pretty sure bectl promotes snapshots to clones when creating BEs, so they are independent (and freebsd-update obviously uses bectl to create the BE on updates)

kernel: MCA: CPU 0 COR (1) MS channel ?? memory error

hm. well that can't be good.

well it happened yesterday and on feb 2 .. i'll just write it off :)

Macer: happens if i run something like poudriere bulk -J16 on an i7 ;)

or overclock the cpu then run it with normal heavy load

there's a port IIRC that helps deal with MCA errors

yeah i was doing a scrub on a pool which kind of makes the thing go into hot mode

it says bank 8 but i'm not sure if that actually means an 8th bank dimm

conf thought all BSDs used beastie

I corrected them, and then they spelt my username wrong

well, it is a generic BSD mascot, its just that openbsd additionally has its own mascot (do dragonfly and net have their own mascots? I couldn't find any with a quick search, but maybe it was just too quick)

Macer: look for other things like cooling. also check yr bios settings it shouldnt do that in normal high load where theres no overclock. a scrub shouldnt do it

ive only had MCA errors where the config has deliberately been pushing the envelope

well. this is in my ancient dual xeon 5670 that i "modified" so it didn't sound like a jet taking off nonstop .. so maybe with 28 or so disks in it and using noctua fans in the front instead of the molex ones is causing too much heat

i changed out the original fans and put hsfs on the cpus

nice! ^^^

my proliant stack is in a wooden rack ;) yay ikea

what does yr temp monitoring say

*sigh* pf keeps exploding…

how do i make sure that things like wireguard get started properly before pf is started? or even better: how do i keep pf from just exploding if a specified interface doesn't exist? this keeps breaking setups for me to the point where pf is the single most common point of failure for me…

also the firewalling part doesn't seem to really work anymore? i have "block log all" as the basis of my pf.conf and the prometheus instance on the host is just publicly reachable ¬_¬

can a failing backup/receive disk lead to the source/sending disk getting marked as having errors?

my backup disk died. replaced it yesterday, ran a backup, and scrubbed both pools. now zroot's 21 errors are gone, but i didn't clear them.

okay, something is fucky here… stopped prometheus and it's *still* reachable o_O

ah, pf in my local gateway forwards *:9090 to the local monitoring jail, so firewalling *does* still work, just didn't look like it for me.

Hi, I'm bothering you with a question because I've been searching for hours and haven't found anything specific.

I have an "Intel HD Graphics 3000 (Sandy Bridge GT2)" graphics card that loads the "i915kms" module correctly, but it has very limited support in Wayland because it only allows up to a resolution of 1024x768@60Hz. From what I've found, it's possible to force the EDID so that Wayland, through KMS, can see the resolutions supported by the graphics card and monitor. The resolution I need is 1366x768@60Hz to be able to use Wayland.

Using the "EDID override" option, I was able to create a custom binary EDID file with the resolution I need.

What I need to confirm is whether the configuration in the "loader.conf" file that allows the kernel to load the binary EDID file is as follows:

drm.edid_firmware=LVDS-1:/boot/edid/1366x768_60Hz.bin

?

GeoBarRod, I am just going to mention that I read your question so you know it is visible but I don't use Wayland and so have nothing of value to add other than moral support and wishing you good hunting.

GeoBarRod: Where did you get that sysctl from? I'm not saying it's wrong, only that I want to read the documentation that you rèad so I can understand what is happening.

@rwp: THX

hello guys, can anyone give me a hint for pouriere, how to compine databases/p5-DBD-mysql - with mysql80 flavor ?

LXGHTNXNG: I checked now with sysctl and the variable does not exist. What I wanted to know is whether it is possible to force the EDID as it is possible in GNU/Linux.