some fail-over provisions should be in place, its applications presumably would be set up to checkpoint and restart
It's not always quite that simple. That sort of thing can work well for services with relatively little state, accessed over the network, but it doesn't work so well for things like desktop systems that tend to have applications with lots of unique in-memory state, persistent network connections etc.
Now, you might think that rebooting a personal desktop isn't such a big deal, but imagine it's a terminal server with sessions running for tens of users, and an update comes out for a local root hole. You've either got the choice of chucking all your users off (quite possibly to another system, but it's still disruptive) or leaving the hole unpatched, neither of which options are terribly appealing. This sort of live patching approach offers a potential way out.
Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds