6 Signs You May Be Outgrowing Your Virtual Backup Solution
This post will give you answers about how to improve your virtual backup performance, but first, some questions...
How often do you create new virtual machines? Will there be more data in your enterprise tomorrow than there is today? What about next month? Next year? On the subject of next year, do you think you'll have more or fewer applications and business systems, and do you think they will be more or less sophisticated?
Enterprise systems, data and virtual environments all tend to grow in size and importance over time. With that in mind, consider just two more questions: A year from now, when systems, applications and retained data are all larger, will there be more than 24 hours in a day? As systems grow, is your backup capability keeping pace?
You can easily outgrow your virtual backup & recovery solution because virtual environments can grow so quickly, especially now, when VDI and self-provisioning are growing. Businesses outgrow their backup solutions all the time -- often without even realizing it. They also don't realize the inefficiency and risks that come from using a virtual backup system that is not appropriately scaled to the environment.
Here are six warning signs that your backup and recovery solution and processes aren't the right fit for your virtual environment, and how you need to remedy them:
1, Backup Windows are Harder to Hit
The inability to complete regularly scheduled backups in the time available is a sure sign you may be outgrowing your backup solution. If backups are consistently taking longer than they used to, it is an indication that the solution doesn't have the processing power and features needed to handle all the VMs and data in the environment. Taking longer to back up is not necessarily a problem, but often does cause other problems. For example, backups that are scheduled to run during low-use periods could cause system slowdowns and inconvenience users if they exceed their backup windows and spill over into higher-use work periods. This is a big concern with methods that require VMs to be powered down to be backed up. Resource allocation conflicts, which occur more frequently when backup windows are missed, can cause backup jobs to fail.
Remedy: There are many things you can do to improve backup speed. Shrinking the VM size using compression and/or deduplication reduces the backup time (scroll down for previous posts on how to use these methods effectively). A virtual appliance architecture can also save time. Solutions that can back up multiple VMs simultaneously are also highly advantageous. Carefully think through your backup strategy -- maybe you are conducting full backups when incrementals would be acceptable.
2. The Proxy Server is a Bottleneck
Slow-performing proxy servers are sometimes the culprit behind the inability to hit backup windows. Virtual backup is typically configured to execute on a proxy server, which then transfers the backup files to the storage target. The proxy server is thus a hub for all backup activity, and can easily become a bottleneck as systems scale.
Remedy: Reconfigure your system for direct-to-target backup that bypasses the proxy server.
3. Throughput Concerns Are Limiting Virtualization
Are you finding yourself reluctant to virtualize more servers or mission-critical applications because you have concerns about your ability to back them up completely and recover them quickly? That's a sign your backup solution isn't robust enough. The backup products and processes you used in the early days of your virtual environment may not provide enough responsiveness or granularity for larger and more essential systems.
Remedy: Migrate to a backup solution with true enterprise features such as file-level restores (or individual message restores e-mail systems are virtualized and will be backed up), user-defined VM grouping and backup scheduling, and most of all the scalability needed to grow with you.
4. Backups Fail Because of Storage Shortages
You know that ad campaign that goes "When lenders compete, you win?" That's not true in virtual environments, and the growing use of thin provisioning is bringing this problem to light. Thin provisioning "lends" more space than is actually available, on the supposition that not all VMs within an environment will simultaneously require all the space allocated to them. Regardless of whether thin provisioning is used, keeping up with storage space needs is a persistent problem. Insufficient space is a leading cause of backup failures. If this is occurring in your environment you need to reduce your backup space requirements -- and backup solutions can help.
Remedy: There's a better option than expanding the SAN. At a minimum, backup solutions should offer a choice of full, differential or incremental backups. They should also provide compression support integration with deduplication appliances (as referenced earlier). Other advanced features include the ability to skip over and not back up white space within VMs, which helps minimize the backup file size. If you're using or considering thin provisioning, be aware that many early-version virtual backup solutions do not support it.
5. You've Gone Hollywood -- Writing Scripts Is A Big Part of Your Job
If writing scripts to configure and manage backup execution is taking up more of your time than it used to (and cutting into your ability to get out of work on time go to a movie), maybe it's time to move up to a backup solution that automates a lot of these tasks.
Remedy: Enterprise-class backup solutions have many time-saving features that automate tasks that previously required scripts. These solutions also support integration with vCenter and provide PowerShell access, so if you do continue scripting, it can be much easier and unlock more powerful capabilities.
6. Virtual Backup Isn't Part of Your DR Plan
There are a lot of very strong reasons for incorporating virtualization into disaster recovery plans (I feel another blog entry coming), but unfortunately today many administrators simply do not trust their backup solutions to meet their recovery needs. If you don't think your virtual backup solution is strong enough to back up your mission-critical environment completely, and enable it to be restored quickly, your backup system is a candidate for an upgrade.
Remedy: Evaluate scalable backup & recovery solutions with the enterprise-friendly features described above. These include: deduplication and other features that reduce backup sizes and times, support for direct-to-target backup; object-level recovery; the ability to back up and recover multiple VMs simultaneously; image-level backup (so you can recover application and OS settings along with data); flexible, automated scheduling; and of course scalability.
It's easy to see how virtual systems grow, but much harder to see if they've outgrown their backup protection. If any of these signs apply to your virtual environment, consider if you've outgrown your current backup methods and whether your solution can scale to meet your current and future needs.
Posted by Jason Mattox on 12/22/2010 at 12:49 PM