thanks for all the help

have a good night!!

So what was the conclusion, if any, of mrkiko's EFI issue?

he would need access to the efi loader, freebsd loader is too late

rtprio, Thanks

dpaste.com//9T93QACYZ

Title: dpaste: 9T93QACYZ

There is "efibootmgr" which could be used to change the order (use at own risk, etc); then reboot (to see if new order is used to boot in the desired OS)

hello

has anybody ever got ipv6 to work without using one of their images

has anybody ever got ipv6 on ovh to work without using one of their images

sn00p: who's images?

ovh's i misstyped

oh. see what their image is doing and do that with your own

they dont have freebsd nomore

i used to have instructions on how to set it up , but i lost them

i see fantastic unsolved thread from 2017 on the forums

url?

Title: ipv6 -host route to on-link gateway fails | The FreeBSD Forums

jello

hey, does drill behave incorrectly if IPV6 nameservers are listed?

so it has a 128?

this link says just use route advertisement

i'm guessing you tried that

yes

this other link from ovh says you should disable route advertisement....

awesome

how do I disable route advertisement

i don't beleve it's enabled unless you have a line like ifconfig_lagg0_ipv6="inet6 accept_rtadv"

any ideas on how to fix this

so you enabled rtadv and it didn't get an ip?

i have enabled it already

(and net.inet6.ip6.accept_rtadv=1 ?)

let me try that

its necessary for rtadv to work, but i _thought_ the rc system did it for you

ifconfig_ix0=ipv6="inet accept_rtadv 2001:41d0:1:5db0::1 prefixlen 128

i have this in the rc.cofn

'rc.conf

if you're manually assigning an address, you don't want/need accept_rtadv

it's either or

ok

it still doesn't wo rk

are you `sh /etc/netstart` or rebooting or what

rebooting

but `ifconfig ix0` should show the ip

can you ping the router?

idk which one is the router

they don't tell you in your control panel? no i suppose they wouldn't

hmm

i edited

with prelinlen 32

i get ping6 yahoo.com

but it takes a long time to resolv

without a router i don't think that'd work

where did you get your ipv6 address

ovh

any other information with it?

no

not even a prefixlen?

prefixlen is 64

but I tried 64 then I changed it to 32

it pings it but doesn' tcome back

you need a router

i have a router

what is the address

sec

is it set in rc.conf as ipv6_defaultrouter ?

ipv6_defaultrouter="2001:41d0:0303:a2ff:00ff:00ff:00ff:00ff

"

yes

and does it answer to ping?

how do I ping it, just type that address in?

ping6 2001:41d0:0303:a2ff:00ff:00ff:00ff:00ff

it pings it

no resposne

uh

i type that full line

what is `ifconfig_ix0` now?

network inferface

i know it's the network interface dummy, paste it because above you have 'inet' instead of 'inet6'

i have inet6

not inet

inet6 2001:41d0:1:5db0::1 prefixlen 128"

right?

ifconfig_ix0_ipv6="inet accept_rtadv 2001:41d0:1:5db0:1 prefixlen 32

inet6

i mean'

sorry I cant paste i have to type, im doing this kvm

i said take out accept_rtadv

ok

also why aren't you ssh'd into this box

cant login root disabled

... PermitRootLogin prohibit-password

or create a user that's a member of wheel

don't make things hard on yourself

ok give me a sec

2001:41d0:0303:a2ff:00ff:00ff:00ff:00ff

but if that's the gateway they gave you and ip they gave you, maybe paste it in to verify it's correct

this is what ovh says is my gateway

that what it says in the dashboard

but that ip address doesn't appear to be in the network

ok,

are those 0's or O's

zeros

add net default: gateway 2001:41d0:0303:a2ff:00ff:00ff:00ff:00ff

root@sn00p:/etc # ping6 google.com

PING6(56=40+8+8 bytes) 2001:41d0:1:5db0::1 --> 2a00:1450:4007:807::200e

-- google.com ping6 statistics ---

12 packets transmitted, 0 packets received, 100.0% packet loss

and I have it set 32

if you can't ping the gateway, it's extremely unlikely anything else will work

i cant ping the gateway

i dont know which one the gateway is

i thought you said it's in the control panel

is it

are there other things in the control panel that will give a clue

it is

just the ivp6 address

and the aother ip4 address

i think it's still prefixlin 64, at least that's what the linux documentation says

if I changes it it says network unreachable

i change it to 64 and it says next unreachable

dd net ::0.0.0.0: gateway ::1 fib 0: route already in table

route: writing to routing socket: Network is unreachable

add net default: gateway 2001:41d0:0303:a2ff:00ff:00ff:00ff:00ff fib 0: Network is unreachable

i took out default gateway

ping6 yahoo.com

PING6(56=40+8+8 bytes) 2001:41d0:1:5db0::1 --> 2001:4998:44:3507::8000

ping6: sendmsg: No buffer space available

the numbers don't seem to line up, man

i cant get it fixed

The default gateway for your IPv6 block (IPv6_GATEWAY) is usually xxxx.xxxx.xxxx.xxFF:FF:FF:FF:FF.

that's on the ovh page

i'm getting really pissed at ovh

so your ip would need to be in that network

such as 2001:41d0:0303:a2ff:a:b:c:d prefixlen 64

I may as well install debian again

if you did that, i'd see how they do ipv6 on their image

it wont matter

i tried that already

long ago

i suppose ovh will tell you to go to hell if you approach them with freebsd

and of course ovh

doesn't have a frebsd image anymore

you don't need one, you just need to know how the image provisisioned ipv6

yea

i just opened up a support ticket

helllo

hello mictty

hello Helper

I am trying to support FreeBSD for a compiler project

I don't have an admin environment nor the idea of the layout where packages managed by own users so that users have their own paths for include or sysroot for claang

that is to say I made the compilation works with --sysroot=/ (duh) but I wonder how it is gonna be like in terms of regular user

why not the freebsd handbook? zoidberg!

i know it goes slightly different that depends on how admin configured. So i thought the best is to ask to a bsd admin

Title: BuildingFreeBSDWithClang - FreeBSD Wiki

llvm/clang foobar

ports build from source and pkg binarys built by some one else

netbsd pkgsrc even more ISA's and os's

its all about abstraction

i say focus on amd64 and risc-v cause arm/uboot/devtree is for the birds

besides the broadcom/softbank ip licensing foo

I mean I need to know how admins typically sandbox users. I need to know how users will get their own packages and place somewhere. I need to know how users would assign their environment variables to use their clang. I need to know typical paths for such an occasion.

I need to make sure users don't really need to configure anything once they try to run this build system

mictty, Unix machines have been used for decades in universities and other teaching environments. You can keep things simple. Or make them as complicated and strong for security as you need.

mutiuser,mutitasking, perhaps UID/GID

The simple case is just to count on regular Unix security in the classic way.

Or you could set everyone up inside their own jail where they are strongly restricted.

The choice is up to you.

3 users imgur.com/RWy9j4x sun 2/120 1984 $16,900.00

Title: Imgur: The magic of the Internet

focus on the specs of what was offered

atime,ctime,mtime, solaris bsm audits

That advert from 1984 is so quaint now.

i use to telnet for ada compiler cause my college could not afford them

insecure telnet/ftp

gopher/usenet rocked on amber display

telnet, ftp, rlogin, those were the best available because encryption was a munition restricted by the ITAR. Using encryption then was illegal.

ill beat the dead horse, ive said before in 1995 solaris 2.5.1 for pentium mmx/pro was $600.00 no compilers, meanwhile walnut creek cdrom for $50 came with freebsd/linux and compilers

and sunsite.edu dump

yeah the print the pgp in book to get around government

free speech vs weapons grade encryption

cat file | gpg -e -o file.gpg -r foobarmypassphrase

cat file.gpg | gpg -d -o file.decrypted

pgp/gpg

before all that dmcrypt/luks,truecrypt...etc

rennj: you don't need to cat...

dont be pedantic

hehe

example of pipes is always nice, and im old

let me scream at the clouds

i really need to shadow a bsd admin for a week

eating together

openbsd finding a null deference i want say in ping after 20+ years

fuzz the binary with file

that news

Title: Fuzzing ping(8) … and finding a 24 year old bug.

submited 11 December 2022

24 years!!!!

haha like lp account on sgi irix no passwd

classic

spam the print server

no paper left

besides ink,toner

Title: Fuzzing ping(8)

Freebsd 12.3 , install and I can't seem to input / in the guided partitioning.  What am I doing wrong?

rewrite in memory safe lang

the pain

What  language?

golang,rust,jvm/kotlin

or some such effort. like nix and reproducible builds

What are you  making?

avoid the supply chain attacks log4/pip

log4j

nothing different than oxide.computer trying to provide

Title: Oxide

from the server bsm all the way up the abstraction stack

Man my only coding has been hello world in two different assembly languages.

Okay, installing when you delete a partition, with a file system, how does that show up in the guided partitioning? Does it have to like refresh or something?

Eggshell with Romalian type. What do you think?

"I Have To Return Some Videotapes."

With an installing how do you check if you have internet.  I'm at fetching  base.txz and its on 98%

at 98% you have not figured out yet weather you have internet connection

cause 0% vs 98%

ping,traceroute,netstat

check the man pages

tcpdump/pcap if you really want to dig down

So the fact I have any thing fetching  mean I have connection

yeah the download perhaps timeouted..worse case

tcp/ip keepalive failure

still at 98% ?

Is there a way in the FreeBSD installer, to go back a step and pick a different site to fetch from?

Yes 98% still

oh so close, well take heart, back in the day just to download 700MB iso was 3days on 56K/modem

perhaps try again..no clue on my end

i remember bitching to OEM to include damn checksum on iso

I really do understand, why open source has so many install guides.

md5 let alone sha,des,aes

email them hey guys how about you include a checksum on the iso download

hehe

hmm 56k modem suckage days

the x-modem,z-modem,y-modem,kermit days of yore

you got the ppp account to your isp doing the dialup

hi

hi

:P

i like doing this to my kernel: strip /boot/kernel/kernel

I appear to create a partition, I selected the mount point and hit the backslash and I don't see anything. My keyboard appears to be recognized, I did the little test and typed a bunch of stuff including the backslash, so why can't I see it in the partitioning guided installation? Am I not understanding how the options actually work?

backslash as in "\"?

parv

mictty, Yes?

do you use CXXFLAGS

No, for I do not compile software often enough, outside of The Ports, where that would matter

do you have your users you administrate

I don't. I am not the person who could advise you on your project (from your earlier posts)

thank you parv. it still helps

Hello

Do you know how can I apply the aspell filters for "HTML" and "url" in the API?

in command line mode there's "aspell -a --mode=url"

cadmio: you mean such an util?

it seems there exists a filter bpa.st/5QLQ but I don't know how to use it

Title: View paste 5QLQ

cadmio: it looks like a set of interface...

do you think there's a possibility to apply the filter in the API?

cadmio: we can script to work between your file and aspell...

this is what I've at the moment bpa.st/RGMA#1L52

Title: View paste RGMA

it works well but it corrects urls like ".html" words

cadmio: what makes you stick to c for this kind of task?

I written the whole program now; I simply wanted to add a spell checker

cadmio: you can always syscall

cadmio: stdlib has system

but it's a pity to resort to syscall when there's the library which in the end works well

cadmio: I thought you do once and throw away code

parv: that was a really cool talk

why mysql has no just 'mysql' like sqlite3?

who did this

mictty: of course it does

you'll have to install one of these codeberg.org/FreeBSD/freebsd-ports/…tabases/mysql56-client/Makefile#L20

Title: freebsd-ports/Makefile at main - freebsd-ports - Codeberg.org

codeberg based

I installed 12.3 and everything seemed to be fine and then I went into the bois and I could not find the install in the menu to boot it. Is that a know problem? I thought since the them drive for the install was seen the bois would see my install.  I have Windows, Linux .  Does freebsd have to have its bootload first or something?

tarel2, you using grub from linux to boot windows and linux ?

tarel2, and witch distro linux you use

hi all

I have a hdd ssd/gpt/ntfs with dual boot windows 10 and linux. I can't mount it in freebsd.

Freebsd is installed on another hdd, separate, all ssd and all gpt.

HDD with windows and Linux I put them with a usb system.

So, I change the hdds between them...

I can't mount the hdd. I can't even visualize it.

I'm not using grub, I say the BIOS because that's what I boot into, I manually select whatever operating system from there. And it doesn't appear in the list. That wasn't so in the past.

I come from the Linux World! I want to stay with FreeBSD! :D

The big I have ran frebsd and openbsd I have enjoyed how everything works.  I was on a pi so the hardware support was lacking .

But arguably, the Raspberry Pi is not a normal piece of hardware.

So I would expect the hardware support not to be there.

yo9fah: :D

mictty: :P

is there a way to explore a file with the extension .core?

is there a way to mark a package as required, so it won't be *falsely* autoremoved?

you can do pkg lock package

Helper: will that still allow to upgrade it tho?

nop

it wont

the only way to upgrade it it will be if you do pkg unlock package.

and then running pkg upgrade

Helper: yes, with lldb, but it's somewhat complicated

oh I see.

I have muliboot windows and Linux , I use the bio to chose the os and boot  alot .  Its older machine , Asus m5 a97 le 2.0.  I think I got the newest 13 and my machine just froze the asus logo.  I installed with 12.3 but I could not find freebsd in the bio boot menu.  12.4 , download for the install got stuck at 98% and I tried more then one

server.  Is my computer not something that works with Freebsd?

what are you downloading that gets stuck?

Helper: would be cool if pkg-lock(8) allowed a Version match, too. so you could go: pkg-lock FreeBSD-fetch -s >=13.1

i never knew about those packages

rtprio: i provide a repo for people to try and feel the same pain as i do: alpha.pkgbase.live

Title: Unofficial FreeBSD pkgbase repository

i guess i eithr used the distrubution files or build from source

that's interesting tho

I've seen people build extremely small images / jails with PkgBase, and I'm kinda impressed

(i usually put Puppet as default on everything, which instantly ruins any chance of it being smol)

I need find a better way to do configuration management.

or maybe i just give up on that, and use buildah to build a fresh image every time there's a change

_xor: could i run nomad on my laptop to run a (podman) deployment on a cloud server?

tarel2, when you boot linux do you get grub hoot menu if yes update grub it will find freebsd and add to boot menu than you can boot it from there

meena: `pkg set -A 0 ...` ? check the examples of the pkg manpage for how to query those

ridcully: thanks!

Hi, could you help me understand this error? I just finished a system upgrade from 13.0 (aug21) to 13.1(dec22) and attempted 'portmaster -af', but it immediately dies with : not foundMk/Scripts/ports_env.sh: set: Illegal option -o pipefail

this is a 2020 issue, my previous OS's ports already had this flag in that file, so I'm extra baffled why this is showing up.

ultramage: i would try `pkg-static upgrade portmaster` and see if it still happens

ah, never mind, I see the problem now. the files got messed up during transfer. sh can't parse CRLF endings

it wasn't the command itself, it was the fact that it was the first non-comment line in the first file that was being executed by portmaster.

dos2unix it?

yup that was the quickfix I went with.

what the hell were you using to copy that didn't respect line endings?

ultramage, of course it can, that's what the $IFS environment variable is for.

I was forced to switch from svn to git for obtaining ports; I didn't have git bootstrapped on this thing yet, so I mirrored it to my pc and copied the files over. I didn't relize git crlf-ies by default.

CrtxReavr: I didn't know. I looked it up just now. Seems like it's a bash-only variable, ports run on sh. Sh seems to only handle LF line endings, and gets confused by the extra stray CRs

ultramage: i'd tar or zip it up the next time you need to do that, for safety

git by default will use the native system line endings. Same as for svn before. So... CRLF on Windows systems. :-(

wouldn't help, it was CRLF from the moment it was checked out, due to default git behavior. That behavior can be changed, but I'd need to re-checkout. git diff doesn't report everything as changed after I changed the setting, so it seems to auto-ignore that by default.

It's crazy but sh now includes bash'isms like pipefail and also has command line editing. Who knew?

you just need to flip core.autocrlf to 'input'. I was just tricked by all of git's meta files being LF

plus the fact that midnight commander only shows the ^M's in edit mode, it doesn't render them in view mode, which I was using to examine the problematic files

I wonder if the CR is the reason why the error message is rendered in such a weird way, without a space between the message and the file path : not foundMk/Scripts/ports_env.sh

ultramage, Bourne/POSIX shell definitely supports $IFS. Had to manipulte it many, many times.

or don't go copying around repos to windows systems >:|

does /bin/sh? it was fed a CRLF sh file and it crapped out on it, as I posted earlier. Then again, I have no idea how IFS works, so maybe it does if you actually set it

although, I don't think I'd want the OS configured to accept both, it's handy to be able to catch file encoding mistakes like that

Title: View paste ZM2A

ok, gotcha. can you use this to accept both LF and CRLF as a single line ending? My first guess is that if you put multiple things into that string, each becomes a separator, and CRLF would produce an empty line

IFS is not a good solution to line ending problems.

YOu can change it at will in a script.

rtprio: I was hasty. Also, regarding svn, svn by default copies the files as-is, unless you tag everything with eol-style property. 'native' would cause this same problem.

I somehow convinced myself by looking at files in .git/, that git runs fully on LF, forgetting that it doesn't and you can tweak that via .gitconfig

that said, it was really confusing how portmaster and Mk responded to a mis-encoded ports tree. I wonder if a safeguard against this could be added into the beginning of the process, so that when it trips, the error message isn't so misleading

first custom kernel success B)

Congratulations MantisOrientalis! Does this imply that there were earlier failures?

sigh... for kernel builds, I used to use MODULES_OVERRIDE= to only build the drivers I needed. Then one time the boot failed because freebsd split off something into a new mandatory module. Now I just build everything.

Hello Friends. I disabled sendmail (via rc.conf) and shutdown the process via the "service" command, however according to my /var/log/mail log, a sendmail process attempts to send an email to root@localhost with a log file daily at 3:00am. Any idea how I can surpress this email because it's filling up the log unncessicarily

sounds like the periodic daily summary report

RoyalYork: That's probably the periodic(8) output, you can change whether it gets mailed or put in a logfile in /etc/periodic.conf

ohh, that's an idea. I got so used to the daily e-mails, I didn't think about having them go elsewhere. Which log do they go to?

Catie: Thanks for this. Im new to BSD and have been scouring the man pages unsucessfully.

(though having the daily report in your face like that can help catch some issues early)

ultramage: I set mine to go to /var/log/{daily,weekly,monthly}.log, but keep the security mails coming straight to me

RoyalYork: Yeah, no worries! There's a whole damn lot to try and get used to

RoyalYork: the periodicity implied it was scheduled. Freebsd uses the usual /etc/crontab to start its periodic scripts, they default to 3am daily, 4am weekly 5am monthly I think.

Oh it looks like the default newsyslog.conf knows about the logfile names I used, which is probably why I used them in the first place

oh I see, makes sense. I went and deleted and commented out all unused log files, very early on, so I didn't realize this was a thing

maybe weird question: when I try to copy a directory from windows via smb, to a share of /tmp, I get "Failed to set folder attributes. The request is not supported.". This was before I updated the OS, and I haven't updated samba nor my file manager nor my OS in at least a year. I don't remember how long it's been doing this, just that in the past, it didn't.

I can process and network trace it later to see what exactly it's trying to do, just asking if it's something obvious. I didn't know tmpfs has operation restrictions

Catie: follow up to periodic if i may. Im reading the man page on periodic.conf(5). Says the conf resides in /etc/defaults but can be overidden with a file of the same name in /etc. Is there any downside to editing the file within /etc/defaults or is it best practice to use a separater override file in /etc?

RoyalYork: Best practice is to use the override file

Understood. Just to exercise the mind, whats the downside of editing the original?

RoyalYork: I think probably it's just that nobody will expect you to do so, honestly. Also there's more work for merging configuration files when update time rolls around, but that's not the biggest deal in the world

Thank you

I've always used the override files myself, so I couldn't really tell you if there are specific issues with editing the OS default

I guess its similar to what I do with rc.conf

ultramage: thanks for your help as well. looks like i accomplished my task. we'll see what the logs say at 3:00am

the override file is in the standard location in /etc, and it starts out empty so you can put in exactly the adjustments you need, and you can put in #comments describing why you put it in there. I use sysctl.conf the same way.

i'm on a ovh server and trying to get ipv6 setup with freebsd every time i get this ping6 google.com

ping6: UDP connect: No route to host

pastebin.com/1QUFDaWi is my periodic conf. I think I sat down one day and went through all the options and picked everything I found useful. There's a lot.

Title: # combine into single e-maildaily_status_security_inline="YES"weekly_status_ - Pastebin.com

sn00p: check ifconfig and netstat -arn, to see if your routable ipv6 address is there and that a default v6 route is there

how does ovh do ipv6? static configuration that they e-mailed you, placed right next to the static ipv4 configuration in rc.conf?

ultramage: very through conf file. i'll have to do that myself one day

ls

the more I play with freebsd, the more I realize man pages were the original google

having all the options and instructions right there, in a consistent format, came in handy countless times. Though there's also the freebsd handbook pages, and web search usually also helps.

Hello, I am trying to setup ipv6 on my ovh dedicated server, i get add net default: gateway 2001:41d0:0303:a2ff:00ff:00ff:00ff:00ff fib 0: Network is unreachable

and here is my rc.conf file pastebin.com/FH5wRY5m can anybody tell me what i'm doing wrong

Title: sendmail_enable="NONE"hostname="sn00p"ifconfig_ix0="DHCP"ifconfig_ix0_ipv6 - Pastebin.com

sn00p: you made a typo

also you might want to be consistent with the zero padding (303 or 0303)

I suspect there's a typo in the fourth part, your ip says a2df but your gateway says a2ff

ultramage, "In a Unix man page every word counts, sometimes twice." :-)

try ping6-ing your gateway, or maybe tracerouteing it. Network unreachable probably means that your host tries to neighbor-discover the gateway via broadcast, and nobody replies

ultramage, what typo did I make i put 32 instead of 128 and it seems to be working but really slow

your ip says a2df but your gateway says a2ff

thats what in the ovh control panel

ah alright then, if these two values are exactly what they should be, then I guess...

i ping6 google.com i get response but the connection seems slow

I'm actually not sure how that all works

normally, your gateway needs to be inside your / broadcast scope to actually work as a gateway, right?

idk

ovh is really weird setup

but I hear /128 is a thing... so right now I don't have a proper mental model on how that works exactly

the slowness might be an unrelated problem to this.

ok i just did a reboot

everything sseems to be working fine

did you put the /128 back in?

it didn't work on 128 i put 32

and it worked

that is not right

I refreshed my info a bit, I think how it works is that you get a /128 ip for your host, and then you either receive a router advertisement for a /64 on-link subnet, or you hardcode it into your route table

does ovh's instructions page cover the part about configuring the route? I also can't remember if you need to turn on a listener daemon for those RAs, or if the kernel handles that automatically

in my tunnel script, I have a line that says "route add -inet6 default -interface gif0". maybe that's all it takes?

one of the ifconfig optional parameters is 'accept_rtadv', btw. not sure if ovh wants you to define a static default route, or let them announce it

ultramage, doubtful you'd get a /128

sn00p: ipv6_default_interface="ix0" is probably what you're missing in that config

Normally you'd get an IP on a /64, and they'd provide you a /64, /56, /48, or maybe a /64 for use on your inside network.

how do I bind host for ipv6 with znc

the routing script calls 'ifconfig ix0 inet6 defaultif'. In my script, I just do 'route add -inet6 default -interface ix0'. Both seem to accomplish the same.

CrtxReavr: that's if you're a business customer, or are talking to a provider that hands out /64's. Apparently DHCPv6 does /128. And from what I've seen, everyone except those that care will only give you a /128

32 and 48 are the only ones that work

sn00p: set a /128. Add that missing rc.conf line that I just gave you. Then it should work.

i tried 128 it didn't work

64 and 128 ddidn't work

Right now it doesn't work because you didn't give it any route. You only configured your ip, and the gateway ip.

sn00p: ipv6_default_interface="ix0"

you are missing this. you forgot to define a default interface. it doesn't know which direction the gateway is

it is also possible that the gateway is announcing routes. in that case you could add that accept_rtadv flag to your ifconfig line. not sure about that.

Well, you realize that /128 is exactly one IP?

/128 is *NEVER* the right answer.

/127 can be for point to point connections.

But a segment or subnet should always be a /64.

And anything with a smaller prefix length, is just multiple /64s to use.

i added that line and reboot

No. You get one ip, one /128. That's all you get. You cannot host a routable subnet, you cannot have multiple IPs, you cannot do autoconf. Just one ip, same as in v4.

WTF are you talking about?

The thing is, routing is decoupled from your ip and prefix. I've honestly never thought about this, so I'm not sure how it works in v4 land. I'm used to hosts having a single v4 ip, a subnet, and the gateway being on the subnet. But apparently you can also just have a subnet of /32, say that the gateway is somewhere on the other end of re0's ethernet cable, and that's enough.

If you use a prefix smaller than a /64, you break and lose features.

a /64 is a one-size fits all prefix. . . every device in the world can be on the same /64

wtf am I talking about? that /64 is the ipv6 ideal dream, but /128 is the harsh reality. No service provider will give you a /64, unless they're are specifically there for giving out /64s to network operators.

You are very wrong.

Most IPSs that support v6 use DHCPv6-PD

you only need a /64 if you're going to run your own subnet. If all you're doing is hosting a webserver that only needs one IP, then that's what you'll get

They configure your public interface, and provide you a block of /64s to configure your internal interfaces with.

well, that is a nice service provider, probably business-tier

ultramage, just stop talking. I've been doing v6 professionally for at least 15 years. . . probably more.

I'm 99.99999% sure that my shitty residential ISP hands out /128s. OVH, the multinational server hosting company, apparently also hands out /128s only. Maybe if you ask for /64 they'll provide it

I've been running a v6 tunnel and networking over v6 for 18 years now... I have a /64 (had a /48 at one point too). It's just that I've seen some shit.

the main reason why I'm still tunneling is that when I asked my ISP about v6, they heavily implied that they don't do dual stack - that if I requested v6, they'd drop my public ip and NAT all my v4 traffic through their CGN; and that I'd only get a single /128 v6 address for my home router, all the PCs behind it would have to do 4to6 or 6to6 nat.

it doesn't seem to like 64

you are supposed to put /128 in there, as per instructions.

i still get network unreachable

you do not have a subnet. your subnet is /128, containing 1 ip - yours. to talk to the router, you blast packets out the default route over your wan interface (once you configure it right). Well, first it'll broadcast neighbor solicitation to find the mac of the router.

ok, how do I fix

i added ipv6_default_interface="ix0"

if the network is more complicated than that, you need a proper subnet and route. either statically configured (apparently not since it doesn't seem to be in OVH's instructions), or you get it via router broadcast - which you probably need to turn on

this is ovh's setup and i'm not using their image i'm using a custonm image

ifconfig_ix0_ipv6="inet6 2001:41d0:303:a2df::1 prefixlen 128 accept_rtadv"

tbh I don't know what exactly it'll take to make it work. All the internet guides and standard documents use /64 with the gateway on-link, or rely on RA and don't configure anything. So I don't have first-hand experience with a /128 situation.

64 and 128 still do not work

Greetings,

have you heard of an equivalent of etckeeper for freebsd , afaics it's not available in ports

sn00p: are you sure your gateway address is right? have you tried prefixlen 48?

i think i tried 48 i'll try again

sn00p: I think ipv6_default_interface was incorrect.

Hi, I would like to connect a laser printer to my FreeBSD, ideally installing as less things as possible. Currently I have a matrix printer attached. I print with lpr

I think what you need is a static route, saying that the gateway ip is on interface ix0

now I have read this:

Title: Problems with my Samsung USB printer | Page 2 | The FreeBSD Forums

would that mean that I do not have to install anything? just point to the IP address of the printer?

sn00p: very early on I suggested using netstat -arn to look at your routes. Your goal is to make the gateway ip appear there, with 'gateway' column saying Link#, and Netif saying ix0

sidh, I have not heard of etckeeper for FreeBSD but would also like it myself. Though zfsutils auto snapshots are an acceptable alternative.

'route add -inet6 <gateway ip> -interface ix0' would be the script, not sure how to express it via rc.conf. Maybe using the ipv6_static_routes section?

rwp, I don´t think you can follow the changes with zfs snapshots

I mean, the comment of the changes

sidh, I know it isn't the same. But it does create a time machine of sorts where one can reach back into the snapshots and diff files.

Since etckeeper has not been ported. But etckeeper seems reasonable to port. It should work with just some tweaks.

Title: etckeeper: pkgng/README | Fossies

sidh, I look forward to your report after you have tried it! :-)

rwp, just found it

Meanwhile... I have been relying upon "zfs list -t snapshot | grep zroot/ROOT/.*auto-snap" to get a listing of zfsutils automated snapshots.

Then "diff -u /.zfs/snapshot/zfs-auto-snap_monthly-2022-11-01-00h28/etc/rc.conf /etc" to see what has changed.

Then using "zfs diff zroot/ROOT/$SNAPSHOTNAME | grep /etc" to see what files have changed.

It's not as nice as having the full git available. But it is possible to see what I buggered up. :-}

rwp, yes , but when you want to centralize your conf files to provision other jails etc., using snapshot can be quite tedious

furthermore if you don't have a etc /local_etc dataset

I mount a DVD using mount -t cd9660 /dev/cd0 /mnt/cdrom, it mounts it correctly -no errors and mount shows it mounted- but then ls /mnt/cdrom shows nothing

any tip?

is there a dvd inside? is anything written on it? is it one of those multisession cases where everything was deleted?

it is a dvd (the printer driver dvd)

it is not a recordable dvd

Title: View paste 2XRQ

Title: View paste 5OAQ

sorry, this is the one: bpa.st/CLDQ

Title: View paste CLDQ

apparently everything is properly mounted but no files

uskerine: unmount it and what's `fdisk /dev/cd0` say?

i haven't mounted a cd in a long time

Title: View paste XBFQ

it is actually a DVD

only thoughts I have is to use ls -a in case there's hidden stuff, and maybe check /dev in case the dvd is a funky one with multiple partitions

I have tried other DVD and it works

i was too poor for a dvd drive

but this DVD is relatively new

(7/8 yr, brand new, stored in the printer box)

that looks weird; does it look the same with one of your dvds that works?

I am going to check

maybe it just means it's all zeroes?

oh, question, does cd9660 actually work on dvds?

yes it does, I mounted the other PC DVD with that

mounted fine and you could ls

okie

i thought cd9660 was for dvds too

Title: View paste TGBA

this one is the dvd that mounts and can be browsed

ok, maybe fdisk is the wrong tool for this

gpart?

worth a shot, yeah

I'd probably image it raw and examine it with some dedicated utility. but it's just supposed to be a driver disk...

yes it is a bit weird

is it a pressed vendor-branded driver disk, or something made in a dvd-rw burner?

it is a pressed vendor-branded

it comes with a samsung printer

so it shall be a "proper" dvd

anyway, maybe it is faulty

I just wanted to see if the manual was there, I guess I can google

btw, the printer seems to use SPL, and I was wondering if there would be something to convert ps into SPL without using cups

I went pretty deep on the cd spec, which can have crazy stuff like multimode, audio tracks, xa mod (playtation), and all sorts of other crap. but I feel like dvds should be much more straightforward

probably the most straightforward thing is to put it into a windows pc and examine it there. but again, idk why it should be special. unless they borked the filesystem in a particular way that freebsd can't recover from

ultramage, i'm trying to setup znc for ipv6 every time i try to connect to server it says cant resolve bind hostname

well I was just thinking the way I was mounting was the wrong one, but apparently not, so it is not worth the effort

Title: Other - FreeBSD thinks my CDs and DVDs are empty | The FreeBSD Forums

I already read that one

ls -ltr

cat /etc/printcap

cat /etc/printcap

slow down boss

before you rm -Rf yourself

oops sorry

uskerine: (reading backlog) did you try mount_udf for your dvd?

no I did not

I can try

oh yes I did

mount -t udf

it did not work either

i'm trying to rack my brain as to what other file format it could be

uskerine: is ther anything other than /dev/cd0... like cd0c perhaps ?

nope

I think there is soemthing fishy with that DVD or with that dvd config, in any case, other DVD is being read ok

now I am trying to get my Samsung laser printer working

if possible without CUPS

udf is the fs for dvds (usually), and when mounting it as cd9660 it would present a README saying use UDF :)

well it seems I have a brick instead of a printer given that I can not use it