Comment by nosolace

oh it's definitely a bit absurd, but i felt i needed a workout.... hence this post. multiple sites, hardware diversity, staggered upgrades, cold storage, are all really good ways to improve this, as they pull failure events closer to independence