Yeah, maybe you're right. I actually had this conversation with my PM the other day - I work on a fairly well known database. They were asking us to improve the startup time of the product.

If it's something like a galera cluster then not really, I think. You want the updated machine to become operational again in a reasonable amount of time, because during the time it's down you have fewer points of failure. But the difference between 2 minutes and 5 minutes in this case is not a big deal, in my opinion.

actually, i want it backup up concurrently.

for instance, the systems better checks all memory banks and disks before it finishes booting.

Having a system boot slower is an annoynance when downtime occurs, but having a system boot incorrectly is far worse.