Disk Snapshots¶

Core Concepts¶

Snapshots provide incremental, point-in-time backups of persistent disks. Understanding snapshot architecture is essential for designing cost-effective backup and recovery strategies.

Key Principle: Snapshots are incremental after the first full copy, making them ideal for frequent backups.

Incremental Architecture¶

How Snapshots Work¶

First Snapshot:

Full copy of all disk data
Stored in Cloud Storage
Time proportional to disk size
Baseline for future snapshots

Subsequent Snapshots:

Only changed blocks since previous snapshot
Significantly faster creation
Space-efficient storage
Automatic deduplication

Chain Management:

Google maintains snapshot chain automatically
Deleting middle snapshot is safe
Required blocks preserved
No user management needed

Storage Efficiency Example¶

Scenario: 1 TB disk, 5% daily change, 30-day retention

Traditional Full Backups:

30 copies × 1000 GB = 30,000 GB

Incremental Snapshots:

Day 1: 1000 GB (full)
Days 2-30: 50 GB each
Total: 1000 + (29 × 50) = 2,450 GB
92% storage savings

Snapshot Storage Locations¶

Regional Storage¶

Characteristics:

Stored in single region
Same cost as multi-regional
Faster restore (same region)
Region-specific availability

Use Cases:

Fast restore requirements
Region-bound compliance
Cost optimization (same-region egress free)
Non-critical data

Limitation: Single region failure risk

Multi-Regional Storage¶

Characteristics:

Stored across multiple regions
Geographic redundancy
Same cost as regional
Protected from regional failure

Use Cases:

Disaster recovery
Cross-region restore capability
Critical data protection
Compliance requirements

Trade-off: Slightly slower initial creation, cross-region restore incurs egress costs

Architecture Decision Matrix¶

Data Criticality	Restore Frequency	Storage Location	Rationale
Critical	Daily	Multi-regional	DR protection
Important	Weekly	Multi-regional	Balance cost/protection
Normal	Daily	Regional	Cost optimization
Development	On-demand	Regional	Lowest cost

Scheduled Snapshots¶

Snapshot Schedules¶

Benefits:

Automated backup lifecycle
Consistent backup timing
Retention policy enforcement
Reduced operational overhead
Guaranteed backup frequency

Schedule Types:

Hourly: Every N hours (critical data, 4-hour RPO)
Daily: Specific time (standard backups, 24-hour RPO)
Weekly: Specific day and time (less critical, 7-day RPO)

Retention Policies:

Automatic deletion of old snapshots
Maximum retention: ~7.5 years
Compliance-driven retention
Cost control

Design Patterns¶

Tiered Backup Strategy:

Hourly snapshots: Retain 24 hours (critical systems)
Daily snapshots: Retain 7 days (standard)
Weekly snapshots: Retain 4 weeks (long-term)
Monthly snapshots: Retain 12 months (compliance)

Pattern: Multiple schedules on same disk for different retention tiers

Snapshot Best Practices¶

Application-Consistent Snapshots¶

Problem: Snapshot at disk level, may capture inconsistent state

Solutions:

Quiesce filesystem: Flush pending writes
Database checkpoint: Ensure consistent state
Application-level backup: Use native tools first
–guest-flush flag: Windows VSS integration

Architecture Consideration: Snapshots are crash-consistent by default; application consistency requires coordination

Backup Windows¶

Considerations:

Schedule during low-usage periods
First snapshot has performance impact
Subsequent snapshots minimal impact
Background operation

Pattern: 2-4 AM local time for daily snapshots (minimize business impact)

Recovery Strategies¶

Point-in-Time Recovery¶

Capabilities:

Restore to any snapshot point
Multiple recovery points (based on schedule)
Granular recovery options
Test restores without affecting production

RTO Considerations:

Disk creation from snapshot: 10-30 minutes
VM recreation: 5-15 minutes
Application startup: Varies
Total RTO: 30-60 minutes typically

Partial Recovery¶

Options:

Restore single disk (not entire VM)
Attach restored disk as secondary
Copy needed files
Detach and delete disk

Use Case: Accidental deletion, specific file recovery

Cost Optimization¶

Snapshot Cost Structure¶

Storage Pricing:

~$0.026/GB/month (regional and multi-regional)
Incremental charges (only changed data)
No retrieval fees
No operation charges

Cost Factors:

Frequency of snapshots (more = more changed data)
Change rate (higher = more storage)
Retention period (longer = more total storage)
Number of disks

Optimization Strategies¶

Retention Policies:

Delete old snapshots automatically
Balance recovery needs vs cost
Different policies for different data tiers
Regular review and adjustment

Snapshot Consolidation:

Fewer snapshots = lower cost
Balance with RPO requirements
Use snapshot schedules for consistency

Right-Sizing Frequency:

Critical data: Frequent snapshots justified
Non-critical data: Less frequent acceptable
Test environments: Manual/infrequent

Disaster Recovery Architecture¶

Cross-Region DR¶

Pattern:

Snapshots in multi-regional storage
DR plan documents restore procedure
Network and firewall rules pre-configured in DR region
Regular DR testing (quarterly)
Runbooks and automation

RTO Targets:

Tier 1 (Critical): RTO 1 hour
Hourly snapshots, multi-regional
Automated restoration scripts
Pre-warmed network configuration
Tier 2 (Important): RTO 4 hours
Daily snapshots, multi-regional
Documented procedures
Tier 3 (Standard): RTO 24 hours
Daily snapshots, regional
Manual restoration acceptable

RPO Considerations¶

Snapshot Frequency = RPO:

Hourly snapshots: 1-hour RPO (max 1 hour data loss)
Daily snapshots: 24-hour RPO
Weekly snapshots: 7-day RPO

Architecture Decision: Balance frequency cost vs acceptable data loss

Snapshot Limits¶

Operational Limits¶

Limit	Value	Impact
Snapshots per disk	Unlimited	No practical limit
Snapshots per project	10,000	Request increase if needed
Snapshot creation rate	1 per 10 min per disk	Affects backup frequency
Snapshot size	64 TB	Matches max disk size

Architecture Implication: Large environments may need multiple projects for snapshot quotas

Comparison: Snapshots vs Machine Images¶

When to Use Snapshots¶

Advantages:

Incremental (cost-effective)
Frequent backups (daily/hourly)
Per-disk granularity
Fast creation (after first)
Low storage cost

Use Cases:

Regular data protection
Point-in-time recovery
Database backups
Frequent backup requirements
Cost-sensitive environments

When to Use Machine Images¶

Advantages:

Complete VM configuration
All disks in one operation
Fast full restoration
VM migration

Use Cases:

Weekly/monthly full backups
VM cloning and migration
Disaster recovery base
Configuration preservation

Architecture Pattern: Use both - snapshots for frequent backups, machine images for complete system backups

Security Considerations¶

Encryption¶

Default Encryption:

All snapshots encrypted at rest
Google-managed keys
Transparent operation
No configuration needed

Customer-Managed Keys (CMEK):

Control key lifecycle
Regulatory compliance
Key rotation policies
Cloud KMS integration

Trade-offs:

CMEK: More control, more complexity
Google-managed: Simpler, less control
Performance impact minimal

Access Control¶

IAM Roles:

roles/compute.storageAdmin: Create/delete snapshots
roles/compute.viewer: View snapshots
Principle of least privilege
Separate snapshot admin from VM admin

Snapshot Sharing:

Cross-project via IAM
Service account access
Controlled sharing
Audit access

Compliance and Retention¶

Regulatory Requirements¶

Common Requirements:

HIPAA: 6-year retention, encryption, audit logs
SOX: 7-year retention, access controls
GDPR: Data location restrictions, deletion capability
PCI-DSS: Encryption, access controls, retention

Implementation:

Snapshot schedules with appropriate retention
Storage location selection (region/multi-region)
CMEK for enhanced security
Audit logging enabled

Retention Strategy¶

Pattern:

Short-term: Daily snapshots, 7-30 day retention (operational recovery)
Long-term: Monthly snapshots, multi-year retention (compliance)
Archive: Consider exporting to Cloud Storage for very long-term

Monitoring and Alerting¶

Key Metrics¶

Monitor:

Snapshot creation success/failure rate
Time to create snapshots
Storage costs trending
Age of oldest snapshot
Recovery testing results

Alerts:

Snapshot creation failures
Storage quota approaching limits
Unexpected cost increases
Missing scheduled snapshots
Compliance violations (retention)

Exam Focus Areas¶

Design Decisions¶

Snapshot vs machine image trade-offs
Frequency and retention policies
Storage location selection (regional vs multi-regional)
Cost optimization strategies

Backup Architecture¶

Tiered backup strategies
RPO/RTO planning
Application-consistent backup methods
Cross-region DR design

Cost Management¶

Incremental snapshot cost model
Retention policy optimization
Storage location cost impact
Snapshot cleanup strategies

Operations¶

Scheduled snapshots configuration
Automated lifecycle management
Monitoring and alerting
Testing and validation

Compliance¶

Retention requirements
Encryption options (default vs CMEK)
Access control design
Audit logging