Disaster recovery

• Disaster recovery overview
  • Replica is unhealthy but cluster maintains quorum
  • The majority of replicas are unhealthy
  • All replicas are unhealthy
• Repair a single replica
  • Split-brain scenario
  • Remove replicas
  • Join replicas
  • Where to go next
• Repair a cluster
  • Diagnose an unhealthy cluster
  • Perform an emergency repair
  • Where to go next
• Create a backup
  • Data managed by MSR
  • Back up MSR data
    • Find your replica ID
      • MKE web interface
      • MKE client bundle
      • SSH access
    • Back up image content
    • Back up MSR metadata
    • MKE field prompts
    • Test your backups
• Restore from backup
  • Restore MSR data
    • Remove MSR containers
    • Restore images
    • Restore MSR metadata
      • DTR 2.5 and below
      • DTR 2.6.0-2.6.3 (with experimental online garbage collection)
    • Re-fetch the vulnerability database
  • Where to go next