Commvault® Validated Reference Design Specification

Commvault HyperScale™ Software on HPE Servers

Introduction to Commvault HyperScale Software

With Commvault HyperScale™ Technology, you can build a unified, modern data protection and management platform that delivers cloud-like services on-premises. The purpose of this technical specification is to detail HPE server components for the Commvault Validated Reference Design. By building these services on a scale-out infrastructure and leveraging Commvault capabilities, you’ll enable:

  • Cloud-like agility, resiliency and availability to on-premises data and applications
  • Greater end-user efficiency through automation and self-service capabilities
  • Improved hardware utilization and optimized costs from general-purpose hardware
  • Seamless storage scalability with predictable performance without requiring forklift upgrades
  • Better, more secure data protection, utilization and movement by eliminating point product and data silos

By shifting the secondary storage and data management infrastructure to this architecture, enterprises can go a long way in transforming their data centers to be as operationally efficient, resilient and scalable as public cloud infrastructure. Lower hardware costs, operational efficiencies and simplified support allows the replacement of limited and legacy backup tools with a modern cloud-enabled data management solution at the cost of replacing legacy purpose-built backup appliance (PBBA). More important, this architecture, which extends into public cloud, allows enterprises to offer consistent sets of services to all workloads running on-premises or in public cloud, independent of the underlying infrastructure for true cloud-based data management.

General availability designation

These configurations are classified as general availability designs, meaning the servers have been tested and validated as per the Commvault Validated Reference Design Program. These configurations are subject to change due to updated part numbers or replacement hardware due to hardware life cycle. Validated reference designs are developed to provide desired resilience and performance at optimal costs for customer data-protection workload. Commvault collaborates with HPE to create fully supported design specifications. Substitutions or modifications to validated design specifications could result in unsupported configurations. Any substitution or modification to validated configurations must be approved by both Commvault and HPE. This configuration is currently orderable for customer deployments and supported through Commvault support channels.

How to use this document

This document details the necessary design components of the Commvault HyperScale Technology architecture, providing the key components required when purchasing and configuring the infrastructure for a Commvault HyperScale Software solution. Commvault Reference Designs deliver validated configurations with leading hardware vendors to provide validated designs complemented by best practices that will accelerate ROI, reduce complexity and add customer value.

The document is broken into a high-level component section detailing the configuration and specific component options that can be selected to satisfy storage capacity and density requirements. The reader is referred to a HPE link for details on individual server configuration, validated by Commvault HyperScale Software. This document does not cover the overall architecture and design of the Commvault HyperScale solution and should be considered as a supplement specific to HPE.

HPE Gen10 general summary

Server overview
Technical specifications Proliant DL360 Proliant DL380 Apollo 4200
Form factor 3x of validated 1RU servers, each with 4x LFF HDD’s 3x of validated 2RU servers, each
with 6x or 12x LFF HDD’s
3x of validated 2RU servers, each
with 6x, 12x, 18x or 24x LFF HDD’s
Motherboard chipset Intel® C620 series Intel® C620 series Intel® C620 series
Processors Single Intel® Xeon® Silver 4210 or higher CPU Dual Intel® Xeon® Silver 4210 or
higher CPU’s
Dual Intel® Xeon® Silver 4210 or
higher CPU’s
Memory 128 GB RAM 256 GB RAM 256 GB RAM

Boot and metadata storage

Boot storage devices house the operating system and core Commvault HyperScale binaries, while the metadata storage device provides caching areas for such operations as deduplication, indexing and extents. The boot devices are recommended to be mirrored (RAID-1) for node level resilience and the metadata partitions are on fast NVMe card.

Data storage options

Data storage houses the data footprint for the customer environment. Data storage configuration directly impacts the amount of data that each node can accommodate.

When deploying nodes inside of the same block (e.g., 3-node initial configuration), choose identical hard disk drives (HDDs). If the nodes in a block have different HDD sizes, the smallest size will be chosen for the data storage, which would lead to wasted resources on nodes with larger HDDs.

Separate node blocks in the same grid may use different HDDs (e.g., mixing a 3 node 6 TB block with a second 3 node 10 TB block in the same grid).

Overall sizing and retention varies per customer and therefore is beyond the scope of this document. Please refer to Commvault HyperScale Technology sizing documentation to determine the drive size (and node quantity) required for the specific deployment.

Commvault HyperScale nodes can optionally be deployed with partially filled HDD slots. As additional storage is required, nodes can be scaled vertically by filling empty HDD slots. Initial deployment and vertical scaling must be done in multiples of 6 drives per node. All nodes within a block must have the same number of HDD and must scale vertically at the same rate (e.g., start a block of 3 nodes with 6 of 24 drive slots filled in all nodes; expand all three nodes simultaneously by adding 6 drives to each node).

Networking options

A minimum of 2 10 GB ports are required for Commvault HyperScale installs, one for incoming data and one for storage communication between the nodes. It is recommended to have 4 ports, 2 for data, and 2 for storage for network resilience. Where possible, these builds have been designed with this recommendation.

Optional I/O add-on cards

The design includes all core components to work with Commvault HyperScale Technology. Some environments and use-cases may require additional parts listed in the “Optional I/O” section of the “HPE Complete BOM Guidelines: Commvault HyperScale” link. The number of available PCIe slots in each server model determines the number of such options that can be accommodated.

Optional I/O cards for SAS, Fiber Channel and Ethernet connectivity and NVMe for housing dedupe partitions for a cloud copy are included, upon validation. SAS Connectivity is typically used for direct tape integration, while Fiber Channel cards are used for Commvault IntelliSnap® operations or tape libraries. Additional Ethernet cards may be required for a dedicated replication network or to connect to clients in isolated networks. Each of these options will require a free PCIe slot in the server being considered.

Note: Smaller form factor PCIe cards can fit in larger form factor server slots, however larger form factor PCIe cards cannot fit into smaller form factor server slots.

Bill of Material (BOM)

The Bill of Material lists all components required to configure Commvault HyperScale nodes. Each component has been tested and validated. Substitutions cannot be supported. Country-specific components such as power cables are not listed and can be changed as required. Please select the server model of interest from the link titled “HPE Complete BOM Guidelines: Commvault HyperScale” for the list of parts for each supported configuration.

Additional resources

Additional information regarding the HPE Gen10 servers can be found on the following HPE websites:

Commvault HyperScale

Technology integrates with storage arrays, hypervisors, applications, and the full range of cloud provider solutions to support the most diverse and dynamic environments.