Some content besides my political opinions:
If you are using a server or multiple servers, you probably are using virtual machines. This is an intelligent decision, for hardware costs are lower, energy costs are lower, many things are much easier. Probably you are using an UPS system, too. This is a very intelligent decision as well.
Simple question: What happens if the blackout is longer than your UPS can handle? For your servers this is the same thing as to unplug the power supply. I guess I don´t have to explain what this means to your writing processes (or mining processes while bitcoin mining... ). In short: you don´t want to experience this.
One possibility is to use very expensive software that monitors your servers and shuts your systems gracefully down.
This means your writing processes are finished in a clean way and your cache memory will be cleanly persisted.
While using ESXi this can pose some problems, especially the free versions and even more especially using APC UPS systems. As far as I am informed there are only workarounds or very expensive software available. Workarounds can cause crypto-security problems. And well... expensive software is ... expensive.
I found a software (made in Germany) that can handle these problems in an affordable and safe way. It seems at the moment it is limited to 10 systems, but I really like the interface and the easiness of the usability.
You can check it out on www.opmonis.de (there is a 30 day trial, too)
Well, I hope this information helps some to raise security.
Greetz
Pollux