thanks yuripv :)

I would really like an invite if someone has a bluesky account.

meena: Do the `packages` and `update-packages` targets only create packages from the host system or can it take a separate root prefix? (i.e. installkernel + installworld using the install at /usr/obj/tmp)

What's the typical method people use to see how much memory is in a system?

"dmesg | grep memory" is what I have been doing. It works. But it isn't very elegant.

It has the advantage of working from an installation ISO shell without any pkgs installed though.

rwp: `top` is about as basic as it gets for showing a few key pieces of system info, including installed RAM.

runs in shell with nothing installed but a base *BSD

There are (much) more minimalist ways of doing it, but `top` is 3 characters :)

Thanks coreystephanphd. Running top is good. "top -d1 | head -n7" gives me one screen with _just the facts_ about the system.

It does suffer from not being as immediately obvious though as compared to other things.

"dmesg | grep real.memory" says "real memory = 17179869184 (16384 MB)" which is directly the amount of installed memory.

"top -d1 | head -n7" on the same system says "Mem: 673M Active, 10G Inact, 467M Laundry, 2903M Wired, 12M Buf, 1653M Free".

Useful information there certainly. But not as easy to say this system as 16GB of RAM installed.

The swap line from top though is perfectly direct and succinct though.

"swapinfo -h" gives swap info more succinctly readable too.

daemon: sure

rwp, sysctl -n hw.physmem

hw.realmem

_xor: they operate like any other make target in our weird, weird, build system

I have a dumb question. when I type it overwrites the old text instead of inserting new text. how the f do I change it back to normal?

typing into what?

into vi?

just general text. if I type an additional word into a sentence, it overwrites the existing words rather than insert new letters. I think i might have accidently pressed shift cntrl and another key

into what program?

did you know you can actually order weird types of pet frogs online? You can keep them as pets. Im not sure what you can do with them other than wear a glove and sit them on a stick or a leaf.

rtprio: i usually use sysctl, like VVD shows. but… i actually don't know how to read it. my brain struggles with big strings of numbers

Hi all. I have newest FreeBSD 13.2. But I can't install newest GnuPG. My version is 2.3.8 if installed from pkg. If installed from ports - 2.3.3. But here is 2.4.x: freshports.org/security/gnupg

Title: FreshPorts -- security/gnupg: Complete and free PGP implementation

What to do?

are you using the quarterly or latest pkg repo?

RhodiumToad: I don't know. How to get this?

what's the repo location in /etc/pkg/FreeBSD.conf or /usr/local/etc/pkg/repos/FreeBSD.conf ?

url: "pkg+http://pkg.FreeBSD.org/${ABI}/quarterly",

it'll have either /latest or /quarterly in the url

So I should set 'latest'?

latest is more up to date, at the expense of having more version churn

How to switch safely?

edit the file, do pkg update -f, then pkg upgrade will upgrade all your packages to the latest version

OK

tnx

or copy the file to /usr/local/etc/pkg/repos/FreeBSD.conf and edit that, but that might be more confusing

Hi. My system just froze up changing between tty's. Had to give it a power cycle. dmesg does not know anything happened (new log starts at the new boot, old log does not show an error). xorg logs look normal too. Where else could i poke around to figure out the problem?

on freebsd 13.2-release-p3 if it matters

KingShark: did you not say that you had a page fault panic a few days ago while rendering in blender? it's probably the driver that's a bit clunky. does /var/crash/vmcore.last say anything useful if you kgdb it?

dstolfa, yep thats me, nice memory :) no files of today in /var/crash

changing from tty9 to any other tty1 always is laggy before it happens but after that initial lag, changing from tty1-8 usually works fine

just now it froze the whole system, keyboard stopped worked, sound stopped..

its an intel+nvidia setup with drm-kmod

it's *probably* the driver. you have it installed from packages right?

but i think the ttys are also rendered using the nvidia drivers so not sure if that matters

yes

are you running latest or quarterly?

Name : drm-kmod Version : 20220907_1 Installed on : Sun Sep 17 13:43:22 2023 CEST

latest. 13.2-release-p3

did an install from the websites download and later a pkg update / pkg upgrade

default kernel, nothing spectacular configured (:

i don't really know how well the intel + nvidia setup works on freebsd but i'd certainly expect this to be a driver issue. unfortunately when nvidia is involved it becomes a bit harder to debug because there's no source for the driver IIRC

all i changed compared to a plain basic install is usbhid_load="YES" hw.usb.usbhid.enable=1 in /boot/loader.conf to have multimedia keys work on my keyboard but the pagefault error happened before that. besides of that, i do lightdm_enable="YES" kld_list="i915kms fusefs" allscreens_flags="-f ter-u28" in /etc/rc.conf

dont think any of them are related to the issue (:

Hello. What happened to forums.freebsd.org ?

i could remove the allscreens_flags which i use to increase font size on the ttys for a try but i dont think that would cause a freeze

CyberCr33p: up and working, what's the issue?

how could i track if a packet got sent out from my freebsd workstation?

WOL issues :)

the wol command claims everything went fine but the target device does not wake up so.. i would like to confirm the packet ever got sent

guess tcpdump is the way to go

weird, weird. i can see the wol packet go out via tcpdump yet the target device doesnt bother waking up. guess a upd broadcast packet on port 9 does not do what it expects

I've been trying to email freebsd-stable@ but I am getting dozens of DMARC auth error responses from various places. I do see my email in the list archives so it's been delivered to the list server, but apparently a lot of subscribers aren't getting the messages.

The problem seems to be that a message from me (@anduin.net) comes from a *.freebsd.org host, which my SPF/DMARC policy doesn't allow for.

heh, why would they notify you though

need a system and network perfomance for server 1 gb ram, 1 cpu, and 20 gb ssd.

Of course i host on freebsd

Hi

How enable 0-RTT in FreeBSD 14-STABLE with Nginx 1.24, OpenSSL 3.0 and HTTP/3 ?

VVD, sysctl! I had completely forgotten about and not been thinking about "sysctl hw". That's very good! Thanks!

what is the cleanest way to mix one port into my pkg based system? i want to try a higher version of something. the newer version exists as port but pkg has a very old one :)

KingShark: just install it with the port, no extra steps needed

@KingShark Use synth ;)

So, how complicated is it to get bluetooth audio working? Specifically Sony headphones?

when i reboot, after the "all buffers synced" message I see ~25 messages like "pid 6983 (zsh), jid 0, uid 1001: exited on signal 11 (no core dump - other error)"

on 14.0-BETA3. don't recall seeing these on 13.2

markmcb: does that also happen when trying to just shutdown too?

BillyJoeBob: I can try now and find out ... brb

Well...

BillyJoeBob: yes, "shutdown -p now" had the same reslut

result

signal 11 no core dumped?

correct

?

pid 6606 (zsh), jid 0, uid 1001: exited on signal 11 (no core dump - other error)

except sshd

slightly different error

pid 3700 (sshd), jid 0, uid 1001: exited on signal 11 (no core dump - bad address)

"bad address" vs. "other"

Hmm...have you done any updates recently?

i'm fully up to date on BETA3. or do you mean hardware?

if hardware, no recent changes

Hmm, is beta 3 version 15? I only run release versions.

14.0-BETA3

Was there a recent update that you can cooralate with this? I mean FreeBSD/pkg updates

i think it started with the move to 14, i don't recall seeing it prior when on 13.2

That would be my guess or something along those lines. You should be able to downgrade, I think. That's if it bothers you that much.

I usually don't upgrade until it reaches RC3/RC4

it doesn't bother me. i just assume it is indicative of a bug somewhere

Could try here: lists.freebsd.org/subscription/freebsd-bugs

Title: FreeBSD Mailing lists: subscription for freebsd-bugs

as someone who's subscribed to 7863256 bugs (rough estimate), i reckon subscribing to bugs@ is going to be complete overkill

geobarrod: from what I understand, 0-RTT is a feature of TLSv1. 3, and the server implementing it. so i reckon it should be an option in nginx, but i can't seem to find anything

"Additionally, TLS 1.3 supports session resumption, which makes connection establishment faster by eliminating the overhead of repeating the TLS handshake when a client returns to a previously visited site. This is also called 0‑RTT (zero round trip time) resumption"

so does 1.2

meena: Thanks for answer. Look here on this HTTP/3 online checker http3check.net/?host=www.youtube.com

Title: HTTP/3 Check - www.youtube.com

meena: Now look my default site http3check.net/?host=geo.prensa-latina.cu

Title: HTTP/3 Check - geo.prensa-latina.cu

@meena Not 0-RTT in my site and the checker don't release the second connection

geobarrod: the detailed description in both cases says that 0-RTT was successful. but you didn't get a badge.

maybe it doesn't fit, cuz you got too many badges lol

meena: detail description is only for explain the features

aah

;)

Linux have 0-RTT, Why my FreeBSD not? :/

why don't you have a http header output, and why is Youtube's header output bigger than my website?

Now in FreeBSD 14, the support for OpenSSL 3.0 is ready, and this allow use HTTP/3 with Nginx, but not work. Why?

meena: freshbsd.org/freebsd/src/commit/3f3…2a83945c31d17e609dd6e2adcfb099ae7f0 notice anything?

Title: FreeBSD / src / 3f39a2a / ntpd: Support IPv6-only hosts - FreshBSD

Freaky: hrmmm… still says c, maybe my phone is hard caching this

meena: oh, you'll need to be using Split diffs

Unified support is coming

I'll try my laptop as soon as my daughter is asleep. in a years or 13

I backported inline diff highlighting from my dev branch

geobarrod: 0-RTT is a TLSv1.3 feature, and TLSv1.3 for a long time. not just OpenSSL 3.0

Freaky: that's really cool

@meena I mention OpenSSL 3.0 because that version allow to Nginx enable HTTP/3 support

meena: My site have only TLS 1.3 enabled

meena: I love being able to slurp in random .jar's and glue them into Ruby

if only all FFI stuff was that easy

(I think saying that three times is how you summon a GraalVM developer)

Freaky: it's really cool

I think the entire JVM eco system is really freakin cool, even tho I could never warm to Java itself.

anyone got recent benchmarks on linuxulator to compare overhead of linux jail vs running linux directly?

wonder what kinda overhead there is (if any!?)

geobarrod: what doesn't work about it?

meena are you mina galic?

polyex: no, I'm Mina Galić :P

reading about a virtio vsock device driver

:P

what's that driver do?

polyex: it allows for cheap communication between guest and host

through like a uds file path?

rtprio: The online tool http3check.net don't show the feature 0-RTT of my site http3check.net/?host=geo.prensa-latina.cu

Title: HTTP/3 Check - geo.prensa-latina.cu

No, it's just its own socket protocol. The device on one side implements it, and the driver on the other side does, too. An application opens a standard socket(), but with that protocol, and then communication happens.

Or so I hope.

ya isn't that what i said?

not being shitty just honestly confused

oh it's just an open standard in the virtio family duh. looked it up. wow it looks COOL

dude how fun is that project?

oh

polyex: I thought UDS meant that there is an actual unix socket file somewhere

ah nah, just LIKE that

but ya i get it now. gonna be so cool

do jails have a similar mechanism of host <> jail communication?

they don't need it

how can code running in a jail communicate with the host system?

polyex: it's painful. my C is rusty, and virtio is painfully underdocumented, hence: reviews.freebsd.org/D41853

Title: ⚙ D41853 virtio: add introduction to virtio, document virtio.h

meena ah :/ well your work is super appreciated!

polyex: jails see a subset of the host's filesystem, so the host can create local-domain sockets visible to the jail

after you make the virtio man page, maybe add a section on virtio to the handbook's chapter 24 virtualization ?

RhodiumToad so jails can use standard uds to communicate with the host? so sockets for jail and vm communication eventually. that's badass

meena ^ (virtio man page...)

looking at chapter 24 in handbook, a section on virtio would be AWESOME

imo should be 24.2 and push the others down

I just rebooted my FreeBSD server by installing a Debian jail & trying to force install sysvinit-core, WTF

wtf

veg: any crash dumps?

meena: no, it straight when into shutdown, twice

to be fair, I was in a chroot, not jailed

i know /usr/src/release has scripts to make .iso but .img can be made too right?