2025-11-01 01:27:00 <WhyNotHugo> When my host runs out of memory, it just freezes for two minutes until a watchdog timer does a hardware reset. Shouldn't the OOM kill processes and free memory or something?
2025-11-01 01:27:48 <ikke> the kernel oom manager notoriously takes a long time to actually do something
2025-11-01 01:28:04 <ikke> there's earlyoom userspace to have it act quicker
2025-11-01 01:28:42 <ikke> https://github.com/rfjakob/earlyoom
2025-11-01 01:29:52 <WhyNotHugo> Weird, I recall many years ago just seeing random processes get killed by the OOM.
2025-11-01 01:29:58 <WhyNotHugo> (bad, but better than hard reboot)
2025-11-01 01:31:00 <ikke> It all depends on the circumstances
2025-11-01 01:31:22 <WhyNotHugo> My swap it 64GB tho, and 0bytes used.
2025-11-01 01:36:44 <ikke> If the watchdog triggers a restart, doesn't that imply that the kernel is not even responding?
2025-11-01 01:38:42 <longnoserob[m]> WhyNotHugowhat are you doing to maxout ram?
2025-11-01 02:47:09 <rnkn> rendering video eats up a lot of memory
2025-11-01 04:08:20 <rnkn> I've been trying to use Pitivi (video editor) and I am pretty sure it has a memory leak when rendering proxy material, which has lead to a few full system lock-ups. would earlyoom be a good candidate here?
2025-11-01 04:25:04 <longnoserob[m]> maybe sstting an ulimit manually works as well?
2025-11-01 05:20:45 <iggy> 1 chrome tab
2025-11-01 05:31:30 <frag> what web browser solution do you use for alpine? i tried chromium and was suprised how fast it was, bit worried about privacy tho
2025-11-01 05:32:29 <longnoserob[m]> mostly ff with mobile-config
2025-11-01 05:32:39 <frag> also the firefox/librewolf browsers dont seem to open urls in the running instance if you give multiple urls like: firefox <url> <url>?
2025-11-01 05:33:06 <frag> longnoserob havent heard of mobile-config, what is it?
2025-11-01 05:35:27 <frag> maybe this is a good solution https://codeberg.org/zz/hardenChromium/src/branch/main/launchChrom.sh
2025-11-01 09:02:07 <rnkn> Firefox thru flatpak (to get DRM)
2025-11-01 09:29:57 <rnkn> currently reinstalling alpine with FDE, wish me luck!
2025-11-01 09:31:32 <frag> good luck!
2025-11-01 09:56:08 <rnkn> frag: thanks
2025-11-01 09:57:15 <rnkn> okay the list of things I have forgotten: how to remap caps to control on the console; how to have gnome login manager automatically start at boot?
2025-11-01 10:07:16 <rnkn> `# service gdm start` just gives me an empty screen with an underscore at the top left
2025-11-01 10:16:25 <burnix> hi community
2025-11-01 10:16:37 <burnix> is there someone here ?????
2025-11-01 10:17:25 <frag> seem to be a bit quiet atm
2025-11-01 10:19:00 <burnix> sleeping ????
2025-11-01 10:19:12 <quinq> Let-me call everybody to see what they're actually doing right now
2025-11-01 10:20:01 <burnix> i have a question on how i can make boot partition smaller on sdcard on raspberrypi
2025-11-01 10:20:26 <burnix> i'm searching, but i'm not a linux king
2025-11-01 10:22:35 <frag> why smaller? how smal do you need it? dunno if you have apk cache you can clean out there..
2025-11-01 10:23:43 <burnix> i want to have a maximum of space on my sdcard (1go sdcard)
2025-11-01 10:24:18 <rnkn> oh I needed to run setup-wayland-base
2025-11-01 10:24:21 <burnix> setup installer create a boot partition that is huge (300mb) with nothing on it (22mb)
2025-11-01 10:25:03 <burnix> so i want to get free space for the "system" partition
2025-11-01 10:26:27 <frag> oh, mine is 2G dunno how much smaller you can make it, sdcards are cheap tho
2025-11-01 10:29:58 <burnix> yes but i have a big stcok of 1gb sdcard (approx 100)
2025-11-01 10:32:42 <frag> oh, maybe someone will wake up and answer later :]
2025-11-01 10:33:08 <burnix> okay. so i'm going to search on my side
2025-11-01 10:33:15 <burnix> goodbye!!
2025-11-01 10:35:03 <frag> bye!
2025-11-01 11:01:58 <rnkn> I'm back baby! now with new and improved FDE!
2025-11-01 11:11:05 <ssm> rnkn: now that I know that it's decrypted (I'm a fed) I'm breaking into your house and stealing your device while it's online
2025-11-01 11:19:38 <rnkn> ssm: you'll need a UPS for that
2025-11-01 11:31:14 <ssm> now I know it's a desktop muhahahaha
2025-11-01 12:31:42 <rnkn> always one step ahead
2025-11-01 13:29:31 <WhyNotHugo> longnoserob[m]: I'm testing out translation models. They use a lot of ram and need access to all of it, so can't easily swap.
2025-11-01 13:29:53 <WhyNotHugo> It doesn't seem like the issue is RAM tho, `free -h` indicates that there is still 26GB free.
2025-11-01 13:30:06 <WhyNotHugo> But the system freezes and reboots (i presume due to watchdog) 2 minutes later.
2025-11-01 13:30:28 <longnoserob[m]> time to run some memtest I would say
2025-11-01 13:46:10 <p_6f3Ik7Suw> if/when burnix comes back recommend them https://github.com/raspi-alpine/builder - and to rebuild the image themselves with their own config
2025-11-01 15:32:03 <invoked> krisk: https://github.com/OpenVPN/openvpn/blob/v2.7_rc1/Changes.rst#new-features food for thought
2025-11-01 20:01:08 <nzjsjsnoqsjaah188162[m]> Hey there! I wanted to let you know that I have a channel where I share some amazing Verified sauce and soft cashout methods. I also provide a Full WalkThrough to make things easier for you.... (full message at <https://matrix.org/oftc/media/v1/media/download/ATAzsGxmGymw9zBE5GKyc0zoCLQS-c20JvSfDu91HhAZuhQ0mHysyDrzh57LSPonxrCLSMDbILmkxUyUgB0xvspCeaiRtyCgAG1hdHJpeC5vcmcvTk55cFJIcVdsZWZTamNJb3NDSnhJckpv>)
2025-11-02 14:11:16 <Random3> Morning.. APKBUILD question, applying a patch to Unbound.. 
2025-11-02 14:12:25 <Random3> I've added the aaaa-filter-iterator.patch to the APKBUILD and updated the checksum.. it *looks* like the patch has been applied.. and the package builds.. but when I run unbound with the aaaa-filter option set.. it says unrecognized.. 
2025-11-02 14:14:09 <longnoserob[m]> did you install the new package after builing it?
2025-11-02 14:14:13 <Random3> running abuild -r -K (leaving things behind..) again it looks like the iterator module has been patched.. in the src dir.. 
2025-11-02 14:14:39 <Random3> yes.. and build date/time on unbound (/usr/sbin/unbound) is the same as the pkg dir.. 
2025-11-02 14:15:21 <longnoserob[m]> ok, maybe the patch requires additional dependencies?
2025-11-02 14:15:36 <Random3> [I] root@alpine-800g3 /h/b/a/m/u/p/u/u/sbin# ls -al unbound -rwxr-xr-x    1 builder  builder    1178544 Nov  2 09:03 unbound*
2025-11-02 14:15:58 <Random3> (not arguing..) but it compiled.. 
2025-11-02 14:16:33 <Random3> (and.. this is 1.24.1) and the patch was updated for 1.24.0 b/c it didn't apply clean) 
2025-11-02 14:16:34 <longnoserob[m]> maybe open a new shell and run it from there?
2025-11-02 14:23:20 <Random3> https://github.com/NLnetLabs/unbound/issues/312
2025-11-02 14:23:42 <Random3> whatever is in the patch for how to use it.. seems to be wrong :/ 
2025-11-02 14:24:03 <Random3> drill -p 5353 -t aaaa www.google.com @192.168.88.252 -Q 2607:f8b0:4006:820::2004
2025-11-02 14:24:14 <Random3> drill -p 53531 -t aaaa www.google.com @192.168.88.252 -Q
2025-11-02 14:24:17 <Random3> (nothing) 
2025-11-02 14:24:26 <Random3> 53531 is the patched.. 
2025-11-02 14:24:52 <Random3> [1762093446] unbound[9117:0] debug: sanitize: removing public name with private address <www.google.com.> 2607:f8b0:4006:803::2004#53
2025-11-02 14:28:30 <Random3> yes.. that seems to do it.. :P 
2025-11-02 14:30:13 <longnoserob[m]> good
2025-11-02 14:39:01 <Random3> .. stock unbound w/o the patch.. that config line seems to do the same *sigh* 
2025-11-03 01:47:36 <ahmadraniri[m]> Hello, iucode_tool failed to be built due to lack of limits.h file (built using gcc 15), upstream gentoo has a patch to fix that issue. Patch link : https://gitweb.gentoo.org/repo/gentoo.git/plain/sys-apps/iucode_tool/files/iucode_tool-2.3.1-limits-include.patch 
2025-11-03 07:01:50 <krisk> invoked: noted :) that looks like going out to strongswan project first to have similar flexibility... will take a look at some point
2025-11-03 10:19:39 <longnoserob[m]> fossdd: do you know any plans to rollout gdm 49.1?
2025-11-03 12:12:53 <achill> nope, its completely broken without systemd
2025-11-03 12:13:09 <achill> don't think I have to even look into it
2025-11-03 12:25:42 <longnoserob[m]> oh ok
2025-11-03 12:26:22 <rnkn> bad omen?
2025-11-03 12:27:51 <longnoserob[m]> no, did not know about the brokeness w/o systemd
2025-11-03 12:35:25 <rnkn> this doesn’t sound good
2025-11-03 12:35:26 <rnkn> https://blogs.gnome.org/adrianvovk/2025/06/10/gnome-systemd-dependencies/
2025-11-03 12:47:32 <invoked> "GDM will work with them for now" "Apologies for the short notice" (read between the lines)
2025-11-03 14:48:45 <WhyNotHugo> It'a amazing how hostile the GNOME devs are to people using their DE in different platform or in different ways, and folks just insist patching it up and ignoring its hostility.
2025-11-03 14:49:14 <WhyNotHugo> You'd think that by now people would actively avoid GTK and GNOME, but it's quite the opposite.
2025-11-03 14:54:01 <Sheila> still upsetting to me personally that they don't want to put in the work to make the high-contrast theme usable for everybody who can't ue the default theme. (I filed a bug a decade ago; closed for timeliness, ISTR; I asked back-channel recently, and was told "patches welcome".)
2025-11-03 14:54:53 <Sheila> idk. I'm just going to use KDE instead, at least they explicitly support my not developing a migraine trying to use my computer visually.
2025-11-03 16:39:13 <SyntheticBird[m]> rnkn: lol you are late
2025-11-03 16:39:16 <SyntheticBird[m]> It has rolled out already and Artix dropped support
2025-11-03 16:39:31 <SyntheticBird[m]> Some user made some patch for Gentoo, so openrc support has a PoC
2025-11-04 07:52:00 <rnkn> what's the likelihood that Alpine will continue to support Gnome?
2025-11-04 07:52:08 <rnkn> I want to prepare for heartbreak
2025-11-04 08:15:21 <grawity> honestly I think the likelihood is higher that Alpine will have systemd
2025-11-04 08:16:10 <frag> is gnome that popular? :S
2025-11-04 08:16:15 <rhizoome> I haven’t read the recent drama, but I stopped using anything from ecosystems that do strong coupling between their components. some ecosystems do it out of ignorance others maliciously to dominate. 
2025-11-04 08:22:16 <rhizoome> It is amazing how far get with things like mdevd and a little glue-code.
2025-11-04 08:24:19 <frag> rhizoome tell us more
2025-11-04 08:28:59 <rhizoome> Well I have a little daemon that mounts disks, keeps a list of the mounts and lets you select a disk via rofi, which then gets unmounted. for example.. the daemon does more.
2025-11-04 08:29:54 <rhizoome> mdevd calls a little C program that informs the daemon about new disks via a unix socket
2025-11-04 08:31:01 <rhizoome> It does also sleep and inhibition of sleep depending on what program currently are running.
2025-11-04 08:32:00 <frag> neat
2025-11-04 08:34:59 <rhizoome> Maybe except from the socket stuff it is really simple. but even the socket code, is just a slightly modified version of standard example you get if you search unix sockets in C.
2025-11-04 09:53:53 <rnkn> for me one of the main attraction to Alpine was that it doesn't have systemd
2025-11-04 10:03:46 <ahmadraniri[m]> <ahmadraniri[m]> "Hello, iucode_tool failed to..." <- Done : https://gitlab.alpinelinux.org/alpine/aports/-/issues/17677
2025-11-04 11:22:21 <toric> rnkn: Funny, for me thats the one thing im not a fan of in alpine, a lack of systemd.
2025-11-04 11:27:28 <ayex[m]> for me the attractors are: no systemd, being small snappy and secure due to decreased musl attack surface.
2025-11-04 11:29:44 <toric> Ah, for me its being relatively small without having massively out of date packages (looking at debian), and the declarative package management.
2025-11-04 11:30:06 <toric> But if I really want small, I routinely build smaller images with systemd using yocto.
2025-11-04 11:34:30 <toric> Oh, and being availible for ARM, or id probably just be using arch like I do on x86.
2025-11-04 11:41:27 <rnkn> toric: rare to hear from someone who *likes* systemd, as opposed to just accepting it. what do you like about it?
2025-11-04 11:46:09 <toric> rnkn: Im an embedded engineer, you could say I build distros. And im a sysadmin as a hobby. Systemd is really, really powerful in how it lets you describe dependencies between things, setup complex network interfaces declaratively, isolate services from one another (you basically dont need docker), and more stuff that ive probably forgot makes it *really* powerful. Like yah, is it more featurefull than what came before? Yup. But its not like the k
2025-11-04 11:46:09 <toric> ernel 'follows the unix philosophy' either...
2025-11-04 11:46:44 <toric> And the documentation in the man pages is genuintely quite good.
2025-11-04 11:50:14 <rnkn> toric: while that sounds powerful, it also doesn't sound like stuff a regular user needs, especially not on PID 1
2025-11-04 11:52:27 <toric> rnkn: Depends on what you think by needs, I think. Expect a regular user to modify? Maybye not. Have a regular user rely on becase a distro maintainer is a able to use all of it? Absolutely. Same reason most distro kernels include drivers that 90% of users wont need, you dont want to make the user compile their own kernel because they want to use a joystick for a flightsim.
2025-11-04 11:52:48 <Imbus64> Got the same experience tbh, systemd brings a lot to the table, but the cli is atrocious
2025-11-04 11:53:36 <toric> Imbus64: CLI is the worst part of systemd, sure, and even then its nothing compared to expecting a user to modify bash scripts.
2025-11-04 11:53:59 <Imbus64> True, i absolutely agree on the shell part of traditional init
2025-11-04 11:54:37 <Imbus64> Shell is simply not a good language, and the declarative part of systemd is my favourite part
2025-11-04 11:54:53 <Imbus64> Dinit also does it that way, without the 20 million lines of systemd
2025-11-04 11:55:16 <toric> Id happily try a systemd replacement as long as it was declarative. (Im actually a NixOs stan for my servers)
2025-11-04 11:55:17 <Imbus64> Then again, dinit does not have the same scope as sysd
2025-11-04 11:55:44 <Imbus64> Im on debian on all my servers because debian, im always a fan of debian
2025-11-04 11:56:13 <Imbus64> I tried chimera, which lifts APK from alpine, but introduces dinit and the core userland from freebsd
2025-11-04 11:56:25 <Imbus64> I must say i find it quite tasteful
2025-11-04 11:57:27 <toric> Imbus64: Actually the reason Im on #alpine-linux RN is because I got frustrated with limited package availability with debian on an Rpi... Sorry.
2025-11-04 11:57:34 <rnkn> shell script may not be a fancy language, but it's simple, and allows users to have agency over their computer
2025-11-04 11:57:48 <rnkn> systemd obfuscates this in a "we know better" approach
2025-11-04 11:57:51 <Imbus64> Oh i use alpine on my rpi's, forgot that. Simply nothing compares imo
2025-11-04 11:57:59 <toric> Ill have to try out chimera, that sounds nice!
2025-11-04 11:58:36 <Imbus64> Yeah chimera is a breath of fresh air for sure, it feels a bit like alpine because of apk
2025-11-04 11:58:52 <toric> rnkn: Id rather lua than shell script. *almost* anything other than shell. (ok, maybye not yaml-turned-into-a-fully-turing-complete-lang, thats even worse)
2025-11-04 11:59:00 <Imbus64> It also revamped aports into 'cports', which uses python as an in-tree build system instead of cports
2025-11-04 11:59:10 <toric> And lua is even smaller than bash.
2025-11-04 11:59:26 <rnkn> shell script != bash
2025-11-04 11:59:28 <Imbus64> Im not all too familiar with neither aports nor cports so cant comment on the technical merit, but both work fine.
2025-11-04 12:00:06 <toric> rnkn: POSIX shell is even worse.
2025-11-04 12:00:12 <Imbus64> turic got taste
2025-11-04 12:00:19 <Imbus64> i agree on the lua comment
2025-11-04 12:00:26 <rnkn> toric: POSIX shell is fine, I write all my script with /bin/sh
2025-11-04 12:00:41 <rnkn> your init scripts shouldn't be fancy cool languages
2025-11-04 12:01:03 <Imbus64> my problem with shell is how unreadable it is for anyone that isnt a shell expert
2025-11-04 12:01:38 <rnkn> the same is true of any language though?
2025-11-04 12:01:39 <Imbus64> And i absolutely agree, no fancy cool flavor of the month languages for init scripts, or scripts in general
2025-11-04 12:01:49 <rnkn> it's going to be unreadable if you can't read it?
2025-11-04 12:02:11 <Imbus64> Lua is pretty idiot proof, and its also fairly old and boring, which is an important feature imo
2025-11-04 12:02:28 <rnkn> the point is not which language one likes best, but which language will be most democratising for the human beings on the other end
2025-11-04 12:02:38 <toric> and if you really dont like it, there are a good few languages that compile to lua. Even a lisp!
2025-11-04 12:02:52 <toric> rnkn: and posix shell is absolutely the wrong tool for taht.
2025-11-04 12:02:57 <Imbus64> Im not a fan of transpiling stuff, its smells like duct tape
2025-11-04 12:03:07 <rnkn> there is *way* more resources on the www for sh than lua
2025-11-04 12:03:10 <rnkn> and I like lua
2025-11-04 12:03:37 <rnkn> but I know not to say someone else needs to like/learn lua to control their system
2025-11-04 12:03:56 <Imbus64> youre assuming they already know shell
2025-11-04 12:04:11 <rnkn> if they're typing at a console they're typing in shell
2025-11-04 12:04:43 <toric> Other shells exist than SH.
2025-11-04 12:04:48 <Imbus64> Yeah, well thats true in a sense, but i feel like there is an acute difference of using a shell interactively versus writing a script
2025-11-04 12:05:07 <rnkn> but the point is that you learn shell by simply using your computer
2025-11-04 12:05:19 <rnkn> there is no barrier to entry
2025-11-04 12:05:30 <Imbus64> yeah thats a fair point
2025-11-04 12:06:37 <Imbus64> Ive seen so many instances of replacing tried and true tech with new hotness, with no regards for longevity or quality
2025-11-04 12:07:06 <Imbus64> I also think that a description of a service should not be turing complete
2025-11-04 12:07:43 <Imbus64> In its nature it is a declaration of a very simple thing, and systemd gets this just right imo
2025-11-04 12:08:56 <toric> Imbus64: Imma push back *slightly* on the 'new hotness', as replacing things is the only way to actually make any progress improving things in the long term. C was a 'new hotness' at one point.
2025-11-04 12:09:32 <rnkn> if systemd was so good it would be optional, and users would choose it on its own merits
2025-11-04 12:09:58 <rnkn> it would also be portable
2025-11-04 12:09:59 <Imbus64> Dinit does as well, its C++ though, about 20k sloc. Its kinda ok, wonky formatted, but it passes the smell-test imo
2025-11-04 12:10:08 <Imbus64> @rnkn hard agree
2025-11-04 12:10:18 <Imbus64> portability matters
2025-11-04 12:36:56 <Imbus64> I have an issue with ddclient on alpine, it complains about ownership. 
2025-11-04 12:37:15 <Imbus64> https://paste.centos.org/view/edee5547
2025-11-04 12:37:20 <Imbus64> Any idea where i can report this?
2025-11-04 12:38:44 <Imbus64> nvm i rtfm'd, must have missed it
2025-11-04 12:48:42 <Imbus64> https://gitlab.alpinelinux.org/alpine/aports/-/issues/17680
2025-11-04 14:29:12 <bsmntoid> hello! i'm having an issue with a fresh install of alpine linux where ifup fails during boot. first ip fails (ioctl 0x8913 failed: No such device), then udhcpc fails (ioctl 0x8933 failed: No such device), then ifup fails (failed to change interface wlan0 state to 'up'). the weird part is udhcpc and ifup both work fine after logging in
2025-11-04 15:16:20 <prabu> the issue might be due to race condition between services. check here https://wiki.alpinelinux.org/wiki/Iwd#%22ip:_ioctl_0x8913_failed:_No_such_device%22
2025-11-04 15:41:16 <bsmntoid> that definitely sounds right, although i'm using wpa_supplicant instead of iwd. shuffling around the dependencies so networking starts before or after wpa_supplicant doesn't seem to have any effect and (expectedly) removing networking from the boot level as the iwd article suggests just results in no attempt being made
2025-11-04 15:57:52 <prabu> since i only use iwd, i've documented it that wiki page. once you find a proper solution, i'll appreciate if you can drop a note either here on wiki.
2025-11-04 16:00:56 <prabu> just fyi. https://tpaste.us/xDwW 
2025-11-04 16:20:11 <invoked> i use wpa_supplicant in runlevel boot, networking in default, and that's with wireguard also. anecdotes being what they are worth
2025-11-04 16:20:59 <invoked> iwd is fine, it's just a little antithetical to what i want from my environment
2025-11-04 17:13:18 <bsmntoid> no luck unfortunately, error happens with networking in default too. i'll see if iwd has any sort of similar issue because if it doesn't i might just switch and be done with it
2025-11-04 17:25:43 <invoked> bsmntoid: you could try adding a sleep in /etc/interfaces for wlan0 (eg, "pre-up sleep 5") to buy more time for iwd to settle. see interfaces(5)
2025-11-04 17:26:36 <invoked> /etc/network/interfaces, that is
2025-11-04 17:55:05 <bsmntoid> ok, iwd works perfectly out of the box. i also tried what you suggested with /etc/network/interfaces but using wpa_supplicant and that works too! it takes quite a while but it's no longer than iwd, it's just in the foreground. the guide on the wiki for starting services after the login prompt worked, although it still prints to the screen
2025-11-04 17:59:20 <invoked> excellent
2025-11-05 15:10:25 <agent0range> quit
2025-11-05 15:11:22 <invoked> quit failed successfully
2025-11-05 15:11:45 <Imbus> peak irc