DRPolicy CRD

DRPolicy CRD

Overview

The DRPolicy custom resource defines the disaster recovery topology and replication configuration between peer clusters. It is a cluster-scoped resource created by administrators on the OCM hub cluster that establishes the DR relationship between two managed clusters.

A DRPolicy specifies:

  • Which two clusters participate in DR
  • The replication schedule for Async (Regional DR)
  • Storage class selectors for volume replication
  • Whether to use async (Regional DR) or sync (Metro DR)

Lifecycle: DRPolicy resources are created once during DR setup and are typically long-lived.

API Group and Version

  • API Group: ramendr.openshift.io
  • API Version: v1alpha1
  • Kind: DRPolicy
  • Scope: Cluster

Spec Fields

Required Fields

drClusters ([]string)

List of exactly two DRCluster resource names that participate in this DR policy.

Requirements:

  • Must contain exactly 2 DRCluster resource names
  • DRCluster resources must exist on the hub cluster (see DRCluster )
  • Immutable after creation

Example:

drClusters:
  - east-cluster
  - west-cluster

Optional Fields

schedulingInterval (string)

Defines how frequently to replicate volume data between clusters for Async (Regional DR).

Format: <number><m|h|d> where:

  • m = minutes
  • h = hours
  • d = days

Behavior:

  • Empty string ("") = Sync (Metro DR)
  • Non-empty = Async (Regional DR)
  • Immutable after creation

Examples:

schedulingInterval: "5m"   # Every 5 minutes
schedulingInterval: "1h"   # Every hour
schedulingInterval: "12h"  # Every 12 hours
schedulingInterval: ""     # Synchronous (no scheduling)

Typical values:

  • Production Async (Regional DR): "10m" or "30m"
  • Testing: "5m"
  • Sync (Metro DR): "" (empty)

replicationClassSelector (metav1.LabelSelector)

Label selector to identify VolumeReplicationClass resources for Async (Regional DR).

When to use: Set this for Async (Regional DR) when using VolumeReplication with schedulingInterval and there are multiple VolumeReplicationClass that match the requirements.

How it works:

  • VRG uses this selector to find the appropriate VolumeReplicationClass for each PVC
  • VolumeReplicationClass must exist on managed clusters with matching labels

Example:

replicationClassSelector:
  matchLabels:
    ramendr.openshift.io/replication-class: rbd-replication

Immutable: Cannot be changed after creation (presence/absence only).

volumeSnapshotClassSelector (metav1.LabelSelector)

Label selector to identify VolumeSnapshotClass resources for Async (Regional DR).

When to use: Set this for Async (Regional DR) when using VolSync with schedulingInterval and there are multiple VolumeSnapshotClass that match the requirements.

How it works:

  • VRG uses this selector to find the appropriate VolumeSnapshotClass for snapshots
  • VolumeSnapshotClass must exist on managed clusters with matching labels

Example:

volumeSnapshotClassSelector:
  matchLabels:
    ramendr.openshift.io/snapshot-class: csi-snapclass

Immutable: Cannot be changed after creation (presence/absence only).

volumeGroupSnapshotClassSelector (metav1.LabelSelector)

Label selector to identify VolumeGroupSnapshotClass resources for grouped snapshot operations with VolSync in Async (Regional DR).

When to use: For Async (Regional DR) workloads with multiple PVCs that need consistent point-in-time snapshots and there are multiple VolumeGroupSnapshotClass that match the requirements.

Example:

volumeGroupSnapshotClassSelector:
  matchLabels:
    ramendr.openshift.io/vg-snapshot-class: csi-vg-snapclass

Status Fields

The DRPolicy status is automatically populated by the Ramen hub operator.

conditions ([]metav1.Condition)

Standard Kubernetes conditions indicating the policy validation state.

Common condition types:

  • Validated - DRPolicy has been validated successfully

async (Async)

Status of async (Regional DR) details between clusters.

Fields:

  • peerClasses ([]PeerClass) - List of common StorageClasses with async relationships

sync (Sync)

Status of sync (Metro DR) details between clusters.

Fields:

  • peerClasses ([]PeerClass) - List of common StorageClasses with sync relationships

PeerClass Structure

Discovered peer relationship information between peer clusters:

  • storageClassName - Common StorageClass name across peers
  • replicationID - Common replication ID from VolumeReplicationClass labels
  • groupReplicationID - Group replication ID for volume groups
  • storageID - Storage instance IDs (singleton if shared, distinct if separate)
  • clusterIDs - The two cluster IDs (kube-system namespace UIDs)
  • grouping - Whether PVCs can be grouped for replication
  • offloaded - Whether replication is managed externally (not by VRG)

Examples

Example 1: Async (Regional DR)

This policy configures Async (Regional DR) between geographically distributed clusters:

apiVersion: ramendr.openshift.io/v1alpha1
kind: DRPolicy
metadata:
  name: regional-dr-policy
spec:
  # Replicate every hour
  schedulingInterval: "1h"

  # Two clusters in different regions
  drClusters:
    - us-east-cluster
    - us-west-cluster

  # Select VolumeReplicationClass
  replicationClassSelector:
    matchLabels:
      ramendr.openshift.io/replication-class: rbd-replication

Example 2: Sync (Metro DR)

This policy configures Sync (Metro DR) between clusters in the same metro area:

apiVersion: ramendr.openshift.io/v1alpha1
kind: DRPolicy
metadata:
  name: metro-dr-policy
spec:
  # Empty scheduling interval = synchronous
  schedulingInterval: ""

  # Two clusters in the same metro
  drClusters:
    - metro-cluster-1
    - metro-cluster-2

Note: Sync (Metro DR) uses synchronous replication at the storage level. VolSync is used for Async (Regional DR) and should not be used with Sync (Metro DR).

Troubleshooting

DRPolicy Not Validated

Check the DRPolicy status:

kubectl get drpolicy <name> -o yaml

Common issues:

  1. Clusters not registered in OCM

    Verify managed clusters are registered:

    kubectl get managedclusters

  2. DRCluster resources not validated

    Verify DRCluster resources referenced in the DRPolicy are validated:

    kubectl get drcluster <cluster-name> -o yaml

    Check for Validated condition in status:

    status:
  conditions:
    - type: Validated
      status: "True"

No PeerClasses in Status

Cause: Classes available on managed clusters may not have required labels or annotations, or replication isn’t configured.

For details on required labels and annotations, see DRClusterConfig .