PF-related question (again, due to no response....) from what I can see, match rules don't create states - If multiple match rules exist - they are all evaluated. The question is - why do these rules allow, for example: match in on ix0 inet proto tcp from any to 11.22.33.44 port = 60000 flags S/SA keep state (source-track rule, max-src-conn 1,

overload <maxconn>, src.track 10) match in on ix0 inet proto tcp from any to 11.22.33.44 port = 60000 flags S/SA keep state (source-track rule, max-src-conn-rate 1/1, overload <connrate>, src.track 10)   but nothing gets overloaded in any of these tables?

quiet as a tomb ...

does samba exists on bsd?

bzimage: yes

thank you

ridcully: can i set up domain contollers and everything like in linux?:)

never tried it bzimage

thank you

did something change in sudo or something else lately? i used to be able to type 'sudo myscript' and it would find and start ~/.local/bin/myscript. but now it says command not found. i didn't make any changes to PATH in my ~/.login_conf or anything like that. just did a pkg upgrade, rebooted, bam doesn't work. 14.1-p1 fwiw

i'm wondering to which home is `~` expanded to?

~ refers to the user

like /home/l00py/.local/bin/myscript

aaah, i remember now. it's expanded by the shell, before providing it to sudo. i was wondering if it would try the home of root

maybe something to do with secure path? seems new in sudo

nothing with ~ was passed to sudo

all i type is `sudo myscript` and it used to work (until an hour ago)

ya it looks like secure_path is now enabled by default. the sudo docs suck, looking for how to disable it

so what's the syntax to disable secure_path and return to old behavior? Set secure_path false in /usr/local/etc/sudo.conf didn't work

l00py: in the link you pasted: You can disable using secure_path by deleting or commenting it out from the sudoers file. In this case, sudo falls back to the old behavior.

i had to add Defaults !secure_path to /usr/local/etc/sudoers/DEFAULTS

if i run su in a dougrabson/freebsd14.1-small podman container, i'm getting Shared object "libbsm.so.3" not found, required by "su". how do i install this library?

bjorn3: on the off chance podman uses pkgbase, you need FreeBSD-libbsm package - if not, no idea (i don't use podman)

ivy: that worked. thanks!

nice

Hi, I have noticed that the terminal does not work well in vim when I am remote logged (rlogin) to a colocated server (I have a workstation + a couple of rack servers colocated). If I do LANG=C then it works well, LANG=C.UTF-8 by default. 1) where do I define how LANG is configured in FreeBSD? 2) does this resonate to anyone? C.UTF-8 causing issues with remote terminals?

uskerine: I didn't know anyone still used rlogin. But check your .login_conf

or /etc/login.conf

i have a vague memory that freebsd rlogin/rsh only supports obsolete cyphers (DES?) even though Kerberos supports newer cyphers, which would make me wary of using it

rlogin etc aren't in base any more and the man page from 13.2 says "The rlogin utility will be replaced by telnet(1)in the near future." :-D

it's not in base but i believe cy@ still maintains it in ports

Is there any way to *embed* the GELI passphrase on FreeBSD so that it won't ask for it anymore each time the system boots?

tercal: that, if possible, would sidestep the whole point of having GELI

I think you can, once booted, remove the GELI requirement, but I don't know how to do that

i got a bsdinstall script that started failing. there's a line in it that runs 'route add default 10.1.1.1' but errors with invalid argument. i found a freebsd forum post talking about the syntax changing wrt fibs or something?

14.1 btw

Yo. Where would one look for packages made via "make package" for a given port? I remember that there were some .tbz or .tgz files created somewhere in /var in the past. Can't find any such package files anywhere after `make package` issued in port directory - dpaste.org/dMAfU/raw

nie to okienko

Yo. Where would one look for packages made via "make package" for a given port? I remember that there were some .tbz or .tgz files created somewhere in /var in the past. Can't find any such package files anywhere after `make package` issued in port directory - dpaste.org/dMAfU/raw

It may be possibly related to the fact that I've used freebsd-update and buggy etcupdate. My ports are telling me "Ports Collection support for your FreeBSD version has ended, and no ports are guaranteed to build on this system. Please upgrade to a supported release." while I'm up to date with /usr/ports (I can see fresh commits from today), and uname tells me that I'm on 14.1-RELEASE-p5

regis: it's in <portdir>/work/pkg/

ivy: I can only see work-dirs needed for compilation there. Not the resulting package file.

are you sure the build suceeded? i just tested it and it worked here

[61!] hemlock ~/src/ports/lf/security/dehydrated [lf/main]% ls -l work/pkg

total 41

-rw-r--r-- 1 lexi users 38770 Oct 28 14:02 dehydrated-0.7.1.6.pkg

evilham: the build succeeds. `make package` succeeds as weel (can't find resulting package archive though). BUT, I'm on 14.1-p5 done fia freebsd-update and on `main` ports branch, yet I'm getting silly "Ports Collection support for your FreeBSD version has ended"

I'm trying to build Dovecot from ports because the package dis compiled w/o MySQL support and my setup depends on both Postfix and Dovecot talking to MySQL.

It's quite silly. I've never had these issues with source building and mergemaster. freebsd-update and etcupdate seem to be half-baked every time I try to do the stuff that worked before.

And I'm talking about release versions, while sporting CURRENTs on my ThinkPad.

stl ivy in a LAN environment where you have several servers they are still useful. They cna be installed separately, unfortunatelly there is a bug so rcp does not work. A header in base points to /usr/bin/rsh instead of /usr/local/bin/rsh

i assume you find them useful because hosts.equiv is easier to set up and shosts.equiv for ssh?

s/up and/up than/

i/29

sorry

yes it is pretty straightforward, also forwarding X11 in rlogin is straightforward

if you are developing something in a controlled LAN segment, there are no security concerns and there is little to gain with ssh

I would like to be able to use rcp, when I have a free slot I will try to learn how to submit a bug or a patch for that header

Hey folks! How hard is it to get sound to work, on Lenovo P50 (Skylake, snd_hda probably)? mpv plays stuff without errors, but I don't hear any output. I had a look at the docs, but it looks scary.

cizra: you mean the internal speakers?

(have P51 and it doesn't seem to have any issues like that)

yuripv: Internal speakers, cable output, or Bluetooth headset. I haven't gotten around to thinking of the Bluetooth yet, so far I'd like at least something to work. Thanks for the info.

OK, I'm having a major WTF moment. Here, a reboot later, with no changes, sound works. Maybe the difference is between being docked or not docked?

it's been over a decade since I had a thinkpad dock, but don't they also route analog audio through them, once connected? (has its own headphone socket or whatever)

could be, yes

Yes, the dock has a headphone socket. Very comfy for leaving the headphones at work desk.

Is there a mechanism for choosing/switching the output?

cizra: i don't do anything fancy with sound but there's a couple of sysctl that are interesting. first cat /dev/sndstat to see what devices there are

pcm0: <Realtek ALC298 (Analog 2.0+HP/2.0)> (play/rec) default, and pcm0=microphone <- and that's regardless whether I'm docked.

then you have hw.snd.default_auto (which defaults to 1), which will try to choose it automatically (plug detection i guess?)

and you can set hw.snd.default_unit to the one you want to use

hw.snd.default_auto was 1. hw.snd.default_unit was 0 (that's a pointer to my pcm0, I suppose?)

yeah. sounds like you only have 1 device? which might make sense for a laptop. but mine gives me 8 pcm devices (3 hdmi, 2 analog, 3 USB devices...)

I don't know how it chose the correct unit (default_auto=1) for me, but it did and I'm glad, I don't need to touch it ;-)

I occasionally see some playback issues. Often it is caused by a desktop application hat is connfig'd to use Pulse Audio rather than OSS

mate "sound preferernces" shows several devices available, though. You can pick default on the input and output tabs. YMMV with gnome/kde/other-box

for me hw.snd.default_unit i 0, may be affected by the mate sound settings thingy

let's say in loader.conf i have kernel=kernel.FOO; foo_bar=YES; bar_load=YES; now i'm at the "OK" prompt and i want to boot kernel.BAR instead, but i still want to load my modules. what do i type?

s/foo_bar/foo_load/

I've never actually tried specifying a non-GENERIC kernel from loader.conf like that... 'man nextboot' maybe?

i don't see anything immediately relevant there sadly

I'm trying to get about 50 png files that are 28.5MB into an email (gmail) that only can hold 25MB how can i do this and how can i make them alphabtical order or is this even possible ?

oxbar, The usual thing is to upload them to an image site and then send the URLs to the images.

You might have heard someone say that email is not a file transfer protocol. Though we do abuse it for one often. It's okay for small things.

If you must email photos as attachments then only send one at a time, or three at a time, until you have transmitted all of them.

When I need to do something like this I upload them to my own web site and then send the URL in the email.

Hi

If they must be a bundle then I tar them up for Unix folks or zip them up for MS folks and post a URL to download the file bundle from my web site.

rwp: my wife is not techincal at all. she wants them in an email so she can see them download them put them in order/rename them etc ..

Of course! Most people can handle links okay. But for what you described I would send them attached either one at a time or attached three at a time.

hello

i use freebsd as a nat/router service among other stuff, it has 2 nics, one external one internal. I run a game server on port 27500 which listens on all interfaces, the problem is that i cant connect to the internetip, just the internal one. for others outside my lan the external works. can i fix this with some route or pf trickery?

sopparus: freebsd got a discord also.. they got channels like #networking and #helpdesk.. you might get a answer there faster

just fyi

if you can't get *OS* support without using discord that doesn't seem great

sopparus, I am somewhat surprised that does not Just Work already. It Just Works for me. I just checked. Therefore it must be something you are doing in your pf.conf file that other people are not doing. You would need to post it.

mewt: you can get OS support here or on the mailing list, no one is saying you *have* to use Discord

it was just an option.. this place dosen't answer right away like hours or so depends on who is here i guess

i'd love to answer the question but i do everything i can to avoid hairpin NAT because it's so gnarley, but i'm sure other people know the answer if OP hangs around long enough

i don't think it's something an open-source project should really...endorse by use at all but i've apparently been able to forget it existed til now, so shrug

my bad then

i do agree that in general freebsd inc. should not endorse discord

but... it does... and there are people there who could maybe help

rwp: dumpinen.com/ZGHxrge3et-

i do prefer to point people at the mailing lists before discord though

I personally don't like web forums. They just don't work well for my work flow. Mailing lists are good. IRC is good. For other people it is other things.

well, unless i'm compelled to use it for some reason i'll continue pretending i don't know about it

sopparus, Maybe. This is at the edge of my comfort. But if you want to allow your LAN try adding "set skip on $intif" on the line after "set skip on lo0" and see if that allows it.

That will of course allow anything on your LAN side interface. But I think that is what you are wanting to do. I allow it here too. Some people like things otherwise though.

different strokes for different folks

kevans: i love you but this is a terrible opinion

discord is literally trying to privatise all real-time discussion on the internet

you can't just "well you do you" that, what they are doing is literally killing open discourse

i mean, i'm not a fan

you can't really deny that irc, discord and matrix all have likely relatively disjoint userbases, and it'd be silly to put all your eggs into one basket when you're seeking help (from the perspective of someone seeking help)

kevans: indeed... matrix is at the polar opposite end of the spectrum of "conversation privacy"

I miss the "happy medium" days....

it's sad that all we can offer in opposition to Discord is Matrix which is fucking terrible

and IRC which only sad old millennials like me use

yeah. i get pissed every time I look at my matrix tab in thunderbird

it could have sucked less with better clients

"it'd be silly to put all your eggs into one basket when you're seeking help" - i agree that if FreeBSD Discord exists, freebsd users should go there to seek help, my objection is that FreeBSD Inc. should not officially endorse Discord as a platform

ivy: hey... I use irc.....and I'm before the millennials....

I like irc...

I am one odd one though who prefers using a Mac as his input machine...

ivy: as far as I'm aware there's no such project endorsement

and I wouldn't say "should go there" without qualification

kevans: here is a website owned by FreeBSD Inc. that explicitly endorses Discord wiki.freebsd.org/Discord/DiscordServer

wiki is community maintained, much like the discord server

ok, you are going to reply that they didn't put out a press release saying "all FreeBSD users should join discord", but this is the first ddg link for "freebsd discord" and the first link is the discord itself

as far as I'm concerned the project only endorses mailing lists and matrix, both of which are services it hosts

brb

kevans: did you see my latest panic

sopparus, Meanwhile... Try that change I suggested. Let us know if it solves the problem or not. Also let me know if leaving it out was intentional or accidental.

ok will try

no, it didnt help

i cant remember, i made the cfg several years ago :)

kevans: Eh? The project endorses IRC I thought.

I see a Discord link here: freebsd.org/community

ivy: There are other proprietary platforms listed there too. And ironically I don't see Matrix listed.

yeah it's weird FreeBSD Inc. links to a website run by a Nazi, but maybe they just didn't get around to removing those

sopparus, Hmm... Sorry that was my best shot. It's really just outside my comfort range of knowledge about it. Here is my current config: paste.debian.net/plain/1333744

sopparus, Strike that. I realize now that it is probably working for me due to my local routing actually hitting the $ext_if side of that interface not the $int_if side. My confusion. It's an odd setup here. Meaning it might not be working for me using that configuration.

ivy: no sorry, very busy concreteing lately

kevans: is that a git thing? git apply-concrete-ref?

I suspect it involves work gloves, shovels, and back breaking labor.

isn't the whole point of knowing C that you don't have to do that?

ivy: no... its called part of home renovation (at least for me)

wait, you mean you do it on purpose?

Yep

... are you American?

Yes but why does that matter?

i think only Americans do this :-)

for fun, that is

Who said it was for fun? Its called being responsible for my place of living..

ivy: no it's a giant hole in my backyard and 16,000 pounds of concrete in my shop

=)

ober: If I could "safely" multiboot this one.. I'd put it on here lol

This MBP is a behemoth

you have a shop? i am so confused

owning property must be a novel concept to ivy

ober: he claims to but he won't make freebsd work on my mac mini, so obviously lies

the_oz: i'm a millennial, we literally invented the renting generation

one of the discord folks finally unblocked the interrupt work that we needed to make progress with apple silicon

unfortunately the only other person working on it passed away after the last bsdcan, but I think I have all of his latest patches for upstreaming

now it's just a matter of finding time

gotcha

the contributor on discord has done some additional work, but it'll be a little more challenging because he's rust'y

i  got a bsdinstall script that started failing. there's a line in it that runs 'route add default 10.1.1.1' but errors with invalid argument. i found a freebsd forum post talking about the syntax changing wrt fibs or something? 14.1 btw

FreeBSD has a discord?

That was news to me too

Administration got political and violated platform promise.

Discord admins, namely.

"add net default: gateway 10.1.1.1 fib 0: invalid argument"

the_oz: you mean they banned people for being homophobic or something, i guess?

What's the difference between discord and well...here (other than all the problems associated with discord like bots, spam and ease of account loss)?

ivy, why do you feel the need to defend what you had no part in?

100py: you mean route with option '-fib 1'?

the_oz: so, yes. good to know

TommyC: we got rid of these problems 20 years ago

Without any reasoniung at all you will besmirch complete strangers. Is it that hard to believe that power gets abused?

100py: I am sorry. I did not quite catch up. are you trying to have one more routing table?

luke_jobless_sb no my bsdinstall script has exactly 'route add default 10.1.1.1' and it normally works but on this new machine i'm setting up it errors while running saying "add net default: gateway 10.1.1.1 fib 0: invalid argument"

i got disconnected

hello bsdi baddies! Help me

geom disk list shows me that two hard drives have same ident

100py: I don't know. try: route add default 10.1.1.1 -fib 0

how can I add them to zpool by diskid then?

100py: I am just answering mundane according to the error message since I do not know any detail

luke_jobless_sb i can try that but the man page for route says my syntax should work :(

i'll try tho

l00py59: Can you show us that bsdinstall script?

dh it's pretty long. lemme reduce it to the smallest

noooo not again panic: got NULL turnstile on rwlock 0xfffff80010b1b890 passedv 1 v 1

i heard I can gpart label disks with GPT and such. But my freebsd install has flacky gptid

Sometimes /dev/gptid wont show causing troubles importing zpools

please help. freebsd forum is under maintenance

hi

ScrewDriver1337: I don't know. I have never needed gptid for importing zpools

what happends if I do zpool create ... /dev/da0 /dev/da1

and then I reconnect drives in different order? will ZFS recognize who is who?

ScrewDriver1337: You mean after creating table and partition?

yes ig zpool creates gpt table and such

because drives are clean completely

ScrewDriver1337: So you wanted to ask how to partition and formatting disk?

NO

ScrewDriver1337: you may?

no

I always use GPT labels for my ZFS pools. So that when a drive fails I can identify the drive. But ZFS drives may be imported in any mix of devices.

even if I format disk with gpt table and partition, and reference it by unique gpt id

the issue is that /dev/gpitid does not work :(

I have kern.geom.label.gptid.enable="1" in /boot/loader.conf and still see no /dev/gptid

What is /dev/gptid? I have not heard of it.

is it linux?

no... freebsd

how do you create zpool then?

format to zfs and create?

I have a pool that was created like this: zpool create -o altroot=/mnt -O compress=lz4 -O atime=off -m none -f zroot raidz2 gpt/z0K1G42J7B gpt/z1K1G4UZ7B gpt/z2K1G496BB gpt/z3K1G3T97B gpt/z4K1G4UZXB gpt/z5K1G4G4RB

gpt is /dev/gpt thingy

That was after partitioning the drives and setting them with GPT labels.

and I dont have /dev/gpt for some reason

The FreeBSD kernel allows access to a disk device by only one path exclusively. If you access a disk by /dev/ada0 then all other paths are immediately removed.

so I cannot have both /dev/diskid and /dev/gpt ?

What is /dev/diskid?

kern.geom.label.disk_ident.enable="1"

The BSD installer set up my systems with kern.geom.label.disk_ident.enable="0" kern.geom.label.gptid.enable="0" in /boot/loader.conf and I never knew enough to change those settings.

And so that explains why I never knew about /dev/diskid or /dev/gptid and never needed them I guess. I am using GPT labels instead.

thats interesting

can you point me to the man page? because forum is down and I cant find any information

I would start here: wiki.freebsd.org/MasonLoringBliss/ZFSandGELIbyHAND

It appears what you are using is from glabel(8) but those are from before my time with FreeBSD. gpart(8) is the way that has been recommended to me since I started using it.

do a `zpool status` and see what's going on. If you rebooted and something failed or it got accessed via a different means than normally that could explain the missing expected previous method

thing is

Unfortunately I only know of a book to get a good overview of this, namely Advanced ZFS by Alan Jude, Michael Lucas

my primary hard drive is reference by diskid

diskid/DISK-WW63PMYJ ONLINE 0 0 0

so I cannot turn off disk_ident

but to /dev/gpt work disk_ident needs to be turned off

under the "ZFS Best Practices" chapter

I have never used glabels as I explained but I have used device names /dev/da0 or /dev/ada0 and so on and have converted from those to /dev/gpt/* names. One must boot other media, such as the installer disk, and then import the pool telling it to use the desired names, whichever names those happen to be. Then export the array. Then reboot normally and the new names will be used.

In the "man zpool-import" page it says: -d dir|device Uses device or searches for devices or files in dir. The -d option can be specified multiple times.

Which means that an import of "zpool import -R /mnt -d /dev/gpt" booted from the installer would import using the new device paths. Then one can reboot and it should use the new device path.

Going the reverse direction to glabels should work too.

ok

gpart populated /dev/gpt

now lets see if it persist by reboot

yeah

ScrewDrtabtabtab

welp

ScrewDriver1337, files.catbox.moe/72wrwv.png

starts at "The Importance of Labels"

+1 for good documentation!

ScrewDriver1337, Please see and read the_oz's reference above.

the_oz: thanks!

dh still here? finally got script pared down

it's termbin.com/2op7

Thanks rwp

oxbar, Good luck!

=)

anyone know why my simple 14.1 bsdinstall script termbin.com/2op7 makes the error "add net default: gateway 10.1.1.1 fib 0: Invalid argument"?

fwiw in the ifconfig outputs, the ip i assign the bge0 if never shows up

termbin.com/8o5c includes my etc/rc.conf

l00py: I don't see any bge0 configuration attempts in that script

ya it's in the file

What are the best time/activty for this channel ? cause it can go for a while without a response

I have to remember timezones also

l00py: Not in the excerpt you pasted

dh no i mean it's in the system.tgz. have you used bsdinstall? you combine an installerconfig file and a system.tgz fs overlay

I have used it and have developed install scripts. However I have no idea what system.tgz is or what it contains.

all it is is files that are expanded into place over the base OS install

I assume what ever network configuration it contains, is not actually functional.

well i've used this exact setup on other machines recently and it worked fine. and the man page for route supports what i'm saying

not sure what "Invalid argument" refers to in route add default

You said that bge0 isn't actually configured

ya

So .. networking configuration doesn't actually happen as you expect for whatever reason.

think it would need to be configured with an inet for route add default to work?

Yes

ok guess i'll start there

if i put ifconfig_bge0="inet 10.1.1.14 netmask 255.255.255.0 group langrp group pubgrp" \n ifconfig_bge0="up" in /etc/rc.conf.d/netif and run service netif restart, shouldn't that add the inet config to bge0? it doesn't show up in ifconfig

maybe that 2nd up line is the prob

the second line is overwriting the first

Howdy, folks!

ya

o/