Plugins for Google Cloud Platform (GCP)

Overview

This repository contains these plugins to support running Velero on GCP:

• An object store plugin for persisting and retrieving backups and restores on Google Cloud Storage. Content of backup includes log files, warning/error files, CSI related resources list, Velero native snapshots list, Velero PodVolumeBackup list, k8s resources list, k8s resources YAMLs and files created by uploader (Restic and Kopia). Content of restore includes restore logs, warning/error files and coming k8s resources list.

• A volume snapshotter plugin for creating snapshots from volumes (during a backup) and volumes from snapshots (during a restore) on Google Compute Engine Disks.

  • Since v1.4.0, the snapshotter plugin can handle the volumes provisioned by CSI driver pd.csi.storage.gke.io.

You can run Kubernetes on Google Cloud Platform in either:

• Kubernetes on Google Compute Engine virtual machines
• Google Kubernetes Engine

For common use-cases, take a look at the Examples page.

Compatibility

Below is a listing of plugin versions and respective Velero versions that are compatible.

Plugin Version	Velero Version
v1.7.x	v1.11.x
v1.6.x	v1.10.x
v1.5.x	v1.9.x
v1.4.x	v1.8.x
v1.3.x	v1.7.x

Filing issues

If you would like to file a GitHub issue for the plugin, please open the issue on the core Velero repo

Setup

To set up Velero on GCP, you:

• Create an GCS bucket
• Set permissions for Velero
• Install and start Velero

You can also use this plugin to create an additional Backup Storage Location.

If you do not have the gcloud and gsutil CLIs locally installed, follow the user guide to set them up.

Create an GCS bucket

Velero requires an object storage bucket in which to store backups, preferably unique to a single Kubernetes cluster (see the FAQ for more details). Create a GCS bucket, replacing the <YOUR_BUCKET> placeholder with the name of your bucket:

BUCKET=<YOUR_BUCKET>

gsutil mb gs://$BUCKET/

Set permissions for Velero

If you run Google Kubernetes Engine (GKE), make sure that your current IAM user is a cluster-admin. This role is required to create RBAC objects. See the GKE documentation for more information.

Create Google Service Account (GSA):

To integrate Velero with GCP, create a Velero-specific Service Account:

1. View your current config settings:

   gcloud config list

   Store the project value from the results in the environment variable $PROJECT_ID.

   PROJECT_ID=$(gcloud config get-value project)

2. Create a service account:

   GSA_NAME=velero
   gcloud iam service-accounts create $GSA_NAME \
       --display-name "Velero service account"

   If you'll be using Velero to backup multiple clusters with multiple GCS buckets, it may be desirable to create a unique username per cluster rather than the default velero.

   Then list all accounts and find the velero account you just created:

   gcloud iam service-accounts list

   Set the $SERVICE_ACCOUNT_EMAIL variable to match its email value.

   SERVICE_ACCOUNT_EMAIL=$(gcloud iam service-accounts list \
     --filter="displayName:Velero service account" \
     --format 'value(email)')

Create Custom Role with Permissions for the Velero GSA:

These permissions are required by Velero to manage snapshot resources in the GCP Project.

```bash
ROLE_PERMISSIONS=(
    compute.disks.get
    compute.disks.create
    compute.disks.createSnapshot
    compute.snapshots.get
    compute.snapshots.create
    compute.snapshots.useReadOnly
    compute.snapshots.delete
    compute.zones.get
    storage.objects.create
    storage.objects.delete
    storage.objects.get
    storage.objects.list
    iam.serviceAccounts.signBlob
)

gcloud iam roles create velero.server \
    --project $PROJECT_ID \
    --title "Velero Server" \
    --permissions "$(IFS=","; echo "${ROLE_PERMISSIONS[*]}")"

gcloud projects add-iam-policy-binding $PROJECT_ID \
    --member serviceAccount:$SERVICE_ACCOUNT_EMAIL \
    --role projects/$PROJECT_ID/roles/velero.server

gsutil iam ch serviceAccount:$SERVICE_ACCOUNT_EMAIL:objectAdmin gs://${BUCKET}
```

Note: iam.serviceAccounts.signBlob permission is used to allow Velero's Kubernetes Service Account to create signed urls for the GCS bucket. This is required if you want to run velero backup logs, velero backup download, velero backup describe and velero restore describe. This is due to those commands need to download some metadata files from S3 bucket to display information needed, and the Velero server has access to GCS but the CLI does not.

Grant access to Velero

This can be done in 2 different options.

Option 1: Using Service Account Key

This involves creating a Google Service Account Key and using it as --secret-file during installation.

1. Create a service account key, specifying an output file (credentials-velero) in your local directory:

   gcloud iam service-accounts keys create credentials-velero \
       --iam-account $SERVICE_ACCOUNT_EMAIL

Note that Google Service Account keys are valid for decades (no clear expiry date) - so store it securely or rotate them as often as possible or both.

Option 2: Using Workload Identity

This requires a GKE cluster with workload identity enabled.

1. Create Velero Namespace This is required because Kuberenetes Service Account (step 2) resides in a namespace

   NAMESPACE=velero
   kubectl create namespace $NAMESPACE

2. Create Kubernetes Service Account This is required when binding to the Google Service Account. Namespace is already created in step 1 above.

   KSA_NAME=velero
   kubectl create serviceaccount $KSA_NAME --namespace $NAMESPACE

3. Add IAM Policy Binding for Velero's Kubernetes service account to a GCP service account

   gcloud iam service-accounts add-iam-policy-binding \
       --role roles/iam.workloadIdentityUser \
       --member "serviceAccount:[$PROJECT_ID].svc.id.goog[$NAMESPACE/$KSA_NAME]" \
       [$GSA_NAME]@[$PROJECT_ID].iam.gserviceaccount.com

In this case:

• [$NAMESPACE/$KSA_NAME] are Kubernetes Namespace and Service Account created in step 1 and 2.
• PROJECT_ID is the Google Project ID - Step 1 and
• GSA_NAME is the name of the Google Service Account - Step 2.

For more information on configuring workload identity on GKE, look at the official GCP documentation for more details.

Install and start Velero

Download Velero

Install Velero, including all prerequisites, into the cluster and start the deployment. This will create a namespace called velero, and place a deployment named velero in it.

If using a Google Service Account Key:

velero install \
    --provider gcp \
    --plugins velero/velero-plugin-for-gcp:v1.6.0 \
    --bucket $BUCKET \
    --secret-file ./credentials-velero

If using Workload Identity:

You must add a service account annotation to the Kubernetes service account so that it will know which GCP service account to use. You can do this during installation with --sa-annotations. Use the flag --no-secret so that Velero will know not to look for a key file. You must also add the GCP service account name in --backup-location-config.

velero install \
    --provider gcp \
    --plugins velero/velero-plugin-for-gcp:v1.6.0 \
    --bucket $BUCKET \
    --no-secret \
    --sa-annotations iam.gke.io/gcp-service-account=[$GSA_NAME]@[$PROJECT_ID].iam.gserviceaccount.com \
    --backup-location-config serviceAccount=[$GSA_NAME]@[$PROJECT_ID].iam.gserviceaccount.com \

Additionally, you can specify --use-node-agent to enable node agent support, and --wait to wait for the deployment to be ready.

(Optional) Specify additional configurable parameters for the --backup-location-config flag.

(Optional) Specify additional configurable parameters for the --snapshot-location-config flag.

(Optional) Customize the Velero installation further to meet your needs.

For more complex installation needs, use either the Helm chart, or add --dry-run -o yaml options for generating the YAML representation for the installation.

Create an additional Backup Storage Location

If you are using Velero v1.6.0 or later, you can create additional GCP Backup Storage Locations that use their own credentials. These can also be created alongside Backup Storage Locations that use other providers.

Limitations

It is not possible to use different credentials for additional Backup Storage Locations if you are pod based authentication such as Workload Identity.

Prerequisites

• Velero 1.6.0 or later
• GCP plugin must be installed, either at install time, or by running velero plugin add velero/velero-plugin-for-gcp:plugin-version, replace the plugin-version with the corresponding value

Configure GCS bucket and credentials

To configure a new Backup Storage Location with its own credentials, it is necessary to follow the steps above to create the bucket to use and to generate the credentials file to interact with that bucket. Once you have created the credentials file, create a Kubernetes Secret in the Velero namespace that contains these credentials:

kubectl create secret generic -n velero bsl-credentials --from-file=gcp=</path/to/credentialsfile>

This will create a secret named bsl-credentials with a single key (gcp) which contains the contents of your credentials file. The name and key of this secret will be given to Velero when creating the Backup Storage Location, so it knows which secret data to use.

Create Backup Storage Location

Once the bucket and credentials have been configured, these can be used to create the new Backup Storage Location:

velero backup-location create <bsl-name> \
  --provider gcp \
  --bucket $BUCKET \
  --credential=bsl-credentials=gcp

The Backup Storage Location is ready to use when it has the phase Available. You can check this with the following command:

velero backup-location get

To use this new Backup Storage Location when performing a backup, use the flag --storage-location <bsl-name> when running velero backup create.