this is weird does no one reveal chipset names anymore?

sysctl net.wlan.devices ?

that took longer then it should of

ax201

well it seems the driver wasn’t available but is now

i’m wondering if whatever version that’s on the iso is out of date?

fbsd 14 ?

yup

it is possible that your hardware is not supported.. if thats the case, look into wifibox xyinn.org/md/freebsd/wifibox

Title: Fast & Stable WiFi with wifibox on FreeBSD | @fearedbliss

according to the bug report i found it is as of october

in wifibox you will need to open the wifibox console and add your wpa_supplicant info inside the wifibox vm

Title: 244261 – Add support for Intel AX200 (iwx) wireless devices

it’s using iwlwifi which is correct

is there a way to see if the device is up? ifconfig just gives me the loop back

ok i see the issue

even though the driver is being loaded for some reason my device isn’t being seen

and i can confirm it is supported

maybe it got turned off somehow? my laptop does have function keys to do that but i already tried that

is there a rfkill in bsd?

markboston: docs.freebsd.org/en/books/handbook/…twork/#basic-wireless-configuration

Title: Chapter 7. Network | FreeBSD Documentation Portal

did you add the rc.conf bits?

i did not

the installer should say hey don’t try looking for network until you

configure the card*

Title: Chapter 7. Network | FreeBSD Documentation Portal

:p

thabks for your help

but yes yes rffm

I can't remember what the installer does with wifi rn. On the laptop I have it on I configured it later, and used ethernet to install.

good luck; hope it works!

i am using openvpn and nating with pf from via my fbsd vps, i get only 10mbit downloads, any clue why?

i want it to serve with the full bandwith

well, i am nating via openvpn with pf on freebsd vps

i am getting low bandwidth on the clients

how can i set it to normal speeds?

mane: tried switching openvpn to udp (or to tcp if using udp now)?

it is on udp :((

On a 10Gbps connection using udp, the tunnel interface (the fastest option at the time), and 40k jumboframes (after that, it started getting into diminishing returns), I was able to push north of 400Mbps - I expect that ovpn(4) will improve things somewhat by not having to context switch as much, but fundamentally, OpenVPN isn't made for speed.

It's one of the reasons I've stuck with IPsec for as long as I have, with the other reasons being that 1) I know how to configure it because I had to learn back in the bad old days when there was no alternative, and 2) setup properly, you can give people a profile that they can import into their OS' implementation of IPsec, and not have to spend time configuring their systems.

hi FreeBSD People !

Hi.

i really need to fix this issue XXD

the whole project relies on this XDD

mane, do u test speed with iperf?

i tested it with ookla

I don't know what is it.

server goes 860/450 no problem

but my clients go 10/40

ookla is that popular speedtester

speedtest.net

mane, use iperf for network test

benchmarks/iperf3

that's irrevelant

i get bad performance, bad download on clients

any settings for tun0 i could set? i tried mtu, didn't make a differenfce

irrevelant testing network bandwidth in your software

mane: the clients net connectivity is via lan or wifi?

1st of all you must test raw bandwidth

elirco coax lan 350/60

raw is fine

ah ok so that is not the issue

there has to be either pf or tun0 or kernel setting i need to change\

run iperf and check CPU usage

it's chillingm, i checked htop

chilling

on both sides

clients are chill

yesterday there was debian on this vps, there were no issues

If you don't want to do what you were advised to do, then why are you asking for advice?

i did all of that already :s

XDD

nio vpn - full speed - vpn cliejnts slow, tested with speedtest, two different devices, clients and the server are not in stress, server maxes out the connection outside the tunnel

Ookla wont let you test the connection between your two machines

Also, its expected to have a much slower bandwith with a userspace vpn, but 10 Mb/s sounds very slow even for an old machine

it will if it bottlenecks on vpon :V

yesterday it was debian there and it was fine, but i want fbsd

What crypto suite are you using ?

babz it's just openvpn

With which algorithm ?

hm

cipher AES-256-CBC

Does your cpu has AES-NI ?

You might need to load its driver aesni(4)

Also, i think it's slightly more efficient with AES in GCM mode

kldload: can't load aesni: module already loaded or in kernel

:DD

also, uploads are fine, it's the downstream that is sluggish

(upstream for the server)

Bandwith thst you've not tested without your vpn, amirite ?

YES

x((

Hi there

how can I change flavor in poudriere ?

ok never mind, fixed

but can I have multiple flavor for specific ports in poudriere pkg_list file ?

Yes.

look for sets

ok... could anyone give me sample how to clean certain package from poudriere jail ?

trying this - "poudriere pkgclean -j freebsd_13_2_amd64_nox -C lang/php82" but it doesn't seems to work ...

never mind -p <tree> was required ...

thanks

Failed, I have enabled php82 in the ports and now php82 is a default flavor, cant roll back to previous state

how can I force to "forget" poudriere php82 option ?\

I have removed all mentions from the list file

ok now easier thing, i'm trying to connect to smb4 from my macos client, i get auth error, i did everything right according to the handbook\

iperf3 via openvpn show 20.0 Mbits/sec to my office with uplink speed 30Mbit/s

7% cpu on my very old Core 2 Duo E6750⊙3 (oveclock) - openvpn client

10-15% CPU on openvpn server with very old Xeon X3430⊙2

plz, test bandwidth via openvpn with iperf first in both directions and check CPU usage during tests

direct without openvpn I got same 20Mbit/s to same server

this is going to sound like a weird question.. if i am changing compression on zfs pool from gzip to lz4 should i expect a "delay"

as in.. it will take time for the system to switch over to that compression model?

i read material and it seems to assume you have no data on the pool to begin with..

voy4g3r2: no, changing compression doesn't do anything with the existing data, it only affects newly written data

thank you unixwitch

you would need to rewrite the existing data if you want it compressed with the new algorithm (zfs send | zfs recv should work for that, i think)

good thing to try with my next time machine backup.. right now gzip is doing a 1x

lz4 is unlikely to be better than gzip, its advantage is that it offers comparable compression ratio but much higher performance

good to know, i will monitor as it looks like i have to update this pool to not consume my whole hard drive array.. it is already at 223gig of space..

voy4g3r2: Time Machine is supposed to automatically delete older backups when the storage becomes full, perhaps setting a quota on the filesystem would force it to do that? (i haven't tried that)

yeah, that is what i need to do.. while figuring that out, i saw this compression stuff

figure a quota of 500 gig should work.. then time machine does its magic

and something like zfs set quota=500G storage/backups

is sooo easy

:DD

still do not understand how 1 gig of data was "udpated" in a day on a laptop..but that is something different

has anyone had issues with ixgbe(4) not coming up at 10GbE? the link on the switch comes up then immediately goes down every 30s... i'm going to ask the switch vendor but maybe someone ran into a problem like this. i don't think it's a cable issue, and 1G works fine

unixwitch, anyway try to test with other cable - copper 8P8C 6+ cat or fiber?

VVD: copper, Cat6A STP. i'm planning to test with another cable but i need to physically move the system (or maybe make up a temporary cable...)

i did change the patch cable at both ends and also rewired the socket at the host end since it looked a bit iffy, but no difference

I tested cat5e - some cables work fine with 10GE, some not.

Also you can test in different ports in switch and if u have different switch.

BTW, what switch?

VVD: i've tried two different switch ports, but i don't have another 10G copper switch around to test with. it's a mikrotik CRS312-4C+8XG

might try with a 10GBASE-T SFP, i think i have one lying around somewhere

(i hate these though because they generate a ridiculous amount of heat)

unixwitch, if ur network card support u can try to force 5GE and look how it going

this one only does 1G/10G unfortunately

ah, ye, ixgbe doesn't support 2.5 and 5GE

I have ix and ixl - both doesn't support 2.5 and 5GE.

I started up my headless FreeBSD machine today and SSH wouldn't log in after ages. I decided to plug a monitor in and I've got a bunch of errors that look like disk-read errors. The last one is `Mounting from ufs:/dev/ada0p2 failed with error 22: Invalid fstype`. Everything I _really_ care about is on my ZFS mirror set, ada1+ada2. Do I just yank the disk with ada0 on it, buy a new one on the Internet, and reinstall FreeBSD?

I have a `mountroot>` prompt, if that helps.

adiabatic: i would start by booting from the install media (DVD/USB/whatever) and see what's going on with your current root disk. if it's really broken, then yes, you can just replace it and reinstall

How would I see? Figure out what ada0 is under the boot disk, and then run fsck on it?

start with 'gpart show' (don't use fdisk, you never use that in freebsd) then try mounting it

I think I'm running 13.something (I was going to check and see if I have an upgrade to 14 pending or if I'd already done it). Can I just make/use a 14 boot disk, or is that inadvisable?

adiabatic: 14.0 should be fine, although it seems reasonable to use the same version that's already installed

…so I booted from USB stick. I want to use the live CD, right?

there is a memstick image on for download

is 'use the live CD' one of the options it presents on boot? if so yes, whichever option gives you a shell

well, it said "live CD" in the installer text. I wasn't confused by that…

well, there was _also_ a shell option that was called that

so I figured "installer shell" vs. "generic shell"

Neither ZFS thing is showing up. I suppose that's to be expected…

yes, ZFS pools don't show up unless you manually import them (which you shouldn't do here unless you have a reason to)

I tried `mount -o ro /dev/ada0 /slash`. It complained there's no directory to attach to. I tried making that directory, and it complained the file system is read-only. Tried that in root's ~, same error. Is there a decent place in /var that I can mount my probably-busted drive onto?

adiabatic: you can't mount ada0 since that's the entire disk, use 'gpart show' to find the appropriate partition, e.g. ada0p2. there should be a /mnt directory you can mount on

Oh, right. Thanks.

`mkdir: /mnt/slash: Read-only file system`

just mount it on /mnt directly

fair 'nuff

OK, this looks familiar:

(ada0:ahcich0:0:0:0): READ_FPDMA_QUEUED. ACB: «space-separated hex digits»

(ada0:ahcich0:0:0:0): CAM status: Auto-Sense Retrieval Failed

(ada0:ahcich0:0:0:0): Error 5: Unretryable Error

Time to go disk shopping?

if it was working before then probably yes. you could check the disk's SMART status to confirm, but i'm not sure how you'd get smartctl onto the live image for that...

First dead disk I've had in a while. Thanks a bunch for the handholding.

We don't know yet if the disk is in trouble or if the file system bits are simply corrupted. It could be either given the data so far. The disk might be okay. Or might not be.

You can test the zfs pool condition using "zfs import" and confusingly it won't actually import anything without any options but will just report the pools available for import.