Disaster recovery is one of those buzz words that get thrown about and not everyone is necessarily using it right. There is often a misconception that Backup = Disaster Recovery, for example. Backup is not DR. Although backup is important, Disaster Recovery today is much more available, fast and useful than the old days of restoring data from a backup tape. Today, you can easily flip a switch and now all of your data is being stored and hosted in a secure datacenter.
Many backup solutions also make the matters worse by marketing that they have DR capabilities, when they really don’t. If you are unable to fail over to a datacenter and fail back to your main server easily without losing data, then it is NOT a DR solution. Backup is important. It is necessary. However, DR is a separate function altogether. If you want high availability to your Voice, Data and other services, this takes much more design and implementation work and testing to make sure this goes off without a hitch. If you are stuck, don’t hesitate to ask us about creating a DR phone system in our cloud, or our low cost DR as a service solution.
But let’s dig a bit deeper and look at a few things that you should be doing today, even if you only have a backup solution. At the bare minimum you need to be doing quality backups of your data so that even if you do have to wait hours or days for a restore, you can do it eventually.
- Make sure you are backup up your entire image rather than just the files. File level backups are a thing of the past. Now with virtualization and Bare Metal Backups you can back up your entire server as a whole. In the old days of backups you could only get files to back up. This kind of backup would require you to first restore the operating system, the backup software and then restore the files, settings and other things. And obviously this was hit or miss. Nobody ever would believe this would work well. And it didn’t. Today, the software takes a snapshot of everything as a whole. Windows, all the installed apps and databases, data, files, settings and everything is captured as a point in time. So, when you restore you don’t have to go and fetch your old cd’s and configuration files. You get the whole enchilada as it was when it was running live. This system can still do file level restores, but the ability to restore the whole server or virtual machine as it was when it was working live is essential.
- What you back up to matters. Obviously you should never back up to the same Storage Area Network that you have your main data on. If you lose that SAN, you lose not only your servers, but all your backup data as well. If you are still running tape backups, it is time to run to disk-based backups. Tape is terribly slow and unreliable. So many tapes have failed in the past. It is not a viable backup medium, if it ever was. Also, make sure you have your backups somewhere else. At minimum, that means taking nightly tapes off site. A better solution is to set up an off-site datacenter target for your backups so that your backups can automatically be off-site.
- Make sure to test your backups. This is where many businesses fall short. First, when IT buys servers it often only buys enough space to run the actual servers. They often don’t plan to double or triple space so that they can do a full test restore to see if their backups are legit. However Cloud Grid has identified several great pieces of Backup software that allow you to test restores on our infrastructure easily. Even if you haven’t virtualized your server into VMware yet, you can still do a server backup test in a live scenario using our backup testing architecture. If you want to do it yourself, you just have to make sure you don’t restore to a live system. Restores should occur in a walled off test environment. Often you should have a desktop as well configured so you can access the data as a user would to really see if the services are running. Testing backups takes design work and planning, but without it, you’ll never know whether your backups are good. How often should you do a backup test? You are only as good as your last backup test. So, if you can afford to roll back to a month old data, then do DR tests once a month. If a week is important, then a weekly test is in order.
- Consider live DR replication as a complement to backup. Backups are great but they are meant to be point in time snapshots that have to be restored. Restoration takes time. DR solutions copy data over to a recovery site off premise and the goal is to easily restore services if something happens to the main data site. Modern DR solutions, like Cloud Grid’s Hosted DR as a Service solution allow you to have replication points and restore points every 15 minutes if you like. So, you could restore your servers at a point in time 8:45, or 9:00. Each restore is different. The restore is a live working set of data and only takes minutes to turn on even if a normal backup restore might take hours or days. DR is set up to be highly available. Backup is not.
- Make sure your DR has a way to get back to your live system. This is called Fail over and Fail back. Some top Backup solutions have built in partial DR into their software. If something bad happens such as an outage of a line, a bad virus or malware hits your system, etc. you can fail over to a datacenter or cloud solution. However, most backup solutions don’t have the ability to fail back to their original system once the outage has been cleared up. DR solutions do this differently. Even if you fail over to the datacenter, you can just as easily fail back to your main site without losing data.
- Don’t hold the Phone. Invariably when we analyze most customer’s disaster plan, they often leave out phone. Take a scenario if the office communications to the main office goes down. They may have a nifty way of bringing up service in the cloud for their servers, perhaps with Citrix or Terminal Servers in the cloud to turn on at a moment’s notice. But more often than not, they don’t have a way to restore service to their phone system. This alone could be devastating to a client. How long can you go without phones? The answer is not long. We have great redundancy built into our solutions, allowing clients to extend their phone system into our datacenter. Use it when you need it.
Customers demand resiliency built into YOUR business. Your reputation is on the line. There are many threats that can cause a company to implement a DR solution. Fire, Theft, Water Damage from a pipe or another tenant above you, Tornados, Hurricanes, Hardware failure, or Service outages are just some of the real threats facing you today. Over 60 percent of businesses that experience a disaster never recover. Don’t be a statistic. Make sure your business can weather the storm before it hits.