where do ppl point XDG_RUNTIME_DIR to?

like ~/.run or ~/.local/run or?

I thought it goes into /tmp/ and then $USER or soemthing

handbook has /var/run/user/uid

seems weird to put user dirs outside their home tho

another freebsd blog i read used /tmp/uid-runtime-dir

The /tmp isn't weird

V_PauAmma_V: What is the default?

default from where? not base

I am also curious about that

I just checked, and (using Xfce) it's unset by default.

It is usually set by Display Managers

Hmm...

On my linux thing it is

/run/user/{UID}

Uh...

which explains i guess why freebsd handbook says /var/run/user/uid

Change {UID} to the UID, usually 1000 is the UID of the first user made other than root

user is just user, not user login

Anyone know which flavor of regex expr uses/implements? Trying to loop over an fstab file for a jail and create the mountpoint directories before starting the jail, but for some reason I'm having trouble matching the first line against \/.+\s+\/.+ to verify it before cutting the 2nd column.

should be BREs

...or alternatively, if there's a way to tell mount to create the target directory if it doesn't already exist, which would be ideal :D

Hmm

I have a patch that adds an -E option to use EREs, but I haven't looked at it in probably five years

Pulls in a lib, I assume?

nah, just non-standard

try something like \/.{1,}[[:space:]]{1,}\/.{1,}

Ah

ugly, unwieldy, but it might work

alternatively, awk might be nicer

Yeah, I might just go with awk.

Hmm, apparently it didn't like that expression either.

expr '\\/.{1,}[[:space:]]{1,}\\/.{1,}' : '/mnt/services/data /cluster/nomad/jails/EXAMPLE/data/jail/mnt/data nullfs rw,noatime 0 0'

That should match, no?

oh wait, remove the escape slashes since I'm single quoting it?

Nevermind, removing them didn't help either heh.

... You have the string and regex reversed.

Quoting from expr(1): The “:” operator matches expr1 against expr2, which must be a basic regular expression.

also, I forgot that bounds have to be escaped

'\/.\{1,\}[[:space:]]\{1,\}\/.\{1,\}'

welp, we're on -p3 for 14.0-RELEASE already...

oh

dch: and that's why more people need to test RCs

ope

found this: github.com/dylanaraps/pure-sh-bible , maybe some one find this useful

Title: GitHub - dylanaraps/pure-sh-bible: 📖 A collection of pure POSIX sh alternatives to external processes.

update on alacritty not running on a vm, over rdp, after a freebsd-update and pkg upgrade a few days ago. tried to ssh into vm and run alacritty on the rdp session: DISPLAY=:10 (my xrdp x11 display) alacritty says mesa: error: zink: failed to load libvulkan.so.1 glx failed to create drisw screen failed to load driver zink

can anyone help me figure that out or fix it or smth?

adding LIBGL_ALWAYS_SOFTWARE=1 and it runs btw, so definitely a bug in the driver

 llvmpipe is better than zink too i guess, author of alacritty says

the reason vulkan got loaded is because of zink, since it's a OpenGL to Vulkan translation layer.

13.2

now on -p3 woo

uh-oh. No updates for pkgbase yet :O

Ugh. I have a bhyve process that's taking all my memory. Can't seem to kill it. Is there a way I can hit it harder than `kill -9`? My `ps auxw` output shows this:

root 3209 0.0 87.9 127644812 117926436 v0- TNC Tue03 297:54.00 bhyve: haos (bhyve)

ghoti: first up, both of those things seem like bugs and should be reported

if kill -9 doesn't work, then i suggest trying a reboot, after getting the info you need to report a bug

ghoti, how did you run the "bhyve process" ?

if you use vm-bhyve package to manage virtual machine run "vm powerof <vmname>"

"vm poweroff <vmname>"

or stock bhyvectl: bhyvectl --vm=vmname --destroy

or bhyvectl --vm=vmname --force-poweroff

nerozero: for that one, I'm using vm-bhyve to launch; it has a pretty basic config, here: sprunge.us/LMP8gU

is it amd machine ?

meena: this is in 13.2, which I recently upgraded to. All was fine in 12.4.

nerozero: nope, CPU: Intel(R) Xeon(R) CPU E5-2670 0 @ 2.60GHz (2600.11-MHz K8-class CPU)

yah ...

?

vm-bhyve is being run on the host, right?

Yes.

My issue with bhyve is that windows guest machine usint 200% cpu core doesn't matter is guest doing something or completely idle

I haven't run Windows in years.. This was a HAOS appliance, Linux..

And this wasn't CPU, I can afford that; it was failure to release memory.

From top: Mem: 27M Active, 776K Inact, 111G Laundry, 9417M Wired, 1397M Free

same but office need one accounting software which is run only on win 7

The machine has 128GB RAM. Everything plus swap got used.

never had ram issues with bhyve

in case of zfs your zfs will try to grab as many ram as it possible

this is normal

could you please share the "top" head ?

what is in the ARC field ?

Hmm, I was running an older patch level of 13.2. One sec while I upgrade and test again..

nerozero: here's the top of `top` when I'm not running that bhyve VM: sprunge.us/JYyNAM

Sigh.. Seems to bebehaving after the upgrade. Ain't that always the way? Anyway, thanks for your attention...

ghoti, sorry, working in parallel ..

try this link: forums.freebsd.org/threads/bhyve-100-cpu-usage.58398

Title: bhyve 100% CPU usage | The FreeBSD Forums

it may have some useful information

nerozero: indeed, there are some useful diagnostic strategies there. But still, my problem was not CPU usage, it was laundered memory not being released. And I haven't been able to replicate it since upgrading to 13.2-p8.

ghoti, could you please create a new thread in freebsd forum and post the link here?

I'm also a active bhyve user so ...

also you can find a #bhyve channel here on libera, but its kinda dead, never respond anyone there ...

nerozero: yes, I gave up on #bhyve a while ago. Tell you what, I'll create a thread if I can replicate the problem. If I can't, I think I just chalk it up to randomness that went away with the update.

ghoti, good luck!

Or should I make a thread basically saying "this happened; what do I do if it happens again?"

IMHO, if problem still exists, it is good to share the issue and steps taken... If someone else will experience same issue - the thread will grow.

In almost 2 decades of FreeBSD, I've never seen this before, which makes me think it can't be too common. :)

I'm with BSD from bsd 3.0

Dang, I'm counting wrong. I started with 2.2 I think. So a quarter century I guess.

:) remember deco commander? on Hercules monochrome screen? :)

squid and mgetty, IDSDL for USRobotics Courier modems ... yah...

720 × 348 was my favourite resolution. :) I didn't use deco though. I came from 4DOS/4OS2 and SCO Xenix. Different lifestyle...

yeh ....

what a time...

386 with 16MB or ram ...

40MB hard drive

that was a server :D

custom kernel compilation took ~1 day

actually I do remember 2.2 BSD

the longest supported bsd was 4.22 if I recall

Chronologically, the 2.2 branch had the longest stretch of releases, with 2.2.9.

freebsd 4.x had a long production lifetime because of the whole 5.x issue (5.x was really a development-only branch for making SMP work)

yes. fbsd has long had a reputation for putting experimental/broken stuff into odd-numbered releases. Seemed to go away with fbsd 9.

I think I skipped over either 7 or 8 or both, I don't recall

I think the idea was that the odd numbers were "current" rather than "stable", but I wasn't paying so much attention to development back then

Is this 404 for you too? mail-archive.freebsd.org/mail/archive/2022

It 404 for me

For me too

Any list admins here to look at that?

mail-archive.freebsd.org/mail/archive/2021 does work, so likely wrong owner/rights of the directory

Title: Index of /mail/archive/2021/

Yep, most years work.

Wouldn't wrong owner/rights be a 403 rather than a 404?

V_PauAmma_V: oh! You are right! Sorry.

michaeldexter, if no list admin answers, email postmaster@? I think that's also the web archive server admin group.

Mystery maybe solved: lists.freebsd.org/pipermail/freebsd-current

Title: The freebsd-current Archives

Certain archives are there.

But not 2022. Hmm

I have mailed postmaster@ asking how that is intended to be accessed.

at least the directory 2022/ with a file named “no archive for 2022” so people do not wonder what happened

the lizards did it

maybe it was advanced lizards…

like… birds.

Those: almanach.worldofgothic.de/index.php/Echsenmenschen

Title: Echsenmenschen – Gothic Almanach

I think technically they (birds) are not lizards, even in a broad sense, since "lizard" refers to a different branch of the reptiles

I'll believe the advanced lizard alchemists theory, but there not that birds exist, no way, did my own research on that one.

yeah, I guess that's like saying a shark is a fish, which is true in the most abstract sense

uh

a shark definitely is a fish?

it's just not a bony fish

Hi.

RhodiumToad, Re leap-seconds.list, Yes, Is there a plan that I might know about how FreeBSD handles such things? (I know I can manually update it.) I have so many questions. Is the plan to update from 13 to 14? Is it updated in 14? To install the port of ntpd, ntpsec, or openntp?

Just finished installing 14.0 at work. This will be our clang machine (why? because I love it).

(our Linux devs are very excited :))

Hey... not sure where to look... everytime I open a term, I got this: bash: SRCDIR: unbound variable

I am not sourcing anything in bashrc nor .profile....

not sure where to start looking 8-) any clues?

cybercrypto, Sounds like a shell with -u set which will error when any unset variable is referenced.

But I can't create a reproducer for that exact message.

Ah, yes, try this: bash -uc 'echo $SRCDIR'

Regardless, I am NOT an advocate for the use of -u and I recommend not using it. Along with -e. Both of those are not helpful. Not like people wish them to be helpful.

cybercrypto, If it is every time you open a terminal then it must be in the environment file somewhere. If not the default ~/.bashrc then $ENV. But put a "set -x" at the top of the .bashrc file for a debug run and that should trace through it.

You don't have to open a new terminal. Just stack another shell. Invoke "bash" again in the terminal.

join #c

whoops...

i'm reading about ipsec but the example the handbook contains is showing how to connect two lans over ipsec

i'd like to see how to make kinda client - server solution if possible