BEST PRACTICES GUIDE FOR DATA PROTECTION WITH FILERS RUNNING FCP

Similar documents
Best Practices Guide for Data Protection with NetApp Filers Running iscsi

Server Fault Protection with NetApp Data ONTAP Edge-T

StoreVault and NetApp Snapshot Technology

Chapter 11. SnapProtect Technology

WHITE PAPER: ENTERPRISE SOLUTIONS

Technology Insight Series

SnapManager for SAP

Monitoring and Reporting for an ONTAP Account

A Thorough Introduction to 64-Bit Aggregates

NS NetworkAppliance. Network Appliance Solution Architect-Backup and Restore (New)

The NetBackup catalog resides on the disk of the NetBackup master server. The catalog consists of the following parts:

Building Backup-to-Disk and Disaster Recovery Solutions with the ReadyDATA 5200

IT Certification Exams Provider! Weofferfreeupdateserviceforoneyear! h ps://

SnapMirror Configuration and Best Practices Guide for Clustered Data ONTAP

A Thorough Introduction to 64-Bit Aggregates

Volume Disaster Recovery Express Guide

Virtualization with Protection for SMBs Using the ReadyDATA 5200

Data ONTAP 7-Mode Administration Course D7ADM; 5 Days, Instructor-led

WHITE PAPER: ENTERPRISE SOLUTIONS. Disk-Based Data Protection Achieving Faster Backups and Restores and Reducing Backup Windows

Data Sheet: Storage Management Veritas Storage Foundation for Oracle RAC from Symantec Manageability and availability for Oracle RAC databases

A Close-up Look at Potential Future Enhancements in Tivoli Storage Manager

VERITAS Storage Foundation for Windows FlashSnap Option

Benefits of Multi-Node Scale-out Clusters running NetApp Clustered Data ONTAP. Silverton Consulting, Inc. StorInt Briefing

Oracle Backup and Recovery with NetApp Storage. Klaus Wagner Solution Architect

Midsize Enterprise Solutions Selling Guide. Sell NetApp s midsize enterprise solutions and take your business and your customers further, faster

Data ONTAP 7-Mode Administration (D7ADM)

The VERITAS VERTEX Initiative. The Future of Data Protection

Veritas Storage Foundation for Oracle RAC from Symantec

Outline: ONTAP 9 Cluster Administration and Data Protection Bundle (CDOTDP9)

WHITE PAPER. Oracle Backup and Recovery Essentials INFORMATION THAT EVERY ORACLE DATABASE ADMINISTRATOR SHOULD KNOW

Designing Data Protection Strategies for Oracle Databases

Information Lifecycle Management with Oracle Database 10g Release 2 and NetApp SnapLock

Meeting Today s Enterprise Storage Challenges: A Better Backup and Recovery Solution

NetApp Solutions for Oracle

Accelerated NCDA Boot Camp Data ONTAP 7-Mode (ANCDABC87)

Database over NFS. beepy s personal perspective. Brian Pawlowski Vice President and Chief Architect Network Appliance

Data ONTAP 8.2. MultiStore Management Guide For 7-Mode. Updated for NetApp, Inc. 495 East Java Drive Sunnyvale, CA U.S.

CONTENTS. 1. Introduction. 2. How To Store Data. 3. How To Access Data. 4. Manage Data Storage. 5. Benefits Of SAN. 6. Conclusion

Veritas InfoScale Enterprise for Oracle Real Application Clusters (RAC)

Designing Data Protection Strategies for Oracle Databases

Oracle E-Business Availability Options. Solution Series for Oracle: 2 of 5

Installation and Administration Guide

IBM Storage Sales for N series Version 1.

NETAPP - Accelerated NCDA Boot Camp Data ONTAP 7-Mode

Veritas NetBackup 6.5 Clients and Agents

VERITAS NETBACKUP 6.0 TECHNICAL OVERVIEW UNABRIDGED VERSION

Data ONTAP 7.3 Data Protection Online Backup and Recovery Guide

Chapter 12 Backup and Recovery

Accelerated NCDA Boot Camp Data ONTAP 7-Mode Course ANCDABC87; 5 Days, Instructor-led

Infinite Volumes Management Guide

STORAGE CONSOLIDATION WITH IP STORAGE. David Dale, NetApp

NetworkAppliance NS Network Application FAS900 Pre-Sales Engineer. Download Full Version :

Netapp Exam NS0-155 Data ONTAP 7-Mode Administrator Version: 6.2 [ Total Questions: 173 ]

Exam Name: Midrange Storage Technical Support V2

EMC CLARiiON CX3-40. Reference Architecture. Enterprise Solutions for Microsoft Exchange 2007

Chapter 10 Protecting Virtual Environments

VERITAS Storage Foundation 4.0 TM for Databases

BrightStor ARCserve Backup for Windows

Symantec Design of DP Solutions for UNIX using NBU 5.0. Download Full Version :

SAFEGUARD INFORMATION AND ENSURE AVAILABILITY WITH THE NETAPP BACKUP AND RECOVERY SOLUTION

SAP HANA Disaster Recovery with Asynchronous Storage Replication

Simplify and Improve DB2 Administration by Leveraging Your Storage System

VERITAS NetBackup 6.0 Enterprise Server INNOVATIVE DATA PROTECTION DATASHEET. Product Highlights

Network- Appliance EXAM - NS NetApp Certified Implementation Engineer - SAN, Cluster-Mode. Buy Full Product

Protect enterprise data, achieve long-term data retention

This course is intended for those who provide basic support for and perform administrative functions of the Data ONTAP operating system.

NetApp SnapManager 2.1 for Hyper-V on Clustered Data ONTAP 8.3

Veritas Storage Foundation for Windows by Symantec

Hosted Microsoft Exchange Server 2003 Deployment Utilizing Network Appliance Storage Solutions

VERITAS Volume Manager for Windows 2000

CA ARCserve Backup. Benefits. Overview. The CA Advantage

SnapManager 7.2 for Microsoft Exchange Server Administration Guide

Netapp Exam NS0-510 NCIE-Backup & Recovery Implementation Engineer Exam Version: 7.0 [ Total Questions: 216 ]

SnapManager 6.0 for Microsoft SQL Server Installation and Administration Guide

INTRODUCING VERITAS BACKUP EXEC SUITE

Microsoft E xchange 2010 on VMware

Business Continuity and Disaster Recovery. Ed Crowley Ch 12

INTEGRATING FILENET IMAGE SERVICES WITH NETAPP UNIFIED STORAGE FOR WINDOWS PLATFORM: SQL SERVER

Copy-Based Transition Guide

iscsi Technology Brief Storage Area Network using Gbit Ethernet The iscsi Standard

NCIE-Backup & Recovery Implementation Engineer

Backup Solutions with (DSS) July 2009

White Paper. A System for Archiving, Recovery, and Storage Optimization. Mimosa NearPoint for Microsoft

Data Domain OpenStorage Primer

MONITORING STORAGE PERFORMANCE OF IBM SVC SYSTEMS WITH SENTRY SOFTWARE

SQL Server Consolidation with Server Virtualization on NetApp Storage

Technology Insight Series

7-Mode Transition Tool 2.2

Symantec Storage Foundation for Oracle Real Application Clusters (RAC)

IBM N Series. Store the maximum amount of data for the lowest possible cost. Matthias Rettl Systems Engineer NetApp Austria GmbH IBM Corporation

iscsi Technology: A Convergence of Networking and Storage

Titan SiliconServer for Oracle 9i

Chapter 2 CommVault Data Management Concepts

The Data Protection Rule and Hybrid Cloud Backup

STORAGE CONSOLIDATION WITH IP STORAGE. David Dale, NetApp

IBM System Storage N3000 Express series Modular Disk Storage Systems

XenData Product Brief: SX-550 Series Servers for LTO Archives

Disaster Recovery Solutions with Virtual Infrastructure: Backup and Recovery. Paul Nashawaty Session SLN104B

VERITAS Database Edition for Sybase. Technical White Paper

DELL EMC DATA DOMAIN EXTENDED RETENTION SOFTWARE

Transcription:

BEST PRACTICES GUIDE FOR DATA PROTECTION WITH FILERS RUNNING FCP Nick Wilhelm-Olsen, Brett Cooper October 1, 2002 TR3202 TECHNICAL REPORT Network Appliance, a pioneer and industry leader in data storage technology, helps organizations understand and meet complex technical challenges with advanced storage solutions and global data management strategies. Table of Contents 1 Introduction 3 1

2 Data Protection Overview 3 2.1 Snapshot Technology 3 2.2 SnapRestore 3 2.3 SnapMirror 3 2.4 SnapVault 4 2.5 SnapDrive 4 2.6 Native Tape Backup and Recovery 4 2.7 Network Data Management Protocol (NDMP) 4 2.8 Host-Based Backup and Recovery 4 3 Configuration Models 5 3.1 Simple Host-Based Backup and Recovery Models 5 3.1.1 Non-Snapshot Capable File Systems 6 3.1.2 Snapshot Capable File Systems 7 3.1.3 Raw Partitions 7 3.1.4 LUN Snapshots 7 3.1.5 Logical Volume Managers 7 3.2 Simple Filer-Based Backup and Recovery Models 9 3.2.1 Snapshots 10 3.2.2 SnapRestore 11 3.2.3 SnapMirror 11 3.2.4 SnapVault 12 3.2.5 NDMP and Native Tape Backup and Recovery 12 4 Summary 13 5 References 13 1. Introduction Enterprise corporations around the globe are facing a wide variety of challenges. Challenges from competitors, economic conditions, resource limitations, and other factors make it difficult for a corporation to thrive and grow. For these companies to succeed, they rely on their enterprise storage vendors to provide constant data availability and absolute protection from data loss and corruption. 2

Network Appliance has always been a leader in providing the highest level of availability and protection to corporations, as well as educating customers on the best methods for protecting data in their unique environments. Network Appliance is committed to providing fast, simple, highly available, and flexible solutions. In keeping with this tradition, Network Appliance has now added support for the Fibre Channel protocol (FCP) to its core line of filers. Now, SAN and NAS technologies are available together on the same physical device with the same level of performance, availability, and simplicity that Network Appliance has provided for over a decade. The purpose of this paper is to educate the reader on the best methods for protecting filers utilizing the Fibre Channel protocol. The paper opens with a brief discussion of available data protection technology from Network Appliance. Next, it outlines potential configurations for data protection of FCP-only filers as well as filers configured for both FCP and NAS data. 2. Data Protection Overview In order to discuss models for data protection, it is necessary to first describe the technologies used in these models. This section briefly describes all of these elements. Further explanation as well as command syntax and examples can be found in the Data ONTAP Data Protection Guide, available online at the Network Appliance NOW Web site registration is required. 2.1. Snapshot Technology The Network Appliance Write Anywhere File Layout, or WAFL, supports Snapshots, a unique feature that allows administrators to maintain multiple read-only versions online per filer volume. Snapshots, an included component of Data ONTAP software, allow users to recover accidentally damaged or deleted data by utilizing a version of the data from within a Snapshot directory. Snapshots are described in detail in the paper entitled SnapMirror and SnapRestore : Advances in Snapshot Technology at http://www.netapp.com/tech_library/3043.html. 2.2. SnapRestore SnapRestore software leverages the Snapshot feature of Data ONTAP software by restoring a file system to an earlier preserved state. It can be used to recover from a corrupted database or application, or damaged file system. The system administrator can restore a single logical unit number (LUN) or the entire volume containing the LUN to any existing Snapshot. Restored LUNs are available for full production use without rebooting the filer, having returned to the precise state that existed when the selected Snapshot was created. Restoration will take less than a minute regardless of the size of the LUN or volume being restored. 2.3. SnapMirror SnapMirror provides a fast, flexible enterprise solution for asynchronous mirroring or replicating data from one filer to another over local or wide area networks. SnapMirror transfers Snapshots from specific points in time to other filers or NearStore appliances. These replication targets can be in the same data center connected via a LAN (local area network) or distributed across the globe connected via MAN (metropolitan area network) or WAN (wide area network) networks. Since it operates at the changed block level instead of transferring entire files or file systems, it greatly reduces bandwidth and time requirements for replication. 2.4. SnapVault 3

SnapVault software builds on the asynchronous, block-level incremental transfer technology of SnapMirror by providing archival technology. This allows data to be backed up via Snapshots on a filer and transferred on a scheduled basis to a destination filer or NearStore appliance. These Snapshots can be retained on the destination for many weeks or even months, allowing recovery operations to occur nearly instantaneously from this destination system to the original filer. 2.5. SnapDrive TM SnapDrive software offers a rich set of capabilities to virtualize and enhance storage management for Microsoft Windows environments. It is tightly integrated with Microsoft NTFS, providing a layer of abstraction between application data and physical storage associated with that data. It facilitates creation and management of virtual devices on the filer, and allows the creation and management of Snapshots either from the Microsoft MMC or via a command line interface. Storage managed by SnapDrive logically appears to the Windows host as locally attached storage. In reality, the capacity comes from a centrally managed pool of networked storage equipped with enhanced attributes. SnapDrive runs on a Windows 2000 host and complements native NTFS volume management with virtualization capabilities. It allows administrators to easily create virtual disks from pools of storage that can be distributed among several storage appliances. SnapDrive is included at no additional charge with the NetApp Windows Host Attach Kit 1.0. 2.6. Native Tape Backup and Recovery Network Appliance filers support backup and recovery from local, Fibre Channel, and Gigabit Ethernet SAN-attached tape devices. Support for most existing tape drives is included as well as a method for tape vendors to dynamically add support for new devices. In addition, the RMT protocol is fully supported, allowing backup and recovery to any capable system. Backup images are written using a derivative of the BSD dump stream format, allowing full file system backups as well as nine levels of differential backups. 2.7. Network Data Management Protocol (NDMP) The Network Data Management Protocol, or NDMP, is an open standard for centralized control of enterprise-wide data management. The NDMP architecture allows backup application vendors to control native backup and recovery facilities in Network Appliance filers and other file servers by providing a common interface between backup applications and file servers. NDMP separates the control and data flow of a backup or recovery operation into separate conversations. This allows for greater flexibility in configuring the environment used to protect the data on Network Appliance filers. Since the conversations are separate, they can originate from different locations, as well as be directed to different locations, allowing greater flexibility in designing NDMPbased topologies. Available NDMP topologies are discussed in detail in the Data Protection Solutions Overview at http://www.netapp.com/tech_library/3131.html. 2.8. Host-Based Backup and Recovery Host-based backup and recovery is the most common method for protecting data stored within Fibre Channel LUNs. Backup and recovery operations are performed by systems that have access to the LUN, rather than the filer where the LUN resides. Since the host or application server can understand and interpret the data within the LUN, this will be the recommended method for data protection for most applications as it allows more precise backup and recovery operations. 3. Configuration Models 4

Presented below are some of the possible models for data protection with FCP-enabled filers. This is not intended to be a complete list, but rather a collection of some of the best models for most environments. Using the guidelines and recommendations presented in this document as well as other documents from Network Appliance, it should be possible to adapt these configurations to each unique environment. Initially host-based data protection models are examined, specific to the type of data accessed via FCP. Next is a discussion of filer-based protection models and situations where they should be implemented. Finally, models using a NearStore appliance as a staging area for backups are presented for best performance. 3.1. Simple Host-Based Backup and Recovery Models Determining the best model of host-based backup and recovery is entirely dependent on the class of applications in use and how they are configured. Applications are considered below by the method used to organize data, such as UFS, NTFS, Snapshot image-capable file systems such as VERITAS File System (VxFS) or raw partitions. Logical volume managers such as VERITAS Volume Manager (VxVM) will also be discussed. The diagram below represents a network and storage design for the servers and devices used in these models. For simplicity, this diagram does not reflect the Ethernet management network that must exist for managing Network Appliance filers. Application Application Backup Figure 1) Host-Based SAN Backup/Recovery Diagram The above diagram depicts a SAN with two application servers, two filers, and a backup server all redundantly attached to two fabric switches, avoiding single points of failure. The backup server is also attached to a separate SAN for tape backup. For simplicity, this SAN is not depicted with redundant connections although that is recommended for a highly available backup environment. Zoning can be configured to support attaching tape devices to the SAN if using a third switch is not desired. 5

Tape libraries and drives can either be SCSI- and/or ACSLS (Automated Cartridge System Library Software)-attached to the backup server or attached through a Fibre Channel SAN as depicted above. 3.1.1. Non-Snapshot Capable File Systems Applications that are implemented on top of ble file systems that are not capable of Snapshots, such as UFS or NTFS, represent the simplest scenario from a backup and recovery perspective. When backing up these applications, they must first be quiesced, or taken offline, to avoid open files or files changing during the backup operation. Then, file system caches must first be committed before the backup operation commences. The application remains quiesced or offline until the backup is completed, at which point normal application operation can resume. This can result in a significant period of unavailability for the application. Some applications have a built-in hot backup mode, allowing a backup to occur while the application operates at a reduced efficiency and often limited capabilities. This type of mode is typical in messaging and database applications such as Microsoft Exchange and Oracle. It will result in higher overall application availability than without using hot backup mode. However, it can still potentially result in a long interval of reduced efficiency and limited performance. Further instructions for backup and recovery of databases utilizing Network Appliance SAN configurations can be found in the Network Appliance Technical Library. One final alternative is to use an open file manager from a backup software vendor. These applications are designed to handle backup operations on files that are still locked by an application. They work well for simple applications such as home directories and shared documents. However, they should be avoided with complex applications, such as messaging applications or databases. 3.1.2. File Systems Capable of Snapshots The previous section discusses some methods to avoid potentially long periods of application unavailability, but that may result in reduced performance and function during backup. Utilizing a file system such as VERITAS VxFS that allows the creation of Snapshots of the active file system can help decrease application downtime. By backing up a Snapshot of the file system, the original file system can remain online. This minimizes any application downtime or periods of reduced functionality required during backup operations. File systems capable of Snapshots are frequently used for applications that do not provide a hot backup mode. These applications must normally be taken offline for the duration of the backup operation, usually referred to as a cold backup. Cold backups are also preferable in database environments where rollforward recovery is not possible or enabled. With file systems capable of Snapshots, the application or database is only taken offline for the duration of Snapshot creation instead of the entire backup operation. The Snapshot file system can also be mounted and used directly should the original file system become damaged or corrupt. The Snapshot of the file system will take little time to create and restore. The Snapshot occupies little additional disk space, since it is only a copy of the changed blocks from a point in time of the active file system. For most file systems, a volume or LUN containing approximately 15% of the original file system size will be adequate to hold the Snapshot and any Snapshot iterations. When a Snapshot is backed up, the data that has not changed is read from the original file system. Any changed data is read directly from the Snapshot. This may result in a small degradation of application 6

responsiveness since both blocks from the active file system and changed blocks in the Snapshot are read during backup. 3.1.3. Raw Partitions Some applications have the ability to work directly with raw device partitions, allowing significant performance benefits. In order to back up this application data, however, it is necessary to use specialized modules from a backup software vendor tailored for each application. Without these modules, there is no means of knowing if the data is in a consistent state for backup. When backing up raw partitions, it is usually necessary to place the application in hot backup mode to ensure that the data is consistent. As there is no file system exposed outside of the application, there is no method of performing Snapshots from the host system. Thus, the application must remain in hot backup mode for the duration of the backup operation. Further instructions for backup and recovery of databases utilizing Network Appliance SAN configurations can be found in the Network Appliance Technical Library. 3.1.4. LUN Snapshots Data ONTAP provides another method to increase application availability during backup operations. It is possible to use the WAFL Snapshot technology to provide Snapshots for the LUN, regardless of whether or not the file system on the LUN provides Snapshots or the LUN is used as a raw device. First, the application is quiesced or placed in hot backup mode, then file system caches are committed. Next, a Snapshot is created on the filer volume containing the LUN. Normal application operation can now resume. A new LUN is created that points to the LUN within the created Snapshot. This LUN can now be backed up at any convenient time. When backup is complete, the LUN and Snapshot are destroyed. Since WAFL Snapshots occur almost instantaneously, this solution offers a significant benefit in availability, as the application is only quiesced, in hot backup mode, or, during Snapshot creation, in cold backup mode. In addition, LUN Snapshots can be combined with SnapMirror or SnapVault software to transfer these Snapshots to an alternate filer to completely offload from the source filer any overhead from data protection on the LUN. 3.1.5. Logical Volume Managers In the previous four configurations, file systems and partitions were restricted to one LUN and, therefore, one filer. With a logical volume manager such as VERITAS VxVM, it is possible to create file systems and partitions that span many LUNs and even filers. This can be done to increase performance by striping multiple LUNs together to appear as one larger LUN. It can also be done to increase availability by mirroring LUNs. Both operations can even be done simultaneously to mirror concatenated or striped LUNs for performance and availability. Mirroring can also be used for performing backups. Mirroring increases application availability and can be used to offload backup overhead from application and file servers. When using a logical volume manager for backup and recovery, it is suggested that a mirror be initiated when a volume is created. This will reduce the time required to mirror the data across volumes before backup operations commence. The difference in performance can be significant with large databases. For example, a 500GB database residing on a volume can take upwards of 15 minutes to mirror to another volume. Using mirroring, the application is quiesced or placed in hot backup mode, then file system caches are committed. The mirror relationship is broken and normal application operation resumes. At this point, 7

the split mirror can be backed up with a host-based backup application at a convenient time. Once the backup is completed, the mirror can be resynchronized to capture any changes to the data that may have occurred while the mirror relationship was broken. This process is known as split mirror resynchronization. It will degrade the performance of the primary volume being mirrored until resynchronization is completed. As a result, these activities should be scripted and scheduled during periods of low activity, such as late in the evening or on the weekend. Application Application Backup Filer 1 Striped Volume Filer 2 across two filers 01 01 using logical volume manager 01 01 LUN0 LUN1 (volstrip1) LUN2 LUN3 (disk01-01)(disk02-01) (disk03-01)(disk04-01) Mirrored Volume on Filer 2 Split for Backup Resync'd After Backup (mirvol1) 01010101 Filer 2-LUN4 (disk05-01) Figure 2) Backup Using a Logical Volume Manager and a Split Mirror 3.2. Simple Filer-Based Backup and Recovery Models In certain situations where file systems or raw partitions consist of a single LUN, it will be desirable to perform backup and recovery operations directly on the filer instead of through the SAN. Using filerbased backup and recovery offloads the majority of backup overhead from the host system as well as 8

the application SAN. In addition, full backup and recovery operations are significantly faster when performed native to the filer instead of through the application SAN. Filer-based data protection should not be considered in situations where file systems span multiple LUNs via VERITAS Volume Manager or a similar product. In addition, they should not be considered when selective restoration of files within the LUN is a requirement. Filer-based data protection models always operate at the LUN level. The filer has no method of interpreting the data within the LUN itself. Incremental operations will properly transfer all changed blocks within the LUN. The diagram below represents a network and storage diagram for the servers and devices used in the filer-based data protection models below. For simplicity, this diagram does not reflect the Ethernet management network that must exist for managing Network Appliance filers. Application Application Legend Application SAN NDMP Tape SAN Ethernet Backup Network Backup Figure 3) Filer-Based SAN Backup/Recovery Diagram The above diagram depicts a SAN with two application servers and two filers redundantly attached to two fabric switches, avoiding single points of failure. The filers are also attached to a second SAN for tape backup. The backup server and the filers are all connected to a private Ethernet network for controlling backup and recovery operations. Zoning can be configured to support attaching tape devices to the SAN if using a third switch is not desired. The library and drives could alternatively be SCSI-attached to the filers or attached via Gigabit Ethernet if an NDMP-compliant tape library is in use. 3.2.1. Snapshots 9

Snapshots are the most basic element of all filer-based data protection scenarios discussed in this paper. All of the methods discussed below operate with either explicitly named Snapshots, or will create, use, and destroy Snapshots as required. Since the internal scheduling mechanism for Snapshot creation within Data ONTAP has no means of ensuring that the data within a LUN is in a consistent state, scheduled Snapshots should be disabled for all volumes that contain LUNs. Either SnapDrive or batch files and scripts should be used on a host that has administrative access to the filer. These scripts must first make the data within the LUN consistent. This could be accomplished by quiescing a database, placing an application in hot backup mode, or even taking the application offline in extreme cases. Next, use the rsh or ssh command to create the Snapshot on the filer. Snapshot creation will take only a few seconds, regardless of volume size or utilization. Finally, the application is returned to normal operation. The UNIX cron or Windows Task Scheduler service can then be used to execute this script at specified intervals. To recover a LUN from a Snapshot, the LUN can be cloned as described in the Data ONTAP SAN System Administrator Guide available online at the Network Appliance NOW Web site registration is required. Cloning a LUN from a version contained in a Snapshot will cause that Snapshot to be locked. Thus, the Snapshot can only be deleted if the LUN is destroyed and recreated. LUN cloning is typically only done when the cloned LUN is needed for a short amount of time. If the LUN needs to be restored to a previous version permanently, the SnapRestore feature discussed in the next section should be used. 3.2.2. SnapRestore SnapRestore will revert a LUN or an entire volume containing LUNs to a previous state in only a few seconds, regardless of size or utilization. Furthermore, the SnapRestore operation takes effect immediately, without requiring a reboot of the filer. When restoring a single LUN, it must be taken offline or unmapped prior to recovery if the LUN has not been destroyed. This makes SnapRestore ideal for environments where very large LUNs are in use, or environments where transaction logs are maintained, such as database applications. Host-based methods would require that all the data within the LUN be recovered block by block or file by file. This will be extremely time-consuming for large LUNs. On the other hand, SnapRestore can revert the LUN to a previous version nearly instantaneously. SnapRestore dramatically increases application availability, providing extremely fast recovery times for business-critical applications such as databases and groupware. 3.2.3. SnapMirror SnapMirror asynchronously transfers volumes or qtrees from a source filer to a destination filer or NearStore. This involves an initial baseline transfer, and then block-level incremental updates thereafter. As a result, SnapMirror works very well in LAN and WAN environments where bandwidth is limited. Normally, SnapMirror Snapshot creation and updates are controlled by a schedule local to the filer. Due to the issues discussed previously, Snapshot creation with LUNs must be controlled by host systems. Scripts are set up to create Snapshots as previous mentioned and then initiate the SnapMirror update to the remote filer or NearStore appliance. These operations can be scheduled on the host system using the UNIX cron utility or Windows Task Scheduler service. 10

SnapMirror is often used for disaster recovery (DR) planning. Here, volumes containing LUNs would be mirrored asynchronously to a filer at a DR facility. Preferably, application servers would be mirrored to this facility as well. In the event that the DR facility needs to be made operational, applications can be switched over to the servers at the DR site and all application traffic directed to these servers as long as necessary to recover the primary site. Once the primary site is online, SnapMirror can be used to transfer the data efficiently back to the production filers. After the production site takes over normal application operation again, SnapMirror transfers to the DR facility can resume without requiring a second baseline transfer. SnapMirror can also be used for backup consolidation or offloading tape backup overhead from production servers. This facilitates centralized backup operations, reducing backup administrative requirements at remote locations. It can also dramatically reduce overhead from stressful backup operations caused by small backup windows on production filers. Since backup operations are not occurring on the production systems, small backup windows are not as important. 3.2.4. SnapVault SnapVault builds upon the Network Appliance SnapMirror software by providing greater flexibility in the retention of Snapshots on the destination. This makes SnapVault ideal for backing up or archiving Snapshots to a filer or NearStore device. SnapVault can be used in this manner to perform disk-based backups of LUNs to a second filer or NearStore appliance, greatly reducing the time to back up or recover these LUNs when compared to tape-based backup and recovery. Keep in mind that SnapVault as well as all host-based methods can only back up and recover an entire LUN. SnapVault will work well with applications that maintain transaction logs, such as databases. The LUN is restored and then transaction logs are rolled forward to bring the LUN to the desired state. With SnapVault, there can be two separate schedules maintained. One schedule controls creation of the Snapshots on the source filer. The second schedule controls transfer of these Snapshots to the destination filer and the amount of time to retain them. It is recommended that Snapshots on the source filer be created manually when FCP data is contained on the file system. Scripts should be used to quiesce the application or place it in hot backup mode before creating a Snapshot. Once the Snapshot is created, the application should be returned to normal operation. The UNIX cron or Windows Task Scheduler service can then be used to execute these scripts at specified intervals. 3.2.5. NDMP and Native Tape Backup and Recovery Tape backup and recovery operations of LUNs should generally only be performed on the filer for disaster recovery scenarios, for applications with transaction logging, or when combined with other filer-based data protection elements, such as SnapMirror and SnapVault. All tape operations local to the filer operate on the entire LUN and cannot interpret the data or file system within the LUN. Thus, LUNs can only be recovered to a specific point in time if transaction logs exist to roll forward. Where finer granularity is required, host-based backup and recovery must be used. If the operator does not specify an existing Snapshot when performing a native or NDMP backup operation, the filer will create one before proceeding. This Snapshot will be deleted when the backup completes. When a file system contains FCP data, one should always specify a Snapshot that was created at a point in time when the data was consistent. As mentioned earlier, this is ideally done in a script by quiescing an application or placing it in hot Backup mode before creating the Snapshot. After Snapshot creation, normal application operation can resume and tape backup of the Snapshot can occur at any convenient time. 11

When attaching a filer to a Fibre Channel SAN for tape backup, it is necessary to first ensure that Network Appliance certifies the hardware and software in use. A complete list of certified configurations is available in the Network Appliance Data Protection portal. Redundant links to Fibre Channel switches and tape libraries are not currently supported by Network Appliance in a Fibre Channel tape SAN. Furthermore, a separate host bus adapter must be used in the filer for tape backup. This adapter must be attached to a separate Fibre Channel switch that contains only filers, NearStore appliances, and certified tape libraries and tape drives. The backup server must either communicate with the tape library via NDMP, or have library robotic control attached directly to the backup server. Figure 3 in section 3.2 illustrates a Fibre Channel tape SAN with the tape library controlled over NDMP. 4. Summary There are many methods to back up data stored on a filer via FCP. Host-based methods provide the finest granularity of data protection, allowing selective backup and recovery of data within a LUN. Filerbased methods provide very fast and efficient disaster recovery solutions. There is no single best method meeting all data protection requirements. One must prioritize requirements and combine methods to meet and exceed these requirements. Remember that data protection is a constantly changing thing. As needs and usage change, so must data protection policies and technologies. Network Appliance is committed to providing industry-leading data protection solutions now and in the future. As needs and technology change, Network Appliance will continue to exceed customer expectations, providing the constant data availability and absolute data integrity they rely upon for success. 5. References Network Appliance Data Protection Portal http://www.netapp.com/solutions/data_protection.html Data Protection Solutions Overview http://www.netapp.com/tech_library/3131.html SnapMirror and SnapRestore : Advances in Snapshot Technology http://www.netapp.com/tech_library/3043.html Oracle9I for UNIX Backup and Recovery Using a NetApp Filer in a SAN Environment http://www.netapp.com/tech_library/xxxx.html Oracle9i for Windows Backup and Recovery Using a NetApp Filer in a SAN Environment http://www.netapp.com/tech_library/xxxx.html Network Data Management Protocol home page http://www.ndmp.org/ Network Appliance, Inc. 2002 Network Appliance, Inc. All rights reserved. Specifications subject to change without notice. NetApp, the Network Appliance logo, SnapMirror, SnapRestore, and WAFL are registered trademarks and Network Appliance, Data ONTAP, NearStore, NOW, SnapDrive, Snapshot, SnapVault, and The evolution of storage are trademarks of Network Appliance, Inc. in the U.S. and other countries. UNIX is a registered trademark of The Open Group. Oracle is a registered trademark and Oracle9I is a trademark of Oracle Corporation. Microsoft and Windows are registered trademarks of Microsoft Corporation. All other brands or products are trademarks or registered trademarks of their respective holders and should be treated as such. TR-3202 Rev. 10/02 12