An important lesson to learn.

Yesterday at 23h40 UTC, the data center of OVH located at Strasbourg, France, caught fire. It burned the whole SBG2 lot and SBG1 partially (there are 4 "SBG" areas). Because of security precautions, the electricity was cut.

OVH plans to restart parts of the datacenter slowly, starting Monday, 15th. This means a 6 days downtime at best with no ways of accessing the data stored on these servers. There is also the possibility that these data are completely gone.

My company, PDFShift, had all the servers located at that datacenter. In retrospect, this was a bad idea, but we were careful when building our architecture. We implemented load balancers, redundancy, and backups. You could kill one of our servers and we would not flinch at all.

I never expected a whole data center to go down.

But it has.

I spent the entire ultra stressful day of yesterday moving all our service to Amazon, learning their infrastructure, how they work, and the relation between their service. We set up two Elastic Beanstalk, Lambda functions, SQS, and a database in a day. It. Was. Intense.

What am I saying all that?

Don't do the same mistake we did. Save your data elsewhere. Even better, at another company! I know this means an extra cost. I know this is hard to consider when you are an indie-bootstrapped entrepreneur. But it might save you. This is your insurance! And like all the insurances, they are useless until they aren't.

I'm aware that some of you consider this to be evident and are surprised that it wasn't the case for me. But I'm pretty sure I'm not the only one, and I'm hoping that by sharing my story here, some of you will do the right thing and be ready.

Is your service ready if your data center burn?