> Or just run -current in production, like we do.[0]
If you develop, it's probably best to do that against current [1], but if I'm running a web, mail, file, database, etc, server there is IMHO very little advantage to doing so. Most folks aren't trying to push >400Gbps.
Seems like the reason is to catch new bugs, fix them and upstream the fixes promptly, with a team of 10 doing that. Sounds awesome, but I could see other people just passively consuming stable.
While I also use -current, I don't think this is good advice to the kinds of people who don't know if they should be running 14.4 or 15.0. There are caveats to running -current (for example, you need to disable the built-in debugging stuff on -current to get decent performance but the debugging stuff is already disabled on actual releases), so I think for new people it's best to recommend they use the latest release (15.0) and they can discover -current when they are more familiar with FreeBSD.
"I don't think this is good advice to the kinds of people who don't know if they should be running 14.4 or 15.0."
You don't need to wonder about this because FreeBSD has an official, documented position on this topic[1]:
"... include work in progress, experimental changes and transitional mechanisms that may or may not be present in the next official release ..."
"... whether or not FreeBSD-CURRENT sources bring disaster or greatly desired functionality can literally be a matter of which part of any given 24 hour period you grabbed them in!"
"(is not) In any way ``officially supported'' by us."
GP works for Netflix. The team that maintains their FreeBSD stack includes FreeBSD committers, as noted in the linked presentation. Bit of a special case.
With that said, I've quickly upgraded to every production release, including .0 releases, on my personal infrastructure boxes for decades and have never been bitten in the ass or spent more than a few minutes making required configuration changes, and have run -CURRENT on development boxes, where it usually works fine.
As a rough analogy, -CURRENT is a bit like Debian Sid. You probably wouldn't run it directly in production, but it's not an unreasonable option if you have the resources to maintain an internal fork (or, for that matter, as the upstream for a downstream distro).
Side note: Netflix support for FreeBSD is one reason I've continued to subscribe through price increases and periods of low use. Keep up the good work!
P.S. As -current is essentially running the latest master branch commit it should be evident that FreeBSD will be unstable, liable to break at any point and may have horrible bugs.
So -current is good for experimentation but probably not too much more than that (unless you're Netflix with team of FreeBSD experts who famously like to run -current -- see https://news.ycombinator.com/item?id=47322830 ).
Yeah but nobody else has as many a FreeBSD developers on staff to fix stuff when it breaks. Or, you know, to run monthly stabilization weeks and extensively QA before deploying to a herd of cattle.
There are so many factors in favour of Netflix running 16.0 which don't apply elsewhere.
If you develop, it's probably best to do that against current [1], but if I'm running a web, mail, file, database, etc, server there is IMHO very little advantage to doing so. Most folks aren't trying to push >400Gbps.
[0] https://www.youtube.com/watch?v=q4TZxj-Dq7s
[1] https://www.youtube.com/watch?v=GQ0mvmZtbaY