-region>.amazonaws.com"
    AWS_BUCKET: "<name-of-your-bucket>"
    AWS_REGION: "<your-aws-region>"
    AWS_FORCE_PATH_STYLE: "false"

In this guide, we provide an example of how to use Azure Blob Storage as a Mender Artifact storage.

Create a new Storage Account in Azure Blob Storage, then a Storage Container in the Storage Account.

As a reference, here's your Helm Chart customization to make Mender use Azure Blob Storage as the storage backend:

global:
  storage: "azure"
  azure:
    AUTH_CONNECTION_STRING: "BlobEndpoint=https://<name-of-your-storage>.blob.core.windows.net;SharedAccessSignature=..."
    CONTAINER_NAME: "<name-of-your-container>"

In this guide, we provide an example of how to use Cloudflare R2 Storage as a Mender Artifact storage.

Create a new Storage Bucket in Cloudflare R2, then create an API Token with Object Read & Write permission on the Bucket.

export AWS_ACCESS_KEY_ID="replace-with-your-access-key-id"
export AWS_SECRET_ACCESS_KEY="replace-with-your-secret-access-key"
export AWS_REGION="replace-with-your-aws-region"
export STORAGE_BUCKET="replace-with-your-bucket-name"
export STORAGE_ENDPOINT="https://replace-with-your-cloudflare-account-id.r2.cloudflarestorage.com"