HPE 3PAR Storage Architecture and Administration

3 Nov 2026, 17:5319 min readby ITYantra Admin

Deep dive into HPE 3PAR storage architecture, administration, CPG, VVsets, thin provisioning, deduplication, and replication.

# HPE 3PAR Storage Architecture and Administration

Introduction

HPE 3PAR StoreServ is HPE's mid-range to enterprise storage array, designed for mission-critical workloads requiring high performance, scalability, and advanced data services. The 3PAR architecture uses a unique ASIC-based design with node pairs, chunklets, and Common Provisioning Groups (CPG) to deliver predictable performance and high availability.

This guide provides a comprehensive deep dive into 3PAR architecture and administration.

You will learn:

  • 3PAR hardware architecture and node pairs
  • Chunklet-based storage layout
  • Common Provisioning Groups (CPG)
  • Virtual volumes (VV) and VVsets
  • Thin provisioning and deduplication
  • QoS and priority optimization
  • Remote copy and snapshots
  • CLI and management interface

Technical Background

3PAR Hardware Architecture

**Node Pair Architecture:**

  • 3PAR arrays consist of pairs of nodes (controllers)
  • Each node pair acts as a single system
  • Multiple node pairs can be added for scale-out
  • Active-active across all nodes
  • Cache is distributed across all nodes

**Node Components:**

  • Processor (Intel Xeon)
  • Memory (cache)
  • Custom ASICs (Gen5 ASIC in current models)
  • Front-end ports (FC, iSCSI, NVMe/TCP)
  • Back-end ports (SAS to drive enclosures)
  • Cluster interconnect between nodes

**Drive Enclosures:**

  • 2U, 24/25 drive enclosures
  • 12 Gb/s SAS backend
  • Drive types: SAS, SATA, NVMe (flash)

Chunklet-Based Architecture

3PAR's unique architecture breaks drives into 256 MiB chunklets:

**Chunklet Distribution:**

  • Each physical drive is divided into multiple chunklets
  • Chunklets are distributed across all drives in a CPG
  • RAID is performed at the chunklet level, not the drive level
  • A virtual volume is composed of chunklets from multiple drives

**Benefits:**

  • Automatic load balancing across all drives
  • No hot spots
  • Fast rebuild (only affected chunklets, not entire drives)
  • Online expansion without performance impact

Common Provisioning Groups (CPG)

CPGs are the fundamental building block of 3PAR storage.

**CPG Characteristics:**

  • Group of physical drives with same characteristics (type, speed, RAID level)
  • All chunklets for volumes in a CPG come from drives in that CPG
  • Supports thin provisioning (deduplicated CPGs)
  • Supports deduplication (dedupe CPGs)
  • Growth by adding drives to CPG

**CPG Types:**

  • **Standard CPG:** Thick provisioning
  • **Thin CPG:** Thin provisioning (space allocated on write)
  • **Dedupe CPG:** Thin + deduplication (data blocks deduplicated)
  • **Adaptive Optimization CPG:** Multi-tier (SSD + SAS + SATA)

Virtual Volumes (VV)

VVs are the logical devices presented to hosts.

**VV Characteristics:**

  • Created from a CPG
  • Can be thick or thin
  • Can be part of a VVset
  • Supports snapshots and clones
  • Can be converted between thick and thin

VVsets (Volume Sets)

VVsets group multiple VVs for consistent management.

**VVset Benefits:**

  • All VVs in set share same characteristics
  • Snapshots applied to entire set
  • QoS policies applied to set
  • Simplified management for related volumes

Quality of Service (QoS)

3PAR QoS controls I/O priority for volumes.

**QoS Controls:**

  • **Priority:** High, Normal, Low
  • **Min IOPS:** Guaranteed minimum
  • **Max IOPS:** Maximum limit
  • **Min BW:** Guaranteed minimum bandwidth
  • **Max BW:** Maximum bandwidth

Data Reduction

**Thin Provisioning:** Space allocated only when data is written.

**Deduplication:** Duplicate data blocks eliminated at the block level.

  • Inline dedupe (during write)
  • Post-process dedupe (scheduled)
  • Dedup savings depend on data type (VMs, databases = high savings)

**Zero Detection:** Zero blocks not written to storage.

Remote Copy

3PAR Remote Copy provides asynchronous or synchronous replication.

**Modes:**

  • **Asynchronous:** Periodic snapshot replication (RPO > 0)
  • **Synchronous:** Write acknowledged only after remote write (RPO = 0)
  • **Streaming:** Continuous asynchronous replication

**Target Types:**

  • Same array (local)
  • Remote 3PAR array
  • Cloud (3PAR CloudStore)

Prerequisites

Hardware Requirements

  • HPE 3PAR StoreServ array (8200, 8400, 8440, 8450, 8480, 20800, 20850)
  • Drive enclosures with SAS or NVMe drives
  • Front-end switches (FC or Ethernet)
  • Management workstation

Software Requirements

  • HPE 3PAR StoreServ Management Console (SSMC)
  • HPE 3PAR CLI (SSH access)
  • Host operating system with 3PAR support
  • 3PAR OS (current version recommended)

License Requirements

  • Base array license
  • Thin Provisioning License
  • Deduplication License
  • Remote Copy License
  • Adaptive Optimization License
  • Dynamic Optimization License

Step-by-Step Implementation

Step 1: Initial Array Setup

**Physical Installation:** 1. Rack mount node pairs and drive enclosures 2. Install drives 3. Connect SAS cables (node to enclosure) 4. Connect front-end ports (FC or Ethernet) 5. Connect cluster interconnect between node pairs 6. Connect management ports 7. Connect power to redundant PDUs 8. Power on

**Initial Configuration:** 1. Connect to array via SSMC or CLI 2. Run initial setup wizard 3. Configure:

4. Accept EULA and licenses

  • System name
  • Network settings (management)
  • NTP and DNS
  • Administrator accounts
  • Email alerts

Step 2: Create Common Provisioning Group (CPG)

**Via SSMC:** 1. Navigate to Storage → Provisioning Groups 2. Click "Create CPG" 3. Configure:

4. Select drives 5. Create

  • Name
  • RAID level (1, 5, 6)
  • Drive type (SAS, SATA, NVMe)
  • Growth increment
  • Thin provisioning (yes/no)
  • Deduplication (yes/no)

**Via CLI:**

# Create thick CPG
creatcpg -R 1 -t RAID1 CPG_SAS_RAID1

# Create thin CPG
creatcpg -R 5 -T CPG_SAS_Thin5

# Create dedupe CPG
creatcpg -R 6 -T -D CPG_SAS_Dedupe6

# Add drives to CPG
addcpg CPG_SAS_Thin5 -d 1:1.1,1:1.2,1:1.3,1:1.4

**Verification:** CPG shows "online" status, capacity displayed.

Step 3: Create Virtual Volumes (VV)

**Via SSMC:** 1. Navigate to Storage → Volumes 2. Click "Create Volume" 3. Configure:

4. Create volume

  • Name
  • Size
  • CPG
  • Type (thick, thin, deduplicated)
  • Format (standard, RC copy)

**Via CLI:**

# Create thick volume
createvv -T thin -c CPG_SAS_Thin5 VV_DB_Data 500GB

# Create thick volume
createvv -c CPG_SAS_RAID1 VV_OS_Boot 200GB

# Create volume with QoS
createvv -c CPG_SAS_Thin5 -prio high -minIOPS 1000 -maxIOPS 5000 VV_Performance 1TB

**Verification:** Volume shows "online", accessible to hosts.

Step 4: Create Host and Export Volume

**Via SSMC:** 1. Navigate to Storage → Hosts 2. Create host:

3. Create host volume set (HVS) 4. Add volumes to HVS 5. Assign to host

  • Name
  • OS type
  • Initiator ports (FC WWN or iSCSI IQN)

**Via CLI:**

# Create host
createhost -i "10:00:00:00:c9:2e:40:10" -i "10:00:00:00:c9:2e:40:11" -t VMWare ESXi_Host1

# Create host volume set
createhvs -a ESXi_Host1 VV_DB_Data,VV_DB_Log

# Export
exportvvlun -a ESXi_Host1 VV_DB_Data 1
exportvvlun -a ESXi_Host1 VV_DB_Log 2

**Verification:** Host sees LUNs.

Step 5: Configure QoS

**Via SSMC:** 1. Navigate to Storage → QoS 2. Create QoS rule 3. Configure:

4. Apply to volume or VVset

  • Priority (high/normal/low)
  • Min/Max IOPS
  • Min/Max Bandwidth

**Via CLI:**

# Set QoS on volume
setvv -prio high -minIOPS 2000 -maxIOPS 10000 VV_Performance

# Remove QoS
setvv -prio normal VV_Performance

Step 6: Configure Snapshots

**Create Snapshot:**

# Create snapshot
createsnap VV_DB_Data@snap1 VV_DB_Data

# Create snapshot with expiration
createsnap VV_DB_Data@snap2 VV_DB_Data -expire 24h

# Remove snapshot
removesnap VV_DB_Data@snap1

**Snapshot Schedule:** 1. Navigate to Data Protection → Snapshot Schedule 2. Create schedule:

3. Apply to volumes

  • Frequency (hourly, daily, weekly)
  • Retention count
  • Start time

Step 7: Configure Remote Copy

**Create Remote Copy Group:**

# Create RC group
addrccopygroup -t async RCG_Async

# Add volume to RC group
addrccopy -t VV_DB_Data -g RCG_Async

# Configure remote array
addrcremotedevice -name RemoteArray -i <remote-ip>

# Start replication
startrcopy -g RCG_Async

**Verification:** Replication status shows "synchronized" or "consistent".

Configuration Examples

Complete 3PAR Setup for Database

# Create CPGs
creatcpg -R 1 -t RAID1 CPG_OS_RAID1
creatcpg -R 5 -T CPG_Data_Thin5
creatcpg -R 6 -T -D CPG_Archive_Dedupe6

# Create volumes
createvv -c CPG_OS_RAID1 VV_OS 200GB
createvv -c CPG_Data_Thin5 VV_DB_Data 2TB -prio high -minIOPS 5000
createvv -c CPG_Data_Thin5 VV_DB_Log 500GB -prio high
createvv -c CPG_Archive_Dedupe6 VV_Archive 10TB

# Create host
createhost -i "10:00:00:00:c9:2e:40:10" -i "10:00:00:00:c9:2e:40:11" -t Linux DB_Host1

# Export volumes
createhvs -a DB_Host1 VV_OS,VV_DB_Data,VV_DB_Log
exportvvlun -a DB_Host1 VV_OS 1
exportvvlun -a DB_Host1 VV_DB_Data 2
exportvvlun -a DB_Host1 VV_DB_Log 3

# Create snapshot schedule
creatsnapsch -name DailyDB -freq daily -count 7 -time 02:00 VV_DB_Data,VV_DB_Log

QoS for Mixed Workload

# High priority for production database
setvv -prio high -minIOPS 5000 -maxIOPS 20000 VV_Prod_DB

# Normal priority for development
setvv -prio normal -maxIOPS 5000 VV_Dev_DB

# Low priority for backup
setvv -prio low -maxIOPS 1000 VV_Backup

Verification

Array Health Verification

# Show system status
showsys

# Show node status
shownode

# Show CPG status
showcpg

# Show volume status
showvv

# Show host status
showhost

# Show task status
showtask

**Expected:** All components "online", no failed drives, tasks complete.

Performance Verification

# Show performance statistics
showstats -t VV -d VV_Name

# Show QoS statistics
showqos

Capacity Verification

# Show CPG capacity
showcpg -capacity

# Show volume capacity
showvv -capacity

Troubleshooting

| Scenario | Symptom | Possible Cause | Diagnostic | Corrective Action | Verification | |----------|---------|----------------|------------|-------------------|--------------| | Volume not accessible | Host doesn't see LUN | Host not exported, zoning issue | `showhost`, `showvv` | Export volume, check zoning | Host sees LUN | | Pool/CPG full | Cannot create volume | CPG capacity exhausted | `showcpg -capacity` | Add drives to CPG, delete unused volumes | Capacity available | | Slow performance | High latency | Wrong QoS, CPG on slow drives, node overload | `showstats`, `showqos` | Adjust QoS, move VV to faster CPG | Latency reduced | | Node failure | One node down | Hardware failure, network issue | `shownode`, event log | Replace node, verify failover | Both nodes online | | Dedupe not working | No space savings | Data not dedupable, dedupe not enabled | `showvv -dedupe` | Verify dedupe CPG, check data type | Dedup savings shown | | Remote copy fails | Replication error | Network, license, version mismatch | `showrccopygroup`, event log | Fix network, install license | Replication synchronized | | Thin provisioning alert | Pool near full | Actual usage high | `showcpg -capacity` | Add drives, delete snapshots, thin provision | Capacity OK | | Rebuild slow | Degraded CPG long rebuild | Drive size, system load | `showtask`, `showcpg` | Reduce load, wait | CPG returns to online |

Best Practices

1. **Use thin CPGs for most workloads** — Better capacity utilization 2. **Use dedupe CPGs for backup/archive** — Significant space savings 3. **Separate CPGs by drive type** — SSD, SAS, SATA in separate CPGs 4. **Use QoS for critical workloads** — Guarantee performance for production 5. **Monitor CPG capacity** — Alert at 70%, 80%, 90% 6. **Test failover** — Quarterly node failover tests 7. **Use Adaptive Optimization** — Auto-tier between SSD and HDD 8. **Keep 3PAR OS updated** — Bug fixes and performance improvements 9. **Document CPG design** — Critical for capacity planning 10. **Use Remote Copy for DR** — Asynchronous for remote sites

Conclusion

HPE 3PAR provides enterprise-class storage with its unique chunklet-based architecture, CPG model, and advanced data services. The active-active node pair design ensures high availability, while thin provisioning, deduplication, and QoS provide efficient resource utilization.

Successful 3PAR administration requires understanding CPG design, volume management, QoS, and data protection features. Regular monitoring and capacity planning ensure optimal performance and availability.

References

  • HPE 3PAR StoreServ Management: https://www.hpe.com/info/3par-docs
  • HPE 3PAR CLI Reference: https://www.hpe.com/info/3par-cli
  • HPE 3PAR Best Practices: https://www.hpe.com/info/3par-best-practices
  • HPE 3PAR SSMC: https://www.hpe.com/info/ssmc
  • HPE SPOCK: https://www.hpe.com/info/spock
#hpe#3par#storage-array#enterprise-storage#storage-administration#infrastructure