Any happen to try 14 ISOs with Ventoy? It kernel panic at boot?

Looks like it's been reported already. ventoy/Ventoy #2636

Title: [issue]: FreeBSD 14.x not working workaround included.. · Issue #2636 · ventoy/Ventoy · GitHub

2636 – Bad port of apache or its just me i don't know bugs.freebsd.org/bugzilla/show_bug.cgi?id=2636

... now I wonder if a bug's title had #.... in it.

Great ventoy works with 14 release now.

trying to install an operating system that is from 2018 is not easy

Ventoy looks nice

jb1277976: Yep, it's a nice thing to carry around.

hey guys, I need help with setting up eduroam on FreeBSD-14.0R. I have some SSL error, is it a known bg or a misconfiguration?

0x0.st/Hwq6.jpg, 0x0.st/HwqI.jpg. anyone? it's 4am, and I need the laptop in the morning

if someone knows how to help me, has a bouncer and is just sleeping - please, email me at olk⊙wc or pm me on xmpp: olk⊙pn thanks in advance, I really need to fix that

olk: it means the issuer needs to update their signature algorithm used for their certificate as the ones they use have been deprecated

You can also check if wpa_supplicant can be configured to allow ones that've been deprecated, but that seems risky.

/usr/share/examples/etc/wpa_supplicant.conf:191 seems to indicate that it's possible.

Dam dude left ?

my college uses the same auto just with no certificate can somebody give home this please forums.freebsd.org/threads/howto-wi…ss-wpa2-enterprise-connection.39557

Title: HOWTO: Wireless WPA2-Enterprise Connection | The FreeBSD Forums

Works for me everytime

rwp: i have admitted defeat, i have gone all the way back to 2018 linux distro and still get the issue

voy4g3r2, What is the issue?

the bad block thing still happens22:44 < sublime> [ 567.245967] EXT4-fs (dm-0): Number of reserved GDT blocks insanely large: 7491

22:44 < sublime> [ 1030.211334] EXT4-fs (dm-0): Number of reserved GDT blocks insanely large: 7491

uhh pasting mess

i have spent all day trying to get the disk operational

so i can continue and MAYBE get it working or take the hit

I thought that when we last left our heros that you had mdadm assembling the array into a /dev/md2 device. And then you would need to activate the VG and get access to the LVs. No?

yup

got to the point, lvs done

mdadm all setup, then when i try to mount.. that messages comes up

which is a result of kernel 4.7 and above

the oldest working linux i can get with raspberry pi is ubuntu 18.04

I found this that you probably also found: superuser.com/questions/1470173/rec…-ext4-number-of-reserved-gdt-blocks

Title: linux - Recovering data from Synology NAS – RAID5, ext4, "number of reserved GDT blocks insanely large" - Super User

There it notes that Ubuntu 15.10 would be the preferred system with the preferred kernel to use to recover.

Ubuntu 15.10 would have had Linux kernel version 4.2.

which does not have an arm release

i am not sure if ppc64 = arm64.. my brain says no

You don't have available some old laptop or desktop or other random bit of x86 junk that could take an Ubuntu install?

ppc is not arm, no, that's Power PC.

i may have something.. pc wise which may work out better.. i got tons of machines.. 2 work laptops (can not use them) 2 pis, 5 macs

so yeah..

i did see that article and stopped because of the no arm thing

..

steps seem straight forward to get the support i would need.. thanks for getting me out of rut

i am searching and searching focusing on the wrong thing

I don't recall if Ubuntu has a 15.10 live boot image available in the archive but if so then you could use any machine as booting the live image would not require installing it upon the internal drives. Just run the recovery on the live system. I have done that multiple times booting USB and SD card boot.

hrm worth seeing if i can find an old debian live cd that MAY support arm

Title: About Debian 8 Jessie | Freexian

Looks either I don't recognize these as live boot images or they didn't have one then: old-releases.ubuntu.com/releases/wily

Title: Index of /releases/wilyUbuntu 15.10 (Wily Werewolf)

voy4g3r2, They do have PowerPC images there though.

hehe i got rid of my last powerpc based mac in 2012 :)

that was an awesome machine..

Since you mentioned PPC I assumed you had one of those. (shrug). I would see about scrounging an x86 and then booting 15.10 and seeing if that avoids the problem enough to recover.

nah, my ppc reference was just mentioning, my mistake

well i got the willy release and will try to pray this old x86 works

i have not booted it in years

Times like this are why I am totally a craphound and keep a lot of old equipment around. Sometimes it is just what is needed.

hehe, i still have my first computer.. over the years i have thrown out things as i had no use other than moving from house to house

and the reason for the high mac count, i got sick of fixing peoples computers.. you need a computer for unviersity.. buy a mac, it just works

and they last a real long time, this machine i am using now is from 2012 and still going strong.. even has ventura (n-1) from latest os x

I am routinely using a Thinkpad X220 from 2009 all of the time. Best keyboard they ever made. Still going strong. Still have an x201 from before that and it is still working too.

The Thinkpad x201 is limited to 4GB of RAM which is the limiter there. I have demoted it to being one of my testing machines.

i loved the t42 i recommended to a family member back in 2004

one o fmy work machines is a lenevo, blanking on the model

they are solid machines and nothing beats old school thinkpad for that

if a pc is required that is portable, i usual recommend to people thinkpads

the best is: i can get a dell for cheaper.. yes you can but you will be buying one every 1-2 years.. spend he extra money and this will last 10x longer

i spent 2800 on this laptop.. it is now 11 years old.. same price as that 200 buck dell :)

it has been a long day, thanks rwp for tips.. we shall have some status update tomorrow.. last day of work and i get a 5 day weekend :)

I loved my T42. The usual limiters there but the T42 is an IDE disk machine. More difficult to upgrade to a SATA SSD.

I am still maintaining two of them for dedicated training software though. They do what they do now exactly as they did before. I did find IDE SSDs for them though.

The T43 was the transition machine. It had a SATA controller and a SATA to IDE converter chip and shipped with IDE drives. It's possible to bypass the IDE converter and upgrade it to a SATA. But my T43 died a graphics hardware death so I moved on to a T60p. My T60p is now my coffee table machine.

Good luck voy4g3r2 on your new direction. It still feels like you are on the recovery path okay. Remember you haven't lost anything until you have stopped trying.

FreeBSD on vbox is better than hyperv imho. Mouse support. :D

Hey guys, i want to attempt to build pulseeffects. but it has been deleted from ports and i cant find it in /usr/ports/audio/pulseeffects -- how can i get a copy of the latest state of this port ?

Beladona, "helloSystem" project was|is working on a GUI (Python, Qt5) to manage Bluetooth devices: github.com/helloSystem/Utilities/tr…Construction/Bluetooth%20Wizard.app (I have not tried either of the two).

Title: Utilities/Under Construction/Bluetooth Wizard.app at master · helloSystem/Utilities · GitHub

hernan, Search on FreshPorts, freshports.org ; or, search the Git commit history of "ports" tree (see "Using the Ports Collection', docs.freebsd.org/en/books/handbook/ports/#ports-using , on how to have a local "ports" repository )

Title: FreshPorts -- The Place For Ports - Most recent commits

hernan ... from there collect the URL of source files, or binary blobs, as the case may be, along with FreeBSD patches; then compile self

Also possible that "pulseeffect" may have the source publicly available (which may or may not compile as is)

parv thanks !

Hi. its fine to use gitup instead of git for updating ports? there is any downsides or any other recommended alternatives? Thx!

pad.bsd.to is giving cloudflare 502 bad gateway, btw

I could use a tip regarding this ipv6 gif tunnel issue - pastebin.com/raw/9XZf16Xu

my isp had a brief network outage, but the dynamic ip didn't change. I just now discovered that my ipv6 gif tunnel is not working anymore, even though when I look at the various command outputs, things look the same as usual. I could fix this with a reboot, but I'm also interested in understanding what's going on here.

ultramage: how did you build it

meena: er, long ago when I registered the tunnel, I wrote a rc script that runs at startup and sets up the gif tunnel. I don't seem to have added any additional parts that would react to dynamic ip changes.

ifconfig $he6_if create mtu 1480; ifconfig $he6_if inet tunnel $my4_ip $he4_ip; ifconfig $he6_if inet6 $my6_ip $he6_ip prefixlen 128 deprecated; route add -inet6 default -interface $he6_if; ifconfig $my6_lan_if inet6 $my6_lan_ip alias

I've experienced many isp outages over the years, but I don't think I've ever seen the host malfunction this particular way.

what I find most strange is that the IP6 assigned to my side of the gif tunnel still appears to be assigned, and the interface is marked as up, and the route is there, but when I try to send a manual ping from this ip, it says it can't assign the requested address. so it seems as if the networking part of the kernel is destabilized. But it could be something else that I'm just missing or not aware of.

or I could just be selectively forgetting that this has been happening forever, and I just thoughtlessly rebooted the host after every outage. I don't remember.

stupid question, on 12.x what is the difference between boot1.efi and loader.efi? which one should be copied to EFI/BOOT/BOOTX64.efi ?

mage: man boot1.efi -> boot1.efi has been deprecated loader.efi(8) handles all its former use cases with more flexibility. boot1.efi loads /boot/loader.efi from the default root file system and transfers execution there.

(this is 13.1, check your 12.x man page just in case.) I guess previously the efi boot was broken up into multiple stages and multiple executables, but now it's all handled together

uefi(8) describes the boot process

"boot1.efi(8) is used to load loader.efi when it is placed within the system at /boot/loader.efi"

but there is no boot1.efi manpage on 12.x

it was added when it was deprecated

I'm trapped in buildworld for the rest of the day, so there's lots of time left to figure out that gif0 malfunction of mine. But I have no idea what else to check. I imagine deleting the interface and re-running all the setup commands would solve it too, but won't give any info

rwp: just solid series, the lenevo i am using for work surprsingly does nto have a model number.. but it has core i5 8th gen and 16gig of ram.. crazy how a company woud give me a machine like this when i am mainly in office most days...

now i remember why i hated this machine (x86) laying around.. the dang dlink video port...

dlink ?

some weird display port

ultramage: don't you need to set ipv4 src and dst on gif0 ?

also if you dont have a route for the v4 packet, it will forward the ernno to ping

babz: oh wait...

yea you just reminded me

in ifconfig gif0, there's supposed to be a line saying my wan ip4 -> tunnel broker ip4

(I was going to compare the outputs before and after rebooting, that would have revealed this discrepancy. But it's good to know early.)

so... when the wan interface temporarily goes down, and/or gets its ip rewritten by dhclient, the gif0 interface stays up but loses its tunnel mapping. that is unfortunate. any idea how to deal with that?

sounds like a bug to me

'ifconfig $he6_if inet tunnel $my4_ip $he4_ip'. the v4->v4 tunnel entry created by this command is what disappears.

Title: Some late-breaking FreeBSD 14 breakage

it is entirely possible that the gif driver retains this state, but some builtin rc script deliberately cleans things up

I've experimented with a static private address on a tap interface linked with the far end with an openiked transport setup.

The openiked transport didn't have a problem with the physical link going up and down. I just don't recall if there were any changes of assigned address involved on that layer.

hm I wonder, would using the host's LAN ip work as the tunnel parameter, hoping the host would figure out the proper outgoing interface?

I doubt it.

There may be something you can do with ipfw, pf, or whatever to manage things.

could you figure out the minimal action to make that tunnel setting disappear ?

Don't have any idea what at the moment....

Hurricane Electric's website might have some notes that help. IIRC, they have IPv6 tunnels over gif on dynamic addresses.

i cannot reproduce you problem by bringig the parent interfac down or changing its ip

Where the ipv4 is dynamic.

dhclient might be involved. from the /var/log/messages snippet I posted, it reported (the same) new ip, gateway and netmask, multiple times in some cases

well, LAN ip doesn't work. re-adding the tunnel association with the wan ip restored the connectivity

babz: as expected. which is why I'm suspecting dhclient, or some rc script, is trying to be helpful

"trying to be helpful"

I know that feeling

ha! i got willy ubuntu release on a machine!!! rwp progress!!

babz: I created a dummy lo1 interface, then a gif1 interface tunneling from it. then I destroyed lo1. the tunnel mapping stayed.

so whoever's cleaning the tunnel up, it's not the kernel

rwp: IT MOUNTS!!!!!!

rwp: /dev/mapper/vg1000-lv 8.1T 1.2T 6.9T 15% /mnt/synology

ultramage, you probaly want to be a little more deliberate and pecise.

This might provide a clue: trioptimum.com/~crtxreavr/hev6

CrtxReavr: mine looks like this. I improvised it in 2008 and probably didn't need to touch it since. pastebin.com/raw/H9DT27Uy

CrtxReavr: in both cases, this thing only runs at startup, so it has no effect on my dhclient issue, which occurs at runtime

My response was in regard to your lo1 & gif1 comment.

CrtxReavr: ah, that comment was about testing whether the tunnel is removed automatically by the kernel. it isn't, it stays. there was no concern for safety in that test, I only did it to see if the observed behavior could be reproduced.

my issue is that yesterday, something nuked the v4->v4 tunnel mapping on the gif0 interface after my ISP rebooted my modem and dhclient woke up to refresh leases

ultramage, that's the rc.d script I wrote to manage my Hurricane Electric v6 tunnel when I needed one.

Learned the hard way that tunnels didn't get cleaned-up like you might think.

yea, same, but I'm guessing you're not on a cable modem with a dhcp-assigned wan address.

Well, I was.

hm I'm not sure what you mean. I thought that start() runs at system startup, and stop() runs on system shutdown. does stop() also run on every interface down event?

These days though I use native v6 on Google Fiber.

No, it ran on startup and shutdown because I set it to in rc.conf

Along with the other hev6_* variables.

ok, so, how is it related to my issue that's occurring at runtime?

sorry if I'm missing something obvious

hev6_stop() stop_cmd only runs on system shutdown, if my understanding is correct, and at that point it doesn't matter. The script doesn't run at any other point in time, so it has no effect on any other shenanigans that the builtin rc scripts do in response to system events

from what I can tell, the tunnel is set up correctly by start_cmd, so if anything malfunctions after that point, it is outside of my script's control.

I would also like to ditch this tunnel, but last time I asked my ISP about ipv6 (which they extensively talk about in their faqs), they told me they don't offer dual-stack, and that I would be switched from public dhcp wan ipv4, to a private v4 or v6 address + CGN. So I said no.

and that v6 address I assume would be a /128 meant for my router only, which would then do nat4to6 for the LAN. so why even bother.

voy4g3r2, Woot! That

voy4g3r2, Woot! That's excellent news! (still quite early here and my brain has morning fog)

rwp: same here sir, i start my day at 630 .. europe clients, indonesia.. my sleep pattern is just horrible

I am also getting online so I can join a meeting here too. Very happy to hear your good news though! That's motivating for the day. :-)

rwp: yeah it is, now to move 250 gig worth of stuff

once that is done.. i get to go down openzfs route and that x86 machine is MUCH more powerful than these raspberry pis so that may become to server.. plus it has more usb 3 ports than the pis.. because who doesn't love usb3 hard drives :)

Well... *I* don't love USB persistently attached storage. Seems like it works for a while and then at least once a month I have to power cycle the USB adaptors because they are freaked out. The ones I have just have not been long term reliable. Good for intermittent use sure. Just not good for 24x7 use.

yeah, if i was using it more frequently.. totally get that, i mean i have 8TB of space and was only using 1.2TB

and a good majority of that was just "left overs" so i got options for redundancy and back ups with the 3 drives..i want to try and physically isolate 1 drive for "offsite" backup

rwp: that's been my experience too

spend 50 bucks on 3 enclsoures or 600 on a new synology.. i went cheap but can see what i can do with other options in the future

but i do know once this is done, this ubuntu box is going 14.0 release and openzfs.. as i got a 256 gig internal drive which looks like i can help with "balancing the pool, but anyway.. some more meetings with belgium for the next 3 hours

mmm. belgium

has an ER for an update to github.com/openzfs/zfs/releases/tag/zfs-2.2.1 been discussed yet

Title: Release zfs-2.2.1 · openzfs/zfs · GitHub

tykling: 14 is 2.1.9 and it looks like git src is the latest commit.

skered: the release notes say 14.0 is 2.2

freebsd.org/releases/14.0R/announce "ZFS has been upgraded to OpenZFS release 2.2, providing significant performance improvements."

Title: FreeBSD 14.0-RELEASE Announcement | The FreeBSD Project

14-STABLE might be 2.1.9 now but if the release notes are accurate the release is 2.2

<sigh> right as I try to upgrade to 14, I get what is most likely a hardware or bios issue. >:(

I do have 14 running with wayland and sway on a t430. Runs fantastic on there.

tykling: doh! I keep forgetting I didn't update this machine (where I ran zfs -V) to 14.

It's the machine beside me.

Yeah 14 is 2.2.0 nm.

hey I'm trying to boot the installer and it panics

`panic: vm_fault_lookup: fault on nofault entry, addr: 0xffffffff81fda000`

my CPU is an Intel Core i5-13600KF

does it have anything to do with the P-core/E-core stuff?

let me try disabling the E-cores brb

tykling: from the linked issue, block cloning is disabled by default on freebsd already

exit

wait.

intel does big-little now ?

da.

Regarding jails, is there needed to run a firewall for getting external network inside the jail? Tried jls, iocage and allow rack socket, sysctl settings etc. But I cant get even ping to work. Is there something im missing? tried several methods but cant seem to get it to work

you can add a different ip to a loopback interface, and use firewall translation rules to redirect traffic to the appropriate jails

you can use vnet with epairs, and route ip traffic on the host

or you can bridge the host side of the epair to your external interface

Title: FreeBSD Jails And Networking

also, see the sysctl variables in if_bridge(4), section "packet filtering"

If n3c had stayed around we could have helped with jail networking.

i think i am going crazy here.. freebsd 14.0 installs multiple version of python 3 .. 3.8.10, 3.9, 3.10 and 3.11.. when i do pkg install python. I see tips on how to change versions for /usr/ports install, are there instructions to change version for pkg installed pythons?

There are limited python versions available from the pre-compiled binary pkgs. But you can simply select from among those that are available and install them.

voy4g3r2, On 13 (because I am still on 13) bsd.to/vL8F/raw

Title: vL8F

I can choose among 3.8, 3.9, 3.10, 3.11, with the "python3" package depending upon the latest one.

yeah, it is defaulting to install all of those 3 ones

Oh, also 2.7 too.

eww 2.7... :)

So... What version of python do you need?

3.11

do i just remove the other ones? and i am making this too difficult

You are making this too difficult. Try "pkg autoremove" and see what it wants to remove.

That's a good general cleanup to do regardless.

yeah.. well i am cloning packages which now makes me wonder what other machines are acting up like this

my main dev box is still 13 but the one i have been fighting with.. i put 14 on it

I am still on 13 and quarterly but I only have one version of python installed. bsd.to/ypSA/raw

Title: ypSA

Title: dpaste/nz5a (Plain Text)

voy4g3r2: why do u need to remove other versions?

do they break something?

to ensure i am using a specific version for my programs

It's hard to get excited about having two versions. I think it likely that some port is still using the older 3.9 version.

What does "python3 --version" say for you? 3.9 or 3.11?

it is saying 3.9 right now

i figure once i remove it,, it will defualt to 3.11

Title: Use different Python version with virtualenv

?

yup, i am using venv python -m venv <directory> is my friend

I don't know but I think if you remove it that you won't get 3.11 by default. I don't know but with that info I expect some port is using 3.11 and that pulled in the newer one.

well i can not move 3.9.. irssi depends on it

But you could try it and then you would know. The worst that happens is that you find out it isn't what you need and need to put it back.

yeah, just threw me in a loop when i "cloned" my packages from one machine to antoher and it will pulling in multiple python 3 versions ..

And now you know. But irssi? That's mostly perl not python. Did someone stitch python into it too? Gack.

looks like it..

Title: dpaste/GC2o (Plain Text)

It's just a good time for me to say that I am always hearing about these problems with python but I don't recall having this problem with perl. Just saying!

rwp: haha, well at least i can read my python code after i write it.. perl just hurt my eyes.. but i learned it back in the day of the o'rielly books at v5 before the v6 mess

it was VERY helpful when i had to analyze report outputs on HP-UX land

use a terminal emulator on a windows machine, pull report from HP-UX, do some magic with report -> download to windows, perl program to execute the results and tell me what pages i needed based on conditions

Perl v6 is now officially no longer an upgrade path of perl 5. Perl v6 is now Raku and only "a" language in the perl family. Raku is now a standalone language separate from Perl.

long left that job but man it saved me hours of time and tons of paper.. they use to print out the 500 page report and only need like 75 pages

so i saved a lot of greenbar paper, ink, time and all with about 100 lines of perl

probably could be done more efficiently but man it was "ugly"

Perl has been a strong performer for years. The community hate toward it now I think is not justified.

yes, i agree with .. it is better than awk/sed/grep mess

Tractor feed green bar computer paper. I still have some old printouts of some old programs of mine in my bookcase museum section.

yup!

I will also defend awk/grep/sed too. I blame the programmer not the program!

bbiab

hahah

i was like 18

but i am out myself concert time.. trans siberian orchestra!

rwp: have you been reading Hillel Wayne's posts on Raku? I quite like his takes on it!

meena: Raku as in formerly known as Perl 6 ?

Title: Raku is surprisingly good for CLIs • Buttondown

mns: Raku, the artist formerly known as Perl 6.

thanks meena

oh, we don't have a raku port?

no we don't it seems

meena, I have not read those. I will put them in the queue to read. Thanks!

I really HATE that Raku decided to rewrite the PCRE making it incompatible but without any compelling reason that I can see. Other than Raku's unique regular expressions the rest of the language is interesting. But for me the Regular Expression change just kills it for me.

That is a big change

next step is to get a Raku port going

rwp: you'll be very unhappy to learn that that which we call PCRE across languages is actually not very compatible… ar5iv.labs.arxiv.org/html/2105.04397

Title: [2105.04397] Why Aren’t Regular Expressions a Lingua Franca? An Empirical Study on the Re-use and Portability of Regular Expressions

It's a good paper well worth the read.

bookmarked it for the reading queue

meena, I understand that there was incompatibilities in non-perl things. But PCRE in should work in perl things.

Raku's regular expression syntax is incompatibly changed. Mostly it is a Caesar Cipher of just rotating characters. Which is one of the reasons I particularly dislike it. It's prone to "negative transference of learning". It either needs to be the same or it needs to be quite different.