We’ve proposed to hold off on the upgrades to Ubuntu 18 (Bionic Beaver) until after the v2 rollout, in Q3/Q4. We can start thinking about the logistics of it now though.
We currently manage 12 servers across 8 countries and will be adding a new one soon. Upgrading the OS on each of those servers could be done in 2 ways:
Upgrade via migration: for each server we create a new one, set it up from scratch with a fresh Ubuntu 18 image, then migrate and shut down the old one. I think this is considered best practise?
In-place upgrades: for each server, we schedule brief downtime and upgrade the server via apt.
Doing this for 12 or 13 servers will take a lot of time and resources with the first method, and a lot less with the second method.
Are there any more pros and cons to consider?