Planning for Business Continuity
For the sake of business continuity, we often plan for a disaster that never comes.
In the case of your server backups, you
should be doing a test-restore periodically to make sure your backup procedures actually work. Though, like an insurance policy, you hope you never need it.
One of my larger clients has a Datto Siris2 unit in place. This unit backs up all of their servers hourly and keeps a BOOTABLE image of those servers.
Think about that for a minute. Instead of a “file backup”, which is just your data, this is a snapshot of your entire server at a point in time. Now imagine we can power-up that snapshot and have a running server in a matter of minutes. Datto’s backup of your server can be started as a virtual machine and the users can do business as usual.
Business Continuity – A Success Story
This is not just a “backup”. This is not just “disaster recovery”. This is Business Continuity!
One Sunday night, I was doing a batch of Windows updates on a client’s Windows 2012 server. This client runs 24 hours a day, five days a week, so I always do the updates on the weekend. This time, one of the Windows Updates didn’t go very well. When the server did not appear back on the network, I checked the VMware console to see what was happening. The screen said an update failed. “Last Known Good” did not work. The server would not boot in safe mode. A startup repair did not work. I was running out of tricks!
A little Internet research told me this happened to some other people as well. The fix looked like a long and painful process.
By now, it was quite late on Sunday night. People would start showing up to work bright and early Monday morning. Not good! I was out of time and patience. If this client did not have a Datto Siris2, I would have slept a couple hours, then gone onsite in the morning to work some magic – while dozens of people hover and ask “Is it fixed yet? How long will it be?”
I logged into the Datto Siris2 interface and started the server’s newest backup as a virtual machine. I changed the IP address as required, and THAT’S IT! The server was back in operation and the fail-over was undetectable to any of the users on Monday morning.
This coming weekend, I’ll move the virtual machine from the Datto unit to the production hardware. No lost data, no lost productivity, no lost sleep!