They do.

tuaris, What from that makes you think the kernel is being updated in a jail? I skimmed looking but didn't see anything.

Normally there is no kernel installed in a jail and no need for any kernel updates in a jail. But it is technically possible to copy a kernel file into a jail; it just won't ever be used and will just be bits on disk that are not needed to be there.

blah. occambsd is a brilliant idea but when trying to build 14.1 it's including a bunch of debug files and writing 1.5GB of output when trying to build a jail. and every iteration with it takes an hour and a half, because i'm not wise enough to know what it doesn't need to clean, so i end up rebuilding clang all day

i just want a tiny jail, with all the work done by somebody else

also i am building on spinning disks. - stop laughing at me!

i got a, like, 80MB jail for 14.0. it was super cool.

I don't thiink using spinning rust warrants laughter as much as empathy.

If you want tiny jails, you may want to check out poudriere-image(8) that's part of ports-mgmt/poudriere from packages or ports. Combining it with as many WITHOUT_ option in src.conf(5) as you think you'll benefit from, you can make FreeBSD pretty tiny.

right, so my issue isn't just the devd action (easy to do), but the pcm #

if i have several sound sources, how can i make sure that devices are always on the right pcm #?

debdrup: that's probably a good lead. occambsd is supposed to sort of invert the sense of everything, so by default it will build WITHOUT_ everything, and you add things back in. it's mindbending to write WITHOUT_THING_I_WANT in its config file when you want that thing, and it appears dexter has had a little bit to do for each release to make the right profiles

is freebsd 14 know to be unstable ?   I had a system crash and all I did was try to config kde and xfce .  zfs some was messed up but linux Ubuntu used it just fine.   The  ethernet connection did not work , tried to install xfce again had problems and transmission and deluge both keep crashing .

maybe 13 would be better?

tarel2: worth reporting a bug, 14 is not supposed to be unstable. i've been running it on a number of servers without issue

Well, I used it for my desktop.

Demosthenex: I think there's changes to pcm(4) that'll help you somewhat in this, but you may want to talk to christos@ because he's currently working on freebsdfoundation.org/project/freebsd-audio-stack-improvements

at first everything installed great  and it took no time.

tarel2: it's not meant to be unstable no matter where you run it, unless it's an unsupported/barely supported architecture (which i doubt your desktop is)

tarel2: you're gonna need to rootcause the issue before you report any bug.

The crazy part at first everything was fine.   the best system I had set up in sometime.  install fast , got xfce working fast.  after I guess setting up xfce wrong , then I rebooted and zfs somehow got messed up

jaredj: OccamBSD is certainly interesting, but as far as I understand, it doesn't really do anything that FreeBSDs own build system or poudriere can't do.

in what way did zfs get messed up?

"zfs somehow got messed up" indicates you don't know what went wrong, so start figuring that out.

debdrup: ah, interesting. its not a huge issue, i just thought it might be another devd thing

Linux (Ubuntu ) used the zfs just fine.   It was just freebsd says i/0 error 5 and could not find the kernel

Demosthenex: I don't think devd is related to devices being initialized in a different order

debdrup: i'm also sure not everyone has 6 pcm devices ;]

Demosthenex: I suspect you're right.

security emails are getting duped?

skered: not on my end

tarel2: you're gonna need to be more specific with the error than that..

debdrup: beat me to it :P. tarel2, what setup do you have exactly? how does ubuntu fit into the picture here? what kind of reboot was it? what does "setting up xfce wrong" entail here? did you by any chance delete some files/packages? if so, which? did you modify any /etc files?

I reinstalled the whole os.  Then the  ethernet did not work , xfce did not work and two different bit torrent clients transmission and deluge .

jaredj: like I said, you can do WITHOUT_ in src.conf(5) and it'll affect both FreeBSDs own build process and the one used by poudrere; all that work comes out of the FreeBSD Built Option Survey started some 10 or so years or more ago by phk

tarel2: we can't help you if you just reinstall the OS.

I _could_ maybe make a few educated guesses based on what you've said, but it wouldn't help.

 Well,  I get that , that why I asked if it was unstable .  Like maybe I should have picked 13 .

What I couldn't do is pinpoint the issue and figure out who you should inform about it - so unless you can somehow repeat it, there's no real way anyone can help you.

If it was unstable, it wouldn't have been released.

If it was known to be unstable*

If it was released and was unstable, it's because nobody reported any stability issues to the release engineer (who _will_ hold a -RELEASE if bugs are found).

it's not supposed to be unstable, no. i have quite a few machines running 14.x without issue, so it's either specific to your hardware (i.e. a bug in the base system somewhere), a port/package you used has a bug/was misconfigured or there was some user error that put your box in that state

Before 14.0-RELEASE, it was run as 14-CURRENT for a good few years by a good portion of the developers working on it - so most issues would've been found then. It's possible something got through, but if so then nobody's told anyone about it (even after your mention of it, because your problem description doesn't permit figuring out if it was a bug or something else).

tarel2: for transmission it could be bugs.freebsd.org/bugzilla/show_bug.cgi?id=280644

thanks man , mimaje

Hi there, about: freebsd.org/security/advisories/FreeBSD-SA-24:08.openssh.asc - Is OpenSSH even vulnerable if no blacklistd is used / running?

:qq

would have to look what that does concretly, but as it has to figure out at runtime if blacklistd is running, my guess would be that it is, I don't think you can do that async signal safe

well, ok, I also have no idea how you would do logging code in an async signal safe way, but openbsd seems to insist on doing that and other complex stuff in signal handler and therefor invest time to do that in a safe manner

I wouldn't think connecting to a socket needs anything that isn't async-signal-safe

re: "figure out at runtime if blacklistd is running"

is current booting on alder lake yet?

so Im trying to update an older 13-stable jail to 14-stable, i've got host upgraded already. The world is installed to the jail root dir but I am struggling to update etc with etcupdate, not sure how to make it do it :/

I run etcupdate -D /path/to/jail but it doesnt seem to do anything to the etc in the jail...

althoug if I run etcupdate diff -D path, it shows the differences

you could probably jexec etcupdate

concussious: tried that but I dont have /usr/src in the jail and it seem etcupdate needs it. Anyway I also tried etcupdate build on the host, then etcupdate extract in the jail. Doesnt seem to be working, as when I run etcupdate in the jail after that, it says Failed to build new tree.

jexecing would be a workaround anyway. I'm not knowlegable enough on this, but it seems like you've found a bug. I hope someone more knowlegable comments.

Hello , /usr/ports/net/cjdns is broken , know you a solution to use it please ?

The "broken" message points to hyperboria/bugs #219 , which looks like the port needs an upstream fix to work.