Re: what am I missing RE: freebsd-update?
- In reply to: Steve O'Hara-Smith : "Re: what am I missing RE: freebsd-update?"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Sat, 03 Dec 2022 14:09:35 UTC
On 2 December 2022 5:20:52 pm AEDT, Steve O'Hara-Smith <steve@sohara.org> wrote: > On Thu, 1 Dec 2022 16:44:37 -0800 > paul beard <paulbeard@gmail.com> wrote: > > > Thank you for all the info…it seems nothing is wrong, but what > prompted > > this was a message about this release nearing EOL or end of > support. Not > > sure where I saw it now, perhaps at login but I don't see it now. It was the firsf message on screen in the freebsd-update process, and yes it's rather trigger-happy. > The file with the EOL dates in got updated as part of the update. > The dates are a guess that gets refined as the release process of the > next > release progresses. It doesn't work too well it always alarms people > early. More definitive is https://www.freebsd.org/security/#sup showing 12.3 EoL as 3 months after 12.4-RELEASE (which is immanent, likely this week), and that 12-STABLE will be EoL on 31st December 2023. > > So freebsd-version -kru > > 12.3-RELEASE-p6 > > 12.3-RELEASE-p6 > > 12.3-RELEASE-p10 > > > > means that my installed and running kernel are the same and my > userland is > > a few patches ahead but they should sync up at some point? > > Not quite - what it really means is that p7, p8, p9 and p10 did not > contain any kernel updates and so the kernel version didn't get > bumped in > them. So if there are no more kernel updates before EOL they'll never > "sync up". Without a major security incident, p6 should be it for 12.3 kernel. Maybe a few userland patches if needed ... > This has historical reasons - once upon a time the kernel version > was the version and all updates were source builds but with binary > patch > updates it seems crazy to update the kernel just to change a number > so we > get the concept of userland version and the original version becomes > the kernel version. > > There's a practical advantage to it, scripted updates can check for > a changed kernel version and decide whether to reboot or just restart services. Yes, though you (or scripts) need to know which services need restarting; rebooting is always safer. cheers, Ian