I don't want to get sucked into the "compile everything" mentality that I got when I used gentoo

JetpackJackson: the joys of freebsd is you can selectively compile what you need

I compile security vulns as they can be slow to update

I guess

I just hate rust compile times

How much space do I need for a poudriere build jail if I decide to go that route? Rather than just doing 'make config' in a ports tree

JetpackJackson: for one port its not really word setting up poudriere, but for the most part the space requirement is low

afaik it spawns the jails on demand

Alright

Im planning on doing the build in a VM and my laptop has a bunch of really chonky school projects on it

I should try using bhyve to virtialize the os we used in my os class

does anyone have an idea how I can setup the webcam?

sixpiece: is it recognized?

yes it is recognized

it came build in with my laptop as well

I mean it's recognized in the sense that it knows what it is but it appears their might not be a driver for it or whatever recognized via shell not by like going into firefox and then zoom.us and starting a meeting

check dmesg

look for the device

also there is a command to check but I forgot it

I think its under the wifi or audio section of the handbook

it lists the devices + its driver

iirc

JetpackJackson: nothing wrong with compiling all the things portmaster does a good job doing that

JetpackJackson: just dont be a weirdo and watch it compile heh

poudriere is good for port dev tho

or if you compile a lot of ports

cpet: I watch things compile

I watched buildworld

weirdo

for most of the 6 hours

i have a scrip that just says Compilg X - Done

boring

dont you like looking like a hacker >:)

thats not how you look like a hacker

lol

nah I was kidding

I just like how it looks

bullshit you were

-_-

some people like to do it i dont

tmux new compile args

go one with day

wait for an email saying your sht broke

when I follow the instructions webcam -l does not show up in the list

the more text on my terminal, the more enjoyment I get :P

sixpiece: likely upsupported

sixpiece: did you run webcamd

if you can extract the device info from dmesg it could be useful :)

I would check dmesg lists it first

thats step 1

jesus he watches dmesg as well

cpet: dmesg is useful ffs

hah

everything you need is usually in dmesg

my msg buf is full of log in vain shit

cpet yes I tried to run webcamd it doesn't work

maybe restarting will help?

sixpiece: which webcam is this ?

it's on my laptopp

again which webcam is this

if you dont know your hardware well

but it theory if webcamd doent find anything you should just give up cause its probably not suported

you will learn thats one of the issues with freebsd top of the line isn't good

which is why im surprised this new mini PC worked fine with freebsd with the exception of the ethernet card which required the driver from ports

other than that and the issue with the cable works like a champ

im not a fan on intels E core and P core crap but hey

cpet: I would still suggest reading dmesg as step 1

xD

[769951] TCP: [185.156.73.182]:44276 to [69.92.116.218]:15356 tcpflags 0x2<SYN>; tcp_input_with_port: Connection attempt to closed port

[770067] TCP: [88.210.63.191]:47999 to [69.92.116.218]:15459 tcpflags 0x2<SYN>; tcp_input_with_port: Connection attempt to closed port

ok

not that shit, the boot part lol

oh whatever, you handle this im off to bed

oh you have to be specfic man

ugen0.3: <Lenovo EasyCamera Sunplus Innovation Technology Inc.> at usbus0, cfg=0 md=HOST spd=HIGH (480Mbps) pwr=ON (500mA)

right so freebsd does at least see the hardware

cpet I don't want to paste too much and my webserver is also down having a lot of freebsd issues now but the laptop looks very promising

so the issues will resolve themselves but for example my freebsd server is also down

sixpiece: you are running webcamd as root or as a user in the group right

I tried as both

sixpiece: see if service devd restart does anything

well then its likley unsupported

I think I tried that

you think?

submit an issue to bugzilla giving the dmesg output, and the webcamd output

giving the laptop you are using as well might help

do you have cuse running ?

ok sounds good

er loaded

topic something like "webcamd unable to detect integrated camera on x laptop"

and then wait a few years :p

sometimes you just have to buy a logitech one

the issue with laptops is they are basically un by software within software

and msot of that software is either prop or weirdly licensed

I don't have a bugzilla account but I guess I will go through that process

i dont see your card here

sixpiece: are you in the webcamd user ?

cpet: I used to watch lol

JetpackJackson: jesus you as well?

Keyword "used to"

but you did

do you guys get popcorn and watch it like you would Halloween?

run off when you see compiler errors ?

Nah I would sit and wait in case of error

sixpiece: id username to find out

JetpackJackson: i pipe it to mail and then just wait for a email

Yo foo your shit is borked, fix it and fir it now!

I should figure out how that works

Unixy mail stuff always seems cool

Even though I use Thunderbird lmao

only way to go

old school by using sendmail

and learning m4

which is confusing as hell

Oh m4

Haven't heard that in ages

What man page do I read for mail stuff

man mail?

I actually enjoy the fbsd man pages

Epic thanks

apropos mail

google mail freebsd

chatgpt why is the nastalgic use of mail on unix still popular

Lol

See I pay for mail but now I wanna host my own mail server

JetpackJackson: it easier to just use opensmtpd its one line tog et email going

JetpackJackson: i woudnt bother email is such a pain to setup these days

you have to setup DKIM DMARC SPF

I had configs for notmuch and neomutt and stuff

cpet: true

sometimes even request a whitelisting on the big guys

and if you have a shity ISP like you do

i really doubt port 25 is even open

So it'll send mail to my email right?

so you would have to do some magic to forward to a different port

i personally wouldnt bother but i guess in regards to learning

Or is there like some local mail program that reads some sort of email file

Nah I don't need to host my own email lol

It's too much

you can use fetchmail if its still around or getmail

or you can setup sendmail or opensmtpd to proxy to your ISP's email server

I think I used smtp something or other

i cant use any of those with proton mail

Have you tried hydroxide?

there fancy bridge doesnt work on linux (that well)

I used it for a bit

Oh wait idk if it works on bsd

Oops

Linux brained moment

im sure its possible

What email client do you use for mail

the issue with LInux is most dev will make the program only run on Linux vs making it portable

a lot of times sending patches will get a we only support linux so i just dont use it

this is the case with Enlightenment

Ah

Maybe I should set up notmuch again

Or configure Thunderbird to use mbox or smth

Idk

theres way to many programs that do the same crap as the older ones do

Im sleepy and I get sleepy-brained computer ideas

Wdym

Like mail and stuff?

redefning the tree doesnt really help perfect exmaple is sudo-r its not even a year old and already has some pretty nasty CVE's

Yeah

I just use doas

i use doas on openbsd sudo on freebsd

cpet yes I am sorry I was working on setting up fetchmail and mail to be able to send the request to bugzilla

cpet: oh interesting

cpet I am going to go to the grocery to get food before it closes

I don't have anything to eat and I need time there's ice on teh car etc..

thank you cpet as well be back in around an hour or so if you're still available mostly everything on the laptop is working great and I think it will be just checking the details obviously this has saved the computer

and I would like to make it perform at its fullest if possible and I will need help to get my server up and running again as well

see you in around an hour

ok

o\

hi

off-topic but it's fckin sunday: i.ibb.co/dJtXj3Xz/test.jpg <-- can someone please tell me why this doesn't work, but works fine wtihout the square root? that's obviously wolfram alpha since it wasn't in the capture

heh mathway just says solution graph

wife: "im going out in 5 minutes, make sure ur redy" . /me sets timer for 25 minutes <-- this is tonight's experiment

i got graphs 2

good luck

Matt|home: i see too many mistakes with that

i would just get ready and play the PS5 until shes ready as its to unpredictable on how long she will take, you arent going to win so why bother

ls

ls -a

I have to say "pkg" on 15 feels much snappier. Could be the 10G fiber line I'm now on but whatever, I'm happier! Well done folks

Probably the fast internet helps.

lots of changes were done with pkg

too bad we dont have fiber here

crb: could always join the 15+ email question on why we have clang in base and its not used with ports heh

i have both gcc13 and clang19 installed

exit

cpet: that depends on the program in the ports

voy4g3r2: yeap

how to import legacy pool before fstab?

zfs does not use fstab

zfs detects the fs is zfs and mounts it

mountpoint=none

zfs import yada yada

you use zpool to import a pool

how to import legacy pool before fstab?

zfs does not use fstab

fstab

# Device Mountpoint FStype Options Dump Pass#

/dev/ada0s1a / ufs rw 1 1

zroot/usr /usr zfs rw 0 0

well that could be why it doesnt work

cause that not how its done

# Device Mountpoint FStype Options Dump Pass#

/dev/gpt/efiboot0 /boot/efi msdosfs rw 2 2

/dev/nda0p3 none swap sw 0 0

/dev/nda1p3 none swap sw 0 0

do you see any fstab mount options in there ?

NAME STATE READ WRITE CKSUM

tank ONLINE 0 0 0

mirror-0 ONLINE 0 0 0

nda0p4 ONLINE 0 0 0

nda1p4 ONLINE 0 0 0

I need to import legacy pool before fstab

lets just keep repating it

why?

why before fstab?

set a mount point ad i am sure it will get mounted

zfs does it own thing as long as there are mount set it will mount them

think someones else was having some import issues he fx them by changing the mounts

Chip1972, Why do you need to import the pool before fstab? If you put the dataset's name in fstab are you getting errors because the pool seems not to exist?

because the boot crash and goto shell, because cant find zroot/usr

figure out why it crashed

have you already added zfs_enable="YES" to rc.conf?

yes

in the single user shell do you see the pool when you run `zpool list`?

my understanding of zfs is that mountpoint=none cannot be mounted, mountpoint=legacy can be mounted using mount -t zfs

yeap

however if the pool is in a inconsistent state it wount mount either

how using 'mount -t zfs' before fstab?

you don't; fstab uses the same mechanism as mount -t zfs

in theory single user mode only mounts /

so if you dont mount / the system wont even boot

so your before ftab makes no sense

how run ANY command before fstab?

... that's advanced administrator stuff, you shouldn't need to do that under normal circumstances

does it show mountpoint=none or mountpoint=legacy when you do `zfs get mountpoint zroot/usr`

if it is mountpoint=none then the dataset cannot be mounted

Chip1972: single-user mode

ek: its ok man

I mounted with boot iso

ek, Chip is unable to successfully start a system because /usr from a ZFS pool is not mountable

should still be able to single user boot

now if he messed up his pool well its a simple format and reinstall youre ..

how run ANY command before fstab? like 'zpool import zroot'...

dewd

3 people have told you

stop repating the same thing over and over

you don't.

Trying to catch up here. But, why is single-user mode not acceptable? Commands for mount should work given the correct commands are provided.

it is acceptable

i think Chip1972 is a bot really heh

ek, single user mode should be acceptable, in order to complete troubleshooting tasks

well I can always follow MelMalik advice

how run "echo hello world" command before fstab?

one does not do that.

do what ?

can you describe what did before you encountered the problem

correction: can you describe what you did before you encountered the problem

MelMalik: some people are just a lost cause really

did the system once boot successfully, and now doesn't, and what did you change in between those two states?

ek: :(

I copied /usr to zroot/usr mounted in mnt/usr with compression=zstd

what does one do to crash the system to the point of it failing

sounds like HW to me

the system boot without "zroot/usr /usr zfs rw 0 0" in fstab

/etc/rc.conf run before fstab?

right, so you added that line to fstab

/etc/rc.conf is read before fstab, otherwise `zfs_enable="YES"` wouldn't do anything useful

Why do you have any zroot mounting happening in fstab? Are you using an actual zroot setup? You shouldn't need anything in fstab for zroot mounts.

ek, This isn't zroot, it's /usr on ZFS with / on UFS.

Treat the word «zroot» here as an opaque string.

ek: some people are just a lost cause man

how to add 'zpool import zroot' to /etc/rc.conf?

you DON'T.

that's not how this is done

Chip1972: Well, I'm still very confused, but you could add that command to /etc/rc.local?

ek, would be too late for /etc/fstab, as /etc/rc.local is run after mounts

I mean, there has to be a better answer to all of this (again, a lot to scroll through here).

What is /usr (as mounted by ZFS afterwards) doing that would require pre-rc.conf?

Is this a full /usr directory that isn't being mounted?

Or is this a /usr/home thing?

ek, Correct. This is the full /usr directory that isn't being mounted because the pool isn't imported when rc runs mount -a

full /usr directory

Chip came in with an XY problem. I did not want to front load that because not only would they not understand, if they did understand they'd be quite rightly insulted. When I come across an X-Y problem I go straight to trying to backtrack and figure out what the X is, without initially mentioning that I think it's X-Y.

Okay. So, this is not a root on ZFS system? Just ZFS being used outside of UFS for /usr only?

The X is «I'm trying to move my /usr directory to a ZFS pool (root remains on UFS however) but the system is failing to boot because the zfs pool isn't imported by the time /etc/fstab is processed.»

how 'zpool import zroot' before rc runs mount -a

So, I'm not quite sure this is a possible scenario as UFS will be used for booting which would require /usr (base libs and such) to be available before the ZFS FS is available.

Chip1972: Not possible. You'll need to either leave /usr on UFS for the system to properly boot, or migrate to a root-on-ZFS system.

i like ek's "not possible"

However, you *CAN* move certain things from /usr to a ZFS mountpount/dataset as long as it isn't required by base to boot.

For example, if you're just trying to place /usr/somedir into a ZFS ENV, that would be fine. Or /usr/home. But, /usr/lib, /usr/libexec, /usr/local/anything-at-all will need to remain available on UFS for boot.

i worked with /usr/local

so I lost two weeks for nothing

Chip1972: As a FBSD standard, most *local* is not considered "base". So, /usr/local will likely be fine to be mounted after boot. However, /usr itself (/usr/bin, /usr/sbin, etc...) will be require to boot.

Chip1972: Unfortunately, it sounds that way. Sorry. :(

its possibke to link /usr to /zroot/usr?

no

Chip1972: /usr cannot be excluded from the boot filesystem. It simply will not work with FBSD.

If your boot FS is UFS, /usr must remain as part of the / boot UFS FS. It cannot be separated.

However, /usr/chip1972 *CAN* be divided from /uar as a separate FS via ZFS.

But, /usr as a whole cannot.

if zroot is imported normally, so /zroot/usr is available for boot

If you're using root on ZFS, yes.

I was told you are not?

I will try here

But, /usr isn't a separate mount by default with root on ZFS. /usr is part of / which is mounted by zroot/ROOT/default

While I applaud your effort to get this to work, I don't think it'll be possible given your current setup.

I wish you the best of luck, though!

my understanding is that the zfs tools don't require anything from /usr. to make what chip is happening happen I would hypothetically want to start modifying /etc/rc itself. I've not touched that part of the boot process in a long time though

er, to make what chip wants to make happen happen

ek: ?

Well, if the base system doesn't exist during boot (Ex: /usr) it simply isn't going to be bootable. And, there is no way to have ZFS kick in and mount a pool before the base system is booted.

That's just my logic, though. Perhaps there is a magic workaround somewhere under the rainbow out there.

If there is, I certainly can't fathom it.

no magic required. if the zfs tools require something from /usr, they won't work, and this whole endeavor is doomed to fail; if they don't, they will work.

some silliness with `ldd` suggests that at least /sbin/zpool and /sbin/zfs do not require anything from /usr, if the only problem is libraries

you can mount modules in the loader

MelMalik: Yes. That is understandable. But, there is no way to ZFS import/mount before the OS is loaded. The OS itself requires /usr.

however, the ZFS fault management daemon is in /usr/sbin

So, /usr will not be available to boot the OS successfully in order to use /sbin tools.

most tools are avail in rescue

This is a chicken/egg problem. In this case, the egg is coming before the chicken. I cannot think of a way to boot FBSD without /usr in the available boot FS.

im not following this whole freebsd cant boot without /usr

Academic question: has anyone here - don't answer this one, ek - ever run separate /usr, both / and /usr on UFS?

yeap

back when zfs wasnt so cool

My understanding is that in olden times, enough was on /-except-/usr to mount /usr, if it was on a different volume. That may not be true with FreeBSD today, but as I said, academic question, not of practical import here.

i always liked having seprate slices for things

o\

still uses zsf?

ilooked into writing my own uh

brain work no bad today

zsf - the fs part

minus the z part**

you mean zfs ?

wow... zroot just vanished....

im going to rephrase what i say and pretend sobriety exists: one of my projects was looking into writing my own file system, and i wasn't sure what the benefits of zfs were at the time over other systems

not a singl typeo :D

iirc redundancy was what maked it popular

00:38 < Matt|home> not a singl typeo :D

MelMalik: In the past, having different mount points was somewhat standard (as it still is today), but they weren't terrible separate from the mounting file system. You can still easy specify different mount points today. Should be a problem as long as everything is available at boot time.

yeah don't mind me , 60mg of morphine and im feelin pretty goodbetter~ just tired. but ya does fbsd still use zfs?

yeap

Matt|home: There are many things that ZFS offers to make it a good FS. Redundancy is certainly a plus, but not the only thing.

any particular reason/ ? i had a hard time dealing with openbsd a few months ago trying to get a headless setup and the fs was a huge problem -_-

whats wrong with ffs ?

I thought this was a good article. Especially coming from XDA: xda-developers.com/reasons-zfs-is-the-best-filesystem-for-nas

oh it was just a problem, i forget the details but iric it had to do with transfering stuff via flash drive. fat32 was the only comptaible system

that wasn't the only time ive had probelms with fs cmopatibilty across devices frankly speaking.. so i really should read more up on them, never got a chance to last time. can't tell you the number of times i had to wipe a damn flash drive reformat it and pray it worked that time -_-

thank u ek

ill read it when i can

super off topic but anyone else really like philly cheesesteaks? it's been a while since i had a great one and today i was pleasantly surprised with something halfway decnt. fam hates it for whatever reason ;_;

Matt|home: Just had a Philly Cheesesteak the other day during a big snow. Was amazing.

If you like meat, a Philly Cheesesteak is a great sandwich.

<3 personla opinion: yellow mustard = requirement

I'm somewhat partial to wiz, honestly.

wiz?

I'd be okay with mustard as well.

dijon mustard

wiz = Cheez Whiz.

dijok's great but i liove yellow 4 cheese. oh ew ;_;

ew.. well im glad u enjoy but ew lol

Fake cheese sauce. It's a staple of a "real" Philly Cheesesteak.

are you one of those people who make yuk sounds when you dont like things ?

i got a hankering to make my own now.. beef and chicken are stupid expensive

Matt|home: You'd be surprised how good it is! Everything is worth a try.

eh, im ok with trying new things. just p stubborn bout foods

fake cheese is bad cheese

sadly fam is not :p i like beef & other meat belogs on ground, they like sushi and seafood and stuff

I agree. A Philly is the only thing I ever eat it on. If "wiz" is an option, I'm gettin' it.

cheese wiz is nasty as fuck

i've had it a long time ago i think, i honestly don't remember what it is tasted

Again, I agree. I'd never eat it on anything other than a Philly.

remeber guys so we dont upset the topic people add freebsd every now and then after the non topic stuff

but i mean it's canned cheese and i eat that canned parm when $$ factors in so ya

It's basically like a salty Kraft singles "cheese".

eww kraft singles arent even cheese

u brought up the philly tho ;_;

wtf ek

Matt|home: just saying

I use Kraft singles all the time when I grill my FBSD USB sticks.

there we go man

Matt|home: Actually, you brought up the Philly Cheesesteak sandwiches. lol

i buy authentic cheddar

i'll take ur word for it,

if you continue to have reading comprehension issues youll summon o0x1eef

we dont want that ;/

uh

I just eat cheese. You know, like a 'Merican.

eww

Speaking of 'Merican. I'm going to make chilly cheese dogs tomorrow.

Gotta make the chili first, though. It'll be a day!

i made myself some tacos

hungry now -_- kind of. weird feeling

how can being hungry be weird ?

time for sideways nap. note to self:

eh, appetite surpression side effect

note to self: bring up OS implementation idea later

I had tacos for lunch. They were alright. I wanted hard shells but didn't wanna make 'em. So, I settled for soft.

i eat once a day so i wouldnt know

Matt|home: Enjoy your nap!

I have a 6-hour nap coming up soon.

has someone be able to install chromium on freebsd15 ? after I upgraded to 15 I dont have chromium anymore because it looks like the package timing out during build when I looking on pkg-status.freebsd.org.

cpet, re `man periodic.conf' thanks.

ant-x: irc is not an email man

nwe: chromium-142.0.7444.175 Google web browser based on WebKit

===> License BSD3CLAUSE LGPL21 MPL11 accepted by the user

===> chromium-143.0.7499.40 depends on file: /usr/local/sbin/pkg - found

=> chromium-143.0.7499.40-lite.tar.xz doesn't seem to exist in /usr/ports/distfiles.

=> Attempting to fetch commondatastorage.googleapis.com/ch…/chromium-143.0.7499.40-lite.tar.xz

chromium-143.0.7499.40-lite.tar.xz 1085 MB 102 MBps 10s

think the problem is fixed

cpet: yay thanks for the heads up :D do you have ny rss-feed or something to get this kind of updates?

learn to be patient ?

cpet: meant more about getting the update that the packages had been built. :)

there is no real way to get info about built packages

i ran make -C /usr/ports/www/chromium fetch

cpet, it can used asynchronously with stable connections and/or bouncers. Over on Newnet, members from different timezones regularly exchange walls of text in an e-mail-like manner.

ant-x: anything over an hr I forget, lose interest and forget

cpet, so you don't like e-mail, mailing lists, forums, BBSes, Usenet, and FidoNet? OK.

i like simple emails, i dont like robot talk, and the rest are too old to really matter any more

They matter as long as people love and use them. What's robot talk?

Hello sir, how are you doing today ? I am doing well. The reason for the email ...

robot talk

Ha

Ah, that's plain bullshit: <deathtobullshit.com> .

same applies to re: 4 hours later thank you cpet for the link

i have given up fucks have been given, i forgot about it

who cares

A delayed reply in not the same as bullshit (IMHO).

the way t is expressed it is

irc is not an email

i do not come here for fame, I came here to releive bordism help some people and then take a nap

in that time im tkaing a nap i dont not cre what happens here

up until i wake up and decide to releive bordism again

OK. An expression of gratidude four hours late is just an annoyance you.

there ya go

ant-x: would you like a bill instead ?

the time it took to help you with that took about 30 mins at 100/hr you owe me 50 dllrs

:(

no bitcoin or coffee

No, I thought people were helping each other because helping others makes one feel good and human. By the way, /dev/urandom seems blocking on my FreeBSD machine becuase Subversion hangs during checkout: <issues.apache.org/jira/browse/SVN-2590> .

ant-x: no people just google and then say they helped

Not everybody.

lies

what are the perms to /dev/urandom

cpet, as I said, I have but occasional access to that machine. I cannot do much now excpect collecvt TODO notes to take along with me next time I get access to the machine.

heh

lets ask a question about a system i dont even have access too

I'd have been able to connect it to the internet, use some SSH reverse-tunnel to access it, but no.

Indeed! What I do get to the machine, I must be all prepared :-)

I'll check /dev/urandom for blocking, and save the exact e-mail from save-entropy.

ok

I had FreeBSD 14.1 and expected pkg (install, upgrade) to work as long as The FreeBSD Foundation kept the repositories for 14.1 online. But yetsterday an attempt install Firefox caused a massive upgrade, and ended up with dependency conflicts (Firefox wouldn't start), so I had to upgrade from 14.1 to 14.3.

How can I check beforehand whether I have to upgrade my OS in order to use pkg again?

14.1 is EOL

Yes.

why would there be any packages ?

Why did pkg find them?

cause you upgraded

there is no ABI changes between small releases

No ABI changes, but things may break anyway, unless I upgrade the OS, right?

sometimes you have to learn to fix things yourself

such as I did when I upgraed to 15.0

also reading ports/UPDATING so you know what changed or what needs to be done to fix x

This time, I am trying to understand why pkg (seeminly) considered it OK to upgrade the repos and install packages for a later version of FreeBSD than I had at the moment.

Nowadays, binary packages are also called ports? The Handbook seems to reserver that term for sources.

cause dong a pkg upgrade will upgrade all pkg's doing a pkg upgrade firefox would also upgrade all packages

And will a simple `doas pkg install ...' cause an automatic upgrade?

no at the end it will say y or n

now if you do a pkg upgrade -y

then yes

After prompting, I mean. Semiautomatic. Yes.

if you press yes it will upgrade all packages

your issue is the fact that the difference in 14.1 and 14.3 is about 1.5 years

But the pkg.conf will not have changed?

nope

but then again people do weird things then blame it on the main system rather there edits

and then telling them to remove those to test they refuse so bleh

I thought that if I had FreeBSD 14.1 with its correct pkg.conf, I was entitiled to either get correct binaries or none at all. I never touched my pkg.conf .

ports are basically main

meaning updating them you will get the latest quarterly ot latest packages

Regardless of my FreeBSD version? That may explain my problems.

if you dont want to update for whatever stupid reason you could use portmaster and compile firefox

By the way, when upgrading a long-neglected FreeBSD, is it better gradually to upgrade to each next version, or go to the last one?

last one

do you have a mix of pkgs and ports ?

ant-x: upgrades are supported by release dates

OK, that's what I did yesterday. Thank. Hopefully, I'll fix the /dev/urandom problem and it'll be a working system.

so you should move to the next version which is newer in the series

but not as a jump, I don't think that is supported

No mix, only packages.

"not as a jump"? Ought I to have gone 14.1->14.2->14.3 ?

you can go from 14.1 to 14.3

but i run openbsd on a firewall and it requires to do it one at a time so thats how i do it here

you can go 14.1 -> 14.3, but you couldn't go 14.4 -> 15.0

why not?

because 14.4 will come out after 15.0 and so we don't support that change

you can upgrade majot releases

yeah maybe

we need a whiteboad

whiteboard

well if he goes from 14.1 to 14.3 then 15.0

why would he go back to 14.4 ?

I didn't say that

Whiteboard -- some metaphor?

ant-x: is this a server or desktop? im guesisng since it has firefox a desktop?

it needs a diagram

Yes, a desktop.

[tj], a version graph would be somewhat useful, perhaps in Graphviz. But wouldn't a simple chronological ordering help as well?

freebsd-update is slowly being deprecated in favor of pkgbase

so that whiteboard is rather useless now

or it would be vlaid for the duration of 15.0

because pkgbase will be smarter?

pkgbase will update your system like you would update packages

and freebsd-update would be gone

i tend to build on the fastest machine I have then NFS mount and do a make installworld on each one so fare that has faile once and that was due to switch turning off due to me kicking the power strip

but with many chrry picking and other hacks I managed to fix it

ant-x: you wouldnt have to worry about that until March if you dont decide to update now

personally i dont get why people dont update there desktops every release if its not a server

cpet, you build/compile on a single machie, and the distribute the artefacts?

but bleh im not the one having these issues heh

ant-x: yeap server. builds desk receives

desktop*

on 2.5 gbps its fast enough

cpet, Me? I am simply too afraid, the noob that I am. Yesterday was my first FreeBSD upgrade, and that only a minor one.

as long as you keep important files off your machine

reinstalling is a learning upgrade

i could format right ow reinstall in 5 mins, reinstall all my software and having a working KDE desktop in about 30 mins

why cause all my files are kept out of my machine

now the server is another story that runs my web server, DNS, DHCP, and some other stuff and well i dont touch that often

i think some of the wording that is used somtimes doesnt help in regards to upgrades

such as updating wndows and turning it off wont physically damage your computer but it will damage files or FS

cpet, re: keeping the files in a safe place> I have considered keeping not only my dotfiles, but all the important configuration files I have changed under VCS, but how: 1) make / version-controlled? 2) use an install script to put checked-out files at the right places?

i'm still wondering how smartctl knows a disk is smr... i didn't even notice it tells you if a disk is or isn't and am thinking it goes off of a serial number db or something

i doubt they'd put that in the fw

normally SMR disk are layed out differently the older CMR

and I think the highest drive you can get that is CMR is 6TB due to the areal density limitations or what not

normally I just buy a drive stick it and go on with my day

but I know theres people out there can will question why there drive is 0.000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000015% slower than there old drive

and a lot of times those are the people that buy the cheap crap and expect it to run like the 500 ent stuff

i remeber back in the day when libs were staticly linked

then we had a long discussion on the drawbacks and looking at the ML we are having that same issue in 2025

in the case of smr vs cmr...

it's way slower :)

and there are drives that are larger than 6TB that are cmr

smr definitely does not work well with zfs... it probably doesn't work well with anything other than single disk bulk storage for something that goes a bit on the slow side like maybe camera footage.. and then you'd move the data somewhere else if you really needed it since read speeds are ok

I wouldnt know I havent used a normal HD in years they are either SSD or NVE

my last NAS did have NAS drives but I never cared much to actually bnch mark it as all it did was backup like 500 megs weekly

now on a personal note is it really worth wating so long to en large an array vs creating a new one.

as I have always beleive that checking the drive that so large the chances of the drive dying is high as you are putting my wear and tear on it than normal use

as I dont think your serve an FTP from that

but since I have always liked creating slices zfs makes it a hell of a lot easier to create them than UFS

under normal circumstances this should have taken a day or two

the delay was related to awful seagate smr drives failing

i guess smr drives aren't so bad when you create the pool if you don't expect to expand the vdev and if you have to replace one you can't use another smr drive to replace it otherwise the resilver will either never finish or take 10x the time

183G / 69.2T copied at 696M/s, 0.26% done, 1 days 04:52:43 to go

i just added another drive... and the time is reducing a lot faster than it actually states... so maybe less than a day if it doesn't have a drive fail

obviously more time for larger vdevs / disks .. but given the fact you can still use it and scale it up... it's great imo

scan: scrub in progress since Mon Dec 8 06:31:25 2025 23.5G / 23.5G scanned, 2.95G / 23.5G issued at 1.47G/s 0B repaired, 12.56% done, 00:00:13 to go

10 years ago I wrote a jails bootstrap script that unpacks (in 15.0 case) this download.freebsd.org/ftp/releases/amd64/15.0-RELEASE/base.txz

when pkg tries to install python or whatever goes into those jails now on 15 it says: pkg: packagesite URL error for pkg+https://pkg.FreeBSD.org/FreeBSD:15:amd64/quarterly/meta.txz -- pkg+:// implies SRV mirror type

could someone maybe tell me should I sed delete something out from pkg configuration after base.txz is unpacked?

maybe sed -i -e 's/pkg+:/pkg+https/g' ?

anyone having issues with syslogd in 15.0-RELEASE?

🤔

It really feels like 15.0-release has been problematic in general.

i haven't had any issues with it so far. i vaguely remember having an issue with syslog though a long long time ago and wound up swapping to syslog_ng

there's also rsyslogd

That's a good way to go about it. I feel the same way about using dhcpcd instead of dhclient.

well i mean. it was just one of those things where i was like... "quick and easy" swap, disable, enable, forget

Right.

i tend to stick with the default tools as much as possible, including shell.

i had similar issues with powerd and wound up swapping to powerdxx

i do too.. but powerd randomly stopped working and i'd find that the cpus were maxed out for 3 days before noticing

i should take a look at that again. maybe go back to the base tools and see what happens.

kenrap: 15 has isues despite a warning saying pkgbase is still experimental?

if [ ! "$nonInteractive" == "YES" ]; then

bsddialog --backtitle "$OSNAME Installer" --title "Select Installation Type" \

--yes-label "Distribution Sets" --no-label "Packages (Tech Preview)" --yesno \

"Would you like to install the base system using traditional distribution sets or packages (technology preview)?" 0 0

if [ $? -eq 1 ]; then

PKGBASE=yes

fi

fi

(technology preview)?

howdy all, does anyone have advice for how I could set up a gentoo (linux) vm with networking? I had a hell of a time trying to get it working today.

are you using vm-bhyve ?

yeah, first time

ok so

I tried using it to make a switch but when I boot the vm iso up I don't have a connection

cp /usr/local/share/examples/vm-bhyve/* to /path/to/vm/.templates

recreate the vm using legacy-gentoo

and see if that works

so you would use vm create -t legacy-gentoo yada yada

did you create the switch ?

vm create switch vm-public && vm switch add public nic0

vm create switch vm-public && vm switch add vm-public nic0

wow I didn't know that was a thing with the legacy

you have to copy those files over to the .template dir

cpet: issues like like drm-kmod not working for wlroots compositors even after rebuilding drm-kmod in ports. It can kernel panic when rebar is disabled on some machines. And it breaks wine gaming for some: codeberg.org/Alexander88207/Mizutam…nts-Current-Problems-Limitations.md

I guessing part of the problem is because much of the testing went to pkgbase and little to the actual base system and infrastructure for 15.0

or drm and friends are ports and not part of the base system?

(hence "and infrastructure")

bold of you to assume people actually test when we ask them to

vm switch create public and vm switch add bge3 for me cpet

kevans: :(

spuos: test that template

i would edit that template and change the size or add in -s 128g

or however big you want

think it default to 20g

regardless it wont get the same testing as if it were included

once upon a time it was

now it works perfectly fine with KDE, intel i915 alderlake

kevans: I was told by alexander that prior releases like 13.0 weren't as bad as 15.0 in terms of how things work out of the box.

ever tried 5.0 ?

now that was fun

cpet: no matter what size I keep hanging at [ 8.558105] virtio_blk virtio0: [vda] 41943040 512-byte logical blocks (21.5 GB/20.0 GiB)

as if it's too big every time

how do you mount zfs / in rw in single user mode?

i've done it in my sleep with ufs but have not once in zfs

zfs readonly=off zroot/ROOT/default

thanks

do you have perms on where the vm saves ?

yeah, i suppose i could mdconfig the disk and zpool import it with -M

does your CPU support the VM-x stuff ?

yay i got my new spammed mcaffee license

yes, it does support VMX but not sure why that's relevant

cpet: yeah, root.

rtprio: that was for spuos

oh

also I think vm extensions but how do I check

which CPU you have ?

dmesg|grep CPU:

it's been too long since cpu

so dmesg won't show

or sysctl hw.model

hw.model: Intel(R) Core(TM) 7 240H

yeah it has it

did you init first ?

did you edit the config ?

init first? that was part of the init I think, I left it a while and it freaked out because ram0 didn't exist

edit that config and add in debug=yes

and rerun see if it spits out more

I don't know, I'm getting esoteric zfs related errs, I was at least able to boot a live system before this, minus internet

So i recently ran into a problem where I can't build rust because I don't have enough ram. how do i switch from ports back to packages?

cpet: any chances I'm doing something wrong with the vm switch or something?

is it pf or something?

CrypticSquared: pkg update

cpet just keep using that over the top?

spuos: which os are you trying to install?

rtprio: gentoo linux, I have it booting, but no internet

I'm on the installation iso and grub bootloading to it

spuos: unix.beer/~chris/gentoo.conf

that works

CrypticSquared: it will reinstall any packages which were compiled as ports with different options

if you did not change any options it wont change anything

cpet ah ok thaanks

spuos: thought you said it had issues with blk?

cpet: only with the legacy cfg

10:54 < spuos> cpet: no matter what size I keep hanging at [ 8.558105] virtio_blk virtio0: [vda] 41943040 512-byte logical blocks (21.5 GB/20.0 GiB)

so uefi works fine ?

could be gentoo doesnt have the virtio crap loaded

cpet: from previous attempts of my own, no.

I get stuck on this frame buffer thing

i loaded the kernel and got to root prompt

using that conf i posted

yeah, previous attempts and I've been running that new cfg since the "only with the legacy cfg' line

this line fbuf frame buffer base: 0xd3401800000 [sz 16777216]

IDk what that means, maybe I am on bios freebsd and that's an issue? I forget what I installed with

but I just tried replacing it with grub and it boots instantly

cpet, I see the mac addr in an interface when I run ifconfig now

but no ip, or anything. How would I get the switch to let it connect?

I have the vm-public in the host ifconfig

cpet: Somehow, I managed to get orca slicer working in an arch bhyve vm using flatpak lol

so vm-public should have your main interface and the tap of the vm, is that correct?

rtprio: I think so, I have like 4 taps from opening and remaking vms

Also til that twm is a thing, it reminds me of Rio lol

spuos: so dhclient inside the vm and ... anything ?

spuos: `ifconfig vm-public` to confirm

vm-public is a bridge

the vm would see the nic as vtnet0 or whatever its called in the penguin world

penguin world hehe

nxjoseph: teh penguin world

its raining and its 38C and people ar ebitching about my lawn

JetpackJackson: i would of thought that works in wine

It did but it was a bit janky

rtprio: ifconfig vm public shows an interface

arent we all when we wake up in the morning ?

would I expect an ip?

spuos: no its a bridge

if you see vm-public and memeber of nic0

you shoudl be able to do a dhclient bleh0

it doesn't share the ip wih the host?

cpet: wildebeest/penguin world :)

root@server:~ # ifconfig vm-public | grep member member: re0 flags=143<LEARNING,DISCOVER,AUTOEDGE,AUTOPTP>

root@server:~ #

kenrap: the onyl linux system I can use is Slackware

my members include bge3 (my out interface) and some taps from previous vm attempts

im not a options kiddie or care about making my system .01^pi% faster like in arch and gentoo

I can't tolerate any of them anymore. That's why I'm here :)

I just use gentoo because it supports the software I like on my niche device

spuos: i would remove all that

hey fam. anyone upgraded to FreeBSD 15? anything to be aware of?

kenrap: i just got my degree so im stuck with using them regardless

antranigv: if done correctly not much

kenrap, i can't enjoy them anymore too, just openwrt on cheap tplink router

Heheh

i use linux on my phone so bleh

its not going away

linux, or, android

android

I just pretend I don't see the "penguin" in android ;)

there is droid instead :D

there was, actually, its gone

i tried changing my phone to one of those alternatives but navy fed stops working

and i has to pay my bills

cpet: how so?

I think it doesnt like the version string and gives some weird error that i never cared to fix or google

so I just reverted back

one of those oh cool ok bye momments

think id.me was having some issues as well and i cant login to va without that id.me crap

and you're sure this couldn't be like a pf thing?

well i dont know your setup

and since most people dont like to mention it i stopped bothering to ask

but if you are using pf i would turn it off

i dont think russia will hack ou in 5 mins of pf being off

alright

if it works well now you know to add rules to pf

no idea still

this is an agitating issue

reboot the machine and let vm-bhyve create the configured switch

hi i just installed freebsd this weekend, whats up

acdw: awesome, enjoy your stay :)

thanks!

👍

using it on a laptop which is ok. suspend/resume was a surprise lol

hey acdw :)

hiiiiiiii

oh i have a question --- so i've up to this point installed everything --- kde, emacs, etc etc etc -- with pkg. however, sbcl is only available thru ports i believe. i know that using both pkg and ports is not a great idea, so do i need to go back and rebuild all the software i've installed via pkg with the ports workflow?

acdw: theres no harm in mixing

just build sbcl from ports and it will use the packages that are already installed

oh ok cool - i've seen it said where it can cause problems due to like, dependencies getting out of sync? is that old intel or is it more like, it can be a problem but not til i have a bunch of stuff overlapping?

ports will update / replace whats needed

> Before installing and using the Ports Collection, please be aware that it is generally ill-advised to use the Ports Collection in conjunction with the binary packages provided via pkg to install software

like this is straight from the bsd handbook

i'm just confused i guess

i dont know how the author of that came to that conclusion

There should be no problem with mixing *provided* you don't set default port options when building ports. If you do, yopu may run into dependency loops, install conflicts, and other niceties.

s/default/nondefaukkt/

ohh

like what kind of non-default port optoins?

s/kk/l/

Anything that would add to the direct or indirect dependencies, basically.

ahh gotcha, thanks :)

hey folks, I believe I have a bug report

the report might be that I was trying to do something stupid, but it does panic the kernel

Does it have steps to duplicate?

it does

If you have a FreeBSD Bugzilla account, you should do it there. Include the steps to duplicate, the FreeBSD version, and if you have one the panic dump.

working on it :)

helpfully it seems like dmesg sticks around across the reboot so I was able to recover it

I think I've minimized my test case. I can provide a crashdump if anyone can point me at how to tell freebsd to dump to my swap partition as a liveusb disk

See "man dumpon".

perf, got it

ope, no i don't, I got a "savecore: operation not permitted" booting into my regular environment. hm.

does it? i thought dmesg clears out. although i think you get the same stuff in /var/log/messages

zip: do you have dump device configured ?

The liveusb may have a different swap partition or none.

I do

ah, I needed to unmount swap first

hopefully encrypted swap didn't stomp on it, but it was happy to extract the core

finally installed bsd for the first time on a partition of my arch linux thinkpad, so i guess im allowed to be here

whooboy, compressing this sure is making the fans spin

is it worth popping freebsd-14 on the ol' USB stick and seeing if I can repro it with that?

If you can isolate it to one major version that way, yes. Narrows the search.

I'll need to wait for access to bugzilla

*nod*

repro'd on 14.3, guess I'll go earlier tomorrow

Me: man I am fucking exhausted after a day of fixing bugs in open source software

Also me: oooh a kernel panic let me work out how to reproduce it from a live usb and then take a crash dump that won't have passwords in it

Tired of spending all day looking at the bad rectangle, time to finally kick back and spend the evening looking at the good rectangle, etc etc

dont forget the smaller vertical rectangle!

just realised I have a thinkpad lying around that also has an iwl chipset, repro'd there too. This may actually have the makings of a useful bug report...

There is no rectangle, only Zuul.