Mini Document Title in Header – Double Click to Adjust
Insert your photo here
Hitachi Vantara Object Storage with HDPS
Solutions Reference Architecture & Best Practices
Solutions Engineering & Architecture
v1.1.0 – Feb 2025
© Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
Table of
Contents
Solution Introduction ............................................................................................. 2
Rebranding and maintenance releases................................................................................................................................ 2
Partner certification program ................................................................................................................................................ 2
Prerequisites ........................................................................................................................................................................ 2
Executive Summary ............................................................................................... 2
Reference Model .................................................................................................... 2
Reference implementation diagram ..................................................................................................................................... 3
HDPS Configurations ............................................................................................ 4
HDPS configurations............................................................................................................................................................ 4
Overview of Best Practices for VSP One Object................................................ 4
Overview of Best Practices for HCP .................................................................... 7
HCP configurations .............................................................................................................................................................. 7
Best Practices for Multi-Site Deployments ........................................................................................................................... 7
Using Erasure-Coded Replication to optimize HCP Storage Use ........................................................................................ 7
Best Practices for Synthetic Full Backups ............................................................................................................................ 7
Implementing Combined Storage Tier.................................................................................................................................. 8
Empty directories ................................................................................................................................................................. 8
Overview of Best Practices for HCP for Cloud Scale ........................................ 9
Appendix A – Troubleshooting .......................................................................... 10
Appendix B – Resources .................................................................................... 10
Hitachi Vantara resources .................................................................................................................................................. 10
Commvault resources ........................................................................................................................................................ 10
Appendix C – SRA Document History ............................................................... 11
1 © Hitachi Vantara LLC 2024. All Rights Reserved.
Mini Document Title in Header – Double Click to Adjust
Solution Introduction
With Hitachi Data Protection Suite (HDPS), powered by Commvault, Hitachi Vantara and Commvault together deliver a unified,
modern offering that facilitates the backup, recovery and management of enterprise and application data with industry-leading
object storage offerings: Hitachi Content Platform (HCP) and Hitachi Content Platform for Cloud Scale (HCP CS). The joint
solution offers the reliability required by the world’s largest organizations, while featuring simplicity, cost-effectiveness and
modern capabilities that are needed to remain agile and competitive. Hitachi Vantara’s object storage offerings seamlessly
extend the secured and guaranteed management of long-term data retention at petabyte scale.
This solution is currently supported with the following versions of Hitachi and Commvault offerings:
• Hitachi Content Platform 9.6
• Hitachi Content Platform for Cloud Scale 2.6
• Hitachi Content Protection Suite 11.32 LTS (powered by Commvault)
HDPS 11.32 LTS has been validated for compatibility with the Hitachi object storage offerings listed above.
No specific use cases (workload) have been validated at this time.
Rebranding and maintenance releases
Commvault delivers one Long-term support (LTS) platform release each year. Each LTS platform release is re-branded to
HDPS by Commvault and provided to Hitachi Vantara. A re-branded HDPS release is functionally equivalent to Commvault of
the same version. Commvault maintenance releases (MRs) are not re-branded. HDPS customers will update their system to
the latest Commvault MR through CommServe without any impact to HDPS branding.
Important:
Commvault MR 11.32.38 includes a fix for automatic database sealing that is required for the solution to function properly.
Without this fix, objects in HCP or HCP CS will not be pruned in a WORM scenario. Ensure that your HDPS system is on
Commvault MR 11.32.38 or greater.
Partner certification program
(not applicable)
Prerequisites
This document assumes you have firm knowledge and understanding of:
• Features and functionality of the products and components in the solution
• How to install and configure the products and components in the solution
• Any support infrastructure used as part of the solution
Executive Summary
The validation process for solution compatibility covered basic data lifecycle testing (backup, restore, and delete
functionality) of HDPS with HCP and HCP CS as backup targets.
The backup sources included:
• Microsoft SQL Server
• Oracle databases
• VMware ESXi VMs
• Linux servers
• Window servers
As this is currently a compatible solution, specific use cases, scale, and performance tests were not part of the validation.
Reference Model
The reference implementation used to validate compatibility of the solution included the following components:
© Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
• Backup sources:
o Microsoft SQL Server 2016 v18.8
o Oracle Server v21c
o VMware VM (various OS)
o Windows Server 2019
o Linux Server CentOS 7
• Supporting infrastructure:
o VMware ESXi Server 7.0
o Microsoft Windows Server® Standard 2016 for Active Directory
o 25GbE network with Brocade VDX 6740 switches for front-end and back-end networks
Reference implementation diagram
3 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
HDPS Configurations
This section summarizes the specific HDPS configurations that are important for solution interoperability.
HDPS configurations
Setting Value Notes
WORM Storage Lock Storage Lock WORM functionality within the storage backend.
Requires S3 Object Lock enabled on HCP
namespace or HCP CS bucket on creation.
Compliance Lock HDPS level WORM WORM functionality within HDPS that blocks
premature deletions.
Max Object Size With deduplication = 8 MB These are default values and not end user
configurable.
Without deduplication = 32 MB
Deduplication Block Size 512 KB
Chunk Size 4GB
Media Agent
DataMoverLookAheadLinkReaderSlot 1024
SILookAheadAsyncIOBlockSizeKB 4096
SILookAheadOvlIOThreadLimit 128
SILookAheadOvlIoThreadsPerRdr 128
DedupPrunerThreadPoolSizeCloud 60 Not required but suggested. Default is 10 – 20.
Overview of Best Practices for VSP One Object
This section summarizes the specific VSP One Object configurations that are important for solution interoperability.
HDPS MediaAgent configuration
Before VSP One Object can be configured as a Cloud Storage Library the following procedure must be completed on any
Media Agents which will be used to access the VSP One Object bucket. For this procedure you will need to know the region
that was specified during the VSP One Object installation.
Step 1. Find the Media Agent which will be accessing the VSP One Object system and open the properties window. Select the
Additional Settings tab.
Step 2. Add the sCloudDefaultRegion setting with the following details:
• Setting Name: ‘sCloudDefaultRegion’
• Category: ‘MediaAgent’
4 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
• Type: ‘String’
• Value: <Provide the region that was specified during the VSP One Object Installation>
• Enable: checked
Step 3. Repeat for all MediaAgents which will be accessing the VSP One Object system.
Step 4. (Optional) Add the nCloudServerCertificateNameCheck setting with one of the following values:
• 0: To skip checking the server’s CA Certificate.
• 2: To allow a self-signed certificate to verify the cloud server’s claimed identity.
Step 5. Use the Process Manager to stop and restart the MediaAgent services to apply the new settings.
5 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
Adding VSP One Object as Cloud Storage
When adding VSP One Object as Cloud Storage you can use one of the following options:
• Hitachi Content Platform for Cloud Scale (S3)
• S3 Compatible Storage
When configuring the Cloud Storage type, enter the full VSP One Object URL in the Service
Host field using the following format: [Link]
Occasionally, you may encounter the following error: "The authorization mechanism you provided is not supported. Please use
AWS4-HMAC SHA256. Resource." This occurs because HDPS attempts v4 authentication multiple times and then reverts to
v2. To resolve this issue, restart the MediaAgent services.
VSP One Object configurations
Setting Value Notes
Object lock enabled If using HCP CS enabled WORM. Requires Storage Lock enabled in HDPS.
6 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
Overview of Best Practices for HCP
This section summarizes the specific HCP configurations that are important for solution interoperability.
HCP configurations
Setting Value Notes
Versioning enabled
S3 Overwrite disabled Only if Versioning is enabled. Does not overwrite, new objects are
ingested even if enabled.
Pruning unchecked Only if Versioning is enabled. No new versions are created by HDPS.
S3 Object lock enabled If using HCP S3 object lock. Requires Storage Lock enabled in HDPS.
Namespaces unbalanced
Best Practices for Multi-Site Deployments
There are several ways to implement multi-site deployment of HDPS and HCP. The recommended implementation is to use
the HDPS Auxiliary Copy feature. This allows backups to be transferred to secondary destinations as redundant copies which
is useful for disaster recovery, data archiving, and ensuring data availability across different sites. Backup transfer can be
triggered once a backup is complete or on a schedule. Deduplication Accelerated Streaming Hash (DASH) copy can also be
enabled to optimize the replication process. DASH copy will allow for faster effective data transfer while maintaining data
integrity.
HDPS Auxiliary Copy can be paired with HCP replication within each site. Doing so will provide protection against data
unavailability from a single HCP system outage. A fully connected active/active replication topology can be implemented. This
means that every HCP System will have its data replicated to all other HCP systems. In the event of an outage of the HCP
system, requests to that system will be rerouted to another system to be fulfilled. This topology can have a maximum of 6
systems in it.
When configuring the HDPS storage pools, it is important to distribute the storage pools among the HCP systems. Meaning
that each storage pool will be mapped to a single HCP system instead of being able to read/write directly to multiple HCP
systems. Configuring the storage pools in this manner will prevent a situation where data is written to one HCP system and is
requested to be read from a different system before it has been replicated throughout the topology. When this happens the
request may fail and will have to be retried, which can be repeated depending on which system the request gets sent to next.
Even if there is some replication lag across the replication topology requests can still be fulfilled as whole objects will be
available on the HCP system each storage pool is configured to.
While not recommended, a multi-site deployment can be achieved using only HCP replication. In this configuration, each
HDPS site would be configured to an HCP system and the HCP replication topology will facilitate data replication. Just like with
Auxiliary Copy, it is important that every storage pool be configured to be a single HCP system. This will help avoid a scenario
where data is requested from HCP, which has yet to be replicated to that system.
Using Erasure-Coded Replication to optimize HCP Storage Use
If HDPS is configured to an HCP replicated topology, storage use can be further optimized by leveraging HCP’s erasure-coded
protection. With this feature, after an object is ingested into HCP, chunks for the object are calculated and then distributed
among the rest of the HCP systems in the replication topology. A full copy of the object will remain on the original HCP only.
This means that that HCP system can respond to requests without having to retrieve chunks to reconstruct the object. If this
system were to become unavailable, other systems may fulfill requests by reading the chunks from the other systems to
reconstruct the object. In this configuration the setting called erasure coding delay can be used to set the amount of time which
the whole object will be available on the system which it was ingested. After that the whole object is replaced by a chunk as
well saving additional storage. It is recommended to set this time to be equal to the retention time set in the HDPS storage
policy as it is less likely to be read back by HDPS. In a 3 HCP system topology this configuration can store the same amount
of data using have the storage compared using whole-object distribution.
Best Practices for Synthetic Full Backups
HDPS synthetic full backup is a backup type that has several benefits over full backups or full backup plus incremental.
Synthetic full backups are constructed by combining the latest full backup (synthetic or initial full) with the incremental captured
since, to create a new up to date full backup. Because this process only entails looking at backups already on HCP, it does not
7 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
require communicating with the source device at all. This also means the backup window for the source device is limited to the
time needed for an incremental, except the initial full backup.
Since synthetic full backups work by combining a previous backup with any incremental that have happened after the latest full
backup, the creation of the synthetic full backup requires reading back all that data in addition to writing out the new full
backup. This means that even if there are no recovery operations going on there may be more read traffic to HCP during this
time than expected for taking a backup.
To mitigate the read intensiveness of performing synthetic full backups there are several things that can be done. Options
include:
- Combined storage tiering can reduce the amount of read requests by ~20+%. See Implementing Combined Storage
Tier for more details.
- For deployments leveraging multiple HCP systems with replication, assigning storage pools to use a single HCP
system. See Best Practices for Multi-Site Deployments for more details.
- It is recommended that the frequency of synthetic full backups be equal to or greater than the retention setting in
days. For example, when using a retention period of 30 days, synthetic full backups should be scheduled to occur
every 30 days or more.
Implementing Combined Storage Tier
The combined storage tier feature for HDPS allows multiple storage devices to be used together. One tier, the warm tier, will
contain indexing and metadata that is frequently accessed. The second cold tier contains the actual data and full copies of the
indexes and metadata. For every read and write request from HDPS to HCP there are also metadata/index requests
performed such as HEAD requests. The warm tier will be able to accommodate some of these requests to minimize any
performance impact to the actual data requests to the cold tier HCP. When implemented with synthetic full backups you can
expect up to 20% reduction in read requests coming to the HCP.
It is recommended to implement a NAS storage device, such as VSP One Block, as the warm tier while HCP is used at the
cold tier. For assistance in configuring VSP One Block as a warm tier consult your Commvault sales representative.
Empty directories
Under some circumstances, container delete operations initiated by HDPS may leave objects and directories behind the HCP
system. This is a known issue with HCP and is expected to be resolved in an upcoming HCP release.
8 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
Overview of Best Practices for HCP for Cloud Scale
This section summarizes the specific HCP CS configurations that are important for solution interoperability.
HCP CS configurations
Setting Value Notes
Object lock enabled If using HCP CS enabled WORM. Requires Storage Lock enabled in HDPS.
9 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
Appendix A – Troubleshooting
This section summarizes some of the common errors or issues with potential workarounds.
Error or issue Workaround
Full restore on a windows data source will not restore in use This is expected behavior.
system files. These will generate a warning.
A different data flow can be used to backup and restore
these system files. See below links for guidance.
• Full System Recovery: Windows File System Agent
• Bare Metal Recovery Using 1-Touch for Windows
HDPS shows the following error: “Failed to run Data Aging Restart the HDPS services.
job due to change in system time” and/or data is not
removed from storage by HDPS This is an issue with HDPS that arises after a de-sync
between the Media Agent internal time and the system time.
This can most commonly occur due to system outage or
manual/automatic system time change.
See Commvault community forum post Data Aging - Failed
to run Data Aging job due to change in system time for
additional information.
Appendix B – Resources
This section summarizes the resources available about the solution and/or used to validate the compatibility of this solution.
Hitachi Vantara resources
• Hitachi Vantara & Commvault alliance
• Documentation
o Hitachi Content Platform (HCP)
▪ See Installing an HCP System PDF for configuration and setup guidelines
o Hitachi Content Platform for Cloud Scale (HCPCS)
▪ See Installing HCP for Cloud Scale web document for configuration and setup guidelines
o Data Protection Suite
▪ See Data Protection Suite under Storage Software
Commvault resources
• Hitachi Data Protection Suite powered by Commvault
• Commvault on the Hitachi Vantara alliance
• Commvault HDPS documentation
o See Cloud Storage configuration for instructions to configure HCP or HCPCS with HDPS
o See Commvault Best Practices and CommCell Performance Tuning for guidance on performance
parameters
• Commvault Maintenance Release 11.32.38 with fix for “Auto Sealing DDB's not acting as expected” (hotfix numbers
9900 and 9901) – access to release note requires a Commvault login
10 © Hitachi Vantara LLC 2024. All Rights Reserved
Hitachi Vantara Content Portfolio – Reference Architecture & Best Practices
Appendix C – SRA Document History
Revision history for this document
Revision Date Summary
1.0.0 Feb 5, 2024 v1.0.0 of HDPS + HCP & CS v1.0 SRA (Compatibility only)
1.0.1 Mar 20, 2024 v1.0.1 of HDPS + HCP & CS v1.0 SRA (Compatibility only):
▪ updated whitepaper template; formatting changes only
▪ removed list of test cases as they are not pertinent to the best practices for the solution
1.1.0 Feb 7, 2025 V1.1.0 of Hitachi Vantara Object Storage with HDPS
• Added procedure for adding VSP One Object to HDPS
• Added best practices for configuring HCP replication with HDPS
11 © Hitachi Vantara LLC 2024. All Rights Reserved