disaster-recovery.yaml

# Disaster Recovery Example
# This example demonstrates:
# - Taking a backup before maintenance
# - Performing a full restore
# - Point-in-time recovery
# - Cross-region restore

---
# SCENARIO 1: Pre-maintenance Backup
---
# Take a manual backup before maintenance window
apiVersion: neo4j.neo4j.com/v1alpha1
kind: Neo4jBackup
metadata:
  name: pre-maintenance-backup
  namespace: neo4j-production
spec:
  target:
    kind: Cluster
    name: production-cluster
    namespace: neo4j-production

  # One-time backup (no schedule)
  storage:
    type: pvc
    pvc:
      name: maintenance-backup-pvc
      size: 50Gi
      storageClassName: fast-ssd

  options:
    backupType: "FULL"
    compress: true
    verifyBackup: true  # Verify integrity after backup

  # Add metadata for tracking
  metadata:
    labels:
      purpose: maintenance
      timestamp: "2025-01-20-1400"
    annotations:
      description: "Pre-maintenance backup before Neo4j upgrade"
      operator: "admin@company.com"
---
# SCENARIO 2: Full Database Restore
---
# Restore entire cluster from backup
apiVersion: neo4j.neo4j.com/v1alpha1
kind: Neo4jRestore
metadata:
  name: full-cluster-restore
  namespace: neo4j-production
spec:
  targetCluster: production-cluster

  # Restore all databases
  databaseName: "*"

  source:
    type: backup
    backupRef: pre-maintenance-backup

  # Force overwrite existing data
  force: true

  options:
    verifyBackup: true
    additionalArgs:
      - "--expand-commands"  # Show detailed progress

  # Pre-restore validation
  preRestoreHooks:
    - type: script
      script: |
        #!/bin/bash
        echo "Starting full cluster restore..."
        echo "Target cluster: production-cluster"
        echo "Source backup: pre-maintenance-backup"

        # Verify cluster is ready for restore
        kubectl get neo4jenterprisecluster production-cluster -n neo4j-production
---
# SCENARIO 3: Point-in-Time Recovery
---
# Restore to specific point in time
apiVersion: neo4j.neo4j.com/v1alpha1
kind: Neo4jRestore
metadata:
  name: pitr-restore
  namespace: neo4j-production
spec:
  targetCluster: production-cluster
  databaseName: maindb

  source:
    type: backup
    backupRef: daily-backup-20250119
    # Restore to specific timestamp
    pointInTime: "2025-01-19T14:30:00Z"

  # Create as new database to compare
  options:
    targetDatabaseName: "maindb_pitr_recovery"
    verifyBackup: true

  postRestoreHooks:
    - type: script
      script: |
        echo "PITR restore completed"
        echo "Original database: maindb"
        echo "Recovered database: maindb_pitr_recovery"
        echo "Point in time: 2025-01-19T14:30:00Z"
---
# SCENARIO 4: Cross-Region Disaster Recovery
---
# Source region backup configuration
apiVersion: neo4j.neo4j.com/v1alpha1
kind: Neo4jBackup
metadata:
  name: cross-region-backup
  namespace: neo4j-production
spec:
  target:
    kind: Cluster
    name: production-cluster
    namespace: neo4j-production

  # Backup to S3 with cross-region replication
  storage:
    type: s3
    bucket: neo4j-dr-backups
    path: "us-east-1/production"
    cloud:
      credentialsSecret: s3-dr-credentials
      region: us-east-1

  schedule:
    cron: "0 */4 * * *"  # Every 4 hours

  options:
    backupType: "AUTO"
    compress: true
    encryption:
      enabled: true
      kmsKeyId: "arn:aws:kms:us-east-1:123456789:key/xxxxx"
---
# Target region restore configuration
apiVersion: neo4j.neo4j.com/v1alpha1
kind: Neo4jRestore
metadata:
  name: cross-region-restore
  namespace: neo4j-dr
spec:
  # Restore to DR cluster in different region
  targetCluster: dr-cluster
  databaseName: "*"

  source:
    type: s3
    bucket: neo4j-dr-backups
    path: "us-east-1/production/backup-20250120-0800"
    cloud:
      credentialsSecret: s3-dr-credentials
      region: us-west-2  # DR region

  force: true

  options:
    verifyBackup: true
    parallelRestore: true  # Speed up large restores

  postRestoreHooks:
    - type: script
      script: |
        echo "Cross-region restore completed"
        echo "Source region: us-east-1"
        echo "Target region: us-west-2"

        # Verify databases are online
        kubectl exec -n neo4j-dr dr-cluster-0 -- cypher-shell \
          -u neo4j -p $NEO4J_PASSWORD \
          "SHOW DATABASES"
---
# SCENARIO 5: Selective Database Restore
---
# Restore only specific database
apiVersion: neo4j.neo4j.com/v1alpha1
kind: Neo4jRestore
metadata:
  name: selective-restore
  namespace: neo4j-production
spec:
  targetCluster: production-cluster

  # Restore only analytics database
  databaseName: analytics

  source:
    type: backup
    backupRef: weekly-full-backup-20250114

  # Don't overwrite, create as new
  force: false
  options:
    targetDatabaseName: "analytics_restored"

  # Validation queries after restore
  postRestoreHooks:
    - type: cypher
      database: analytics_restored
      queries:
        - "MATCH (n) RETURN count(n) as nodeCount"
        - "MATCH ()-[r]->() RETURN count(r) as relationshipCount"
---
# SCENARIO 6: Backup Validation Job
---
# Regular backup validation to ensure recoverability
apiVersion: batch/v1
kind: CronJob
metadata:
  name: backup-validation
  namespace: neo4j-production
spec:
  schedule: "0 6 * * 1"  # Weekly on Monday at 6 AM
  jobTemplate:
    spec:
      template:
        spec:
          containers:
            - name: validator
              image: neo4j/neo4j-admin:5.26.0-enterprise
              command:
                - /bin/bash
                - -c
                - |
                  # List recent backups
                  echo "=== Recent Backups ==="
                  kubectl get neo4jbackup -n neo4j-production

                  # Get latest backup
                  LATEST_BACKUP=$(kubectl get neo4jbackup -n neo4j-production \
                    -o jsonpath='{.items[?(@.status.phase=="Completed")].metadata.name}' | \
                    awk '{print $NF}')

                  echo "Latest backup: $LATEST_BACKUP"

                  # Create test restore
                  cat <<EOF | kubectl apply -f -
                  apiVersion: neo4j.neo4j.com/v1alpha1
                  kind: Neo4jRestore
                  metadata:
                    name: validation-restore-$(date +%s)
                    namespace: neo4j-production
                  spec:
                    targetCluster: test-restore-cluster
                    databaseName: "*"
                    source:
                      type: backup
                      backupRef: $LATEST_BACKUP
                    options:
                      verifyBackup: true
                      dryRun: true  # Validation only
                  EOF
          restartPolicy: OnFailure
---
# SCENARIO 7: Automated Failover Configuration
---
# Configure automatic failover with backup/restore
apiVersion: v1
kind: ConfigMap
metadata:
  name: failover-config
  namespace: neo4j-production
data:
  failover.sh: |
    #!/bin/bash
    # Automated failover script

    PRIMARY_CLUSTER="production-cluster"
    DR_CLUSTER="dr-cluster"
    NAMESPACE="neo4j-production"

    # Check primary cluster health
    PRIMARY_READY=$(kubectl get neo4jenterprisecluster $PRIMARY_CLUSTER \
      -n $NAMESPACE -o jsonpath='{.status.conditions[?(@.type=="Ready")].status}')

    if [ "$PRIMARY_READY" != "True" ]; then
      echo "Primary cluster unhealthy, initiating failover..."

      # Find latest backup
      LATEST_BACKUP=$(kubectl get neo4jbackup -n $NAMESPACE \
        --sort-by=.metadata.creationTimestamp \
        -o jsonpath='{.items[-1].metadata.name}')

      # Create restore to DR cluster
      kubectl apply -f - <<EOF
      apiVersion: neo4j.neo4j.com/v1alpha1
      kind: Neo4jRestore
      metadata:
        name: failover-restore-$(date +%s)
        namespace: $NAMESPACE
      spec:
        targetCluster: $DR_CLUSTER
        databaseName: "*"
        source:
          type: backup
          backupRef: $LATEST_BACKUP
        force: true
        options:
          priority: high
    EOF

      # Update DNS/Load balancer to point to DR
      # ... DNS update logic ...

      echo "Failover initiated to DR cluster"
    fi