Persistent Disk - Block Storage for VMs

Complete guide to Persistent Disk - durable block storage for Compute Engine.

---

📋 Table of Contents

1. Introduction
2. Disk Types
3. Disk Management
4. Snapshots
5. Performance
6. High Availability
7. Encryption
8. Cost Optimization
9. Best Practices

---

Introduction

Persistent Disk provides durable, high-performance block storage for Compute Engine VMs.

Key Features

✅ Durable and reliable (99.9999% durability)
✅ Multiple disk types
✅ Up to 64 TB per disk
✅ Snapshots for backup
✅ Encryption at rest
✅ Regional persistent disks
✅ Resize without downtime
✅ Independent of VM lifecycle
✅ Automatic encryption
✅ High IOPS and throughput

Architecture

┌─────────────────────────────────────┐
│    Compute Engine Instance          │
├─────────────────────────────────────┤
│  ┌──────────────────────────────┐   │
│  │  Boot Disk (pd-balanced)     │   │
│  │  - OS and system files       │   │
│  │  - 20 GB                     │   │
│  └──────────────────────────────┘   │
│  ┌──────────────────────────────┐   │
│  │  Data Disk (pd-ssd)          │   │
│  │  - Application data          │   │
│  │  - 500 GB                    │   │
│  └──────────────────────────────┘   │
│  ┌──────────────────────────────┐   │
│  │  Local SSD (temporary)       │   │
│  │  - Cache/temp data           │   │
│  │  - 375 GB                    │   │
│  └──────────────────────────────┘   │
└─────────────────────────────────────┘

---

Disk Types

Standard Persistent Disk (pd-standard)

HDD-backed storage

Characteristics:
- Sequential I/O optimized
- Lower cost
- Good for throughput
- Up to 64 TB

Performance: - Read IOPS: 0.75 per GB (max 7,500) - Write IOPS: 1.5 per GB (max 15,000) - Throughput: 120-240 MB/s

Pricing: $0.040/GB/month

Use Cases: - Batch processing - Sequential workloads - Cold storage - Backup targets

Balanced Persistent Disk (pd-balanced)

SSD-backed storage with balanced performance

Characteristics:
- Balance of performance and cost
- Good for most workloads
- Recommended default
- Up to 64 TB

Performance: - Read/Write IOPS: 6 per GB (max 80,000) - Throughput: 240-480 MB/s

Pricing: $0.100/GB/month

Use Cases: - General purpose workloads - Boot disks - Development environments - Most applications

SSD Persistent Disk (pd-ssd)

High-performance SSD storage

Characteristics:
- High IOPS
- Low latency
- Consistent performance
- Up to 64 TB

Performance: - Read/Write IOPS: 30 per GB (max 100,000) - Throughput: 480-960 MB/s

Pricing: $0.170/GB/month

Use Cases: - Databases - High-performance applications - Analytics workloads - Latency-sensitive apps

Extreme Persistent Disk (pd-extreme)

Highest performance SSD storage

Characteristics:
- Configurable IOPS
- Highest performance
- Premium pricing
- Up to 64 TB

Performance: - IOPS: Configurable up to 160,000 - Throughput: Up to 4,800 MB/s

Pricing: $0.125/GB/month + $0.065/provisioned IOPS

Use Cases: - High-end databases - SAP HANA - Mission-critical applications - Extreme performance needs

Local SSD

Ephemeral high-performance storage

Characteristics:
- Physically attached to server
- Very high performance
- Data lost on VM stop
- 375 GB per device

Performance: - Read IOPS: 680,000 - Write IOPS: 360,000 - Throughput: 2,400-9,600 MB/s

Pricing: $0.080/GB/month

Use Cases: - Caching - Temporary data - Scratch space - High-performance computing

Disk Type Comparison

Type	IOPS/GB	Max IOPS	Throughput	Cost/GB	Use Case
pd-standard	0.75-1.5	15,000	240 MB/s	$0.040	Sequential
pd-balanced	6	80,000	480 MB/s	$0.100	General
pd-ssd	30	100,000	960 MB/s	$0.170	High perf
pd-extreme	Custom	160,000	4,800 MB/s	$0.125+	Extreme
Local SSD	N/A	680,000	9,600 MB/s	$0.080	Temporary

---

Disk Management

Create Disk

# Create standard disk
gcloud compute disks create my-disk \
  --size=100GB \
  --type=pd-standard \
  --zone=us-central1-a

# Create SSD disk
gcloud compute disks create my-ssd-disk \
  --size=500GB \
  --type=pd-ssd \
  --zone=us-central1-a

# Create from snapshot
gcloud compute disks create restored-disk \
  --source-snapshot=my-snapshot \
  --zone=us-central1-a

# Create from image
gcloud compute disks create boot-disk \
  --image-family=debian-11 \
  --image-project=debian-cloud \
  --zone=us-central1-a

Attach/Detach Disk

# Attach disk to VM
gcloud compute instances attach-disk my-vm \
  --disk=my-disk \
  --zone=us-central1-a

# Attach in read-only mode
gcloud compute instances attach-disk my-vm \
  --disk=my-disk \
  --mode=ro \
  --zone=us-central1-a

# Detach disk
gcloud compute instances detach-disk my-vm \
  --disk=my-disk \
  --zone=us-central1-a

Format and Mount

# Format disk (on VM)
sudo mkfs.ext4 -m 0 -E lazy_itable_init=0,lazy_journal_init=0,discard /dev/sdb

# Create mount point
sudo mkdir -p /mnt/disks/data

# Mount disk
sudo mount -o discard,defaults /dev/sdb /mnt/disks/data

# Add to /etc/fstab for automatic mounting
echo UUID=$(sudo blkid -s UUID -o value /dev/sdb) /mnt/disks/data ext4 discard,defaults,nofail 0 2 | sudo tee -a /etc/fstab

Resize Disk

# Resize disk
gcloud compute disks resize my-disk \
  --size=200GB \
  --zone=us-central1-a

# Resize filesystem (on VM)
sudo resize2fs /dev/sdb

Terraform Example

resource "google_compute_disk" "default" {
  name  = "my-disk"
  type  = "pd-ssd"
  zone  = "us-central1-a"
  size  = 500

  labels = {
    environment = "production"
    team        = "backend"
  }

  physical_block_size_bytes = 4096
}

resource "google_compute_attached_disk" "default" {
  disk     = google_compute_disk.default.id
  instance = google_compute_instance.default.id
}

---

Snapshots

Create Snapshots

# Create snapshot
gcloud compute disks snapshot my-disk \
  --snapshot-names=my-snapshot \
  --zone=us-central1-a

# Create snapshot with description
gcloud compute disks snapshot my-disk \
  --snapshot-names=my-snapshot \
  --description="Daily backup" \
  --zone=us-central1-a

# Create snapshot with labels
gcloud compute disks snapshot my-disk \
  --snapshot-names=my-snapshot \
  --labels=env=prod,type=backup \
  --zone=us-central1-a

Snapshot Schedule

# Create snapshot schedule
gcloud compute resource-policies create snapshot-schedule daily-backup \
  --region=us-central1 \
  --max-retention-days=7 \
  --on-source-disk-delete=keep-auto-snapshots \
  --daily-schedule \
  --start-time=02:00

# Attach schedule to disk
gcloud compute disks add-resource-policies my-disk \
  --resource-policies=daily-backup \
  --zone=us-central1-a

Restore from Snapshot

# Create disk from snapshot
gcloud compute disks create restored-disk \
  --source-snapshot=my-snapshot \
  --zone=us-central1-a

# Create disk in different zone
gcloud compute disks create restored-disk \
  --source-snapshot=my-snapshot \
  --zone=us-west1-a

Snapshot Management

# List snapshots
gcloud compute snapshots list

# Delete snapshot
gcloud compute snapshots delete my-snapshot

# Copy snapshot to different region
gcloud compute snapshots create my-snapshot-copy \
  --source-snapshot=my-snapshot \
  --storage-location=us-west1

---

Performance

IOPS Calculation

pd-balanced:

IOPS = min(disk_size_gb × 6, 80000)

Example: 500 GB disk
IOPS = min(500 × 6, 80000) = 3,000 IOPS

pd-ssd:

IOPS = min(disk_size_gb × 30, 100000)

Example: 500 GB disk
IOPS = min(500 × 30, 100000) = 15,000 IOPS

Throughput Calculation

Per VM limits: - Read: 1,200 MB/s (n2, n2d, c2, c2d) - Write: 800 MB/s (n2, n2d, c2, c2d)

Per disk limits: - pd-standard: 240 MB/s - pd-balanced: 480 MB/s - pd-ssd: 960 MB/s - pd-extreme: 4,800 MB/s

Performance Optimization

# Use SSD for high IOPS
gcloud compute disks create high-iops-disk \
  --size=1000GB \
  --type=pd-ssd \
  --zone=us-central1-a

# Use larger disk for more IOPS
gcloud compute disks create large-disk \
  --size=2000GB \
  --type=pd-balanced \
  --zone=us-central1-a

# Use Local SSD for extreme performance
gcloud compute instances create high-perf-vm \
  --zone=us-central1-a \
  --machine-type=n2-standard-8 \
  --local-ssd=interface=NVME \
  --local-ssd=interface=NVME

Performance Tips

✅ Use SSD for databases
✅ Size disk for required IOPS
✅ Use Local SSD for temporary high-performance needs
✅ Use pd-extreme for highest performance
✅ Monitor disk performance
✅ Use appropriate machine type
✅ Enable discard/TRIM
✅ Use multiple disks for higher throughput

---

High Availability

Regional Persistent Disks

# Create regional disk
gcloud compute disks create regional-disk \
  --size=500GB \
  --type=pd-ssd \
  --region=us-central1 \
  --replica-zones=us-central1-a,us-central1-b

# Attach to VM
gcloud compute instances attach-disk my-vm \
  --disk=regional-disk \
  --zone=us-central1-a

Benefits: - Synchronous replication across zones - Automatic failover - 99.99% availability SLA - No data loss on zone failure

Limitations: - 2x cost - Slightly higher latency - Limited to 2 zones in same region

Backup Strategy

# Automated snapshots
gcloud compute resource-policies create snapshot-schedule backup-policy \
  --region=us-central1 \
  --max-retention-days=30 \
  --on-source-disk-delete=keep-auto-snapshots \
  --daily-schedule \
  --start-time=02:00 \
  --storage-location=us

# Apply to disk
gcloud compute disks add-resource-policies my-disk \
  --resource-policies=backup-policy \
  --zone=us-central1-a

---

Encryption

Google-Managed Encryption

• Default encryption
• No configuration needed
• Automatic key rotation

Customer-Managed Encryption Keys (CMEK)

# Create key
gcloud kms keyrings create my-keyring --location=us-central1
gcloud kms keys create my-key \
  --location=us-central1 \
  --keyring=my-keyring \
  --purpose=encryption

# Create disk with CMEK
gcloud compute disks create encrypted-disk \
  --size=100GB \
  --type=pd-ssd \
  --zone=us-central1-a \
  --kms-key=projects/PROJECT/locations/us-central1/keyRings/my-keyring/cryptoKeys/my-key

Customer-Supplied Encryption Keys (CSEK)

# Generate key
openssl rand -base64 32 > disk-key.txt

# Create disk with CSEK
gcloud compute disks create csek-disk \
  --size=100GB \
  --csek-key-file=disk-key.txt \
  --zone=us-central1-a

---

Cost Optimization

Pricing Comparison

Monthly cost for 1 TB: - pd-standard: $40 - pd-balanced: $100 - pd-ssd: $170 - pd-extreme: $125 + IOPS cost - Local SSD: $80

Optimization Strategies

✅ Use pd-balanced instead of pd-ssd when possible
✅ Right-size disk capacity
✅ Delete unused disks
✅ Use snapshots for backups (cheaper than disks)
✅ Use committed use discounts
✅ Use Local SSD for temporary data
✅ Monitor disk utilization
✅ Use lifecycle policies for snapshots

Cost Example

Scenario: Database with 500 GB storage

Option 1: pd-ssd
- Cost: 500 GB × $0.170 = $85/month

Option 2: pd-balanced
- Cost: 500 GB × $0.100 = $50/month
- Savings: 41%

Option 3: pd-standard
- Cost: 500 GB × $0.040 = $20/month
- Savings: 76% (if performance acceptable)

---

Best Practices

Performance

✅ Use SSD for databases and high-IOPS workloads
✅ Size disks for required IOPS
✅ Use Local SSD for temporary high-performance needs
✅ Monitor disk performance metrics
✅ Use multiple disks for higher throughput
✅ Enable discard/TRIM
✅ Use appropriate machine type

Reliability

✅ Use regional disks for critical data
✅ Implement automated snapshots
✅ Test restore procedures
✅ Use snapshot schedules
✅ Keep snapshots in multiple regions
✅ Monitor disk health
✅ Plan for disaster recovery

Security

✅ Use CMEK for sensitive data
✅ Enable audit logging
✅ Use IAM for access control
✅ Regular security audits
✅ Encrypt snapshots
✅ Use VPC Service Controls

Cost Management

✅ Right-size disk capacity
✅ Delete unused disks
✅ Use appropriate disk type
✅ Use snapshots for backups
✅ Monitor costs
✅ Use committed use discounts
✅ Implement lifecycle policies

---

Troubleshooting

Disk full:

# Check disk usage
df -h

# Resize disk
gcloud compute disks resize my-disk --size=200GB --zone=us-central1-a
sudo resize2fs /dev/sdb

Performance issues:

# Check IOPS
iostat -x 1

# Monitor disk metrics
gcloud monitoring time-series list \
  --filter='metric.type="compute.googleapis.com/instance/disk/read_ops_count"'

Cannot attach disk: - Check if disk is already attached - Verify zone matches VM - Check IAM permissions - Verify disk exists

---

Next Steps

• Filestore - Managed NFS file storage
• Storage Comparison - Detailed comparison
• Best Practices - Production guidelines

---

Last Updated: March 2026
Status: ✅ Complete