I’m thrilled to find folks who already diligently backup up their data files any number of common ways, such as:
- Copying files to external drives or USB thumb drives
- Copying files regularly to a file share (online or locally)
- Using special software to backup data to another device in the office
- Backing up data via online data backup services (such as Mozy or Carbonite)
While data backup for key files is important, it will still be a rough few days when your equipment fails or, even worse, a whole site event takes place (like a fire or flood for example). If only your data files are protected, you must perform the following tasks to get back up and running (assuming a single server failure):
- Repair or replace the troubled hardware
- Locate appropriate installation CD/DVDs for your operating system and applications
- Locate licensing information for the above software
- Install the operating system on the new or replaced hardware
- Configure the operating system, disks, network, etc. per the previous setup (there was documentation, right?)
- Install each application previously used on this system, again using the same configuration as before (without great documentation this can take at least a few tries to get right)
- Install backup/recovery software and recover your files, databases, etc.
- Finally, import recovered files into the appropriate applications
This process can easily take a few days even in simple situations, and in some complicated cases can take weeks and cost thousands of dollars in services (see our Real Life Data Recovery Stories for more).
Additionally, key applications such as databases and email servers require additional steps to export key data before backups take place. If the proper export process was not followed, and instead just the data files were backed up, your data may never be recoverable.
So this begs the question: what can be done to ensure a speedy recovery with a minimum of complications?
First, we must ensure that all data files are backed up properly and consistently. Second, we must make sure we follow vendor backup recommendations for databases and other “special” applications and data. Thirdly, we must test the recovery process and carefully document each step to ensure a reasonable recovery time in the future.
This testing process will often identify many issues that can easily be solved today, but that could cost us hours or days in a real recovery scenario. Common issues we may find in the recovery process include missing software media, missing licenses and special configuration changes required for applications to work with our systems in our business.
After testing and documenting the recovery, you have a key part of an IT Disaster Recovery Plan, albeit limited to this one system. Now if anything happens to this server, you can easily follow your document step by step until the system is back up in running.
Does this process seem easy to you? It may seem easy on paper , but in reality this testing can be very complicated and time consuming. There are a few shortcuts you may be able to use, but there is never a substitute for good recovery testing (or an all-out disaster recovery drill). If you fail to test the recovery process, I can promise there will be roadblocks when you attempt to recover under pressure.
If this all sounds like way too much work, I have a couple of suggestions to ease the burden. First, consider an imaging backup product to protect your whole server, such as StorageCraft’s ShadowProtect or AppAssure’s AppAssure (formerly Replay4). These products have the ability to protect your entire server including installed applications, user data, and the operating system. If a server should fail, you can expect to be up in running in just hours after replacement hardware arrives as these software combine steps 2-8 above into a relatively painless one-step process.
Another idea is to virtualize your servers. Virtualized severs using VMware ESX or Hyper-V (for example) paired with the right backup software, behave similarly to ShadowProtect or Replay4 in that everything can be recovered in the event of a server failure. Furthermore in many configurations virtualized servers can easily be moved from one server to another should a failure occur.
Of course with all the advantages these two shortcuts provide there must be a drawback and it comes in the form of additional upfront complexity and cost. Notice that I say upfront. This upfront cost will often save money later when a system inevitably crashes or experiences a hardware failure. The additional upfront cost can be planned for and managed, versus the unplanned emergency crash requiring all hands on deck no matter the cost.
Businesses must choose how much risk they wish to take on by how thorough their backup and recovery plans are. The more involved and complete the plan, the more upfront cost and lower recovery cost and lost productivity due to downtime. With less involved and incomplete backup and recovery plans businesses may be able to recover after an event, but at an increased cost. Our backup service offerings reflect this choice; the more expensive the plan, the less the cost and duration of recovery.