unixwitch: tldr: non-ECC RAM don't magicaly corrupts your on disk data on a zfs scrub

satanist: yeah i mean... that seems reasonable. don't really understand why someone linked that to me though

scrub of death is the idea that with some failing (or evil) part of the ram zfs might caused to override a good block with wrong data in the complete fs

i don't really see how that could ever happen since scrub only overwrites data which is already corrupt. unless bad RAM somehow caused it to read bad data from all disks in the pool which seems rather unlikely

maybe misinterpreted this: "a single error can render the entire pool broken which is not ideal behaviour" as a referenc to the "scrub of death"

but a single error don't render the entre pool broken

In terms of ipfw, what's bdg_forward?

I can guess of course, but wondering if there's any further documentation on it.

I grepped /usr/src and it seems to only be in sbin/ipfw/ipfw.8, the man page.

Is it just a thin packet forwarder used to implement bridging? Based on the diagram in the man page that's what it looks like, and also based on net.link.bridge.ipfw=1.

> Controls whether bridged packets are passed to ipfw.

stand by net.link.bridge.ipfw at SYSCTL VARIABLES

Right, the sysctl desc, but I'm more curious about a further description.

Like, if it's enabled, does ipfw do anything else with it before handing it off to the next layer? I'm guessing not, but the diagram + description is is a bit ambiguous, IMO, because it's implied as being another layer in the stack.

this is about packages which are passed throuh a bridge interface, so if you have two interfaces connected with a bridge packages with goes throuh this bridge take the bdg_forward path

enabling net.link.bridge.ipfw allows you two filter such packages

Right, the explicit description is pretty straightforward, but it can read as if there's more to it than that elsewhere (implied behavior), since there's no other reference to bdg_forward in the man page, but it's shown as a layer in the stack.

I haven't found much more, only the description in if_bridge(4)

Yeah either, no biggie. I'm sure that's all it is.

So, does hostname in /etc/rc.conf expect an unqualified name or a fqdn? I've generally defaulted to an unqualified name, but am finally getting around to reworking my DHCP/DNS setup and this is going to matter. I'm reading about it right now and getting info on both using unqualified vs. fqdn. /etc/defaults/rc.conf hasn't told me much, yet.

Like, for example, this comment: "You've completely disabled sendmail(8) which, to the best of my knowledge, is the only service that requires an actual FQDN for the hostname. It doesn't need to exist (as long as it resolves locally) but it does require a hostname with at least one dot in it."

I know the sendmail alternative in base, but in general, trying to understand the ramifications of using hostname="my-machine" vs hostname="my-machine.my-domain.tld"

Oh, got my answer in rc.conf(5). Hehe @ "Amnesiac"

g

The webcam on my laptop sucks ass lol. anything i can do software wise to fix it ?

Apparently I can use my Android phone camera as a USB webcam. Or so I am told. I haven't tried it yet.

Hi. While installing bash, bash patches time outs to download from sunsite.icm.edu.pl fetch: ftpmirror.gnu.org/bash/bash-5.2-patches/bash52-007: Operation timed out. ping is ~12ms, no loss. noticed not for the first time.

/t

Title: dpaste/M827 (Plain Text)

FaFaS: Works for me, but it gives me a different mirror: mirror.kumi.systems/gnu/bash/bash-5.2-patches/bash52-007

vkarlsen it works, but not always, on random URLs. so i have manualy download missing ones before the installation starts.

FaFaS: Looks like an upstream mirror problem

probably not FreeBSD related how package is looking for the mirrors. Just wanted to share it.

thanks!

Hi, I'm looking for help with an issue with FreeTDS 1.4.2 on FreeBSD 13.2

Better describe your issue straight away. This will save a round of questions from someone who may or may not be able to answer

.

V_PauAmma_V: ok!

the server has a a conf file: /usr/local/etc/freetds/freetds.conf where among other params, one can set the version of the target database, for example: tds version = 7.0

The program freetds includes a utility called tsql that allows connecting to MS SQL Server. For one server I defined "tds version = 7.0" and it connects without issues.

But, from one program (written in FreePascal) the param "tds version" is ignored, in the freetds log, when connecting from my program I see "Connecting with protocol version 7.3".

BUT, if I set the environment variable TDSVER=7.0 before running my program it connects without issues.

Now, I need to understand why the config para tds version isn't read when I connect using my program.

hi all

martinrame: perhaps your program is expecting the config to be in /etc/ and not /usr/local/etc/

32

I'm working on setting up ethersrv to share files to an old 486 over ethernet. Is there a way to create/mount a FAT partition that's just a big ol' file? like a hdd image?

etherdfs says it works best with fat filesystem but i really don't want to repartition.

you can create a memdisk out of an img file

e.g., `truncate -s 256M fat.img; mdconfig fat.img; newfs_msdsofs ... /dev/mdN` then mount /dev/mdN

makefs(1) also supports FAT

sweet

and then the new filesystem that gets mounted I can browse/modify like any other mount?

i'm gonna be playing so many DOS games on my 486.

yes

tyvm

Enjoy DPMS.

What was that mode called that was kind of a hybrid mode between real mode and protected mode on 386?lll

hell if i know... i feel both fortunate and unfortunate that i never got into systems programming back in those days.

Fetching packagesite.pkg: 28% 2 MiB 8.2kB/s hmm... is there something i can do to speed this up? choose a different mirror?

did cgit.freebsd.org/src/commit/libexec…af45488bef649e9a84890e2414ff80b3a00 not get MFC'd to stable/14 or can i just not find the commit?

Title: src - FreeBSD source tree

Looking at the tree, it wasn't MFC'd either to 14 or to 13.

_xor: unreal mode

V_PauAmma_V: do you think it's worth asking the committer about it? the current url doesn't work anymore and the last available version of the leapfile expires in 13 days

I would, if I were you. MFCs sometimes fall through the cracks and this may be the case here.

nice that make installkernel doesn't break ps(1) anymore, apparently it uses the kern.proc sysctl instead. wonder when that changed

Hi i am trying to install freebsd and i got an error saying no wifi networks were found when it did a scan

i seem to remember the last time i installed freebsd on this laptop i got the same error but i forget how i fixed it

is it lacking the driver?

it did see my wifi adapter

so im using its using a driver?

chipset?

let me look it up

odd i’m not able to find it i have a 10870h if that helps