it seems my rpi running freebsd has no local v6 address, even tho it should get one via DHCP

I have not set up more than `ifconfig_DEFAULT="DHCP"` in /etc/rc.conf

do I need to tell it to also get a v6?

afaik freebsd doesn't have DHCPv6 in the basesystem and additionally you need some postfix to those ifconfig_ variables for v6, the postfix-less variables only configure v4 stuff

hi all, is there no SHASUM file for .txz artifacts? download.freebsd.org/ftp/releases/amd64/amd64/14.1-RELEASE

Title: Index of /ftp/releases/amd64/amd64/14.1-RELEASE/

I guess the second field in MANIFEST is a checksum, but no idea which hashfuction was used

it's sha256, but the format is not tool-friendly (sha256, sha256sum)

grep -q "$(sha256 < base.txz)" MANIFEST # yay

how about awk '$1 == "base.txz" { print $2 }' ?

or more generalized awk -v hash=… -v filename=… '$1 == filename { print match($2, hash) }' ?

i have script for those things, altho it's similar

so, if I make changes to /etc/rc.conf, how do I make the system reapply it? only by reboot?

specifically chagnes to the network setup / dhcpv6

service restart netif <interface> can work

service netif restart <interface>

before I dig into that too deeply: does freeBSD support serial console on the rpi? :D

I do have a very genuine ""FTDI"" serial console / usb adapter, I used that with other OSs on the pi before

yourfate: docs.freebsd.org/en/books/handbook/serialcomms

Title: Chapter 29. Serial Communications | FreeBSD Documentation Portal

right, should have looked there first

:)

now i'll have to dig that out of some random box of cables before I mess up the networking on the PI and lock myself out of ssh

but its a lot less faff than connecting a screen and keyboard

can i give a bhyve vm access to the vt-x feature of my cpu? apparently the linux subsystem of windows depends on it and i'm going crazy without grep…

I'm confused about what in the handbook is thin about thinjails? docs.freebsd.org/en/books/handbook/jails/#thin-jail I'd like to have my jails use one single version, and have their own packages added on top. How would freebsd version upgrades happen for one selected thin jail? 1. use handbook to create "syncthing" jail (13.2) 2. install syncthing in the jail 3. upgrade to freebsd 14.1 4.

Title: Chapter 17. Jails and Containers | FreeBSD Documentation Portal

?? 5. now we have a jail running 14.1 and with syncthing

oh, it's actually the nullfs jail that might be a better fit for me

moviuro: no idea about th zfs version, but for nullfs readonly mounted basejail kinda thin-jail, i usually have one location where i mount the basejail and skeleton as rw and do freebsd-update on that to update the base for all thinjails at once.

yup, exactly what I need/want

moviuro: so i have a part of my fstab that looks like this: paste.xinu.at/K39Z2

Title: stdin

phryk: could that be automated in jail.conf(5)?

(update is a jail name, right?)

oh, no, you misunderstand. that is to have a place where the base and skeleton are mounted on top of each other in rw-mode so i can update it with freebsd-update -b /jail/update -d /jail/update/wherever

moviuro: my mount automation for the jails in jail.conf looks like this: paste.xinu.at/Oe85u

Title: stdin

ooooh, shiny

would that not cause a bad reference

voy4g3r2: elaborate?

thin jail = use the baseline operating system and have all the udpates implied

if you are updating the thin jail with those mounts.. are you not just going to thick jail?

no? because there's only one basejail and it's shared among all jails.

if you do a freebsd-update on the host, all the thin jails get those updates

so you have jails inside of jails?

i have no idea what you're talking about, but you're most certainly misunderstanding something.

yes i am.. that is why i am asking

i have never heard of thin jails nullfs-mounting pieces of the actual host basesystem.

it was always one separate base system, for as long as i've been using jails, which is pretty sure over a decade now.

so phryk my understanding is: i need a place for permanent writes (usr/local, home, var, etc), that's your /jail/rw/$name/, which you automount on jail start. The only thing that a jail needs to be defined is: a skel of those permanent directories, an IP in jail.conf(5), and... a management script/program to create/update the jails' OS (wrapper around freebsd-update and fetch+tar)

phryk: i am use to the nomenclature of temmplates and not baseline.. i was reading this interaction as the base host

which now completes the gap in my understanding.. i just put stuff in /usr/local/jails/templates

which is what you were saying?

moviuro: look at the fstab i posted first. if you create that as a structure (i.e. create empty directories were necessary for the nullfs mounts) and then just untar base.txz inside /jail/update you'll should get both the base and skeleton. at /jail/base and /jail/skel. then to create a new jail you just do "cp -r /jail/skel /jail/rw/newjail" and add the config for newjail to jail.conf

voy4g3r2: no idea where that nomenclature comes from, but yes, if you for example run jails of multiple different freebsd versions, then "template" might be a good term for what i call "base". i just always run the same version as the host, so it's just another base-system for me^^

docs.freebsd.org/en/books/handbook/…ils/#host-configuration-directories <-- talks about the concept of templates

Title: Chapter 17. Jails and Containers | FreeBSD Documentation Portal

Ah yes. Thanks.

always an opportunity to learn something new after 10 years of doing jails

Hi, i am looking to deploy a freebsd bhyve cluster so that i can have some sort of fail-over in case of a hardware issue. I guess this would mean having the filesystems always in sync. Any advice on this?

hello FreeBSD, we have a local BSD community and we will be happy to contribute upstream FreeBSD projects

We are located in Türkiye

voy4g3r2: That's exactly what my cellmate and my lawyer say !

L3Fr0g: like a high availability setup?

Gün aydin.

Hello V_PauAmma_V

Welcome.

Are you Türk ?

No, but I've accumulated bits and odds of languages. This one begged to be used. :-)

Nice to hear that I like to learn languages too

phryk: If I try creating zfs subvolumes first before anything else (/usr/local/jails/templates/14.1-RELEASE/{root,etc,...}), tar(1) complains when extracting base.txz: `./root/.profile: Can't create 'root/.profile': Cross-device link'. How am I supposed to handle that? ignore those files? `tar -xf /usr/local/jails/media/14.1-RELEASE/base.txz -C /usr/local/jails/templates/14.1-RELEASE --unlink'

moviuro: sorry, no idea. i moved away from zfs. too many headaches.

Turks make really nice food.

hwpplayer1, anyway: you sound like you're (part of) a user group. If my guess is correct, are you registered in freebsd.org/usergroups ?

Title: User Groups | The FreeBSD Project

No V_PauAmma_V But We want. I am an admin there who has right to commit to our repositories

How can we register ? V_PauAmma_V

I'm searching V_PauAmma_V

Okay I learned it was so clear Thanks V_PauAmma_V

Title: 279644 – FreeBSD Turkey Group

mmm tasty turkey

Tasty kebab.

You're welcome.

my named, which is configured using root hints available here: internic.net/domain/root.zone is suddenly unable to resolve some domains (such as Wikipedia) - they return NXDOMAIN. is anybody else having this issue by any chance? I know it's not really freebsd specific but I doubt #bind or whatever has an active userbase

Greetings. I have a question. When FreeBSD finds a bit of memory that fails ECC or is ECC corrected, does it ignores that bit of physical RAM or does it still try to use it again? Thank you in advance

@voy4g3r2, exactly

L3Fr0g: there are a few ways.. could do zfs send | zfs recv on the datasets

through snapshots

@voy4g3r2, i want everything in sync anytime - zfs send and zfs recv won't do the trick

L3Fr0g: klarasystems.com/articles/introduction-to-zfs-replication

Title: Introduction to ZFS Replication | Klara Inc

hrm.. that is beyond my skillset, unfortunately.. maybe someone else may be able to guide

i utilize snapshots but my data is not mission critical and requires that type of redundnacy

guys, klarasystems is a joke

DynamiteDan: forums.freebsd.org/threads/ecc-error-reporting-functionality.77921 might have some info

Title: Solved - ecc error reporting functionality | The FreeBSD Forums

thanks saper. I will look at it :)

@voy4g3r2, i want something like zfs over drbd with corosync and pacemaker

this?? corosync.github.io/corosync

Title: Corosync by corosync

bhyve is for virtualizatoin of operating systems, that is about application level.. not the same thing. iw ould recommend to write down what you are trying to do and be clear on what is your goal is

yep

@voy4g3r2, two identical physical machines, each capable of starting any of the configured VMs at any time.

so you are wanting high availability of an operating system that will "turn on" in the even the master dies

i think i need to step away if that is your goal, i have not done somthing like that in years and definitely not on bhyve.. it was vmware 3.x in like 2000s

can anybody assist with this kind of setup ?

L3Fr0g, have you checked docs.freebsd.org/en/books/handbook/disks/#disks-hast ? I think that would solve at least part of your problem as I understand it.

Title: Chapter 20. Storage | FreeBSD Documentation Portal

@V_PauAmma_V, i did, thank you

I have two FreeBSD physical machines, each with a distinct public IP address. Each of these physical machines hosts several VMs. My objective is to establish connectivity between the VMs on Host1 and those running on Host2.

Is there a way to get to prison from vnet? The problem is, userspace is always refering to prisons while I only get to see vnets. I didn't find a way to get info about which prison my object was created from.

github.com/Dautor/ng_vnet_hub/blob/master/ng_vnet_hub.c#L149 this is the line where I'd need that.

Title: ng_vnet_hub/ng_vnet_hub.c at master · Dautor/ng_vnet_hub · GitHub

F3Fr0g: you mean a VPN between the VM nets? I do this with WireGuard.

They left.

The only solution I see for now is to have user specify jid after node creation. I see no way to prevent sibling non-vnet prisons from doing vnet netgraph operations on their cousins' nested vnet jails.

dautor8518050867: no, because there's no 1:1 mapping from a vnet back to a prison

Yeah, I guessed it would be the case. Would the proposed solution of having user specify jid after node creation be acceptable?

your vnet_to_prison doesn't make much sense to me as it is, why do you need the prison specifically rather than resolving a jid to the vnet?

It's a hub where user can "connect" it to other (child) vnets. I didn't see a way for user to specify a nested vnet, just nested jail via its jid.

right, ideally they'd specify a jail and you'd resolve that to a prison + vnet, vnet remains constant for the lifetime of a prison

I guess I'll go with user specifying their jid and ng_vnet_hub just verifying that the vnet is correct (the same as its own) then.

I think the disconnect here is that there is no such thing as an ested vnet

there are nested jails, and a nested jail may have its own vnet or that of its parent

so when you say child vnet, that's not a thing in itself

Yeah, I meant child prison's vnet.

My motivation for why I need that - I want VMs to be able to communicate with stuff inside vnet jail without running those VMs inside the jail.

That would be simple if I just used tap interfaces and moved them via ifconfig. However, those VMs are not limited to just ethernet.

What's the difference between memstick and disc1?

it's created for usb drive

.xz

disc1 and dvd1 and iso are for dvd/cds

how it's possible to boot with chainlodaer +1 (hd5,6) freebsd with SSD i don't know

for MBR with normal hdd i know

but ?!

no uefi

no csm csm is disabled from bios

I get that its purpose is that. Is there no way to make them be the same?

Like, having the same image be usable for either usb or optical discs.

memstick is for usb disc/dvd is for optical discs

I get that, but is there a technical reason that is not allowing them to be the same?

since 12.0 or so disc1 is usable for either usb or optical disk

our .iso (for x86, at least) are hybrid iso and can be dd'd directly and probably work on most machines (barring shi-- finnicky firmware)

(to be fair hybrid iso are pretty funky)

Ah, thank you. I never tried testing booting disc image from a usb and was just wondering.

imo we don't advertise it as well as we should

my #1 problem with a lot of linices back when I was younger was that they only published .iso but they never advertised those visibly as hybrid-iso, and I didn't have optimal media to burn it to

lots of missed opportunity to try various linuses because I didn't want to screw with it

I always assumed they worked on both as they only provided one image.

Btw, do you know why there's 200MB extra data in memstick?

no idea, sorry

I think the .img do have a larger ESP, but if I'm reading this right that only accounts for ~30M

Np. Thank you again for answering.

On Linux I used to use DRM within firefox and just use tidal through that... but since I moved to freeBSD that is no longer possible (I shouldn't have given into DRM anyways but thats another sotry)... what solutions do you guys use for music?

streaming music is cheapest... and albums I love I buy a digital copy and download the flacs for it... then I can index them myself

but the streaming section is the one I am curious about... anyone got any tips on it?

polarian: qobuz has a nifty litle web player

jauntyd: DRM?

nope, html5, etc

they also give you discounts on buying physical copies with their streaming subscription :P

who does?

qobuz

i didn't know they sold physical media

i will check it out

My way of doing things is look for music on bandcamp first, this gives the author the most amount of money... plus is usually cheapest :P then if not I look for it on qobuz and buy it there to download

not physical

digital copies

oh ok

This was a weird one that just happened to me. bugs.freebsd.org/bugzilla/show_bug.cgi?id=279653

Title: 279653 – Page fault in in6_selecthlim