how can i see if the default sendmail is running?

the thing that the hardening option is supposed to disable. just wanna confirm it worked

14.4 fwiw

sockstat -4 -6 | grep sendmail

tyvm sig`

i see /etc/hosts doesn't get populated with the hostname we set during install, why is that?

is there any benefit to putting the public ip and hostname of the current machine in its /etc/hosts?

sorry i got d/c

is there a vm-bhyve but for jail containers?

i used to use bhyve manually but vm-bhyve is a really nice wrapper so i switched to it

super minimal, basic abstraction

Has anyone seen some strange output from the latest update using `freebsd-update install`?

[03:34] andrew@apu ~ [1]> sudo freebsd-update install

src component not installed, skipped

Creating snapshot of existing boot environment... done.

Installing updates...install: ///usr/lib/debug/boot/kernel: No such file or directory

install: ///usr/lib/debug/boot/kernel: No such file or directory

[...]

Full output here: gist.github.com/Davis-A/4c8c8bc576689a7b0082c1cd6ad700af

i'm working on setting up a vnet jail. i'm supposed to use an $id in the config the handbook says, and it represents the last octet of the ip. but what if the box has a class b assigned to it, so there are 2 usable ips with the same octet? like 10.10.10.200 and 10.10.11.200

tyvm

back got d/c

the last octet is in this case the 200.

ya

but how can that be $id, and $id be unique, when there will be another jail with 200?

kerneldove: why do you need vnet jail ?

there's nothing that binds you to use that $id: not sure of the handbook, but my setup uses the last octet to also set the epair

jails without inet perform better

s/inet/vnet

btw, running it without inet is also fun, if you have inet6 deployed correctly

I always found the non-vnets a hassle, adding aliases to the host nic and the meddling with 127.0.0.2

but they are just better with regard to network access

I'm running 15 vnets and not having any problems whatsoever. several bridges are connected to different vlans

better how?

wondering as well

joemie do any of your vnet jails do production traffic levels?

yes

nice

maybe mzar wants to elaborate on his claim?

not at all

just less overhead

not convinced ? connect your jail to at least 10G network with/without vnet and run iperf

that would be a good test. joemie you done that?

mzar so you're saying the overhead is low enough you need to push around 10gbps for it to show up?

nah, I'm happy as I am

nope, it will show up on 1G too

OK

btw, I do not refute what you're writing, always willing to learn

how much % overhead did you see on 1G mzar?

kerneldove: I see no overhead since I am almost no using vnet jails, if so, then only in places where network performance is not that important

lol k

kerneldove: if you want to benchmark it and share results, please let us know

why didn't you do that mzar?

I am not much interested in obtaining any detailed results

just seems kinda selfish that's all

if you collected and reported the data when you found it maybe the devs coulda fixed it?

hhe...

just did an iperf3 from the host and a jail, both towards my firewall, and there wasn't any noticable difference.

as a test I did the same from one of my linux hosts.. same

disclaimer: I'm on a 1Gb network, not 10Gb

joemie thx for putting in the effort to test that. really appreciate ppl care enough about the project to put the time in

overhead ≠ regression, so it's nothing to report here

and this is well known fact

sounds like excuses but ok

mzar: you told me to test with iperf, and so I did. What exactly are you blaming vnets now from?

I cannot help it if the results do not fit your earlier claim

joemie: I am not blaming anyone

overhead is overhead, if you drive a car with/without a trailer you know what I am typing about

okay

TBH this Bumsrakete CVE is very nice, works also in jail

ye its wild. I just patched.

anyone know about java classifiers? trying to get a clojure project running on freebsd and it keeps trying to use linux libraries

Hi, I have a question related to ZFS snasphots. I have send snapshot to my usb flash disk. Now I want to do restore so I can rollback to this flash disk snapshot, but it desn't work:

root@bsdlatitude:~ # zfs receive -F zroot/ROOT/default@backup-2026-06-12 < /mnt/usb-flash/zroot--root--default--backup-2026-06-12.zfs

cannot unmount '/': unmount failed

--- It looks to me like I will be not able to do it from live system. Or is there a workaround?

Nothing written about this in Handboook, even in Absolute FreeBSD book is this part missing

JurassCZ: you need to boot from USB stick, PXE, etc then recover - it will not overwrite working system

you can rollback it, but not overwrite

mzar: Ok thanks

perhaps it will be possible to overwrite / but I don't recommend such an approach

s/possible/feasible

Hello. I have installed FreeBSD inside a virtualbox VM in linux host. I have given it ICH AC97 audio driver. The audio seems to be playing in volume levels which I can see fluctuate, but I don't actually hear anything. The speakers are on and host can make sound, but VM can't. I am not sure what to do at this point. Any guidance?

I am now testing the HD Intel audio virtual hardware, maybe that will work. I don't know yet

Still no luck

how do secure ttys work? doesn't getty (or login) check /etc/ttys against the current ctty?

oh i'm dumb, i uncommented the line :D

is it possible to blank the screen using the new vt console? I think the answer is no

no

and yes

if you load proper graphics driver

but not like the old sc console with saver_green

thanks mzar. reason i ask is because wayland sessions only work with console login currently, and if no one is logged in, the screen never turns off.

i left a comment here bugs.freebsd.org/bugzilla/show_bug.cgi?id=233356

I am still using X11/xorg, so can't help much

I;m disappointed from that guy named Michael W. Lucas, who wrote Absolute FreeBSD book. I suspect he also wrote Handbook for FreeBSD. Both, the Handbook and the Absolute Freebsd has not covered very important ZFS use case "I have ZFS snaphot backup on external media and now i want to do receive that snapshot and rollback changes to it." That

Michael Lucas is sayig in his book "Use ZFS, it's more modern that UFS", but in a meantime, his book don't have any single one occurence of the string "zfs receive". Also, in a meantime, this guy has wrote book named FreeBSD ZFS Mastert, so he obviously very much know, how to do the above thing. And so in summary, I suspect this person

intentionally is not telling this, because he want you to buy his Absolute FreeBSD book, and then to buy a second book about ZFS he wrote. I have big doubts this is not intentional.

The ZFS for ordinary FreeBSD user is suppose to be a major backup solution. And of course this means to do backups on external media. And of course this means there must be also a solution to do recovery. But it's missing in his books.

Have you checked the ZFS chapter in the online FreeBSD handbook? (docs.freebsd.org/en/books/handbook/zfs/#zfs-zfs-send )

JurassCZ: this chapter in handbook is very good writeup for ZFS beginners docs.freebsd.org/en/books/handbook/zfs

Yes I have. I know it's there. There is covered zfs receive, however there is not covered how to place this new snapshot to default zpool. So you can't rollback to it.

This very part is, however, partially covered in the Absolute FreeBSD book. i suspect I have to buy the ZFS Mastery book of the same author, who wrote the Absolute FreeBSD and The hanbook, so I will have full picture.

JurassCZ: this? openzfs.github.io/openzfs-docs/man/master/8/zfs-rollback.8.html

voy4g3r2: Not really, I know how to do rollback. You can't do however rollback to your snapshot from external drive. You rist has to mount its pool as "/", but this can't be done on currently running live system.

hrm..i have done snapshot rollback using external devices.. zfs send / zfs recv

maybe someone smarter than me knows your situation.. but it has worked with other datasets.. i also have snapshots both local to the server and replicated to a remote one.

are you deleting the snapshots ON the machine? after a backup to remote machine?

voy4g3r2: then it wasn't a root "/" filesystem but only e.g. /usr/local, maybe that's possible.

voy4g3r2: yes I tried to destroy snasphot in the default pool. You can't unmount default pool when your system is running. THat's the reason you can't rollback to snapshot received from the external drive.

This will backup entire "/":

root@bsdlatitude:~ # zfs send zroot/ROOT/default@backup-2026-06-12 > /mnt/usb-flash/zroot--root--default--backup-2026-06-12.zfs

Then, it is not possible to do this on live system:

zfs receive zroot/ROOT/default < /mnt/usb-flash/zroot--root--default--backup-2026-06-12.zfs

I haven't tried something like this, but I've seen folks talk about it, where you would zfs receive to some dataset from the USB, and then perhaps set that dataset to mount as "/" on the next reboot, like via bectl or something. I guess i'm kinda thinking out loud

JurassCZ: send/recieve works. you have to practice, that's all - you can create a pool from files attached to geom with mdconfig(8) and start this way

mzar: yes, I'm not saying that send receive doesn't work. I can receive the snapshot to pool. But the pool can't be the "default", it must be eg. "default-2". And you can't rollback to default-2, because it's not mounted to "/".

So now I have to go and try to somewhat mount it in Live CD and hopefully, I will be able to make it working. But it's not covered in Handbook and in neither in Abslute FreeBSD.

OK - it's something for ZFS users to figure out on their own

if you are long-standing FreeBSD and ZFS user you actually like it

If you have UFS, you know you will use dump/restore. And you know what to do. But everybody is still saying you "you should be using ZFS, because UFS is old". And so I have ZFS and obviously, the entire backup procedure from external drive (of the entire "/" system) is not covered in any books.

UFS is not as handy as ZFS

if you get spoiled by ZFS once, you'll rather no longer want UFS

Well, I can create UFS easily with GPart, and I don't need pescial knowledge to learn how to do full "/" backup from external drive.

OK

but you will miss BE (boot environments) - it's IMHO great loss

oh yes.. i have had to rollback the BE

it is great

quick research suggests receiving the snapshot to a different dataset, double checking the properties of mountpoint and canmount, and running bectl list to confirm it is seen as a legitimate boot environment

and go from there

are the 15.1-RELEASE builds still tentatively scheduled for tomorrow (2026-jun-13)?

Ok so I figured out the sound issue. It was to select Intel HD Audio in virtualbox for the FreeBSD VM, and selecting pulseaudio as the driver. I got sound now and a nice FreeBSD Desktop with KDE plasma. Very nice

Hello, all. What shall I define for a C project to get available SIGWINCH from signal.h in a portable manner? -- <pubs.opengroup.org/onlinepubs/9799919799/basedefs/signal.h.html>

Looks like varous runtimes define their own FTMs... For example, glivc has _DEFAULT_SOURCE .

ant-x, I don't think you need to define anything. I think it will be defined by default when that file is included.

rwp, with -std=c89 -D_POSIX_C_SOURCE=200809L , SIGWINCH is not there.

github.com/supaplextor/tech-bucket/blob/main/FreeBSD-15/mountroot.md stuck booting. This is after using nearly identicial kernel config to generic, with 'sa' module support. (tape seq. access)

this is from the ip kvm (no relation to qemu)

I figured out booting the old kernel...