Preparing for Hard Drive Crashes and Network Outages
You keep jumper cables in the trunk. You change the batteries in your flashlights and smoke detectors on a regular basis. You keep a list of emergency phone numbers beside the phone. So when was the last time you performed a disaster recovery on your office server? What did you say? What do you mean you have never performed a disaster recovery?
“But I back up the server every night. Isn’t that enough?” No! What good is it to back up all your important data every night if you can’t recover it when you really need it?
Most small businesses fall into two different scenarios. The first is a “Server” that is really nothing more than a regular desktop computer with a big hard drive that just has shared folders and maybe an accounting program on it that everyone in the office can access. The second scenario is an actual Server with Active Directory, Shared Folders, maybe Microsoft Exchange Server and a SQL Server installed on it. These could all be on the same server if the business is small enough, or they could be on two or three servers in a larger environment.
If you have a small Doctor’s office, Bakery or Coffee Shop then you probably fall into the first scenario. In this case, testing your backups is rather routine. All you have to do is restore random files from your backup to a new folder and see if they are readable. If you have a Lawyer’s office, a non-profit, or medium sized office with ten to twenty-five computers, then you probably fall into the second category, and testing disaster recovery is much more involved. Active Directory, SQL Server and Exchange have to be recovered in their entirety.
If your server is setup in a RAID configuration, then at a minimum you will need three hard drives identical to the ones currently in your server to test your backups. Ideally, you should have a second backup server that is similar, if not identical, to your main server. With a second server, you can then take your backups and install them on this server and make sure all your Users are still there, All the user’s Rights are still correct, All your Email is still readable, All your Documents and Databases are still accessible…It’s a little more involved, to say the least.
“Well, how often should this Disaster Recovery be done at my office?” At least once a year. Preferably twice a year or whenever your backup routine changes. Backup tapes wear out, they become corrupt, hard drives can become corrupt, your motherboard could die, or you might simply think you are backing up everything you need, only to discover when you try to recover your data that you forgot to include your Exchange Server log files in the backup, and now you can’t recover Exchange and everyone’s email is GONE! Wouldn’t it be nice to know this in a test scenario, before you actually need it to be there?
There is really no way to be sure that all the time and money you have invested in your backup procedure is worth it unless you test the recovery of that information. How would it affect your business if you came in tomorrow and your Server wouldn’t start? All your files are inaccessible, nobody has email, your calendar, all your contacts…everything is gone? Maybe the Hard Drive in the server died. Do you have a spare Hard Drive to replace it with? If you don’t, you have to order a new one, wait for it to come in, and then have your IT guy reinstall Windows, all your applications, and then recover everything from your Backup.
Best case scenario? Your entire network is down at least half a day. Worst case scenario? You wait a day or two for the new hard drive, install Windows and try to recover your Backup. But wait…something’s wrong…Active Directory won’t recover correctly. This means you have to rebuild Active Directory from scratch. All your Usernames, All their Privileges, your Shared Folders and all their Permissions…This could take a week to get back to where you were! Want to really be scared? What if your backup tape is corrupt? You might be able to recover some data off it, or it could just as easily all be gone! The chances of a total loss are small, but my point is that if you discover this during your Disaster Recovery test, it’s an inconvenience. If you discover it after your server actually dies, your business could be in severe trouble.
Sure it costs a little more to have a second server or a couple spare hard drives on hand, but look at the cost of the alternative. How much would it cost to have everyone in the office waiting for someone to restore your network? An office with 10 people making $18 an hour would cost you over $1400 for one day’s salaries. What if it took three days? That doesn’t even take into account the cost of reproducing all those missing files if your backup tape is corrupt. You can have spare hard drives on hand for a couple hundred dollars. Brand new Dell servers with 2GB of RAM start at $250. Even if you added a couple more hard drives and more memory to it you would still be way under the cost of one day’s lost productivity. That means it’s more expensive to do nothing than it would be to make sure you are protected in case of a crash.
With a secondary server, you could even have it sitting there on a shelf already set up with Windows and all your core programs already installed on it. Then when you do have a hard drive crash or your motherboard dies, you throw your backup tape in the secondary server, restore all your files and have everyone back to work in an hour or two.
A tested disaster recovery plan can change a server crash from a very expensive headache into a mere inconvenience. If you have any questions on backup solutions, disaster recovery or best practices for your server’s health and well being, I can be reached at mark.williams@ptdtechnology.com