This is probably bad form, but is anyone here from Netflix on the Open Connect program? If so please DM

not seeing any of the OCA netflix'ers in here

(sorr)

No worries!

hey anyone have any idea why my enlightenment desktop is restarting continuously and the error log is saying it failed to connect to dbus?

Is dbus installed and running?

pretty sure

yes

you get a pid from service dbus status right?

yes

I run plasma.. and it uses SDDM for the login system. one of the things that does is call a script - plasma-dbus-run-session-if-needed

presumably if the dbus session is already started, that can tell in some way and use it, or just fire up a new one as needed

yeah I think it is something else

something causing issues with enlightenment on FreeBSD

I shall spam the #e channel

and I have a couple dbus-someID= in /tmp

AumShivaya: have you tried "dbus-run-session enlightemnt" (or whatever the executable is called)?

no

did not work, I think it requires startx

do I try putting that in .xinitrc?

just put 'dbus-run-session' before whatever the command is that starts the desktop. it might not make any difference but some applications require it

nah still has issues

oh well

bbl

hey everyone, yesterday I was reporting a ssh "dropped connection" issue with a bastille container.. i did some work, restart container, changed network from vnet -> direct bridge and i am still experiencing "dropped" ssh (which did work at one point in time). From the logs I do not see anything either, is there possibly something in this output, that I am missing? bsd.to/SSdw

Title: dpaste/SSdw (Plain Text)

i am able to login to other boxes just fine

well that is just a horrible paste.. it was pulling buffer

Title: dpaste/8sC3 (Plain Text)

launch ssh with plenty of verbosity, it should tell you why it disconnects

I would also launch an sshd -d in debug mode and see the other end of the connection.

ha!

thank you! PAM initialization fail

well that is odd as i really never touch PAM but at least it is something

Title: dpaste/D5Gk (Plain Text)

What, in ports, do you guys use for basic file hosting? I mean like *very* basic. I just don't want to create a script that generates an html file with <a href="./somefile">somefile</a><br/> (on multiple lines).

TommyC: are you just focusing on a HTTP server or you talking like windows shares, apple file protocol, nfs? there are many ways to "host files"

voy4g3r2: ah my bad, yeah just http(s)

TommyC: i use nginx, but i think openbsd's http server might in ports and it's even smaller than nginx (i can't remember what it's called though)

ah, www/obhttpd

if you want to go down apache route, i typically see nginx used more.. here are steps for apache: docs.freebsd.org/en/books/handbook/…s/#_configuring_and_starting_apache

Title: Chapter 32. Network Servers | FreeBSD Documentation Portal

ok I apologize. I really should've clarified: I don't mean http daemon software (apache's httpd is what I'm using), I meant what do you use "on" the web server. I've never done webdev work so I really don't know what it's called (content management system, maybe?), but for example, I'm not looking for anything as heavy as wordpress.

you're just serving a static set of files from a directory, right? i don't use anything for that, i just enable directory indexing on the server. maybe i misunderstood the question...

or do you want to let people upload files by http as well?

lw: no uploads outside of myself, but the files are going to be updated every now and then

so i just do something like this in nginx: le-fay.org/tmp/30d/9rBMnU.txt

ah ok, I'll try that out

apache can do that with mod_autoindex (which i think used to be enabled by default, but they might have changed that for security reasons) but i wouldn't suggest apache for this unless you already use it

I do use apache but I'm not all too familiar with any http daemon software. Webdev isn't something I've tried to tackle (no interest and I get paid to do something completely different)

if it helps this is my entire virtual host definition in nginx le-fay.org/tmp/30d/5dc9KI.txt but the location /pub is the only important part. you can put that in location / if you want to serve files at the site root

I'll check out apache's autoindex feature. It may be what I'm looking for.

the generated listing looks like this, which is pretty basic but sounds like it's enough for you: le-fay.org/tmp/30d/Screenshot.from.…024-01-17.at.04_05_00.399773358.png - Apache can do more fancy stuff here like including html headers/footers based on files in each directory

lw: yeah that looks pretty good, and that's enough for me

lw, voy4g3r2: Cheers. Sorry for wording my question confusingly.

good luck

I have a DLP-TEMP-G I'd like to use if possible. This is a small USB device that shows up as a FDDI serial interface with an arcane protocol. Anybody familiar with it? It *works*, but I don't have a way to convert its output back into temperatures.

The datasheet is here: dlpdesign.com/usb/images/dlp-temp-ds21.pdf

I mean FTDI, of course.

FDDI temperature sensor :-D

TommyC, I suggest Nginx myself. It's what I use. It's much lighter on resources than Apache. But anything that works for you, works for you! :-)

things that are must aren't

lw: the fingers type what's most familiar. :/

ghoti, I think I have played with one of those DLP-TEMP-G in the past and I think it just prints temperatures to the serial port continuously if it is the one I am remembering. If you plug it into the machine and then connect to the serial port do you find it printing temperatures back to you?

rwp, per the datasheet I linked to, there's a serial protocol and a response that must be interpreted when you send a request for one of the 3 connected devices. You might be thinking of the TEMPHUM.

I might be thinking of TEMPHUM. Or any of several other random temperature sensors. Memory? What's that?

TEMPHUM is notorioutly approximate. I might use numbers from a bank of 5 of them where I toss the outliers, but those units are really only good for recording relative changes, not absolute values.

I'd love to use a DHI22, but I have no GPIO on this server. :/

hmm, why does /usr/local/etc/poudriere.conf not have the option to add overlays automatically?

kenrap: FWIW, there are #poudriere channels on Libera and EFnet if you don't get an answer here.

i wish portlint would pick up the current ports directory automatically

ghoti: seen this? svn.andrew.net.au/viewvc.cgi/svn.an…t_changes=0&view=markup&revision=22

could work with some modification in freebsd as well I suppose

Hi there, where is the proper place to change environment variable TERM for serial terminal ?

need this for bhyve virtual console

nevermind /etc/ttys ...

Fixed! Seems the upgrade wasn't entirely clean and there still references to pam_opie in /etc/pam.d/ files which was causing su to break

there were*

hey with updating ports collection to latest Quarterly branch

do I have to remove the ports in /usr/ports before checking out with git?

I installed them with the bsdinstaller

ok, I have a question.... enlightenment (desktop) depends on EFL (foundation libs)

I updated via pkg enlightenment, but had lots of issues

until I realised that EFL did not get upgraded with enlightenment upgrade

so I updated to latest EFL and issues are fixed

but my questions are: why did it not upgrade EFL when I upgraded enlightenment

and should I let someone know about this so it is fixed?

AumShivaya: Not sure why it didn't get upgraded if it was a necessary dependency, but if you're 100% sure it was an error on the port's part, then the port maintainer(s) should be notified.

yeah I am not 100%

I am trying to figure it out

though I did upgrade e and EFL was not upgraded

but have no way to know why, if I was my fault or... if it has been since fixed

hi, how to make Alt+b/Alt+f work in zsh in freebsd console (not X)? when i press alt+b i get <ffffffff>, pressing Esc+b works fine, for xterm there is a workaround with metaSendEscape, but i need this behaviour in console, i'm out of google terms

mp3: what should Alt+b/f do? move the cursor?

yes

ah.

mp3: well, Alt+b/f work in vi mode as well. my .inputrc isn't special either. i'm afraid i can't tell what's causing your ffffffffff

in console or in X?

ssh with x-forwarding

man zshzle maybe has answers

just tried alt+b/f in console with zsh, works for me

(vt, if that matters)

mp3: also, thank's for the alt-b/f keycombo. that's handy

yuripv, echo $TERM ?

xterm, the default setting

'kbdmap' and then selecting US and it started to work

which kbdmap were you using previously?

anyone knows how can I reset the network stats?

they're counters, you don't reset them (other than when they're reset from boot)

its up to your time-series database that you're hopefully ingesting the data into, to handle that

antranigv: the stats in netstat -s can be reset with netstat -sz, the ones in netstat -i can't as far as i know

rip6: 18446744073709551610 delivered

lw what about the stats in `ip -s` ?

what is ip?

lw oh sorry, wrong chat. I was asking the same question in #linux. I'm debugging a connection issue between FreeBSD and Linux :D

antranigv: check out siftr(4) too

topic

*ahem*

so, i'm sure you've heard this one before… can i `freebsd-upgrade upgrade -r 14.0-RELEASE` directly from 12.2-RELEASE-p15 (last -p release of its series, i made sure of that), or should i go to 13.2-RELEASE first?

(or maybe also to 12.4-RELEASE, the last 12.x, first?)

n-st: so long as the version you're upgrading from is still supported, it should work

debdrup: hmm, 12.2 went eol March 31, 2022, and 14.0 was released on November 20, 2023, so i guess that's not a safe path?

12.4 was still supported on November 20, 2023, though (went eol a month after that), so i assume 12.2 -> 12.4 -> 14.0 would work?

I would probably upgrade through 13 from 12 to 14, but that might be excessive.

12.2 didn't get the fix for EN-23:12 which prevents it from updating certain files correctly when upgrading to 14.0 freebsd.org/security/advisories/FreeBSD-EN-23:12.freebsd-update.asc

so at least the you want to upgrade to 12.4 first

freebsd.org/releases/14.0R/installation/#upgrade-binary mentions "Systems running 12.4-RELEASE or 13.2-RELEASE can upgrade as follows", so skipping 13 is probably fine. but what debdrup said makes sense, there wouldn't be a tested upgrade path originating from a version that was already eol at the time

Title: FreeBSD 14.0-RELEASE Installation Instructions | The FreeBSD Project

so long as the key fingerprints and update servers haven't changed, the actual "apply binary patches to files as needed" should work regardless of the version, i think.

i think not, without the fix for EN-23:12 you will end up with missing files in /usr/include and the C++ compiler won't work

lw: freebsd-update will grab all the patches from the version you're on to the one you're going to, including patchsets, so i'm not sure how that'd affect things - but, well, it'll be interesting to find out, i guess :D

remember: backups don't count unless you follow the 3-2-1 strategy ;)

debdrup: it will download the patches, and it will fail to install some of them, because of a bug in freebsd-update itself which was only fixed in 12.4

I had issues from 13.2-RELEASE-p4 to 14.0-RELEASE. Got stuck with the "missing files in /usr/include and the C++" issues. Rollback bectl, updated to 13.2-RELEASE-p9 (p9?), and it worked.

It was issue #273661 I think

273661 – freebsd-update install: ///usr/include/c++/v1/__string exists but is not a directory bugs.freebsd.org/bugzilla/show_bug.cgi?id=273661

yeah, that's the one

seems like a fun thing

confusingly it still produces a (different) error message even with the fix, but it does upgrade correctly

For me it was error free from 13.2-RELEASE-p9. Not even sure if it's p9. The latest 13.2 patch

it was supposed to be fixed in 13.2-p4 according to the EN

i wonder why you ran it into it on p4

Or it was,.. p3? Sorry.. can't remember :p

p3 probably

ah

maybe it got fixed better later, i think i got the 'new' error last time i did an upgrade from 13.2 to 14.0

that was a while ago so honestly can't really remember

Oh I mean my system was p3 instead of p4.. Yeah p3 I think

oh yeah, no, i just mean when you did the upgrade the second time from a later version maybe a better fix was committed so you don't get the error at all

I see

hmm maybe not. oh well

n-st: anyway the tl;dr is don't upgrade from 12.2 to 14.0 because it'll break your system, upgrade to latest 12.4 first (or latest 13.2 if you like)

If can use bectl, and its not headless, or not remote, usually ok to try anything. I prefer going step by step otherwise

hi, anyone could help me configure rabbitmq ?

I get a 'node 'rabbit' not running at all' error

is there a standalone vnc client for freebsd that can run directly from console and do 1920x1080 ?

or better yet a "dual head" vnc client that can output on two monitors?

i don't think the freebsd console supports graphics, does it?

i'd expect something with builtin generic vga drivers like windows

there's a VGA driver but it's just used to run the text console (vt(4)). in theory you could write a VNC client that uses DRI to draw the display but i would be surprised if that exists

so it all runs under X or something?

X or Wayland

Linux has (had?) something called svgalib that let applications draw graphics on the console but i don't think freebsd has any equivalent

ah well ok.

anyone managed to add a boot splash with efi?

Bheam, running the vncviewer on top of X probably the simplest. Feasible at least. and optionally some lightweight WM like TWM

trying tightvnc

bah. tightvnc is no fun. default fullscreen is maximized across all screens.

vnc in general on *nix is not fun

ssh with X is better in my experience

on a gigabit lan, yes

rdp is better than vnc (with regards to remoting to windows)

i have a gigabit lan on the server side at least

right i got vncviewer (tightvnc) running 1920x1080 showing bhyve. usb, nic and audio passthru. gpu passthru will have to wait.

now. can i add communication between guest and host by dedicating 2 lan ports (one on host, one on guest) and connecting them by physical cable?

stupid but works? better solutions?

hm, pkg query doesn't have some way to query some ports and all their transitive dependencies?

nimaje: what do you mean by "transitive"?

yuripv, transitive. . . like A depends on B, B depends on C, so A depends on C too.

if A depends on B and B on C when I give it A I want to have the query-format evaluated for A B and C

hm, as I generate a dot graph anyway a tool to remove everything not reachable from the given nodes would work too

debdrup: my new mobo won’t even POST did you need to “bootstrap” it somehow to set eg RAM settings or anything?

dch: AM5 motherboards can take several minutes to train memory first time (or if you disable memory context restore, every boot) - it's just an artefact of DDR5 being what DDR5 is.

Think of it like servers used to work. :P

(Servers used to have memory context restore disabled, so training memory could easily take half an hour)

Ah that explains it. Relatively recently, I tried building a PC with an Asus AM5 motherboard and noticed it would take a bit of time to POST with it. And I ended up returning it due to untolerable coil whine.

good call

though I wonder what coil whine is

Title: Electromagnetically induced acoustic noise - Wikipedia

that sort of high pitched noise that a computer can make when starting/running?

Yeah

wow

hmmm I guess this training is less than 10-15 minutes?

more to the point it should POST and beep before it starts that, right?

same here with intel 1700 motherboard (with 14gen cpu), it took several minutes on first power on

(just bought and was worried that something is wrong :D)

i hope i remember reading this the next time i build a DDR5 machine

Well this has been 20m and not a peep so I guess it’s borked somehow

weird problem: i have multiple vservers at a provider in the netherlands, and after upgrading two of them from freebsd 12.4 to 14.0, they consistently boot up with a blatantly wrong system clock, always somewhen on Tue 8 Apr 2025

i can manually set the clock (necessary since ntp refuses to correct such a big discrepancy) and everything runs fine, but after a reboot i'm back in the future again…

That is pretty crazy. Do you have ntpd_enable="YES" and ntpd_sync_on_start="YES" set in your /etc/rc.conf file?

ntpd starts on boot and then balks "Clock offset exceeds panic threshold." into syslog

even the very first message in the dmesg buffer is already timestamped wrong:

Jan 17 21:05:39 beithe syslogd: exiting on signal 15

Apr 8 18:44:34 beithe syslogd: kernel boot file is /boot/kernel/kernel

(reboot takes about 15 seconds on this machine)

i suspected the kvm hypervisor, so i booted into the rescue system provided by the hoster. comes up with Wed 17 Jan 2024 just fine -- so either they have a very fast (and less squeamish) ntpd then freebsd, or the hypervisor clock is correct and it's only going wrong in my OS somehow…

Do I have to nuke /usr/ports before I can git fetch the latest 2004Q1?

n-st: so you don't have ntpd_sync_on_start="YES" which allows the first sync to exceed the threshold

n-st try with time.windows.com or some other timeserver, as a check?

AumShivaya: where does your /usr/ports come from? is it a git repo?

AumShivaya: no, `git fetch origin 2024Q1` should do it, assuming its already a git repo

no, it was installed at time of install (bsdinstaller)

that is the thing

nimaje: ah, i didn't know that it would do that. but that still leaves me with the wrong time at boot, and only once ntpd has started does it overwrite the time with the correct one

vnstat starts logging with the wrong time, for example. so that's not really a solution yet…

Jan 17 21:22:13 beithe syslogd: exiting on signal 15 --- Apr 8 19:01:08 beithe syslogd: kernel boot file is /boot/kernel/kernel

the offset from reality is pretty consistent, which is also odd

oh, and here's the next data point: it is always at that offset *from reality*, not from the time set during the previous boot -- i've just set the system time to February 2023, and it still came back as April 2025 after the reboot. so there is outside influence of some sort.

yeah, ask your provider, using ntpd_sync_on_start="YES" shouldn't be necessary in your case, it is just a workaround until you get it fixed, ntpd_sync_on_start="YES" is for systems that can't hold the time (or not good enought) but in your case it *should* be able to

but (checked again right now) the rescue linux boots up with the correct time, it's not just getting it from ntp…

n-st, Do you have ntpd_enable="YES" and ntpd_sync_on_start="YES" set in your /etc/rc.conf file?

I asked that before but do not see a confirmation one way or the other.

If you don't have those set then that is the problem. Set those and then reboot. If you do have those set then it is worth debugging further.

rwp: no, that is not the problem, as the system starts with a completly diffrent and wrong time after setting a time and rebooting

Okay then.

oh and now we have Jun 18 2025 13:45:11

(this is after rebooting from the rescue linux into my freebsd)

(see the message from 21:31:03UTC)

nimaje: yeah, but what do i tell the provider? "my clock is off on my own OS, but works fine on your rescue system"? :/

nimaje: (also, bold to be referring to timestamps in this situation :P)

that one was as a reference to rwp

n-st: both freebsd 12.4 and 14.0 are official releases or was 12.4 from your provider, if so, do you know if they have any changes to upstream?

nimaje: all official releases, the provider doesn't do *bsd at all afaik

(bootstrapped via ipxe and netbooting back in the day, i don't fancy doing that again anytime soon -.-)

can you boot 12.4 again on a system where you updated to 14.0 to see if you have that problem then too? I have no idea why 12.4 and 14.0 should behave diffrent or how upgrading to 14.0 should have broke the hwclock and on two systems at that

so do I nuke /usr/ports ?

nimaje: eh, with boot envs, maybe? looks like i have them, but i don't know how to switch…

nimaje: ah, from the bootloader, just following the menu instructions

so that shows: it's not the freebsd version, even 12.3 is broken

ok, are you sure it wasn't broken before? maybe you just didn't notice can you check one of your remaining 12.x systems?

AumShivaya: yes, if it isn't a git repo, then git can't use what you already have there

nimaje: the only remaining "old" one is my backup fileserver, so i'd rather not touch it until i've figured this out… it currently has the correct time, but that might just be because it hasn't been rebooted in a long time

i'm pretty confident that the update wasn't actually the cause; it just exposed the problem because it involved a reboot

ty

so it's ticket time… and then waiting until the morning, unless the provider runs a night shift

n-st, i've had different results in a KVM host along the years. From network/clock only working in full virt mode in the past to things just working right in the present.

If your provider don't explicity support freebsd so indeed a good idea asking. Knowing what timer are enabled (pit rtc tsc kvmclock). I dont know exactly which ones are supported in freebsd, or if all of them

'sysctl kern.timecounters' from guest should spit out some clues I think

does anyone run thin jails inside thick jails?

and i wonder what the overhead cost is for the middle jail?