NetApp - News - Tech OnTap Newsletter


 
Love VMware but Hate Backups?

Vaughn Stewart
Systems Engineer and Resident VMware Expert

Vaughn Stewart Vaughn has been a MSSP consultant and systems administrator architecting and supporting a wide range of solutions including VMware virtualization software. Vaughn is currently certified on solutions offered by Microsoft, Cisco, Sun Microsystems, IBM, and Network Appliance. Vaughn joined NetApp in 2000 and during the past six years he's worked with hundreds of customers and assisted directly with numerous VMware implementations and joint solutions.


The below article was originally published in the June 2006 edition of the Tech OnTap newsletter. To receive the newsletter monthly and enjoy other great benefits, sign up today.

Virtual infrastructures provide a simple solution to the challenges associated with distributed server architectures. More than 20,000 enterprises worldwide use server virtualization software, and the launch of VMware Infrastructure 3 with native iSCSI SAN and NAS support makes this type of consolidation a viable option for companies that hadn't previously considered it.

IT teams that deploy virtual infrastructures usually report that server virtualization makes their lives easier overnight by lowering the total amount of physical servers, network ports, floor space, maintenance contracts, and electricity required to run their data center operations. Once the honeymoon is over, however, many of these same IT teams find themselves trapped in a backup nightmare.

Whether you're considering an initial foray into server virtualization or are already feeling trapped in a love/hate relationship, backup and recovery is a critical consideration. To answer some of the top questions posed by NetApp customers, Tech OnTap turned to Vaughn Stewart. Vaughn coauthored the new technical report NetApp and VMware ESX Server 2.5.x: Building a Virtual Infrastructure from Server to Storage (PDF) with NetApp Professional Services Consultant Mike Slisinger.

Q. I've consolidated 10 physical servers onto a single VMware ESX server. At first it was amazing: we cut hardware and increased CPU utilization, and not a single end user noticed the difference. But now the fact that a single service disruption could impact my entire environment is keeping me awake at night. Do I have to live with this level of insecurity?

Vaughn: Unfortunately your concerns are well founded: with any consolidation effort you are in essence placing more eggs in your proverbial basket, so the redundancy and resiliency of your consolidation platforms are critical decisions in your design. The first step to upgrading your virtual infrastructure platform involves switching from local to shared (or networked) storage.

Here are a couple benefits of networked storage for consolidated environments:
  1. Drive failure protection. It's a sad fact of life that drives fail. To get around this, many companies move to a form of data protection that is more resilient than the RAID5 most physical servers are deployed with. There are three key options: RAID10, RAID6, and RAID-DP™.

    Which RAID option is right for you? All three protect against double disk failures without impacting the availability of your virtual infrastructure. Unfortunately, RAID10 requires 100% storage overhead (which doubles your storage costs), and RAID6 suffers from significant performance overhead. The most efficient option is RAID-DP technology, which offers protection from simultaneous double drive failures while providing the high performance of RAID10 and the cost savings of RAID6. (See a related Tech OnTap article about disk drive resiliency.)

  2. Consolidated backups. Since networked storage enables multiserver access, it makes your life significantly easier when it comes to backups and workload management.
The bottom line is that shared storage will let you build in greater protection and take advantage of more VMware features and management tools. Combining VMware VirtualCenter management software with shared (or networked) storage gives you the ability to dynamically move workloads around without service disruption.

By the way, if you're not currently using VirtualCenter, stop reading this article now and download it right away. VirtualCenter is an incredibly useful tool for managing multiple virtual machines, and, if you're a registered VMware customer, a 30-day license is 100% absolutely free. I guarantee you won't want to give it up.

Q. I've been using VMware for 12 months with about 50 virtual server machines. We've got less hardware, and my backup processes haven't changed, but suddenly I'm spending almost all my time managing backups. How do I stop the madness?

Vaughn: Many customers find that their consolidations are so successful they outgrow the backup bandwidth of their virtual infrastructure. Consolidating tens of physical servers onto a single server often creates a disproportionate ratio of data to physical server bandwidth. While not all VMware deployments will initially face backup challenges, it has quickly become a hot-button issue in the virtualization community. Check out conference agendas from VMworld 2005 and note the number of times the word backup is on the agenda.

Unless you want to permanently devote your nights and weekends to backup processes, you're going to have to make some changes to how you approach backups.

Today, users of VMware's ESX Server have two options for backing up virtual machines: backing up virtual storage either as physical storage or as files. Each alternative solves the problem of protracted backups, but both options have drawbacks when it comes to recovery. Some customers are waiting for the release of the new VMware Consolidated Backup product which is part of VMware Infrastructure 3, but the new software suite was not yet available when this article was published. NetApp technology can make any of the existing solutions work today and, if you want to start doing consolidated backups right now, you can.

Q. I'd love to move to disk-based backup, but it looks too rich for my budget. One vendor tells me I'd need as much as 100% additional storage for each backup.

Vaughn: Those vendors are doing the best they can with their traditional disk-to-disk (D2D) backup solutions. If you are backing up the files that make up the VM (the virtual disks and configuration files), most vendors can only offer two choices: (1) a "mirror copy" approach, which offers optimal performance but requires a 100% additional storage requirement for every backup; or (2) a "copy out" snapshot option, which only copies changed data blocks but comes at such a performance penalty that you'll be hard pressed to find any of their customers actually doing it.

The inherent negative features of traditional disk-based backups do not apply in NetApp storage environments. The unique NetApp Snapshot™ technology enables you to make instantaneous, full backups without performance penalties and at a cost of only a fraction of the original disk space. When combining NetApp Snapshot copies with VMware ESX server, customers are able to back up their entire virtual infrastructure within seconds. (Stephen Manley explains NetApp backup efficiency at length in his Tech OnTap article about smarter backups.)

NetApp Snapshot technology opens up a number of other possibilities. Snapshot copies can be backed up to tape and/or replicated to another facility, VMs can be restored, individual files may be recovered, and clones can be instantly provisioned for test and development environments. You can also utilize this solution with VMware SAN LUNs (which are also referred to as raw device mappings, or RDMs) to create a consolidated backup solution.

And before you ask, yes, there is a large difference between NetApp Snapshot technology and implementations based on copy-out snapshot technologies. In a NetApp environment, Snapshot copies do not move data and, as a result, have no performance impact on production storage.

If you'd rather not take my word for it, check out the recent Mercer report Decision-Maker Perspectives on EMC®, HP® and NetApp Storage Solutions in Enterprise Database Environments (pdf). Page 6 documents how a typical database environment utilizing NetApp storage and Snapshot technology would require 1.44x overhead capacity versus 2.33x required with equivalent solutions from EMC and HP.

Q. Now you tell me. I've already invested in another type of environment and don't have the time or budget for a forklift change. Any suggestions?

Vaughn: Absolutely! NetApp provides a solution to backing up both physical and virtual machines with its Open Systems SnapVault® (OSSV) software.

OSSV is a unique solution in the backup industry as it provides individual system backups, which are stored on NetApp storage devices. OSSV provides a traditional file-level backup, yet it does so utilizing block-level incremental updates after the initial full backup has been completed. It can be used in either a local storage environment or with your existing SAN storage, and it enables you to significantly reduce the amount of data being backed up nightly.

You can combine OSSV software with storing your VMware environment on NetApp primary storage for a solution that provides instantaneous virtual machine restores and classic file-level recovery. Implementing OSSV results in a significant reduction of data transferred in and out of the ESX Server, thus allowing backups to complete quickly and eliminating the need for any artificial capacity limits. Restoration of complete VMs or individual files can be completed easily and in a traditional manner. In addition, OSSV is able to back up VMs that are not stored on NetApp storage systems, thus allowing for an easy integration into any existing ESX environment.

Then you can take that expensive third-party storage that you were using for backups and repurpose it for more demanding database or I/O-intensive applications.

Q. So how do I learn more?

Vaughn: I thought you'd never ask. You can find out more about all of these topics by reading NetApp Storage and VMware ESX Server 2.5.x: Building a Virtual Infrastructure from Server to Storage (PDF). This technical paper provides multiple examples and scenarios plus in-depth discussions on:
  • Realizing thin provisioning and extending beyond physical storage limits
  • Backing up virtual machines as physical servers
  • Backing up virtual machines as files
  • Enhanced backups with raw device mappings (RDMs)

The information in this article is useful to me.
1 2 3 4 5
Strongly Disagree Strongly Agree

What other types of information would you like to see included in this article?


Additional comments:

  Thanks! 



 
© 2007 Network Appliance, Inc. | All rights reserved. | Specifications subject to change without notice.
NetApp, the Network Appliance logo, the bolt design, Data ONTAP, DataFabric, FAServer, FilerView, FlexClone, FlexVol, MultiStore, NearStore, NetCache, SecureShare, SnapDrive, SnapLock, SnapManager, SnapMirror, SnapMover, SnapRestore, SnapValidator, SnapVault, SyncMirror, VFM, and WAFL are registered trademarks and Network Appliance, ApplianceWatch, BareMetal, Camera-to-Viewer, Center-to-Edge, ContentDirector, ContentFabric, EdgeFiler, FlexShare, HyperSAN, InfoFabric, NetApp Availability Assurance, NetApp ProTech Expert, NOW, NOW NetApp on the Web, RoboCache, RoboFiler, SecureAdmin, Serving Data by Design, SharedStorage, Smart SAN, SnapCache, SnapCopy, SnapDirector, SnapFilter, SnapMigrator, Snapshot, SnapSuite, SohoCache, SohoFiler, The evolution of storage, Vfiler, Virtual File Manager, and Web Filer are trademarks of Network Appliance, Inc. in the U.S. and other countries. All other brands or products are trademarks or registered trademarks of their respective holders and should be treated as such.

Contact Network Appliance:
New Customers: 800.443.4537
Current Customers: 888.263.8277  Prompt 1 for Service; Prompt 2 for Product
Federal Customers: 888.352.2996  Prompt 1 for Service; Prompt 2 for Product
Resellers: 888.317.6294