Technical white paper Using HP StoreOnce Backup Systems for NDMP backups with Symantec NetBackup Table of contents Executive summary... 2 Introduction... 2 What is NDMP?... 2 Technology overview... 3 HP StoreOnce Backup Systems key features and benefits... 3 HP StoreOnce Backup Systems in small to large data centers... 4 Why use StoreOnce for NDMP-enabled NAS server backups?... 4 Advantages of using StoreOnce VTLs with NDMP... 4 Symantec NetBackup NDMP architecture options... 4 HP StoreOnce Backup Systems for NDMP backups... 9 Capacity planning... 9 HP StoreOnce Backup Systems and NDMP data recovery... 12 Considerations... 12 Recommendations... 12 Conclusion... 13 Call to action... 14
Executive summary In today s business environment, customers are confronted with many different types of configurations regarding the backup and restore capabilities for their network attached storage (NAS) servers. Symantec NetBackup s integrated Network Data Management Protocol (NDMP) backup capabilities enhance the administrator s ability to effectively manage NDMP-enabled NAS server backup and recovery. NDMP customers rely on the most efficient, highest performing, and reliable backup systems. Customers need to protect increasing levels of data while keeping costs under control. In particular, businesses today are concerned about the costs of backing up and archiving important data from mission-critical software business applications. HP StoreOnce Backup Systems provide a disk-based data protection platform while addressing data growth by applying HP StoreOnce deduplication software for efficient, longer-term data retention. Key benefits include: StoreOnce virtual tape library (VTL) devices are excellent backup targets for NDMP backups which benefit from StoreOnce data deduplication. StoreOnce targets experience quickly rising deduplication ratios that are maintained over time. NDMP backups take advantage of faster backup and recovery with StoreOnce systems. NDMP data is sent directly from the NAS device through the SAN to the StoreOnce system which allows greater throughput and performance. Introduction HP StoreOnce Backup Systems are a disk-based backup system that delivers leading price performance and deduplicates NDMP backup data. StoreOnce systems can be used to automate and consolidate the backup of NDMP data onto a single, rack-mountable device while improving reliability by reducing errors caused by media handling. For business environments with remote offices or a disaster recovery site, StoreOnce systems can be used to replicate data to a central data center or remote facility. HP StoreOnce Backup Systems are ideal for NDMP backup data. Proper configuration of NDMP backups with Symantec NetBackup to StoreOnce systems provides outstanding backup throughput performance and data deduplication ratios. StoreOnce Backup systems integrate seamlessly into current IT environments and offer the flexibility of VTL for NDMP backups. What is NDMP? The NDMP is a standardized protocol meant to transport data between NAS devices and backup devices. This removes the need for transporting the data through the backup server itself, thus enhancing speed and removing load from the backup server. With NDMP, heterogeneous network file servers can communicate directly to a network-attached tape device for backup or recovery operations. 2
Technology overview HP StoreOnce Backup Systems key features and benefits HP StoreOnce deduplication, store more data on disk HP StoreOnce deduplication reduces the disk space required to store backup data sets without impacting backup performance. Retaining more backup data on disk for longer, enables greater data accessibility for rapid restore of lost or corrupt files and reduces downtime. Deduplication ratios are determined by two factors data change rate and retention periods of the data on the deduplication appliance. Low data change rates and data retained for longer periods of time yield higher deduplication ratios. Deduplication-enabled replication HP StoreOnce deduplication is the technology enabler for HP StoreOnce Deduplication Enabled replication, which allows fully automated replication over low bandwidth links to a disaster recovery (DR) site, giving Remote Office/Branch Office (ROBO) and small data centers a cost-effective DR solution for the first time. Rapid restore of data for dependable, worry-free data protection HP StoreOnce Backup Systems offer immediate access to backups for rapid restores. StoreOnce deduplication allows more data to be stored closer to the data center for longer periods of time, which offers immediate access for rapid restores. More affordable disk-based data protection With up to twice the price performance of comparable products with deduplication and replication, HP StoreOnce Backup Systems deliver more affordable data protection. HP StoreOnce Backup Systems offer scalability to 512 TB of usable capacity as your data storage requirements grow, with simple and cost-effective capacity upgrades that allow you to increase storage capacity without purchasing additional systems. Automate, simplify, and improve the backup process HP StoreOnce Backup Systems automate your backup processes allowing you to reduce the time spent managing your data protection. Implementing hands-free, unattended daily backup is especially valuable for environments with limited IT resources such as remote or branch offices. StoreOnce systems can backup multiple servers via a standard Ethernet or Fibre Channel network simultaneously to a disk-based solution at peak speeds of up to 2.4 TB per hour instead of sequentially to a tape drive or autoloader, meaning that you can substantially reduce your backup window. StoreOnce systems can be intuitively managed and configured by using the built-in web browser s administrative interface. And for large deployments of replicating StoreOnce appliances, the StoreOnce Replication Manager can monitor multiple backup systems throughout the data center. StoreOnce systems are self-managing backup appliances that require little, if any, routine maintenance. Unlike other disk-based storage devices, StoreOnce systems do not require virus protection or LUN provisioning. 3
HP StoreOnce Backup Systems in small to large data centers Why use StoreOnce for NDMP-enabled NAS server backups? As customers of small to large data centers prepare to upgrade or migrate data protection environments, an opportunity exists to improve functionality, performance, and total cost of ownership (TCO) with HP StoreOnce Backup Systems. Customers using NDMP in their NAS environment may wish to migrate or consolidate disparate small systems into a scalable StoreOnce system. Migration from traditional physical tape devices to a StoreOnce system can be seamless because customers may use existing processes and policies in their new environment. Also, customers with complex backup schedules find that the StoreOnce systems platform provides an opportunity to migrate to a disk-to-disk backup environment maintaining existing processes used with physical tape using VTL. Customers who use NAS file devices on general purpose shares as backup targets may migrate to a StoreOnce system to mitigate contention and performance issues found in I/O intensive backups, business, and other applications sharing the same storage array. Advantages of using StoreOnce VTLs with NDMP HP StoreOnce Backup Systems may be configured with VTL to be used with NDMP-enabled NAS server backups. When backing up NAS storage using NDMP, StoreOnce VTLs have an advantage in that the NDMP backup images can go directly from the NAS device to the VTL across Fibre Channel (FC), whereas sending the backup images to non VTL-based disk storage requires the data be sent across the LAN to the media server before going to non VTL-based disk storage. StoreOnce systems are designed as target storage for backup applications that need rapid restores and reduced backup windows. As such, a StoreOnce system configured as a VTL uniquely offers: Easy setup and maintenance Data deduplication, allowing more backup data to be retained for longer periods Data replication for cost-effective offsite storage Self-managing requiring little, if any, routine maintenance Symantec NetBackup NDMP architecture options There are several different options for backing up NDMP data using Symantec NetBackup with the most common being Local NDMP, Direct NDMP, Remote NDMP, and 3-way NDMP. This document focuses on the benefits of using HP StoreOnce Backup Systems combined with NetBackup to protect and recover NAS devices utilizing NDMP. 4
Local NDMP The most common option for NAS NDMP protection is local NDMP (or LAN-free) backup. This is accomplished by attaching a single tape drive or tape library directly to the NAS and sending backups directly across SAS and SCSI to the tape device or sending backups across the SAN. This tape device can be a standalone drive, library, VTL, or any number of NDMP dedicated drives in a library. While the devices are considered directly connected, all scheduling and management are still handled by the NetBackup master server. Figure 1. Local NDMP architecture Advantages: Backups avoid the LAN (no additional LAN traffic) and media server (no additional load), and take advantage of direct SAS/SCSI or SAN-attached tape or VTL for potentially faster backup and recovery Disadvantages: Requires SAN for shared NDMP backup and recovery Locally attached NDMP drives are only available for NDMP backups 5
Direct NDMP Direct NDMP backup is similar to local NDMP backup, but direct NDMP backup shares SAN tape drives in a library with a NetBackup media server. NetBackup can share tape resources between the media servers and NAS using the Shared Storage Option (SSO). This requires the NDMP host to be attached and zoned in the SAN to detect the library or drives. The master server controls access to the tape device. Figure 2. Direct NDMP architecture Advantages: Backups avoid the LAN (no additional LAN traffic) and media server (no additional load), and take advantage of direct SCSI or SAN-attached tape or VTL for potentially faster backup and recovery Allows SAN tape drives in a library to be shared with a NetBackup media server Disadvantages: SAN for NDMP backup and recovery to shared tape or VTL is required Locally attached NDMP drives are only available for NDMP backups Local or direct backup to FC/SCSI attached tape or VTL only 6
Remote NDMP Remote NDMP backup integrates the same tape device support as direct NDMP backup, but sends the data stream over the network through a NetBackup media server. Figure 3. Remote NDMP architecture Advantages: Implementation is simple Support for remote NDMP backup to disk storage units Take advantage of media server load balancing for increased performance, efficiency, and high availability Support for Media Server Encryption Option (MSEO) and tape encryption Disadvantages: Potentially slower backups due to LAN versus SAN speeds Additional LAN overhead if a dedicated backup LAN is not used 7
3-way NDMP The NetBackup master server initiates the backup. Data travels over the network from an NDMP host to a storage device that is attached to another NDMP host on the local network or across the SAN. Figure 4. 3-way NDMP architecture Advantages: Implementation is simple Support for remote NDMP backup to disk storage units Disadvantages: Potentially slower backups due to LAN versus SAN speeds Additional LAN overhead if a dedicated backup LAN is not used 8
HP StoreOnce Backup Systems for NDMP backups An important part of managing NDMP backups is maintaining a consistent set of NDMP image backups available for recovery. When data is lost due to user error, system failure, or site catastrophe the HP StoreOnce Backup Systems integrated with a well-planned data protection strategy using NDMP backups provide a seamless recovery. Capacity planning The amount of required backup storage for NDMP backups depends on the following: Number of NAS systems and NAS storage capacity Site backup retention policy Type (full, incremental, differential) and frequency of backups Rate of change of the NDMP data The deduplication ratio achieved by StoreOnce systems HP StoreOnce Backup Systems do not deduplicate across VTL or NAS shares. Each VTL and NAS share has a unique hash table used for data deduplication. A unique StoreOnce VTL or NAS share should be created specifically for NDMP images. Note: The rate of change of NDMP data refers to the amount of data that would be contained in a differential backup as a percentage of a full backup. A 100 GB full backup with a subsequent 5 GB differential backup before the next full backup would be a 5% rate of change. A backup administrator may desire to have 2 weeks worth of backups stored on StoreOnce systems for quick recovery access. Data deduplication provides more backup space without increasing the physical capacity of the backup device; however, a dynamic NDMP environment with changing data affects the backup data deduplication ratio. In performing these tests HP used standard Microsoft Windows systems and a customer representative dataset with realistic structure and content. For the chart below, data was updated between each backup until the desired rate of change was reached. 9
Figure 5 shows the data rate of change effect on deduplication ratios when backing up NDMP data to a StoreOnce VTL. Figure 5. NDMP data rate of change effect on StoreOnce deduplication ratios StoreOnce Deduplication Ratio per NDMP Backup Rate of Change Expected Deduplication Ratio 1% 2% 3% 4% 5% NDMP Data Rate of Change Between Each Backup Figure 6 illustrates how StoreOnce backup throughput with NDMP data was not significantly affected by using different block sizes. Figure 6. NDMP backup throughput trend over time NDMP Backup Throughput 5% Rate of Change Backup Throughput 64 KB Block size 128 KB Block size 256 KB Block size 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Number of Full Backups 10
Figure 7 shows how StoreOnce deduplication ratios rise quickly and are maintained over time when backing up NDMP data. Changing the block size settings had no significant effect on the deduplication ratio. Figure 7. NDMP backup deduplication ratio trend over time NDMP Backup Deduplication Ratio 5% Rate of Change Expected Deduplication Ratio 64 KB Block size 128 KB Block size 256 KB Block size 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Number of Full Backups Figure 8 shows how StoreOnce deduplication ratio varies between weekly full backups and weekly full with daily incremental backups with data changes between each backup. Figure 8. NDMP full backups versus full + incremental backups deduplication trend over time NDMP Backup Deduplication Ratio 5% Rate of Change Full Backups vs. Full + Incremental Backups Expected Deduplication Ratio Full Backups Full + Incremental Backups 1 2 3 4 5 6 7 8 9 10 11 12 13 14 Number of Full Backups 11
HP StoreOnce Backup Systems and NDMP data recovery A solid data protection strategy is important to every company. At the enterprise level, customers are likely to invest in a site disaster recovery strategy that includes a quick and reliable solution. HP StoreOnce Backup Systems, along with StoreOnce deduplication, allow more data to be stored closer to the data center for longer periods of time which offers immediate access for rapid restores. For business environments with remote offices or a disaster recovery site, HP StoreOnce Backup Systems can be used to replicate data to a central data center or remote facility. When performing an NDMP restore, keep in mind that NDMP is a command protocol and not a format. Also understand that there is no NDMP server in the HP StoreOnce Backup Systems for tape offloads. NDMP backups can only be restored to an NDMP host and not directly to a Windows, Linux, or UNIX system. Recovering NDMP data with Symantec NetBackup can be accomplished by using the normal restore steps or by using Direct Access Recovery (DAR). DAR is enabled by default in NetBackup (no configuration is required). DAR can reduce the time it takes to restore files by allowing the NDMP host to position the exact location of the file on the media rather than reading sequentially through the whole backup. Only the data that is needed for those files is read. This feature is inherent to all NDMP backup methods, but may not be supported by all vendors. Without DAR, NetBackup reads the entire NDMP backup image, even if only a single restore file is needed. Considerations You would typically use Local NDMP if you have a tape drive or tape library connected directly to the NDMP host, sending backups directly across SCSI or SAN. If backup directly to tape or VTL is a top priority or if sharing of SAN tape drives in a library is a requirement then you would likely use Direct NDMP. If you do not require backup to disk, or if your backups are smaller in size, or if no SAN has been implemented then you would normally use Remote NDMP. You would typically use 3-way NDMP if you wanted to backup from one NDMP host to a tape device connected to another NDMP host. Recommendations Deduplication The StoreOnce Backup System is an excellent solution for NDMP backups utilizing HP StoreOnce deduplication allowing more data to be retained on disk and enabling low-bandwidth replication to deliver a cost-effective offsite DR solution. Backup block size As shown in Figure 7, there was no adverse effect with the deduplication ratio achieved on StoreOnce systems when using any variation of block sizes. With this in mind you can be confident that the NDMP data backed up to StoreOnce systems can certainly benefit from StoreOnce data deduplication. NDMP full backups versus weekly full with daily incremental backups Daily full backups deduplicate at a much higher rate than full with daily incremental backups but require more server and StoreOnce systems processing resources and bandwidth. End-to-end data compaction and StoreOnce data storage required is roughly the same for both types of backup schedules over an extended time period. Daily full backups deduplicate better but weekly full with daily incremental backups send much less data to StoreOnce systems for deduplication processing. If daily full backups are not required HP recommends a backup schedule that includes incremental backups to reduce the resource load required for NDMP backups. 12
Local and direct NDMP Keep the volumes in a large NAS configuration to a manageable size to assist with backup performance. By keeping the number of files on the volumes in check you can help with backup times. For example, if the options are creating one volume with 100 million files or 10 volumes with 10 million files, the latter would be ideal if applicable. Remote NDMP When building large NAS configurations, keeping the volumes in a manageable size increases backup performance. Depending on the size of your NDMP backups, you might dedicate a network for only NDMP backups, commonly referred to as a backup LAN. With NDMP backups larger than 1 TB, 1GbE links should be the minimum, whereas 10GbE should be used for even larger NDMP backups in the 10 TB range. 3-way NDMP Implementing a dedicated backup LAN would be beneficial. Backups would be able to take advantage of decreased LAN traffic. Conclusion Customers using NDMP backups demand an efficient, reliable data growth management backup system environment while keeping costs under control. HP provides a variety of reliable data protection storage solutions that address such requirements. HP StoreOnce Backup Systems are one such solution. These backup systems offer high performance and reliability while addressing data growth through HP StoreOnce data deduplication technology for small to large customers. They integrate easily with leading data protection products to protect important data for mission-critical applications. Combining HP StoreOnce Backup Systems with NDMP backups provides a comprehensive data protection solution. By using HP StoreOnce Backup Systems with NDMP backups you can have confidence that: HP StoreOnce deduplication reduces the disk space required to store backup data sets without impacting backup performance. HP StoreOnce Backup Systems offer immediate access to backups for rapid restores. HP StoreOnce Backup Systems automate your backup processes allowing you to reduce the time spent managing your data protection. 13
Call to action Effectively manage your NDMP-enabled NAS server backup and recovery with HP StoreOnce Backup Systems, to learn how, visit: http://www.hp.com/go/storeonce Additional information HP StoreOnce Backup Systems user guide http://bizsupport1.austin.hp.com/bc/docs/support/supportmanual/c02295179/c02295179.pdf HP StoreOnce Backup Systems Linux and UNIX configuration guide http://h20000.www2.hp.com/bc/docs/support/supportmanual/c02299831/c02299831.pdf HP StoreOnce Backup Systems best practices guide http://bizsupport2.austin.hp.com/bc/docs/support/supportmanual/c02511912/c02511912.pdf Get connected hp.com/go/getconnected Current HP driver, support, and security alerts delivered directly to your desktop Copyright 2012 Hewlett-Packard Development Company, L.P. The information contained herein is subject to change without notice. The only warranties for HP products and services are set forth in the express warranty statements accompanying such products and services. Nothing herein should be construed as constituting an additional warranty. HP shall not be liable for technical or editorial errors or omissions contained herein. Microsoft and Windows are U.S. registered trademarks of Microsoft Corporation. UNIX is a registered trademark of The Open Group. 4AA3-9999ENW, Created April 2012