Understanding Virtual System Data Protection Server virtualization is the most important new technology introduced in the data center in the past decade. It has changed the way we think about computing resources and has impacted every aspect of networks, storage and servers themselves. Due to its popularity, server virtualization is being used for all applications including large database mission critical applications. It is important to consider that server virtualization has changed many things, including data protection. In this paper we will review the history or server virtualization and its impact on data protection. We will discuss the challenge of protecting virtualized systems and review the new data protection technologies developed as a result. In the early days of virtualization, data protection was performed in the traditional manner by installing a backup agent on each virtual machine. The challenge with this approach was performance. Depending on the number of virtual machines running backup at the same time, shared resources on the host server were limited. If too many virtual machines tried to run backup at the same time, it could lead to a server crash. To solve this problem, new backup vendors emerged with products designed for virtualized systems. Without having vendor support API s to use, the new backup vendor had to interface directly with the hypervisor to perform backup. As hypervisors were altered or upgraded, the backup solutions sometimes failed. While this was acceptable to startup vendors, established backup vendors chose to wait for a vendor API to be developed. Today, hypervisor vendors each provide an API to assist with data protection. This means that the backup solutions will work despite revisions to the hypervisor code. All leading backup solutions leverage the hypervisor API s to perform agent-less backup. No longer are agents installed on each virtual machine, except in special situations which we will discuss a bit later. The good news is that reliable and high performance data protection is available for a large number of backup vendors to protect virtual machines. Virtual System Backup Technology In many ways, virtual system backup leveraged existing backup technologies such as image-based backup with snapshots, but in other ways new technology was developed to manage the task. The following is a review of key technologies that play an important role in virtual system backup. Physical vs. Virtual Systems. To begin, let s review the basic question, why virtualize? The most common reason is to reduce cost and to simplify the management of hundreds (or thousands) of systems. The use of virtualization continues to grow. ESG reported in a 2015 survey that on average organizations have deployed more than half (56%) of their servers; but the preference for physical machines will not go away. One of the main reasons that System Admins continue to deploy applications on their own dedicated server is performance. High performance database applications are a good example where not sharing compute resources is preferred. Another reason is risk. If you have ten applications, for example, on a single piece of hardware and if that goes down, it won t be good. All things being considered equal, the needs of each application must be considered before making the decision to deploy as a virtual system. Virtual Machine Backup 1
Image-based backup. Image-based backup refers to the use of snapshot technology to capture the changed blocks for an entire storage volume. Filebased backup, by contrast, copies the entire file from the file system for backup. Image-based backups are popular with database applications where large data files exist. It is more efficient to copy the changed blocks for a large database file rather than the entire file. Virtual machine files are also very large and benefit from image-based backup same as large database files. All virtual system backup is performed as imagebased backup. Agent-less Backup. We just discussed how new hypervisor API s allow for agentless backup, but some vendors still rely on agents installed on virtual machines to help with application awareness. Database applications, for example, use agents to quiesce the application and gain an error-free backup. Agents also are used for granular backup and recovery of individual files, folders or email. Some backup solutions have developed helper applications that mount the backup images automatically and allow for searching and extraction of data objects from well-known applications such as Microsoft Exchange, SQL Server and Oracle database. Changed Block Tracking. Hypervisor APIs have provide capabilities such as VMware s Changed Block Tracking (CBT) that tracks which parts of a virtual machine s image file have changed since the last backup. CBT initially backups up the entire volume. After this first full backup, CBT tracks the blocks changed since the last backup. Other hypervisor vendors have similar technology with names like delta block and blocklevel incremental backup. No matter the name, CBT reduces the amount of data to be copied during backup and allows for more frequent backup. In-place Recovery. For traditional file-based recovery, the first step is to copy the file back to the source. Virtual system backup changes this practice with inplace recovery. Because the backup of a virtual system volume contains complete virtual machines, it is not required to copy the backup image back to the source. The virtual machines can be executed on the backup device. With in-place recovery, the virtual machine and its data can be returned to operation in a matter of minutes. Steps to restore a virtual machine from a backup image vary, but the end result is the same. Keep in mind that it takes a hypervisor and sufficient computing resources to run a virtual machine, so the backup device must be designed with this in mind. Migration. System admins are commonly asked to create clones of production systems for engineering development, QA Test and Support. Configuring new system resources and installing Operating Systems and application software is a long and tedious task. With virtual system backup, this task can be greatly simplified and also be automated. Using virtual system backup, we can copy a virtual machine copy and restore it to any location. We are also free to restore across physical/virtual platforms and across hypervisors. In this manner, clones of production applications can be made quickly for use in virtual labs and backup testing. Advanced Features. Synthetic fulls. In the early days of file based backup, synthetic fulls were useful only in a small number of cases. Weekly full and daily incremental were the normal backup practice. With virtual server backups, the frequency of backup has increased to offer near-continuous backup. Changed blocks from sequential snapshots can be logically re-combined for point-in-time recovery while reducing storage overhead. Virtual Machine Backup 2
Array integration. Virtual server backup is imagebased using snapshot technology based in the hypervisor. For the best snapshot performance, enterprise storage arrays offers their own API s to perform hardware-based snapshots. Backup software that leverage features of specific arrays can provide a performance edge in your environment, provided of course you are using the proper enterprise storage array. Multiple hypervisor support. If you currently use more than one hypervisor, than you must carefully consider how each backup solution supports the hypervisor. VMware vsphere and Microsoft Hyper-V are the two leading hypervisors, but there are many others. Keep in mind, you want full API integration with the hypervisor. Some backup vendors tout support for many hypervisors installed as a guest OS, which is not optimal for performance. Cross hypervisor migration. A very useful feature is to move from back and forth between the two most popular hypervisors, vsphere and Hyper-V. The machines formats for each are different (VMDK and VHD) so a conversion is required. One reason to move between hypervisors is cost. Many end users and service providers have a Microsoft Data Centre Edition license that includes unlimited use of Hyper-V built into their agreements, by leveraging this instead of been forced to purchase expensive Enterprise VMware licenses represents substantial cost savings for DR. Backup Verification. Best practice is to always verify your backups. Virtual system backup makes this task easy and full automated. To verify a virtual backup image, the volume contained in the backup image is mounted. If it mounts successfully, then it is assumed that the data is not corrupt. You can step an additional step and start up a virtual machine, but generally this is not necessary if your goal is to simply verify that the backup data is not corrupt. Physical System support. Does your infrastructure include both physical and virtual systems? The answer is likely yes and as a result you likely deploy multiple backup products. New next generation backup products from the established backup vendors, provide robust virtual server backup in addition to traditional physical system protection. Look for a backup solution that can support both physical and virtual and simplify your backup operations. Deduplication. Deduplication and compression is an important part of modern backup solutions. It not only helps to reduce backup target storage consumption, but it also reduces the amount of data transmitted over your network. Virtual system backup products include some form of deduplication. Pay attention to products that limit dedupe to a single backup job or a single storage volume. True deduplication should perform across your entire network and the dedupe process should begin at the source node for maximum efficiency. Encryption. Encryption both in-flight and at rest, has become commonplace on storage devices. Encrypting backups is a best practice, especially if data is to be moved off-site in any form. As a general rule, encryption is best done at the target hardware level (i.e., the backup appliance or tape drive) to offload the processing from the CPU. Tape support. Not all virtual server backup products support tape directly. If tape support is a requirement, it may be necessary to have two products, one for the virtual environment and one to manage the tape copies. Modern backup solution from established backup vendors will be the best source of robust tape support. Cloud Support. A public or private hosted data center is an excellent complement to your on-premise backup solution for long-term backup retention and disaster recovery-as-a-service (DRaaS). For long-term backup retention, backup images are replicated to the cloud where they are can be stored long-term for business or regulatory needs. Using in-place recovery, virtual machines can be executed in the cloud to provide disaster recovery for critical applications. Always remember that a successful DR plan always includes sufficient planning and regular testing. Virtual Machine Backup 3
7000 WHITE PAPER Appliance Option. Modern backup solutions are increasingly offering appliance options to their backup family. The advantage of an appliance is having a single point of purchase for the entire backup solution. You no longer have to provision the server and storage hardware yourself. Another advantage of an appliance is that the configuration is tested by the vendor for optimal performance. Modern backup solutions that support in-place recovery require the right combination of compute resources (e.g. CPU, RAM and SSD) that an appliance provides. If your infrastructure includes both physical and virtual systems, you likely deploy multiple backup products. New, next-generation backup products provide robust virtual server backup in addition to traditional physical system protection. Best Practice Don t back up virtual machines at the guest OS layer. With traditional backup you install a backup agent on the guest operating system. But this method isn t efficient in a virtual environment because it causes unnecessary resource consumption on the virtual machine (VM) that can impact its performance, as well as the performance of other VMs running on the host. Best practice is to use image-level backups that back up the large virtual machine without involving the guest OS. Leverage the vstorage APIs. The hypervisor APIs are designed to help offload backup processing from the host. Not only do they allow for easier access to the virtual machines, but they also contain features that can improve backup performance like CBT and Virtual Volumes (VVOLs). The hypervisor APIs provide a much more efficient mechanism for backing up VMs and you should use backup applications that take full advantage of them. Know how quiescing works. If you re backing up VMs that have database applications, it s critical that you quiesce the application so it is in the proper state to be backed up. This ensures the application is in a proper state so no data is lost if a restore is needed. Quiescing works with applications that specifically support being told to pause and write pending data when necessary. Make sure you are using the VMware Tools driver or have installed the vendor supplied driver on your VMs to properly perform an application-consistent backup. Don t skimp on backup resources. To ensure that you have optimal backup performance, make sure you have adequate hardware for your backup server so it doesn t become a bottleneck when performing backups. While having adequate network bandwidth is critical, having enough CPU and memory resources is, too. Your backup server is doing more than just moving data from a source server to a target storage device, it s also doing things like in-place recovery and deduplication. These types of processes require a lot of CPU and memory to help keep up with all the data that is flowing through the server. Learn more about virtual server protection at arcserve.com or call +1.844.639.6792. Virtual Machine Backup 4
Introducing Arcserve Unified Data Protection Arcserve UDP delivers comprehensive Assured Recovery for virtual and physical environments with a next generation unified architecture and unmatched ease of use. This feature-rich solution enables organizations to scale their IT environments easily, while delivering against their recovery point and recovery time objectives (RPO and RTO), onpremise, off-premise or in the cloud. Traditional, overly complex backup policies are replaced by modern task-based data protection plans and deep integration with hypervisors to fully automate complex, mundane repetitive tasks. Arcserve UDP is the first solution to combine industry-proven backup, replication, high availability and true global deduplication technologies within one solution and unlike competitive products, does not require expensive professional services. Fast, simple Agent-less backup for VMware and Hyper-V host environments Arcserve UDP provides single-pass backup of all Windows and Non-Windows VMs, without the need to install software agents on each virtual machine (Guest OS). For VMware and Microsoft Hyper-V environments, Arcserve UDP uses a changed block tracking mechanism that tracks changes to the virtual disk file of a Virtual Machine. With changed block tracking, only the changed or used blocks of the virtual disk are actually read, which improves performance significantly, as well as reducing overall storage requirements. Arcserve UDP leverages the VMware vstorage APIs and Microsoft VSS snapshot technology to help ensure file system and application consistent backups. Arcserve UDP offers a wide array of functionalities to enhance data protection easily and efficiently for VMware and Hyper-V host environments: Single-pass backup of all VMs without the need to install software agents on each VM Auto-discovery of VMs on target hypervisor Integrated with VMware vstorage API s; Only backs up changed blocks Application consistent (Exchange, SQL etc.) and transaction-log purge In-place recovery of virtual machines local or remote Easy recovery of individual files and folders from within each VM Centralized node, group and plan management from Arcserve UDP Console Arcserve Cloud Extend Arcserve Unified Data Protection with Arcserve Cloud Service for Cloud storage and Disaster-as-a-Service (DRaaS). Arcserve Cloud ensures that your data is always protected and your business is Always on in the event of disaster. For more information on arcserve please visit arcserve.com Copyright 2015 Arcserve (USA), LLC and its affiliates and subsidiaries. All rights reserved. All trademarks, trade names, service marks and logos referenced herein belong to their respective owners. This document is for your informational purposes only. Arcserve assumes no responsibility for the accuracy or completeness of the information. To the extent permitted by applicable law, Arcserve provides this document as is without warranty of any kind, including, without limitation, any implied warranties of merchantability, fitness for a particular purpose, or non-infringement. In no event will Arcserve be liable for any loss or damage, direct or indirect, from the use of this document, including, without limitation, lost profits, business interruption, goodwill or lost data, even if Arcserve is expressly advised in advance of the possibility of such damage.