2022-09-01 05:57:13 GreyDev: dmvpn, which we use extensively within Alpine, uses strongswan 2022-09-01 07:59:35 Hi. What is the recommended way to change the /bin/sh link from busybox to bash? 2022-09-01 08:01:08 Remove the link and link to bash 2022-09-01 08:01:08 don't? 2022-09-01 08:01:19 Lol 2022-09-01 08:01:21 Yea 2022-09-01 08:01:30 Didn't help anyway 2022-09-01 08:01:45 I want to try to install mcaffe on this box 2022-09-01 08:01:55 O.o 2022-09-01 08:01:56 XY problem 2022-09-01 08:01:59 But it complains 2022-09-01 08:02:01 Yup 2022-09-01 08:02:11 dwfreed: why 2022-09-01 08:03:16 Ermine: I'm not sure why you're asking me why? perhaps you meant to hilight Forza instead? 2022-09-01 08:03:23 However... Nvmd 2022-09-01 08:03:46 https://dpaste.com/6V8GSAGWB 2022-09-01 08:03:55 I wask asking why not relink sh to bash 2022-09-01 08:04:38 Mcafee Linux is rpm based it seems 2022-09-01 08:04:51 They do have dep option as well 2022-09-01 08:05:11 Ermine: because presumably the symlink is owned by some package (busybox, or some base files package), and you're likely to end up having that symlink change back without realizing it 2022-09-01 08:05:18 Not sure any of those would work on Alpine. I did do "apk add rpm" before trying the install 2022-09-01 08:05:23 if something wants bash, it should be using /bin/bash, not /bin/sh 2022-09-01 08:05:45 nothing should be depending on /bin/sh being bash, even if that's always the case in some distro 2022-09-01 08:05:49 Yea. I don't know if it is needing bash or not. I just guessed based on the error message 2022-09-01 08:06:31 don't mix package managers unless you *really* know what you're doing; since you're trying to install McAfee on a Linux machine, i'm not thinking so... 2022-09-01 08:06:37 Anyway, replacing sh with bash did not change the error 2022-09-01 08:07:20 dwfreed: well In this case mcafee is not my choice. It is forced in order to have this machine in the specific situation 2022-09-01 08:07:47 use something that can be compliant then? 2022-09-01 08:13:23 panekj: only Ubuntu =( 2022-09-01 08:13:32 18.04 2022-09-01 08:13:54 Perhaps the problem is musl? 2022-09-01 08:14:38 Installing Ubuntu would be very overkill for this =( blah 2022-09-01 08:14:41 if it's not built completely static, which seems likely, that would definitely be an issue 2022-09-01 08:17:47 Indeed. I would have to change to Ubuntu I think 2022-09-01 08:18:05 On a test-vm it worked fine on Ubuntu. 2022-09-01 08:18:48 you could extract the rpms and do it all manually 2022-09-01 08:20:45 Maybe. This is what happens 2022-09-01 08:20:47 # rpm -i MFErt.i686.rpm 2022-09-01 08:20:47 error: Failed dependencies: /bin/sh is needed by MFErt-2.0-165.i686 2022-09-01 08:20:47 warning: MFErt.i686.rpm: Header V4 RSA/SHA1 Signature, key ID 7a9da814: NOKEY 2022-09-01 08:21:10 Maybe rpm doesn't know that bash/sh is installed? 2022-09-01 08:21:26 plausible 2022-09-01 08:23:00 Forza: try rpm --nodeps :P 2022-09-01 08:24:03 Yup 2022-09-01 08:24:11 :D seems more issues 2022-09-01 08:24:25 I'll try installing on Ubuntu and then moving over the files 2022-09-01 08:24:46 It tries to run, but there are lots of failed dependencies (like systemd) 2022-09-01 08:24:53 install in container and export that single layer 2022-09-01 08:25:31 Not sure how to do that. Seems more complicated than installing ubuntu 2022-09-01 08:26:14 also I'm pretty sure it's linked against glibc so good luck 2022-09-01 08:27:09 https://dpaste.com/4YWBKH9NU 2022-09-01 08:27:12 Yup 2022-09-01 08:27:30 or? 2022-09-01 08:27:38 Libc.so.6 2022-09-01 08:29:30 what on earth needs a mcaffee install "in the specific situation" 2022-09-01 08:29:42 corporate 2022-09-01 08:29:51 Corporate security policy 2022-09-01 08:29:54 :O 2022-09-01 08:30:01 Those pesky details 2022-09-01 08:30:18 +1 on linked against glibc 2022-09-01 08:30:35 The security team fights everyone. If it were up to them we would use chisel and stones to produce stuff instead of CNC machines and computers 2022-09-01 08:31:13 sounds like it's time for a new security team 2022-09-01 08:31:15 If I were security team, I would kick out everyone not using proper distro 2022-09-01 08:32:17 I'd abolish all security theater 2022-09-01 08:35:28 Agreed 2022-09-01 08:35:51 Security has to be complicated so production staff use stickers with logins and passwords :D 2022-09-01 08:36:01 Anyway: the requirements are https://kcm.trellix.com/corporate/index?page=content&id=KB51573 2022-09-01 08:36:08 What would you choose from that list. 2022-09-01 08:36:41 Windows > macOS > Fedora > other Linux 2022-09-01 08:36:56 Mac is not an option 2022-09-01 08:37:00 i would choose retirement 2022-09-01 08:37:36 i choose death 2022-09-01 08:37:38 Personally I like Fedora more than Debian because of newer packages. But other than that 2022-09-01 08:37:40 =( 2022-09-01 08:37:45 I don't want death 2022-09-01 08:37:46 omg 2022-09-01 08:38:38 But Debian might be a better choice in this case 2022-09-01 08:38:49 uhm 2022-09-01 08:39:00 Is there anything that is rhel compatible these days? Alma and Rocky Linux? 2022-09-01 08:39:19 how about a chroot on alpine? or VM on alpine? 2022-09-01 08:39:22 Alma, Rocky, and CentOS are all RHEL compatible 2022-09-01 08:39:23 centos / fedora ? 2022-09-01 08:39:37 aron: can we do do this with glibc in chroot? 2022-09-01 08:39:55 you can certainly install it in chroot 2022-09-01 08:39:56 I have an archlinux chroot on alpine armv7 2022-09-01 08:40:05 Oh hum 2022-09-01 08:40:15 as for how compliant you will be is... 2022-09-01 08:40:18 I doubt your bass-ackwards security team would approve it, though 2022-09-01 08:40:23 I don't think running mcafee in a chroot would make sense (unless your goal is to just check some box and no one really checks what's happening) 2022-09-01 08:40:55 Mcafee would hook into some kernel subsystem to monitor some file and net usage I am sure 2022-09-01 08:41:06 then again, if they're using mcafee, they may not be able to tell :> 2022-09-01 08:41:14 Hahaha 2022-09-01 08:41:37 Mcafee is forced on us on all our Windows boxes. Makes the computers feel like 10 year old machines 2022-09-01 08:41:40 Forza: but it might not be able to find the actual files in its fs 2022-09-01 08:41:50 ikke: yea 2022-09-01 08:42:02 Not sure if it's worth that mess? 2022-09-01 08:42:15 it's likely using fanotify, which I believe is going to be restricted to the chroot it's running in 2022-09-01 08:42:33 so it would be none the wiser if you were doing things outside the chroot 2022-09-01 08:42:56 Let's skip that. I will try get ot of the mcafee. But if not I will go with something else. Maybe I'll use Ubuntu2204 2022-09-01 08:43:07 I am personally not worried 2022-09-01 08:43:46 It's a headless machine with only serice accounts and on an isolated vlan. However it does need access to AD domain controller, which is causing the mcafee requirement 2022-09-01 08:48:19 Maybe someone should tell Mcafee to support Alpine? :) 2022-09-01 08:51:48 you could give a try 2022-09-01 08:52:07 they should open source their stuff 2022-09-01 08:52:28 hehe, they should, but they won't 2022-09-01 08:57:04 Indeed. 2022-09-01 08:57:19 Do we have clamav on Alpine? 2022-09-01 08:57:38 Not that I think Clamav is great, but it is at least open source. 2022-09-01 08:58:03 yes 2022-09-01 09:33:33 Openwall uses rpm 2022-09-01 09:39:26 What's that? 2022-09-01 09:52:10 Are there any known viruses for Alpine? 2022-09-01 09:52:18 Or musl? 2022-09-01 09:52:25 or linux 2022-09-01 09:52:43 Yes there are 2022-09-01 09:53:02 I suppose root kits and such stuff are considered in that league 2022-09-01 09:57:17 I have not heard of anything specific 2022-09-01 10:00:15 :) 2022-09-01 10:00:47 Maybe Oracle Linux is an option to Alma/Rocky?just that Oracle is oracle... 2022-09-01 11:55:28 Is there a package for llvm’s libc++? 2022-09-01 11:55:41 Cant find anything on pkgs 2022-09-01 11:56:30 maybe llvm14-libs? 2022-09-01 12:24:47 for the boot process, any issue with skipping the grub step entirely and boot using extlinux on arm64? Seems the latter just load grub in order to give me a "kernel image not aligned on 64k boundary" error and exit. 2022-09-01 12:26:32 this is a bit strange, as if a boot loader loop of sorts. Where my u-boot image happens to support booting grub without extlinux. 2022-09-01 12:26:55 don't find this docoumented 2022-09-01 12:36:46 is this fixed in edge? this may be related. https://lore.kernel.org/linux-arm-kernel/CAMj1kXEu_C13WRaBML9T5gvXsukxP8B0hzH3bn5ZD6HpNshmZw@mail.gmail.com/t/ 2022-09-01 13:12:16 aron: nope, only libLLVM-14.so 2022-09-01 13:12:46 do this room have a name for you? 2022-09-01 13:19:15 uh ok nheko shows it correctly, so it's a element bug :/ 2022-09-01 13:21:29 consus[m]: I usually see llvm programs linked against it, the .so contains tenthousands of C++ symbols, thus I thought it was what you are looking for :) 2022-09-01 14:31:16 seems u-boot just reports finding extlinux.conf and then decides to run grub instead, a clue given is that the menu says grub if you look, and that extlinux included in alpine doesn't come with any menu support. 2022-09-01 14:32:05 but I am hit by this relocation bug in grub. 2022-09-01 14:33:11 which may be incorrect handling by an older u-boot binary, where support in newer versions of uboot is dodgy with various things not working fully across versions. 2022-09-01 14:45:09 zpiro_: "...boot using extlinux on arm64?" - there is no syslinux/extlinux for aarch64 in Alpine, it is only packaged for x86 and x86_64 2022-09-01 14:51:36 minimal: well, i discovered that it works with u-boot because extlinux.conf is included in the aarch64 uboot image in alpine, so have been using it elsewhere since then, as u-boot scripts is a hassel when trying things out. 2022-09-01 14:52:07 menu doesn't work as only the extlinux.conf file is included. 2022-09-01 14:52:30 zpiro_: u-boot may support using a extlinux.conf file but that does not mean it is actually using extlinux software 2022-09-01 14:53:41 they could have named the file pizza.conf instead 2022-09-01 14:53:51 minimal: yes, i haven't installed syslinux/extlinux on the partition. But menu.c32 etc should work if added in the folder, as u-boot certainly is doing the initial step. 2022-09-01 14:55:29 minimal: I have been booting and testing openwrt with extlinux.conf hint from alpine, so it certainly works. :) 2022-09-01 14:56:48 zpiro_: I'm not saying extlinux.conf doesn't work for u-boot, I merely pointed out that there is no extlinux packages for Arm and so regardless of what the config file that u-boot is using is called then it not using extlinux 2022-09-01 14:57:03 this with u-boot 2019, not sure how common it is to be enabled. 2022-09-01 14:57:23 its likely using something that understands a extlinux-compatible (BootLoaderSpec?) config file 2022-09-01 14:57:57 minimal: the confusion seems to be that alpine aarch64 has extlinux/extlinux.conf in the generic arm64 download. 2022-09-01 14:58:23 yes, and that is u-boot 2019 ^^ 2022-09-01 14:59:53 zpiro_: all i'm pointing out is that the existance of a extlinux.conf file does not mean extlinux itself is being used 2022-09-01 14:59:56 but it does prioritize grub after reporting that it finds extlinux.conf.... which is strange, but anyhow, menu isn't supported so when that didn't work for opennwrt i was a bit confused. 2022-09-01 15:00:14 this is a bit confusing, but yes, may very well be default support in u-boot since before v2019. 2022-09-01 15:01:26 zpiro_: it sounds like you have u-boot chainloading Grub 2022-09-01 15:01:49 related to this, console=ttyS0,115200 didn't work for me when adding it to extlinx.conf to install alpine as it ignored it and used grub, while u-boot reports finding extlinux.conf before loading grub... 2022-09-01 15:02:21 minimal: I thought so to, but the config line in extlinux.conf in alpine is 2022-09-01 15:02:59 MENU LABEL Linux lts,KERNEL /boot/vmlinuz-lts,INITRD /boot/initramfs-lts, FDTDIR /boot/dtbs-lts,APPEND modules=loop,squashfs,sd-mod,usb-storage quiet 2022-09-01 15:03:14 and well, theere is no menu support as its not packaged as you point out. 2022-09-01 15:03:45 this confused me 2022-09-01 15:03:46 zpiro_: u-boot is using the file extlinux.conf as a *generic* bootloader configuration file 2022-09-01 15:04:11 minimal: you keep repeating that, and i keep repeating that i know and explain the details. 2022-09-01 15:04:16 so if u-boot decides to chain load Grub then I assume it will pass on the contents of extlinux.conf to Grub in a form it can use 2022-09-01 15:05:34 zpiro_: I repeated it because you said " but it does prioritize grub after reporting that it finds extlinux.conf.... which is strange" as I don't find that strange if u-boot is using extlinux.conf as config to pass on to Grub 2022-09-01 15:06:02 it doesnt chainload, it reports finding extlinux.conf but choose grub over it, with marginal information about it, which made me think what you are now suggesting, but it doesnt. 2022-09-01 15:08:17 and i love extlinux, can be amazing hack. have in the past made a small disk image to dd onto the boot disk, and reboot into pxelinux to autoinstall through pxelinux. 2022-09-01 15:09:46 so what do you want u-boot to do exactly? not use Grub? 2022-09-01 15:11:00 minimal: i get hit by this bug it seems. https://lore.kernel.org/linux-arm-kernel/CAMj1kXEu_C13WRaBML9T5gvXsukxP8B0hzH3bn5ZD6HpNshmZw@mail.gmail.com/t/ so will just try again later with extlinux. 2022-09-01 15:12:22 and that is the end of the journey, after adding console=ttyS0,115200 to grub.cfg i got showed the error preventing alpine generic aarch64 from booting. 2022-09-01 15:14:10 and for discussion, extlinux.conf is absurdly simple and doesn't require a fat partition. 2022-09-01 15:17:13 and it may very well be that u-boot loads efi and grub from ext4 just fine. 2022-09-01 15:17:43 zpiro_: perhaps there's a patch in upstream Grub to deal with this? 2022-09-01 15:20:20 minimal: maybe, or there seems to be. could also be my old u-boot v2019. I'll play a bit more... in either case, telling people to create a fat32 partition for efi and grub on aarch64 if these things with ext4 support is enabled by default, is uncesdsary for first boot. 2022-09-01 15:21:44 zpiro_: UEFI is not manditory for aarch64 in general, though I believe it is for aarch64 servers (SBSA/ServerReady specs) 2022-09-01 15:22:30 minimal: openbsd seem to now default on that, where direction seem to be EDK2 support for Arm v8.x 2022-09-01 15:22:55 until then, you have this submarine that greedily tries anything. 2022-09-01 15:22:56 so "telling people to create a fat32 partition for efi and grub on aarch64" would never make sense on aarch84 SBCs/machines without UEFI support 2022-09-01 15:23:15 well, x86 bioses aren't linux centric u-boot either. 2022-09-01 15:23:51 doesn't parse... 2022-09-01 15:23:52 so msdos from pure habit 2022-09-01 15:24:16 u-boot reads ext4, doesnt need msdos to load uefi because it only supports that. 2022-09-01 15:25:10 "doesnt need msdos" - you mean doesn't need FAT filesystem? 2022-09-01 15:26:55 yeah, x86 bioses doesn't support anything else, maybe it supports some apple stuff, but whatever. u-boot doesn't need msdos for UEFI. 2022-09-01 15:28:03 u-boot just needs to find aaarch64.efi or whatever its called, so it can load grub that supports ext4 as well. 2022-09-01 15:28:13 assume ext4 and msdos is enabled by default 2022-09-01 15:28:54 openbsd seem to do something else, but bootloaders and this stuff can be quite nitpicky and frustrating. 2022-09-01 15:29:00 "x86 bioses doesn't support anything else"? x86 BIOSes don't support specific filesystems, the just load a boatloader (i.e. from MBR 1st sector). Or did you mean "x86 UEFI does not support anything other than FAT filesystem"? 2022-09-01 15:29:46 you seem to be mixing up terms, BIOS vs UEFI, msdos vs FAT 2022-09-01 15:31:54 calling FAT msdos is pretty common 2022-09-01 15:32:03 i was broadening and sending off, as i should perhaps go and double check. 2022-09-01 15:32:42 minimal: but the grub EFI part generally need to be on fat32 on x86 with UEFI support. 2022-09-01 15:34:03 you don't tend to write UEFI firmware into the x86 BIOS when installing grub, you may however in the bios register, install signing keys and what not. 2022-09-01 15:34:23 that last sentence makes no sense 2022-09-01 15:34:31 KISS concept on x86 is not the same as it is for u-boot 2022-09-01 15:35:34 heat_: depends on your system, but i have boxes that allows me to install my own keys for UEFI on x86. 2022-09-01 15:35:53 zpiro_: the ESP partition (whether containing Grub EFI file or other bootloader EFI files) needs to be formatted with a FAT filesystem. In theory you could use FAT12 or FAT16 or FAT32 (I think the spec only mandates FAT32), I use a 1MiB FAT12 ESP all the time with VMs. However some commercial UEFI implementations may not support FAT12 or FAT16 2022-09-01 15:35:57 for aarch64 and the same, you need EDK2 2022-09-01 15:36:32 zpiro_, how do you write "UEFI firmare into the x86 BIOS when installing grub" 2022-09-01 15:36:42 that doesn't make sense 2022-09-01 15:36:53 minimal: fine, and u-boot can load the grub efi file from ext4 or whatever other filesystem it has compiled in support for, where you can probably disable msdos support from u-boot entirel, ok? 2022-09-01 15:37:38 yes this makes no sense: "tend to write UEFI firmware into the x86 BIOS", either a machine has a BIOS or it has a UEFI ("CSM" mode in UEFI is a BIOS emulator for backwards compatibility) 2022-09-01 15:37:48 im on armada 8040, so could do this instead of u-boot maybe. 2022-09-01 15:38:47 https://github.com/tianocore/edk2-platforms/tree/master/Platform 2022-09-01 15:39:01 u-boot is not proper. 2022-09-01 15:39:28 https://github.com/tianocore/edk2 2022-09-01 15:40:52 u-boot is not proper what? 2022-09-01 15:41:12 so this is kind of like me repeating what minimal was saying about u-boot for extlinux, about u-boot and UEFI 2022-09-01 15:47:05 bootloaders is often, "until you hit a block for what you ideally want", "and do not care, it boots". For aarch64 it is waiting on EDK2 and support for BL2 to not need any binaries and have arm trusted firmware all the way to signed and verified kernel loading. 2022-09-01 15:47:40 or was it BL3... anyways, a few systems around that fully supports EDK2 for aarch64. 2022-09-01 15:53:09 if you have this, you dont need to be concerned about disk firmware, as you have secure boot. and kernel modifying kits in disk firmware should be caught. 2022-09-01 15:53:40 arm64 is nifty, but you quickly get your head turned into mush with the overlap to OP-Tee and Android ecosystem. 2022-09-01 16:03:17 would be neat if tianocore had more traction, perhaps not enough people that do software raid over two partions with grub to prevent images being modified by firmware, and something that really shouldnt be a concern logically, and not normalized lazyness. 2022-09-01 16:04:37 what about RISC-V then? :) 2022-09-01 16:05:19 tianocore has the most traction out of every firmware solution 2022-09-01 16:05:44 it's literally in almost every x86 PC, arm64 servers, riscv64 servers (when those exist anyway) 2022-09-01 16:05:46 never heard of so perhaps no traction enough? ;) 2022-09-01 16:06:09 Apachez, if your computer has UEFI, it's running tianocore 2022-09-01 16:06:42 (et al of course, usually forks with extra code attached) 2022-09-01 16:10:27 last i looked, openbsd seem to be betting on efi for arm64, needing interim support in some botched way or another from more recent u-boot versions. 2022-09-01 16:38:20 Dear elite programmers. My name is Janusz Chmiel. I AM advanced visually impaired user. I have one big plea to your development team. 2022-09-01 16:38:59 I would like to compile Speech-dispatcher for The AArch64 architecture. I Am using Android 8.0, Termux latest stable version 118 and Alpine Linux installed by using Prootdistro package for Termux. 2022-09-01 16:39:27 I Am suffering big issue. Gcc refuse create executables. 2022-09-01 16:39:38 So I can not finish ./configure script. 2022-09-01 16:41:06 My goal is to use Alpine Linux on Android with Mate desktop. 2022-09-01 16:42:21 Why? Because it is possible to use your distro without Systemd. So all services which are not depending stricly on Linux kernel can be activated even from Android and from Termux. As A result, Mate-desktop environment mate-session do not generate so much error warnings related to connection to The system message bus. 2022-09-01 16:43:23 Thank you for your help how to solve GCC can not create executables error. Speech-dispatcher is strongly needed for Orca to provide speech output. Sure, I will also compile Espeak-ng because it is more modern than Espeak. May be, that there is even some responsible packagers here. 2022-09-01 16:44:05 espeak-ng is already in alpine repos 2022-09-01 16:48:32 Unfortunately, very probably not in The community repository for AArch64 architecture. 2022-09-01 16:48:41 it's in testing 2022-09-01 16:48:57 since last yera 2022-09-01 16:48:59 * since last year 2022-09-01 16:50:21 Thank you very very much. May be, that somebody of us could even compile also Speech-dispatcher speech server. Its build is not complicated. Orca screen reader has build all necessary modules excepts Speech-dispatcher. But thank you very much for your kind support. 2022-09-01 16:51:24 https://git.adelielinux.org/adelie/packages/-/blob/master/user/speech-dispatcher/APKBUILD 2022-09-01 16:51:32 I wish you many positive programmers experience and good luck to all. 2022-09-01 17:18:51 heat_: so if I have that disabled it doesnt run tianocore? 2022-09-01 17:19:42 I guess devuan can be an alternative if you want a systemdfree world 2022-09-01 17:19:49 in case alpine doesnt fit 2022-09-01 17:25:27 Apachez, you usually can't disable UEFI 2022-09-01 17:25:41 you can run CSM, which is a BIOS layer on top of UEFI, but still very much UEFI 2022-09-01 17:30:40 Apachez: tianocore et. al. suffer from the traditional institutional and academic mess, convoluted build environments, wrapper scripts and non-standard methods set in motion by an intern tasked with it. U-boot on the other hand is this swiss army knife to get an embedded solution up and running. To some extent, what heat_ referr to with server class solutions can be seen with those things that 2022-09-01 17:30:46 overlap with it. Armada 8040 and its next evolution using the same kernel drivers etc, https://www.solid-run.com/embedded-networking/marvell-octeon-tx2-family/ where solid run seem to make most of the develper platforms for marvell, or easily acessible here. Where some of this have overlap with android ecosystem, implying that network things like iei puzzle and machiattobin etc are backseat 2022-09-01 17:30:52 passengers. 2022-09-01 17:32:06 without looking, tianocore should be easily available here: https://www.solid-run.com/arm-servers-networking-platforms/honeycomb-servers-workstation/ and here: https://www.solid-run.com/arm-servers-networking-platforms/macchiatobin/ where i am complaining about out dated u-boot support for clearfog gt 8k. 2022-09-01 17:32:43 and should look into tiasnocore again for it, but you quickly end up in a blind alley. 2022-09-01 17:34:15 wonder if some third party vendors and other solutions are more streamlined, but dunno. 2022-09-01 17:35:49 silly expensive with 16 core A72 cores - it's arm! 2022-09-01 17:37:32 if you really really want tianocore, it's better to check and be specific beforehand, that there are no binaries or catch-22 for EL3 (exception level, yes, the one relevant for OP-Tee). 2022-09-01 17:39:42 but A72 seem to be more network and server minded cores, a lot less ERRATA in the kernel, maybe one? 2022-09-01 17:42:15 likely to find a binary needed somewhere, certainly is for 8040. https://trustedfirmware-a.readthedocs.io/en/latest/plat/marvell/armada/build.html 2022-09-01 17:43:03 this in part the reasons for all the rage and interest - not quite there yet. 2022-09-01 17:47:48 would be better if marvell open sourced this https://github.com/MarvellEmbeddedProcessors/binaries-marvell.git 2022-09-02 08:05:18 Hi. I accidentally installed x86 instead of x86_64 on x86_64 hardware. Can I just fix /etc/apk/arch and `apk fix` or is it reinstall? 2022-09-02 08:06:41 iirc the arch is also hardcoded in apk, so you would have to fetch the apk.static of x86_64 2022-09-02 08:06:56 then change that arch, then apk fix --reinstall '*' 2022-09-02 08:07:01 i guess, i dunno 2022-09-02 08:07:58 I'd just reinstall it, if it is a fresh install 2022-09-02 08:08:32 psykose: x86 and x64 abis are different a bit afaik, would that work? 2022-09-02 08:09:02 abis of what, it's a static executable 2022-09-02 08:09:10 kernel abis 2022-09-02 08:10:32 ah, perhaps 2022-09-02 08:11:47 taking a look into musl's syscall wrapper: they are different, a lot 2022-09-02 11:28:33 hello, does anyone here use pipewire? 2022-09-02 11:28:34 with alpine linux 2022-09-02 11:32:17 me 2022-09-02 11:44:48 do you use it with edge or 3.16.* 2022-09-02 11:45:01 on* 2022-09-02 11:46:15 It's probably better to just state the problem you're having 2022-09-02 11:47:10 the problem is that I have followed the wiki and It's not working 2022-09-02 11:47:29 but alsa works 2022-09-02 11:47:34 pulseaudio also doesn't work .-. 2022-09-02 11:47:37 That's really vague... What is not working? Do you see any errors? What version of Alpine and PipeWire are you trying? 2022-09-02 11:47:49 3.16.2 latest 2022-09-02 11:47:56 Note that any message you delete is not actually deleted from IRC, only Matrix 2022-09-02 11:48:06 PureTryOut: I see errors and sound isn't working 2022-09-02 11:48:26 PureTryOut: ignore pulseaudio, it wanted polkit so I stop trying to install it 2022-09-02 11:48:36 s/stop/stopped/ 2022-09-02 11:48:53 So post those errors please. And please refrain from using Matrix-exclusive features like replies and edits, they translate quite badly to IRC (which the majority in this channel use) 2022-09-02 11:49:07 oh ok.... 2022-09-02 11:49:54 But without any errors/logs it's hard, if not impossible, to help you 😉 2022-09-02 11:52:38 it doesn't need polkit and the thing is clearly marked as a warning 2022-09-02 11:52:40 :) 2022-09-02 11:53:13 post errors, and also how you start it 2022-09-02 11:55:20 psykose: oh? I guess that I am wrong then 2022-09-02 11:55:33 (don't use replies in this channel please) 2022-09-02 11:55:44 they show up fine 2022-09-02 11:55:49 oh that's new 2022-09-02 11:56:07 if it's not above a certain length it doesn't url it 2022-09-02 11:56:22 or multiline, but that's multiline 2022-09-02 11:56:37 ah. So then it would have to be, "don't use replies or keep your answers short"? 😂 2022-09-02 11:57:03 ime it's better to just not use any fancy Matrix features at all in majority-IRC channels 2022-09-02 11:57:25 agree 2022-09-02 11:57:35 the polkit thing was for virt-manager, sorry :p 2022-09-02 11:57:41 I was mixing things 2022-09-02 11:57:43 my friend 2022-09-02 11:57:53 the *error* 2022-09-02 11:58:42 sorry, I am reinstalling pipewire and the internet is slow.... 2022-09-02 12:18:08 btw what is realtime scheduling 2022-09-02 12:18:09 and do I have to start the pipewire launcher as root? 2022-09-02 12:18:18 because it gave me more errors when I started it as a normal user 2022-09-02 12:19:25 no you're supposed to run it as your normal user 2022-09-02 12:19:36 again, if you don't actually post those errors we can't really help you 2022-09-02 12:20:17 if they have a W and are printed in yellow they are not errors 2022-09-02 12:21:04 it's warning it couldn't get realtime scheduling (you either need to have polkit or do some stuff with groups, but it is entirely optional) 2022-09-02 12:21:11 or, maybe it encrypted your whole drive and wants some bitcoin 2022-09-02 12:21:14 who knows, we can't see it 2022-09-02 12:21:26 oh ok and what is realtime scheduling? 2022-09-02 12:22:37 applications run tasks on 'threads' and threads have a priority- things with higher priorities come 'first' so they have less chance to lag or meet whatever time limits for doing stuff that matters like playing audio (which is done in real time) 2022-09-02 12:22:54 oh ok 2022-09-02 12:22:56 realtime scheduling specifically is a type of thread with higher priority, making one at all needs special permissions 2022-09-02 12:23:14 it's optional and for the most part you won't notice it 2022-09-02 12:23:30 if you like keeping your load at like 100 with only 4 threads though.. perhaps it will let you play music lag free 2022-09-02 12:23:34 it does for me at any load 2022-09-02 12:24:15 the entire subject is at least 100x more complicated than that but it really doesn't matter 2022-09-02 12:25:25 mod.protocol-native | [module-protocol-: 664 init_socket_name()] server 0x7f1a24dc7500: name pipewire-0 is not an absolute path and no runtime dir is found. Set one of PIPEWIRE_RUNTIME_DIR, XDG_RUNTIME_DIR or USEPROFILE in the environment is the first one but I have set XDG_RUNTIME_DIR to /tmp 2022-09-02 12:25:51 cpild not load mandatory module "libpipewire-module-protocol-native" no such file or directory 2022-09-02 12:26:06 could* 2022-09-02 12:26:27 that file should exist merely by having pipewire installed 2022-09-02 12:26:33 and you need an actual xdg_runtime_dir 2022-09-02 12:27:45 doing XDG_RUNTIME_DIR=/tmp would work too i guess, which means you didn't actually set it 2022-09-02 12:28:00 I did export * 2022-09-02 12:28:10 psykose: from a program? 2022-09-02 12:28:21 the place where you ran pipewire 2022-09-02 12:28:32 i dunno how you start it (i asked and as usual you didn't say how) 2022-09-02 12:28:43 from a newly spawned alacritty window 2022-09-02 12:29:02 I have my xinitrc script run another script that sets the environment variables 2022-09-02 12:29:07 like XDG_RUNTIME_DIR 2022-09-02 12:29:51 XDG_RUNTIME_DIR needs to be set in your session. So either export it just before starting PipeWire in the same terminal, or set it from the script you start with xinitrc 2022-09-02 12:30:03 Duck[m]: ok so I did XDG_RUNTIME_DIR=/tmp and some errors are gone but it still shows me this error 2022-09-02 12:30:04 PureTryOut: I did that 2022-09-02 12:30:18 export XDG_RUNTIME_DIR=/tmp in my script 2022-09-02 12:30:23 is `/tmp` writeable by your user? 2022-09-02 12:30:31 I wouldn't recommend that dir for that env btw 2022-09-02 12:30:38 oh- 2022-09-02 12:30:41 which would you recommend 2022-09-02 12:30:54 PureTryOut: I think so, there are alacritty.sock files tehere 2022-09-02 12:30:57 Well on my KDE Plasma system it's automatically set to `/run/user/1000` 2022-09-02 12:31:03 so perhaps there? 2022-09-02 12:31:09 hm. 2022-09-02 12:31:09 (if that directory exists) 2022-09-02 12:31:24 it doesn't exist 2022-09-02 12:31:47 then don't use that 😛 Perhaps somewhere in your home folder? 2022-09-02 12:32:03 just, not something that is accessible for all users on your system 2022-09-02 12:32:15 ok but the amount of errors wow 2022-09-02 12:32:29 after I did export "xdg", I got a lOT of errors 2022-09-02 12:32:36 PureTryOut: ok let me just do that 2022-09-02 12:33:46 PureTryOut: done 2022-09-02 12:34:31 "the place where you ran pipewire" <- where do you run it? 2022-09-02 12:36:16 failed to connect to session bus: /usr/bin/dbus-launch terminated abnormally without any error message 2022-09-02 12:36:17 you should just run it with /usr/libexec/pipewire-launcher 2022-09-02 12:36:28 psykose: in my xinitrc? 2022-09-02 12:38:23 psykose: an interesting thing is that I have another error which is, error loading shared library /usr/libwireplumber-0.4/libwireplumber-module-logind.so: no file or directory Error acquiring session bus address: Failed to execute child process "dbus-launch" (No file or directory) 2022-09-02 12:39:06 that means you are not in dbus 2022-09-02 12:39:19 you said you followed the wiki and it says right there 2022-09-02 12:39:20 your session needs to have dbus running 2022-09-02 12:39:24 $ dbus-run-session -- .. 2022-09-02 12:39:28 so clearly you didn't 2022-09-02 12:39:32 i still don't see any logs either 2022-09-02 12:39:49 I did dbus-run-session startx 2022-09-02 12:39:50 :( 2022-09-02 12:40:05 it should be in the .xinitrc 2022-09-02 12:40:08 dbus-run-session -- wm 2022-09-02 12:40:27 I thought that I had to run it with xorg 2022-09-02 12:40:29 hehe- 2022-09-02 12:44:29 exec command not found 2022-09-02 12:44:45 oops nvm 2022-09-02 12:47:50 pactl no such file or directory is the only red error now 2022-09-02 12:48:10 yeah i think that was required but i forget why we didn't have it added 2022-09-02 12:48:38 PureTryOut: perhaps we should add pactl to pipewire-pulse? 2022-09-02 12:48:50 it's used in some config file somewhere 2022-09-02 12:48:57 psykose: also where should I run the /usr/libexec/pipewire-launcher from? 2022-09-02 12:49:07 running it everytime from alacritty seems inefficient 2022-09-02 12:49:17 in the xinitrc is ok 2022-09-02 12:50:20 after dbus? 2022-09-02 12:50:29 there is nothing after dbus, that is the last thing in the file 2022-09-02 12:50:46 but yes, after it 2022-09-02 12:50:46 hm 2022-09-02 12:50:49 I ran it before dbus, before(in .xinitrc) and it didn't work 2022-09-02 12:50:53 ._. 2022-09-02 12:51:01 i guess instead of the xinitrc in the startup scripts for the wm 2022-09-02 12:51:40 well sound is still not working 2022-09-02 12:52:01 unless my volume is down 2022-09-02 12:53:09 pactl said connection failure: connection refused when I tried to increase the volume- 2022-09-02 12:54:38 I got ARM64 image magic error using extlinux.conf to boot install with the aarch64 generic image. With Solid-Run armada 8040 clearfog gt 8k board. As a provider of developments board, and it being major architecture and chipset for server and network appliances, this may be detrimental and aid in supporting all the desktop distros that more often get first boot due to convinience. Slightly disapointed 2022-09-02 12:54:44 that the kernel and boot process doesn't seem to work for armada 8040. And the whole EL3, Marvell binaries for EDK2 and u-boot for these industrially geared boards - as opposed to Pi associated boards.... As soon as you buy something like this, and alpine doesn't boot, you do cringe about it. 2022-09-02 12:54:46 psykose: ah true, it is. 2022-09-02 12:55:38 o_O 2022-09-02 12:55:42 it uses pactl to load the module-always-sink module 2022-09-02 12:56:42 there is no red error here but sound is still not working hm. 2022-09-02 12:57:47 wpctl status says that it could not connect to pipewire 2022-09-02 12:58:20 this involving support for OP-Tee in armv8. Which is basically a operating system running at exceptionl level three, hidden from exception level 1 and 2. Involving this binary, and without some industry trust and insight, is cringe, but also because it is there and supported. Just saying, any interested young nerds are served by playing with it for alpine. 2022-09-02 12:58:43 sure, if you buy me one i'll happily get it to work 2022-09-02 12:59:04 Duck[m]: can you strace it and post the full log? 2022-09-02 12:59:12 what is strace... 2022-09-02 12:59:47 apk add strace; strace -f -o log wpctl status 2022-09-02 12:59:49 upload the log 2022-09-02 13:00:19 Could not connect to PipeWire 2022-09-02 13:00:30 the log is in the file named 'log' 2022-09-02 13:00:33 and is huge 2022-09-02 13:00:36 pastebin it somewhere 2022-09-02 13:01:06 it actually is 2022-09-02 13:01:07 ok 2022-09-02 13:01:22 zpiro: an image magic error probably implies that the kernel is still compressed or whatever 2022-09-02 13:01:36 that and i'm not sure how well syslinux even works there (assuming that's why extlinux is used) 2022-09-02 13:01:45 aside from that no idea 2022-09-02 13:04:27 psykose: sure, but it is an amr64 marvell boar ddim talking about. tehnically, with developer and os design and support ambitions, or for protyping and open source ideology, you should be able to ask for a discount to work with the platform. 2022-09-02 13:05:14 https://paste.artixlinux.org/0f83aa26 2022-09-02 13:05:45 https://www.solid-run.com/arm-servers-networking-platforms/macchiatobin/ clearfog gt 8k doesn't work with grub, and solid run provided u-boot flashed in SPI, and extlinux fails to boot kernel due to bad arm64 kernel image magic. 2022-09-02 13:05:49 psykose: 2022-09-02 13:05:58 i'm not sure what you are saying, but if you're suggesting for me to go ask them for a discount to buy their board, i'm not that interested :) 2022-09-02 13:06:04 seems it either doesn't know about XDG_RUNTIME_DIR or ignores is 2022-09-02 13:06:44 Duck[m]: that looks like you have no XDG_RUNTIME_DIR at all 2022-09-02 13:07:39 env | grep XDG_ 2022-09-02 13:07:41 probably not there 2022-09-02 13:07:59 psykose: next generation platform is this: https://www.solid-run.com/embedded-networking/marvell-octeon-tx2-family/ 2022-09-02 13:08:11 of development platforms for marvell by solid run. 2022-09-02 13:11:49 export XDG_RUNTIME_DIR=/smt, I did. 2022-09-02 13:11:49 tho 2022-09-02 13:12:08 psykose: I got the runtime dir 2022-09-02 13:12:33 can you post a log of `env` in the same terminal you run wpctl status from? 2022-09-02 13:12:39 zpiro: yeah, i know where the store is 2022-09-02 13:12:44 my budget is $0 2022-09-02 13:13:02 as long as somebody push to u-boot and alpine, I'm certain their hardware support contracts and marvell contacts pay for cheap arm64 boards to distro developers, quite honestly, for group discounts. 2022-09-02 13:14:00 psykose: it took a few seconds but it's working now 2022-09-02 13:14:10 the problem is that I quickly tried to run stuff 2022-09-02 13:14:11 thanks 2022-09-02 13:14:42 psykose: and if you have a commit history to show, you can maybe ask them to senb some boards for free. 2022-09-02 13:16:26 psykose: but pipewire doesn't start from .xinitrc 2022-09-02 13:16:27 for some reason- 2022-09-02 13:17:58 you can't since it needs dbus first 2022-09-02 13:18:15 has to be in a script after, from the wm somewhere 2022-09-02 13:18:53 it is right after the dbus thing but let me see if I can start it with the wm 2022-09-02 13:19:06 there is no right after 2022-09-02 13:19:22 `dbus-run-session -- wm` runs a dbus session for only that and nothing else 2022-09-02 13:19:26 oh- 2022-09-02 13:19:31 if you do `wm &` and then something on the next line it doesn't have it 2022-09-02 13:19:35 I had that on another line 2022-09-02 13:19:37 after dbus 2022-09-02 13:19:38 hehe- 2022-09-02 13:19:48 i told you like three times 2022-09-02 13:22:32 it didn't work from the wm script 2022-09-02 13:22:33 oh well 2022-09-02 13:24:54 things being said and done, and Alpine based around that Bergen guy. Solid-Run as a Israeli company would likely see benefit and industry trust in paying the community for platform support for Linux and ditch debian. 2022-09-02 13:28:11 and maube, hopefully, that would add leverage, foir marvell to open up this binary required to compile EDK2 and u-boot to fill the OP-Tee and MSS part for EL3 at BL2. https://github.com/ARM-software/arm-trusted-firmware/blob/master/docs/plat/marvell/armada/build.rst 2022-09-02 13:30:06 psykose: where do you set your env variables 2022-09-02 13:30:07 Where that is the funny part for the 53K binary, u-boot just reports that MSS doesn't work. 2022-09-02 13:30:36 I only get audio from mpv when I have set XDG_RUNTIME_DIR manually in that specific terminal window 2022-09-02 13:30:44 then you export it wrong i guess 2022-09-02 13:30:58 I thought that just doing export from anywhere would set it globally 2022-09-02 13:30:58 i export them before dbus-run-session 2022-09-02 13:31:07 from? 2022-09-02 13:31:08 how 2022-09-02 13:31:26 export x=1 2022-09-02 13:31:31 dbus-run-session -- 2022-09-02 13:32:09 from xinitrc? 2022-09-02 13:32:21 or from another script that xinitrc calls 2022-09-02 13:32:43 it does not make any difference 2022-09-02 13:40:31 for what it's worth, microprocessor engineers, and their firmware developers working in an academic overlap that included physics simulations at the extreme endd of things, why one binary is stable electrically and another isn't. And for the inherent security concerns of mangement and seveeraly layers of tehnical staff, users, customers and supporters need a beginning that perhaps should involve 2022-09-02 13:40:37 community leaders like that ncopa, whom, if I am not mistaken - was attempted reqruited by microsoft two decades ago. 2022-09-02 13:41:49 take your meds 2022-09-02 13:41:55 and log off 2022-09-02 13:42:20 wrap or wrap your own hats off. 2022-09-02 13:42:30 *warp 2022-09-02 13:46:06 involved in the proposition, is that, you should know. In FPGA development, writing HDL code, debugging involves phystics simulations to understand why the logic circuit you wrote on the board isn't stable or doesn't work. 2022-09-02 13:46:57 meds 2022-09-02 13:49:24 "it does not make any difference" <- apparantly it does for me-?!!? 2022-09-02 13:49:25 um 2022-09-02 13:49:39 which is why you should take your meds, for comercial versus industrial platforms, and what it's worth to a company with an army of monkeys that finds and support combinations that work. 2022-09-02 13:49:43 I am running the script with, sh env.sh 2022-09-02 13:55:19 you have to do `. env.sh` 2022-09-02 13:56:02 which devd is better 2022-09-02 13:56:08 psykose: hehe ok 2022-09-02 13:57:32 psykose: it's not working :( 2022-09-02 13:58:16 can you post all the scripts 2022-09-02 13:59:33 in the env.sh file? 2022-09-02 13:59:48 also I tried to use .xprofile 2022-09-02 14:02:27 so I just threw the environment variables in the /etc/profile 2022-09-02 14:02:28 and it worked 2022-09-02 14:02:31 it's such a headache 2022-09-02 14:02:38 thanks btw 2022-09-02 14:05:43 I am fairly confident, about having met ncopa, but I am not certain. What I am certain of is work at CERN, where colleagues and room mates did FPGA development for LHC experiments, and if not that, were theoretical physistics. For a blue pill argument, there are two types of people, those living inside the machine and those living in theyr libraries. Where perhaps, the quicker and able outside the box 2022-09-02 14:05:49 are in need of a blue pill in the dividee inside the box. 2022-09-02 14:06:20 take your meds 2022-09-02 14:17:12 bl4ckb0ne: friendly reminder to also take your meds if you have any :) <3 2022-09-02 14:17:30 psykose: friendly reminder to go to sleep 2022-09-02 14:17:44 ur not my mom 2022-09-02 14:18:08 that's correct, I'm your dad 2022-09-02 14:18:34 i took my meds this morning 2022-09-02 14:43:21 psykose: at the end of the day, it is not much to ask that core developers have regular meetings with firmwware developers and hardware deesigners. 2022-09-02 14:43:58 will they be paid 2022-09-02 14:47:42 I was paid to talk to Intel on a regular basis, while working on performance optimzation on Intel MIC and AVX instructions at CERN, that lead to AVX2. Paid as a community is a different question, that you should be asking. 2022-09-02 14:48:42 Generally, you'd expect that industrial hardware, reliable designs and parameters. Be paid for by third parties. 2022-09-02 14:49:36 i accept free hardware donations 2022-09-02 14:49:39 also take ur meds 2022-09-02 14:50:10 so you can pay for your meds 2022-09-02 14:51:56 Broad generalisations are difficult! But as a broader community, there perhaps should be industry contacts that allow support for wider network deployments. 2022-09-02 14:53:54 yeah you can contact me and widely deploy free hardware into my mailbox 2022-09-02 14:53:57 also take ur meds 2022-09-02 14:53:59 I am however certain that a wider Arm64 community, that involves engineers and hired profesionals to work on performance. Allow for commnity enagements for security and stability. 2022-09-02 14:54:43 Hello good people! I’m a fresh Alpine user with a problem - i can’t get Network UPS Tools to work… anyone can point me to a guide compatible with alpine? 2022-09-02 14:54:46 you can hire me but i want a close in my contract for you to take your meds on a daily basis 2022-09-02 14:56:00 Horgrim: what are you trying to do 2022-09-02 14:56:05 and what doesn't work 2022-09-02 14:57:03 I have Alpine installed on a Wyse 3290 thin client, running docker with two containers 2022-09-02 14:57:22 I can’t get the NUT to connect to my ups over SNMP 2022-09-02 14:58:08 I’ve set the SNMP prameters in my APC ups already and UnRAID can connect to the ups and read the data 2022-09-02 14:58:24 what did you run, what error did you get 2022-09-02 14:58:27 post some logs etc 2022-09-02 14:58:34 Give me a sec 2022-09-02 14:59:01 zpiro: why not RISC-V 2022-09-02 14:59:45 I’ll reconnect on my PC in a minutę 2022-09-02 15:01:45 so here i am back ;) 2022-09-02 15:02:40 so, i've installed NUT using apk add nut 2022-09-02 15:02:43 it installed OK 2022-09-02 15:03:35 then nano -w /etc/nut/nut.conf 2022-09-02 15:03:50 MODE=netserver 2022-09-02 15:04:21 nano /etc/nut/ups.conf 2022-09-02 15:04:34 [APC] 2022-09-02 15:04:34 port = 10.1.1.12 2022-09-02 15:04:34 driver = snmp-ups 2022-09-02 15:04:34 snmp_version = v1 2022-09-02 15:04:34 community = public 2022-09-02 15:04:35 pollfreq = 15 2022-09-02 15:04:35 desc = Smart-UPS 750 RM 2022-09-02 15:04:50 and when i want to do UPSC 2022-09-02 15:04:54 upsc 10.1.1.12 2022-09-02 15:04:59 the connection is refused 2022-09-02 15:05:41 are you sure you don't have a firewall blocking the connection somewhere 2022-09-02 15:05:47 Does port = 10.1.1.12 make sense? 2022-09-02 15:05:55 nope, it's in my internal LAN 2022-09-02 15:06:04 yes, i went with 10.1.1.0/24 2022-09-02 15:06:21 port equals IP as far as i understand 2022-09-02 15:06:29 ikke: yeah, it's the tool's nomenclature as far as i'm concerned 2022-09-02 15:06:31 upsdrvctl start 2022-09-02 15:06:37 and then i get this 2022-09-02 15:06:38 Duck[m]: that is also good, but I am trying to make a larger point. We are well past a stage where community leaders in IT should be able to take industry contacts and agreements and demand a Professorship. 2022-09-02 15:06:46 because they usually use serial and `port = /dev/tty{something}` 2022-09-02 15:06:47 Network UPS Tools - UPS driver controller 20220328-3481-gccbf3eb6c6 2022-09-02 15:06:47 Network UPS Tools - Generic SNMP UPS driver 1.21 (20220328-3481-gccbf3eb6c6) 2022-09-02 15:06:47 No matching MIB found for sysOID '.1.3.6.1.4.1.318.1.3.2.7'! 2022-09-02 15:06:47 Duplicate driver instance detected (PID file /var/run/nut/snmp-ups-APC.pid exists)! Terminating other driver! 2022-09-02 15:06:47 Please report it to NUT developers, with an 'upsc' output for your device. 2022-09-02 15:06:49 Going back to the classic MIB detection method. 2022-09-02 15:06:49 No matching MIB found for sysOID '.1.3.6.1.4.1.318.1.3.2.7'! 2022-09-02 15:06:51 Please report it to NUT developers, with an 'upsc' output for your device. 2022-09-02 15:06:51 Going back to the classic MIB detection method. 2022-09-02 15:06:53 Please report it to NUT developers, with an 'upsc' output for your device. 2022-09-02 15:06:53 No matching MIB found for sysOID '.1.3.6.1.4.1.318.1.3.2.7'! 2022-09-02 15:06:55 Detected Smart-UPS 750 RM on host 10.1.1.12 (mib: apcc 1.6) 2022-09-02 15:06:55 Going back to the classic MIB detection method. 2022-09-02 15:06:57 [APC] Warning: excessive poll failures, limiting error reporting (OID = .1.3.6.1.4.1.318.1.1.1.9.3.3.1.5.1.1.3) 2022-09-02 15:06:57 [APC] Warning: excessive poll failures, limiting error reporting (OID = .1.3.6.1.4.1.318.1.1.1.9.2.3.1.5.1.1.3) 2022-09-02 15:07:44 https://networkupstools.org/docs/man/snmp-ups.html 2022-09-02 15:08:00 port should be net name or IP 2022-09-02 15:08:03 i went with IP 2022-09-02 15:08:38 since all the tutorials on the net are about debian/ubuntu/homeassistant 2022-09-02 15:08:48 some commands are not compatible 2022-09-02 15:09:28 Interestingly, Professor, as a position, requires no formal education. No academic track recordd. It can be given to anyone thought able to oversee higher studies that provides vocational training. 2022-09-02 15:10:23 and since i want to shutdown the pc when baterry level is low, i don't need a NUT docker but a standalone install on the system itself 2022-09-02 15:11:19 it would probably be easier with USB, but i have a management card and want to use it ;) 2022-09-02 15:12:12 It looks like NUT is not getting any replies from the ups? 2022-09-02 15:12:27 Did you try a manual snmpget? 2022-09-02 15:13:02 There is no, and have never been, a requirement for the position of Professor to have a Ph.D, a masters or bachelor degree -- not formally. 2022-09-02 15:13:28 academia requires you to take your meds 2022-09-02 15:13:47 a requirement for the rank of professor is to ship me hardware 2022-09-02 15:13:50 also take ur meds 2022-09-02 15:13:58 i used nut-scanner -S -s 10.1.1.12 -e 10.1.1.12 2022-09-02 15:14:07 and it gives me output 2022-09-02 15:14:11 Scanning SNMP bus. 2022-09-02 15:14:11 driver = "snmp-ups" 2022-09-02 15:14:11 [nutdev1] 2022-09-02 15:14:11 port = "10.1.1.12" 2022-09-02 15:14:11 desc = "Smart-UPS 750 RM" 2022-09-02 15:14:13 mibs = "apcc" 2022-09-02 15:14:13 community = "public" 2022-09-02 15:14:21 so it can read snmp 2022-09-02 15:14:55 but starting the upsdrvctl shows some errors 2022-09-02 15:14:58 In Norway, where ncopa is from, we have D.Philos as a unsupervised, higher Ph.D. to keep open the age old door for industry figures to take positions in acamdeia. 2022-09-02 15:15:04 as i've shown above 2022-09-02 15:15:18 Norway? that doesn't exist, what are you talking about 2022-09-02 15:15:48 norway is not a real country 2022-09-02 15:16:16 norway is a fictional place from westeros, the continent from game of thrones 2022-09-02 15:20:50 If i met ncopa, it was at WarpCrew in Stavanger. I am sure that many can sign off on, from the age of 15 until 30. That guidance is likely to provide work and paid income, and from classical literature and history; the position earned and warranted is Professor. 2022-09-02 15:21:33 sounds like a spam email 2022-09-02 15:21:51 zpiro is off their meds 2022-09-02 15:22:14 talking about fictional places and fictional cpu archs 2022-09-02 15:22:25 nobody heard of arm, it doesnt exist 2022-09-02 15:23:01 is single command for update and upgrade packages exists? 2022-09-02 15:23:39 apk -U upgrade 2022-09-02 15:24:43 i prefer: apk update && apk version | sort; apk -ai upgrade && update-conf 2022-09-02 15:24:57 bl4ckb0ne: Yes, I don't know french. Even while working at CERN since 2010 until end of 2016. But it dopes not mean that Geneeva doesn't exists, because I was never paid to learn French. 2022-09-02 15:25:17 when are you sending the sern assassins after us 2022-09-02 15:25:32 bl4ckb0ne: you may find your own meds elsewhere. 2022-09-02 15:25:32 never heard of this "Geneeva" 2022-09-02 15:25:42 j'ai pris mes médicaments ce matin! 2022-09-02 15:25:59 is she on onlyfans? 2022-09-02 15:26:45 if arm exists where is leg 2022-09-02 15:27:55 Ermine: you never heard on legv7 and legv8, the famous archs> 2022-09-02 15:28:09 they power a lot of phones, been around since 1985 2022-09-02 15:28:40 isn't norway located near nandway? 2022-09-02 15:28:49 same year i became a professor at cern and met ncopa 2022-09-02 15:33:58 was it also the same year you found your meds? 2022-09-02 15:34:55 LArch64? 2022-09-02 15:35:29 i told you several time psykose i took my meds this morning 2022-09-02 15:35:40 prove it! 2022-09-02 15:35:50 you're not my mum 2022-09-02 15:36:02 wanna bet?? 2022-09-02 15:36:52 n-no 2022-09-02 15:53:20 Oh noes! Not the meds and cern? 2022-09-02 15:55:16 no 2022-09-02 15:55:18 the very same! 2022-09-02 15:56:12 Is this the room for an argument? ;-) 2022-09-02 15:56:25 depends if you took your meds this morning 2022-09-02 15:56:42 I suspeect, that Norway and CERN do not nead each other- The 26 odd Russian newspapers that covered LHC, and all of the nobodies that do not care about founding members inside and outside of CERN. Can haippily accept their differences and take their separate ways. 2022-09-02 15:56:51 its always morning somewhere :-) 2022-09-02 15:56:55 *need 2022-09-02 15:57:03 what you need is your meds 2022-09-02 15:57:17 nead* 2022-09-02 15:57:23 knead* 2022-09-02 15:57:32 what you need is to be able to boot operating systems. 2022-09-02 15:57:45 sure 2022-09-02 15:57:50 thanks 2022-09-02 15:57:56 yw 2022-09-02 15:57:58 bl4ckb0ne: I was referring to the Monty Python "argument" sketch lol 2022-09-02 15:59:22 BlackIkeEagle: I'm off into the sunet with my non-bootable systems, where are the meds? 2022-09-02 15:59:49 https://www.dailymotion.com/video/x2hwqlw 2022-09-02 16:00:55 heck I didn't make time to watch monty python this summer 2022-09-02 16:03:03 oh my, a moot point! 2022-09-02 16:03:31 I see a mount point 2022-09-02 16:05:19 zpiro: i don't think we have met. You are probably mixing me with someone else 2022-09-02 16:07:19 ncopa: likely, but around the late 90's I met a lot of people, including Øyesten Homelien. 2022-09-02 16:09:00 I don't doubt that. but I'm pretty sure you and I never met 2022-09-02 16:19:22 ncopa: I have these perplexions, whom were they. And it does involve homelien, and the difference between sadnes and berge where boney was figure. This extends into the ircop community of EFnet, but there was this talented guy at one of the warpcrew lan parties where a windows or MS guy were interviiewing or offering work, and was suspecting that was you. 2022-09-02 16:22:20 In which case, we didn't meet. I slept on a concreet floor at the LAN party 20 years ago where you was considered for an offer from microsoft. 2022-09-02 16:22:48 I guess bb mount cannot mount ntfs partitions? 2022-09-02 16:24:11 mount -t ntfs 2022-09-02 16:27:52 thx 2022-09-02 16:31:09 (alternatively apk add mount for the actual real working one that can probe the damn fs) 2022-09-02 16:43:11 ncopa: But for all intents and purposes, free stuff from Solid-Run for the wider community. And with regular talks about how to get a complex kernml running withouth bels and whistles. As one of several canditates from the ARM ecosystem desperate to be on the good side of idelogical software -- needing regular talks to make it happen. 2022-09-02 16:44:50 pills 2022-09-02 16:45:04 ^^ 2022-09-02 16:45:16 is there a pill i can take to get solidrun hardware at my doorstep 2022-09-02 16:46:08 intel is not at the edk2 door step unless you have engineering hardware access. 2022-09-02 16:46:19 psykose: a hallucinogenic? 2022-09-02 16:47:06 when i took those all i saw was atomic universes :/ great for i915 drm developers, not good for my lack of Premium Quality solidrun hardware 2022-09-02 16:48:09 psykose: meds 2022-09-02 16:48:30 Knockoff-Quality flexiblerun hardware is cheaper 2022-09-02 16:50:41 bl4ckb0ne: i just took mine 10 minutes ago 2022-09-02 16:51:04 Smart people often get beaten. 5HT receptrors are also involved in neurogenesis. A family of receptors referred to as the serotontin receptors. Some of which are involved in haluctiogneic discourse. The wider implication is that you are better off during mechanisms that help repair your brain cells. 2022-09-02 17:01:12 Not wishing to add credence. But there is probably something not perfect about yuou if you don't have a maths careers and spend time on IRC. 2022-09-02 17:03:28 i dont know how to count 2022-09-02 17:05:58 Little bit general question but should I use xf86-video-amdgpu or modesetting? 2022-09-02 17:07:59 zpiro: can't parse that as you forgot your bracket, is it "((don't have a maths careers) and (spend time on IRC))" or "don't (have a maths careers and spend time on IRC)" ? 2022-09-02 17:17:34 Ermine, arch's wiki says amdgpu, so probably that 2022-09-02 17:17:59 Ermine, https://wiki.archlinux.org/title/xorg#Driver_installation 2022-09-02 17:22:46 Oh hi heat 2022-09-02 17:22:51 thank you 2022-09-02 17:23:09 no problem 2022-09-02 17:23:11 i am everywhere 2022-09-02 17:23:25 there's an update for it even 2022-09-02 17:23:57 in intel case they recommend modesetting over intel 2022-09-02 17:24:28 yes 2022-09-02 17:24:35 but in that xorg they recommend the opposite 2022-09-02 17:24:38 it's confusing 2022-09-02 17:24:44 modesetting works fine for amdgpu too, use what works 2022-09-02 17:24:48 at the end of the day just use wayland 2022-09-02 17:24:48 :P 2022-09-02 17:24:51 oh and test !38530 after you get it working 2022-09-02 17:25:00 and then promptly uninstall it and go use a wayland compositor 2022-09-02 17:27:12 I'm configuring my new alpine daily driver for now, but I'll take a look 2022-09-02 17:27:33 I guess I need to add your repo as remote to get this mr? 2022-09-02 17:28:25 you could just install the artifacts 2022-09-02 17:32:16 ok 2022-09-02 17:37:57 minimal: the former, thanks! 2022-09-02 17:39:01 nerds 2022-09-02 17:41:05 you'd think, but sometimes context matters. And whether or not you reach for your meds or not matter less. 2022-09-02 17:43:14 i told you i took my meds already 2022-09-02 17:43:17 this is getting old 2022-09-02 17:46:30 This to clarify that Mr. XXX, born xx.xx.8x, of Norwegian nationality, was an Associated mem 2022-09-02 17:46:33 ber of the Personnel (User) of the Euopean Organisation for Nuclean Research (CERN) from xx.xx.2010.... 2022-09-02 17:46:51 zpiro: good to know, I'm perfect then as I had a maths career (though it only lasted a very short time before it careered off a cliff lol) and I spend time on irc ;-) 2022-09-02 17:48:59 https://twitter.com/BalderOddson/status/1565705086529146880 I have done the red or blue pill thing already. 2022-09-02 17:58:12 Mr. [REDACTED] 2022-09-02 17:59:09 There is little to no help from looking at compiler commits for aarch64 as a boss with diplomatic class C, or ambassador equivalence at CERN, for physics, software or open source - just because you can. 2022-09-02 18:00:20 zpiro: by any chance, are you the same person that spammed the heck out of #wayland about a year ago 2022-09-02 18:00:31 ranting about computing performances 2022-09-02 18:00:34 Ermine: does the driver work 2022-09-02 18:00:38 you got kicked several time 2022-09-02 18:00:51 Or, I have no benefits from a fellow Norwegian giving me news about capabilities of 64 bits arm throught those means at CERN. 2022-09-02 18:01:48 bl4ckb0ne: not me, but the lunacy of my personal circumstance would or should excuse much. 2022-09-02 18:06:51 Some of us, born with computers, introduced to them, or what else you may. Should be in positions of state sponsored authority. 2022-09-02 18:16:09 Not to take a piss on European history, and family dynasties in science, politics and the like. But if you had access to a computer outside of academia and state offers. You have a father whos an engineer or equivalent. Outside of oil and gas, the access or inhertited culture is celebrated well into autocratic instituitions. 2022-09-02 18:16:51 not to take a piss on you, but your rambling is highly unwanted, if you need to vent about whatever shit you are on, there is #alpine-offtopic 2022-09-02 18:18:23 panekj: right, thanks. 2022-09-02 18:32:32 psykose: driver works for me 2022-09-02 18:36:00 Ermine: neat 2022-09-02 18:55:01 So Alpine is my desktop distro now 2022-09-02 18:58:09 Ermine: nice! It's pretty incredible as a desktop OS 2022-09-02 18:58:18 What DE are you using? 2022-09-02 18:58:22 Or none? 2022-09-02 18:58:27 I'm using Plasma 2022-09-02 18:58:40 There are some problems to be solved, though 2022-09-02 18:58:45 Is plasma Wayland first now? 2022-09-02 18:58:59 I'm on Xorg version 2022-09-02 18:59:01 Did you use setup-desktop? 2022-09-02 18:59:04 yes 2022-09-02 18:59:23 Ahhh, okay. No experience with that yet, unfortunately 2022-09-02 18:59:40 What issues? Maybe I have ran across similar 2022-09-02 19:01:46 I need to configure font rendering, there are not enough cursors installed by setup-desktop, so I see "X" cursor from time to time 2022-09-02 19:03:01 And it asks me for a password each time I log in so it could mount partition, despite I have it in fstab (had same issue on OpenSUSE though) 2022-09-02 19:04:29 For cursors, is the full breeze theme and assets installed? 2022-09-02 19:05:11 I too can not get auto-mount of NAS partitions in fstab without root 2022-09-02 19:05:42 Apparently there is a PAM module, pam_mount, that can do this that we don't have 2022-09-02 19:12:32 maybe 2022-09-02 19:13:51 easily packageable 2022-09-02 19:15:12 Saijin_Naib: i mean, my partition is mounted and visible from dolphin, but kde still wants to mount it 2022-09-02 19:26:24 Tried to run discord with gcompat installed, still failed to relocate some symbols. Should I report to gcompat devs? 2022-09-02 19:31:34 if you want, there's probably already an mr for the same symbols 2022-09-02 19:31:50 it's most likely never going to work anyway 2022-09-02 19:34:08 Ah, __mbrlen is even merged (thanks ptrc) 2022-09-02 19:36:13 psykose: feature Request to maintainer of pam? 2022-09-02 19:36:24 it's not pam related, just a random package 2022-09-02 19:51:45 Ermine: you're welcome :) 2022-09-02 19:53:51 ptrc is basically a professional gcompat developer /s 2022-09-02 19:54:05 fortunately not 2022-09-02 19:55:08 i'm just working on my glibc addiction 2022-09-02 19:55:50 ey ptrc , did you fix the problem with remmina/rdp? I was out this weeks and saw your message on highlights 2022-09-02 19:55:59 what was the issue 2022-09-02 19:56:26 must have been someone else, haven't used rdp in a long while 2022-09-02 20:23:43 when I try to connect to a windows VM I get: 2022-09-02 20:24:06 connect? get? 2022-09-02 20:24:15 "Could not connect to the RDP server... via TLS. See the DEBUG traces from a terminal for more information" 2022-09-02 20:25:49 ptrc: well, you said that it happened to you "way before openssl3..." 2022-09-02 20:25:59 and when you follow that advice? 2022-09-02 20:26:40 ERROR][com.freerdp.core] - transport_connect_tls:freerdp_set_last_error_ex ERRCONNECT_TLS_CONNECT_FAILED [0x00020008] 2022-09-02 20:28:01 is it listening? 2022-09-02 20:29:27 uhM, it seems that yes 2022-09-02 20:29:59 seems? 2022-09-02 20:31:37 yes i was not sure about the port number, maybe I have some firewall problem but I can't dig into it at this moment 2022-09-02 20:31:40 I will check tomorrow 2022-09-02 20:31:46 ok 2022-09-02 20:32:20 maybe worth trying rdesktop, too 2022-09-02 20:32:27 may be 2022-09-02 20:35:32 oh yes, I will test it too 2022-09-02 20:35:35 thanks 2022-09-02 20:40:21 donoban: could be https://gitlab.com/Remmina/Remmina/-/issues/2711 2022-09-02 20:49:18 uhm, I'm not sure but I edited the advanced configuration and set TLS security level to 0 2022-09-02 20:49:42 "0 -- Windows 7 compatible", maybe they changed the default level to some incomptabile 2022-09-02 20:49:51 well, based on everything else the freerdp openssl issues were solved in like 2.7 2022-09-02 20:50:04 so it's either related to that or something similar, not sure 2022-09-02 20:51:55 with Default and debug enabled I see a failure in negotiation 2022-09-02 20:53:29 maybe they just changed the Default to modern windows versions and added that compatiblity mode, I've never touched it before 2022-09-02 20:53:32 a sanity test could be to build it against openssl1.1, but there's a few things in the dep chain so it's multiple packages 2022-09-02 20:54:20 I will ask #remmina before, for my usecase I'm not worried about using a weaker mode 2022-09-02 20:54:44 see you tomorrow I have to close now, thanks all! 2022-09-02 20:55:03 freedrp, libvncserver only actually 2022-09-02 20:55:05 goocluck 2022-09-02 20:55:09 goodluck 2022-09-02 20:55:14 ty 2022-09-03 01:38:11 is there a way to stop a user from using SU? 2022-09-03 01:38:11 hello 2022-09-03 01:39:18 not installing it? :) 2022-09-03 01:39:22 su requires knowing the target user's password 2022-09-03 01:39:30 don't give them the password of the target user? 2022-09-03 01:39:43 aside from that, the classic solution is wheel group 2022-09-03 01:40:06 (make a group of users who are allowed to run su, and mode 4710 owned by root:wheel 2022-09-03 02:10:13 If you have a Gmail account, a computer or a smartphone and you have 2 hours a day, you can earn 345 EUR a day. Ask me how.... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/zHSOkUzguhVQCcOMwvqYWFmQ) 2022-09-03 05:20:56 "not installing it? :)" <- it's included in busybox 2022-09-03 05:21:12 "(make a group of users who are..." <- and how do I do that 2022-09-03 05:21:24 there is the wheel group and only root is in it I think 2022-09-03 05:21:42 "don't give them the password..." <- I don't want any malicious process to try and bruteforce 2022-09-03 05:22:23 remove busybox-suid package? 2022-09-03 05:22:31 it's not part of the base busybox package 2022-09-03 05:22:36 which is not suid 2022-09-03 05:23:21 Duck[m]: if your passwords are bruteforceable, you're screwed anyway 2022-09-03 05:23:44 dwfreed, not if there's no interface that accepts password login 2022-09-03 05:23:49 dwfreed: why? 2022-09-03 05:24:08 dalias: oh 2022-09-03 05:24:10 hm. 2022-09-03 05:24:36 if sshd is configured right (not to accept passwords) then su and local console are the only possible places a password is likely to be usable 2022-09-03 05:24:51 dalias: best not to rely on that 2022-09-03 05:25:02 ?? 2022-09-03 05:25:08 I am not going to use sshd- 2022-09-03 05:25:32 best not to rely on not having random daemons running as root and giving privilege elevation vectors? 2022-09-03 05:25:34 how much entropy should a root password have 2022-09-03 05:26:02 dalias: do you not have busybox-suid 2022-09-03 05:26:10 dalias: best not to rely on 2022-09-03 05:26:13 erm 2022-09-03 05:26:20 hit enter on wrong keybaord 2022-09-03 05:26:21 it depends on what the attack surface is 2022-09-03 05:27:34 if root password is only for console login on a machine kept somewhere in your physical control, where an attacker who got to it could just as easily take the storage device out and access it directly, then the quality of the root password hardly matters 2022-09-03 05:27:42 dalias: best not to rely on "can't use a password" to protect shitty passwords 2022-09-03 05:27:58 just don't pick shitty passwords 2022-09-03 05:28:12 Duck[m]: ? 2022-09-03 05:29:06 well best is not to have passwords at all but then you need to do something special for console/recovery access :-p 2022-09-03 05:29:35 but i don't think our minor disagreements over these things are helpful to discuss in context of what Duck is asking 2022-09-03 05:31:34 Answer is to remove su I think 2022-09-03 05:31:58 Answer is to remove user I think 2022-09-03 05:32:40 i think removing bbsuid is a good idea 2022-09-03 05:32:47 We don't kill people do we 2022-09-03 05:32:49 just removing su doesn't suffice because of how busybox works 2022-09-03 05:33:03 someone can just ln -s /bin/bbsuid ~/su then run su 2022-09-03 05:33:03 dwfreed? 2022-09-03 08:26:04 after boot my raid1 array always end up in auto-read-only mode 2022-09-03 08:26:16 first drive is missing, second drive is okay 2022-09-03 08:26:30 mdstat shows: 2022-09-03 08:26:33 md4 : active (auto-read-only) raid1 sda4[0] sdb4[1] 2022-09-03 08:26:33 453901312 blocks super 1.2 [2/1] [_U] 2022-09-03 08:26:58 but echo repair >/sys/block/md4/md/sync_action 2022-09-03 08:27:00 heals it 2022-09-03 08:27:15 why it is broken at boot time? 2022-09-03 08:27:57 what does dmesg say 2022-09-03 08:29:51 aron: happens on every boot or just first after fresh install? 2022-09-03 08:30:06 dwfreed: https://dpaste.org/SK6GO 2022-09-03 08:30:14 Apachez: every single time 2022-09-03 08:31:20 how does your fstab looks like? 2022-09-03 08:31:29 seems like you got other raid1 partitions aswell is that correct? 2022-09-03 08:31:30 fstab has nothing to do with it 2022-09-03 08:31:44 md1, md3, md4 2022-09-03 08:31:55 3 raid arrays, 2 of them are always good 2022-09-03 08:32:07 md1 and md3 autorecovers during boot but its like it isnt aware of md4 ? 2022-09-03 08:32:42 yeah, weird 2022-09-03 08:33:07 but /etc/mdadm.conf lists is perfectly 2022-09-03 08:33:50 enable openrc logging and pastebin the log? 2022-09-03 08:33:52 are you sure? 2022-09-03 08:34:09 after rebooting again 2022-09-03 08:34:10 perhaps you had a space instead of tab or some other magic character which "looks" right but is interpreted as junk? 2022-09-03 08:34:12 also pastebin mdadm.conf 2022-09-03 08:34:25 mdadm.conf is made by mdadm itself 2022-09-03 08:34:34 https://dpaste.org/VfUzW 2022-09-03 08:35:12 quick googling tells either bad config in mdadm or TLER issues with drive... but I guess your raid1 is partitionbased and not drivebased right? 2022-09-03 08:35:24 so it shouldnt be TLER related if md1 and md3 uses the same drives 2022-09-03 08:35:55 partition-based 2022-09-03 08:36:16 sda -> sda1 -> md1 + sdb -> sdb1 -> md1 2022-09-03 08:36:18 etc 2022-09-03 08:36:52 mdadm --examine /dev/sdX should be helpful 2022-09-03 08:37:41 could it be some superblock issue perhaps? 2022-09-03 08:37:42 https://raid.wiki.kernel.org/index.php/RAID_superblock_formats 2022-09-03 08:37:52 thinking that manual says that the metadata is per device and not per partition 2022-09-03 08:38:10 and you got one which is 0.90 and two 1.2 where only the first works automagically 2022-09-03 08:38:10 they are identical 2022-09-03 08:38:11 https://dpaste.org/RSrcR 2022-09-03 08:38:37 other than that is the UUID correct? 2022-09-03 08:38:43 can that be verified somehow? 2022-09-03 08:39:42 Apachez: "device" just means block device 2022-09-03 08:39:47 https://dpaste.org/xHHM6 2022-09-03 08:39:55 your mdadm is also slightly sparse comparing to example https://github.com/neilbrown/mdadm/blob/master/mdadm.conf-example 2022-09-03 08:40:42 sparse=slim 2022-09-03 08:41:04 what do you have at md4 ? 2022-09-03 08:41:26 thinking if you comment that out and reboot... everything else will work just fine just /dev/md4 wont exist? 2022-09-03 08:42:16 md4 exists, but it is "broken" 2022-09-03 08:42:29 one device is missing 2022-09-03 08:42:36 (reboot) 2022-09-03 08:43:14 dwfreed: if DEVICE lines are missing will the devices still be identified? 2022-09-03 08:44:12 https://www.kernel.org/doc/html/v4.15/admin-guide/md.html should be helpful for troubleshooting 2022-09-03 08:47:30 md4 is not present in fstab, tho 2022-09-03 08:48:32 https://dpaste.org/4tvzo 2022-09-03 08:48:44 ^ rc.log 2022-09-03 08:49:19 "Checking local filesystems" -> no md4 listed there (which is correct9 2022-09-03 08:52:33 you aren't starting the mdraid service in the boot runlevel; you probably should be 2022-09-03 08:52:53 yes, I tried that one too 2022-09-03 08:52:59 didn't helped, so I disabled it 2022-09-03 08:53:18 reboot with it running so we can see if it says anything 2022-09-03 08:53:55 both "mdadm" and "mdadm-raid" services? 2022-09-03 08:54:10 probably just the mdadm-raid 2022-09-03 08:54:26 whichever one contains 'mdadm -As' 2022-09-03 08:55:03 that's "mdadm --assemble --scan", right? 2022-09-03 08:55:16 mdadm-raid:mdadm --assemble --scan --quiet $tostart 2022-09-03 08:55:26 I believe so 2022-09-03 09:00:54 https://dpaste.org/OX4mN 2022-09-03 09:01:04 I see nothing 2022-09-03 09:01:24 rc-update says "mdadm-raid | boot" 2022-09-03 09:01:33 the fact that nothing changed is concerning 2022-09-03 09:02:48 the only difference is that "local" is added (which was me) 2022-09-03 09:02:50 oh, because the initscript skips devices that already exist, and I bet they all existed already 2022-09-03 09:03:14 I think the kernel fetched them all 2022-09-03 09:03:38 (the gentoo script does not do this) 2022-09-03 09:05:57 I think I'll do the poor man's solution by adding the "echo repair > sync_action" in local service 2022-09-03 09:08:15 yeah, it did the trick 2022-09-03 09:08:40 (do surprise here, afterall) 2022-09-03 09:08:45 s/do/no/ 2022-09-03 09:09:35 I need local service, since network mounts are not performed during boot 2022-09-03 09:09:47 it seems like it may automatically go read-write on the first write 2022-09-03 09:10:02 openvswitch starts after mounts 2022-09-03 09:10:24 you can dynamically adjust dependency ordering so netmount depends on openvswitch 2022-09-03 09:10:53 /etc/conf.d/netmount:rc_need="openvswitch" 2022-09-03 09:11:40 aaah, I see 2022-09-03 09:14:55 okay, netmount worked 2022-09-03 09:16:53 hmm, I think I'll go with this now 2022-09-03 09:17:13 repair stays in local, netmount works as expected 2022-09-03 09:17:21 thanks 2022-09-03 09:22:07 hiii 2022-09-03 09:23:14 i'm trying to set up alpine on my vps and i want to get rid of rsh and tftp 2022-09-03 09:23:21 i can't find any related documentation 2022-09-03 10:00:53 seb1g: they should not be installed by default. busybox-extras I believe is what provides these 2022-09-03 10:01:06 ah 2022-09-03 10:01:15 lynis complains about them being installed 2022-09-03 10:04:48 I mean, they're pretty harmless even if they are 2022-09-03 11:07:41 what is difference between ubuntu fonts and ubuntu nerd fonts? 2022-09-03 11:13:09 https://pkgs.alpinelinux.org/contents?branch=v3.16&name=font-ubuntu-mono-nerd&arch=x86_64&repo=community vs https://pkgs.alpinelinux.org/contents?branch=v3.16&name=font-ubuntu-nerd&arch=x86_64&repo=community 2022-09-03 11:25:03 Ermine: nerd fonts have ligatures and other math and programming symbols 2022-09-03 11:33:33 Saijin_Naib: those ligatures are strange: https://i.imgur.com/kxj39Sz.png 2022-09-03 11:34:21 I don't use them for my UI fonts. I use Serif 2022-09-03 11:34:30 They only get used in my terminal it seems 2022-09-03 13:43:26 How do you people run Syncthing on Alpine? I use it as a user service on my Fedora system, but Alpine doesn't have that. 2022-09-03 13:45:33 Oh, there's an OpenRC service. nice 2022-09-03 14:50:33 Did anybody try launching electron apps in flatpak? zypak-wrapper complains about dbus session bus and aborts 2022-09-03 14:52:54 are you in a dbus session 2022-09-03 14:53:15 yes, plasma launched it afaik 2022-09-03 14:54:35 echo $DBUS_SESSION_BUS_ADDRESS 2022-09-03 14:55:09 Not set 2022-09-03 14:55:29 I guess export $(dbus-launch) ? 2022-09-03 14:56:26 no, the way your wm starts is wrong 2022-09-03 14:57:18 ok, will investigate 2022-09-03 14:58:16 no idea how you start it but for me for instance on wayland it's dbus-run-session -- sway 2022-09-03 14:58:22 with plasma i assumed something already did that for you 2022-09-03 14:58:32 with a login manager i especially have no idea if one is being used 2022-09-03 14:58:44 but if you are typing some version of startx you can throw that in where appropriate 2022-09-03 17:44:15 how is it webkit2gtk have zero patches while other browser stuff like firefox and qt5-qtwebengine require a great deal of patching? is webkit2gtk that good at portability? Or cases requiring patching just haven’t surfaced yet, but in theory there are as many of them as in other browsers? 2022-09-03 17:45:19 it is that good at portability, yes 2022-09-03 17:45:40 there used to be a few musl patches or whatnot, but everything was upstreamed 2022-09-03 17:48:32 cool, it tells a lot about the project. worth trying to use something based on it 2022-09-03 17:55:44 it's what safari uses, just ported to gtk 2022-09-03 17:56:00 so you can expect a similar level of quality (nothing works) 2022-09-03 17:56:27 it's pretty good but every time i try it there's random minor bugs because nobody tests things on safari to begin with 2022-09-03 17:59:28 epiphany (gnome web) uses it, one of the browsers you can install pwa's with, which is convenient if they work in it 2022-09-03 18:04:09 well, i'm quite tired of qutebrowser's sluggishness. it's time to try out similar browsers based on webkit2gtk 2022-09-03 18:07:46 qtwebengine is chromium based so i have bad news for you about the speed of things based on webkit2gtk, haha 2022-09-03 18:08:48 it's mostly ok, and jscore is faster than firefox's js engine even, but everything else is not great 2022-09-03 18:12:36 qutebrowser is so slow on my machine. it seems to me that qutebrowser is slow due to the fact that it is written in python - rendering hints, searching history - all slow. i endured it for a long time, but i can no longer. the last straw was that i learned about patchwork for qt5-qtwebengine, and about the ease of webkit2gtk 2022-09-03 18:22:02 well, there's always netsurf :) 2022-09-03 18:24:14 afaik it doesn't support modern bloats of the web which makes it not very suitable for daily driver( 2022-09-03 18:38:28 true it does not, but be the change you want to see in the world! 2022-09-03 18:38:33 (mild sarcasm) 2022-09-03 18:49:01 forgot to say -- netsurf's goal (or, one of the goals) is to support all relevant web standards, so in theory it will inevitably get some of the stuff that it doesn't currently have, hopefully in a less bloaty way. but it's the absence of that support which incentivizes me (personally) to use netsurf today 2022-09-03 18:49:52 it does have some js support but for things like ecommerce/etc it usually fails on 2022-09-03 18:50:52 i hope that when they fill out the html5/full web stack stuff, that it could be switchable at build time. 2022-09-03 19:10:32 there should be a mandatory link to Drew's article on complexity of modern web 2022-09-03 20:27:49 Hi 2022-09-04 01:04:25 how can I make use of musl-dbg to debug a segfault in ld-musl? 2022-09-04 01:07:49 use a debugger and print a backtrace 2022-09-04 01:09:08 if i had to bet money i would bet the segfault is not actually in 'ld-musl', even if that's the last frame 2022-09-04 01:09:45 psykose: there isn't anything helpful in the stacktrace so I wanted debug symbols or something which I hoped musl-dbg would provide 2022-09-04 01:10:10 you need debug symbols for the actual application too 2022-09-04 01:10:11 I agree that it wouldn't be in ld-musl but that's the only thing I have to go off of at the moment 2022-09-04 01:10:45 unfortunately it is a proprietary application, so no debug symbols there 2022-09-04 01:11:03 also it goes through gcompat, should I compile a debug version of that? 2022-09-04 01:11:22 i doubt that would help in any way 2022-09-04 01:11:47 unless it's a bug in gcompat in some fashion 2022-09-04 01:12:07 is this a 'segfault doing something in the application' or merely starting it "segfaults" instantly 2022-09-04 01:12:40 if it's the latter then it would most likely be some incomplete thing in gcompat, but it's not really possible for me to guess anything 2022-09-04 01:19:47 it's the former, basically when one client is on the server it seems to be ok, but as soon as a second one joins it segfaults instantly 2022-09-04 01:25:10 sounds like it would be thread/fork related then 2022-09-04 01:25:23 which most definitely does not play nice with gcompat things, haha 2022-09-04 01:25:56 the amount of those that 'work fine on glibc' but don't on musl (even if you had the code) is quite high 2022-09-04 02:42:03 hihi, trying to install the extended iso following the directions per the semi-automatic installation wiki page. in particular my current issue is that I can't run ethernet to the system in question right now but I can't get any other available interface besides eth0 when running setup-interfaces 2022-09-04 05:11:42 How release system works in alpine? 2022-09-04 07:09:32 capybara_[m]: what information are you interested in? 2022-09-04 07:31:07 ikke: I was told a lot that on some distros with stable release model will result in not patched vulnerabilities for months. I want to try to understand if it is distro specific or the model itself is insecure. Were there issues with this on alpine? Actually I just want to find for me a distro for server purposes. 2022-09-04 07:31:54 capybara_[m]: security fixes of software in the repositories are not tied to releases 2022-09-04 07:32:25 Releases are mostly for the iso's and docker images 2022-09-04 07:33:27 We aim to quickly update packages with known and fixed vulnerabilities 2022-09-04 07:33:34 So releases are just fully stable snapshots of repositories? 2022-09-04 07:33:45 Not even that 2022-09-04 07:33:49 And usually everyone is in rolling release? 2022-09-04 07:34:22 The repositories for stable releases get updates, but mostly only bug / security fixes 2022-09-04 07:34:27 edge is our rolling release 2022-09-04 07:35:39 ikke: how to switch between those release models? 2022-09-04 07:36:05 It's decided by which repositories you have enabled in /etc/apk/repositories 2022-09-04 07:36:12 https://dl-cdn.alpinelinux.org/alpine/ 2022-09-04 07:36:26 There you sse the repositories of all releases and edge 2022-09-04 07:43:01 ikke: is it possible to install alpine on btrfs or zfs? 2022-09-04 07:44:27 I am actually surprised why such security focused project is so … why there is no guide to install alpine the gentoo way? I relatively know how to do it but it is the only way to get a custom bootloader and file system. 2022-09-04 07:45:56 * not surprised but disappointed 2022-09-04 08:04:15 capybara_[m]: Because it takes a lot of effort to write such a guide and keep it up-to-date 2022-09-04 08:05:04 capybara_[m]: you can use the installer, but select none for disk 2022-09-04 08:05:18 then setup btrfs or zfs, and then run setup-disk 2022-09-04 08:05:52 ikke: why do you guys still use grub and not efistub? _simplicity_ 2022-09-04 08:07:15 To support non-efi systems I suppose 2022-09-04 08:10:24 because efistub doesn't solve or accomplish anything 2022-09-04 08:12:48 it's just kicking the can down the road and pretending it's simple 2022-09-04 08:13:23 how do you change the kernel cmdline at boot? you don't, unless your $shittymotherboardefifirmware lets you 2022-09-04 08:13:27 same for all the other booting things 2022-09-04 08:13:54 better to have some semi consistent booting thing instead of 500 variants of awful motherboards to do something as basic as select a second boot option 2022-09-04 08:14:57 of which most can't even handle cmdlines correctly in the first place in my experience, either truncating them somewhere, or something explodes for no reason.. 2022-09-04 11:00:24 hey, is gcompat package broken? i'm trying to use the .net grpc build tool which is glibc, but it throws a really weird error 2022-09-04 11:00:27 > terminate called after throwing an instance of 'std : error : :system_error' 2022-09-04 11:00:33 > what() : error : No error information 2022-09-04 11:00:44 let's say it has limited compatibility 2022-09-04 11:01:47 ah 2022-09-04 11:03:24 what is the .net grpc build tool 2022-09-04 11:04:10 it's just called Grpc.Tools 2022-09-04 11:09:03 believe it or not 2.46 from https://github.com/grpc/grpc-dotnet builds fine, haha 2022-09-04 11:09:20 the examples don't but the actual dotnet-grpc does 2022-09-04 11:12:41 hmm fairly sure the build tool itself is in grpc/grpc, not the dotnet repo 2022-09-04 11:12:57 i'm not actually using grpc, just the protobuf build tool 2022-09-04 11:13:12 wouldn't know, couldn't find anything else, since nobody linked anything 2022-09-04 11:16:09 no, that says it was moved to the repo above 2022-09-04 11:19:14 i did some research and the protobuf compiler is still in the main repo here , the grpc-dotnet repo contains the runtime 2022-09-04 11:19:20 i'm trying to build compiler myself rn 2022-09-04 11:21:32 is there a way to make setup-disk not install bootloader? 2022-09-04 11:25:38 what is it named 2022-09-04 11:25:51 the protobuf compiler 2022-09-04 11:26:33 ifreund: I don't think so 2022-09-04 11:27:48 psykose: "protoc", it's in third_party/protobuf according to 2022-09-04 11:27:54 you can just install protoc 2022-09-04 11:28:10 apk add protoc 2022-09-04 11:28:10 wow 2022-09-04 11:28:11 oh i didn't think of that 2022-09-04 11:28:27 (it also builds fine) 2022-09-04 11:28:41 (and doesn't have anything to do with .net, funnily) 2022-09-04 11:29:35 yeah, i'm remembering all of this now 2022-09-04 11:29:43 been a bit since i did stuff directly with protobuf 2022-09-04 11:30:07 i guess what i have to do is symlink system protoc into .net package manager cache, lol 2022-09-04 11:30:31 lol that worked 2022-09-04 11:31:15 ikke: yeah, I think I might be better off ignoring setup-disk and using a manual chroot per https://wiki.alpinelinux.org/wiki/Alpine_Linux_in_a_chroot 2022-09-04 11:31:27 you can also delete the installed bootloader after 2022-09-04 11:32:15 true, I think I also need to figure out how to get zfs in the initramfs though 2022-09-04 11:32:51 trying to get an alpine install bootable through ZFSBootMenu 2022-09-04 12:53:35 do i need to do anything other than to modprobe simpledrm in order to get graphics working on the initramfs? when i boot the iso on a VM, it works perfectly, but when i do it on my real machine, it doesn't say much, gets a little buggy and freezes 2022-09-04 12:55:13 it says something like: 2022-09-04 12:55:15 [ 1.791924] simple-framebuffer simple-framebuffer.0: [drm] drm_plane_enable_fb_damage_clips() not called 2022-09-04 12:55:17 [ 1.793605] Console: switching to colour frame buffer d 2022-09-04 12:55:34 and then it gets glitchy and doesn't show much 2022-09-04 12:55:47 user1: add simpledrm to /etc/mkinitfs/mkinitfs.conf 2022-09-04 12:57:06 the first error also shows on my normal alpine install, so shouldn't be something to worry about i guess 2022-09-04 12:57:28 i already have it on the initfs features (i'm creating the initramfs manually) 2022-09-04 12:57:34 INITFS_FEATURES="ata base bootchart cdrom ext4 mmc nvme raid scsi squashfs usb virtio simpledrm" 2022-09-04 13:01:10 i tried modprobing simpledrm and opening a shell right after installing the busybox symlinks, it doesn't get to the shell (only on the virtual machine) 2022-09-04 13:01:32 there's no simpledrm feature tho and it's always moved 2022-09-04 13:01:48 sounds like your gpu is buggy or some other modules are missing 2022-09-04 13:39:14 i'm trying to mount a newly created LVM partition on the alpine standard image but get the error 'mount: mounting /dev/vg0/root on /mnt failed: Invalid argument' when using 'mount /dev/vg0/root /mnt'. is my command wrong? 2022-09-04 13:40:37 you probably need -t for the fstype 2022-09-04 13:43:15 what type would that be? '--help' doesn't list the FSTYPEs 2022-09-04 13:43:32 the type you formatted /dev/vg0/root as 2022-09-04 13:43:54 blkid would tell you 2022-09-04 13:43:56 i think 2022-09-04 13:44:02 ah, so not a special type for lvm 2022-09-04 13:44:08 thanks psykose, that worked 2022-09-04 13:44:23 yeah, busybox doesn't do autoprobe (unless you add that same type into /etc/filesystems) 2022-09-04 13:44:36 even then it's a little subpar- it modprobes the entire list sequentially until it (possibly) works 2022-09-04 13:45:08 if you apk add mount (which is util-linux mount) it works as you would expect with the usual magic 2022-09-04 13:45:29 i see, that's good to know 2022-09-04 13:48:16 https://lists.busybox.net/pipermail/busybox/2021-November/089343.html fixes it, but of course it was ignored 2022-09-04 13:53:02 'setup-disk' runs correctly but gives me the message 'You might need to fix the MBR to be able to boot'. what tool would i use for that? 2022-09-04 13:59:57 fdisk 2022-09-04 15:20:15 thanks 2022-09-04 20:09:16 i'm trying to install alpine linux on a uefi-enabled qemu guest with lvm on luks, but i'm unable to boot. steps followed are https://docs.alpinelinux.org/user-handbook/0.1a/Installing/manual.html, with a 512MB unencrypted /boot partition using grub as my bootloader 2022-09-04 20:10:29 grub was installed using setup-disk with the env vars DISKLABEL=gpt BOOTLOADER=grub USE_UEFI=y 2022-09-04 20:12:19 the error i'm getting while booting: https://0x0.st/opFZ.png 2022-09-04 20:16:51 i'm not sure what other info is needed so please ask if you need more 2022-09-04 20:41:14 Good eventing all, I looked at the wiki and I saw something about APK supporting "Plugable Layouts", but nothing in the code referencing them 2022-09-04 20:41:16 https://wiki.alpinelinux.org/wiki/TODO:apk 2022-09-04 20:41:19 Here's the page 2022-09-04 20:41:57 Is this something that might be added or? 2022-09-04 21:27:53 witcher01: do you have separate ESP and /boot partitions? 2022-09-04 21:34:34 minimal: i'm not sure - i mounted partition 1 at /boot and ran setup-disks. i guess that means no? 2022-09-04 21:34:50 or /mnt/boot, i guess. so /boot in the chroot 2022-09-04 21:42:25 i'm off to sleep, i'll reply to some suggestions/questions in some hours, if there are any. thanks in advance! 2022-09-04 23:00:10 hi 2022-09-04 23:01:28 im trying to install alpine and trying keyboards layouts for my language and none give me the correct layout? 2022-09-04 23:01:32 weird right? 2022-09-04 23:02:26 whats the best approach? 2022-09-05 04:46:53 Seect US layout and change it after installation 2022-09-05 06:10:17 I was going to ask how to delay a script until time sync using ntpd, but I figured it out right after joining 2022-09-05 06:10:58 the issue was that I was trying to use `ntpd -nq` but the ntp pool needs to be specified, so `ntpd -nqp pool.ntp.org` works perfectly 2022-09-05 08:22:53 does alpine-base depend on bash? it seems that apk thinks so 2022-09-05 08:23:06 World updated, but the following packages are not removed due to: 2022-09-05 08:23:08 bash: todo.txt-cli pass pm-utils wireguard-tools alpine-base tlp 2022-09-05 08:25:22 our base system never depends on bash 2022-09-05 08:25:39 yeah I suppose that, so apk is confused by somehting? 2022-09-05 08:26:01 apk will wrongly show packages as depending on if you pass them in cli 2022-09-05 08:26:38 uhM, I just did: apk del -s bash 2022-09-05 08:28:37 I don't remember the details but there was something wrong with how apk display packages when removing something 2022-09-05 08:29:31 you can try apk info --rdepends bash 2022-09-05 08:29:54 oh thanks 2022-09-05 08:30:06 4 packages instead 6 2022-09-05 10:34:48 Hi. I want to enable fq_codel on my Alpine machines. How do I do this best way. Is there something in /etc/networking/interfaces I could do to enable the qdisc, or should I make a simple shell script to enable it? 2022-09-05 10:36:22 Forza: You can always use a *up command in /etc/network/interfaces 2022-09-05 10:37:06 Ah. That is a good idea 2022-09-05 10:53:20 So something like this:? up tc qdisc add dev $IFACE root fq_codel 2022-09-05 10:54:09 Or probably it should be 'post-up' instead of 'up' 2022-09-05 10:54:41 yeah 2022-09-05 10:56:27 I'll try 2022-09-05 11:00:43 Seems to have worked correct :) 2022-09-05 11:01:10 nice 2022-09-05 11:01:52 Perhaps I should do test -x /sbin/tc && tc.... To check for tc first 2022-09-05 11:02:39 post-up test -x /sbin/tc && /sbin/tc qdisc add dev $IFACE root fq_codel 2022-09-05 11:05:11 This worked too. Better like that I think. Thanks again. 2022-09-05 14:44:15 has anyone tried using testing/stack (haskell tool)? `stack build` fails on `make install` with "not found", probably because utils/ghc-cabal/dist-install/build/tmp/ghc-cabal seems to be linked against libHS* and libffi.so.7 which ldd doesn't find 2022-09-05 14:50:56 vyivel: do you have ghc installed? 2022-09-05 14:52:06 nope, but it should be installed by stack locally, no? 2022-09-05 14:54:14 apk add ghc 2022-09-05 14:58:42 nothing changed 2022-09-05 14:59:08 iiuc i shouldn't really install ghc as it kinda ruins the point 2022-09-05 14:59:35 i think stack can be configured to use system ghc but ehhh 2022-09-05 15:00:35 well 2022-09-05 15:00:53 libffi.so.7 doesn't exist (libffi is .so.8), as for the other no idea 2022-09-05 15:01:10 i'm not sure how stack would work with random ghc's downloaded off the internet, as i imagine most of them aren't usable on musl to begin with 2022-09-05 15:01:14 and if they are, you now have that issue 2022-09-05 15:02:53 iirc stack has ghc musl builds 2022-09-05 15:02:53 these specific-compiler-version build tools like bazel don't work on musl distros because nobody makes statically-linked compilers or compilers linked against musl, so most of the time the things they download can't run at all 2022-09-05 15:03:10 in this specific case, the issue is probably just the dynamically linked lib versions, if they have musl builds 2022-09-05 15:03:34 (that .so.7 instead of .so.8 2022-09-05 15:03:36 ) 2022-09-05 15:03:53 why people distribute non-static versions and expect them to work is beyond me though 2022-09-05 15:05:58 https://downloads.haskell.org/~ghc/latest/ 2022-09-05 15:06:06 >alpine3_12 2022-09-05 15:06:08 meh 2022-09-05 15:06:19 that explains .so.7 2022-09-05 15:06:39 yeah, <=3.13 has .so.7 2022-09-05 15:06:54 they are not even distributing for the oldest supported version 2022-09-05 15:08:14 actually i downloaded one of those (ghc-9.4.1-x86_64-alpine3_12-linux-static) and the compiler is fully static 2022-09-05 15:08:40 where do you get the ghc-cabal 2022-09-05 15:08:59 good question 2022-09-05 15:10:23 >Downloading from http://downloads.haskell.org/~ghc/9.0.2/ghc-9.0.2a-x86_64-fedora27-linux.tar.xz to /home/kira/.stack/programs/x86_64-linux/ghc-tinfo6-9.0.2.tar.xz 2022-09-05 15:10:26 stack what are you doing 2022-09-05 15:10:42 its best, or even not 2022-09-05 15:11:28 haha 2022-09-05 15:11:32 fedora 27 lmao 2022-09-05 15:11:40 what's next, pulling a ghc compiler for VAX 2022-09-05 15:16:55 Where compiler for SVR4 2022-09-05 15:31:01 lmao 2022-09-05 15:31:12 https://git.alpinelinux.org/aports/tree/testing/stack/cabal.config lists fedora-dists and fedora-haskell-tools 2022-09-05 15:34:03 afaik that's for the build of stack itself 2022-09-05 15:34:06 but i wouldn't know 2022-09-05 15:35:32 like it's that literal lts haskell list https://www.stackage.org/lts-19.20 2022-09-05 15:36:01 unless i'm not looking where i should stack packages in other distros don't have a single mention of fedora 2022-09-05 15:36:21 fedora-dists Library for Fedora distribution versions 2022-09-05 15:36:24 i have no idea then why it thinks that alpine edge = fedora 27 2022-09-05 15:36:36 that's just a cabal freeze list 2022-09-05 15:48:12 ah, fedora 27 choice comes from stack-setup-2.yaml 2022-09-05 15:56:47 success! with `stack setup --ghc-build=musl`, and faked /usr/lib/libtinfow.so.6 (not libtinfo.so.6), it seems to work 2022-09-05 16:39:20 the w there is because we only have the w variant proveded as the default 2022-09-05 16:39:44 or was it 2022-09-05 16:39:44 hm 2022-09-05 16:40:06 or perhaps we don't have it at all and i am remembering something else 2022-09-05 19:07:04 when I try to restart lxd I got : umount: can't unmount /sys/fs/cgroup/systemd: Resource busy same issue if cgroups is started or not any clue ? 2022-09-05 19:30:27 it tries to unmount that on restart 2022-09-05 19:30:34 which is probably not possible due to something using it 2022-09-05 19:31:02 i don't think anyone thought that through when writing it :) 2022-09-05 19:31:14 you'd have to delete the stop_post in it 2022-09-05 20:10:10 could it be changed to use a lazy unmount? 2022-09-05 20:18:08 probably 2022-09-05 20:43:16 Hello, I was advised to ask here: I'm running musl version of Void Linux, and I have a lot of trouble with Element, as in I manage to turn it on like 1 out of 10 times, otherwise it just crashes 2022-09-05 20:43:59 I tried to run it with `--inspect`, did not find anything particularly useful, besides it giving a bunch of `FILE_NOT_FOUND` errors 2022-09-05 20:47:03 I don't think channel for Alpine Linux is appropriate place to ask about Void Linux issues 2022-09-05 20:52:37 what element 2022-09-05 20:52:58 element-desktop I presume 2022-09-05 20:53:15 "I don't think channel for Alpine..." <- fair enough, wanted to have a try though because the issue has been bothering me for qutie a while now 2022-09-05 20:53:18 s/qutie/quite/ 2022-09-05 20:53:22 panekj: correct 2022-09-05 20:53:26 > <@_oftc_panekj:matrix.org> element-desktop I presume 2022-09-05 20:53:26 * correct, the matrix client 2022-09-05 20:53:37 anything that uses electron will be a pain 2022-09-05 20:53:47 where did you install it from 2022-09-05 20:54:03 (from voidlinux pkg repo I presume) 2022-09-05 20:54:06 it's probably the ancient electron that void has 2022-09-05 20:54:06 on Void Linux, just through the package 2022-09-05 20:54:16 void has 19 2022-09-05 20:54:29 ptrc: it's actually electron19 2022-09-05 20:54:30 oh, that's news to me 2022-09-05 20:54:30 not sure how anyone here is meant to know the issue though, this isn't void linux 2022-09-05 20:54:47 (i said that) 2022-09-05 20:55:06 the thing is, I'm not sure it's specific to Void Linux, and not even sure it's musl per se, because I know two people on Arch who have a similar issue (and use glibc) 2022-09-05 20:55:14 did anyone from void tell you to try flatpak element first before sending you to other distros 2022-09-05 20:55:22 but maybe there is some musl phenomenon known in the lore 2022-09-05 20:55:25 invoked: no 2022-09-05 20:55:30 I'd rather not 2022-09-05 20:55:32 if it's happening on glibc then talk to element upstream? 2022-09-05 20:55:37 ACTION shrugs 2022-09-05 20:55:48 I tried to search for issues on github but didn't find aything 2022-09-05 20:55:50 invoked: there is a stigma with flatpak, like with systemd 2022-09-05 20:55:57 moltylollipop[m]: this was mainly my intention 2022-09-05 20:55:59 you can open issues btw 2022-09-06 02:01:39 Anyone have USB Type C working under their Alpine Edge/LTS setup? Intel host controller... Just trying to see where I need to chase stuff down 2022-09-06 02:14:24 Saijin_Naib: do you have the type-c related kernel module(s) loaded? 2022-09-06 02:18:45 my type c ports always worked normally with nothing done 2022-09-06 02:19:04 including type-c displayport, funnily 2022-09-06 02:50:05 minimal: no. I was hoping a reboot and the OpenRC hardware detect service at boot would autoload 2022-09-06 02:50:32 So, gotta manually modprobe and write to a file? 2022-09-06 02:51:09 psykose: is your controller Intel as well? Maybe it is another pain in the ass quirk of this reference platform my laptop is built around 2022-09-06 02:51:27 probablee 2022-09-06 02:51:31 random dell laptop 2022-09-06 02:51:52 there's always the usual list of lsusb/lspci etc etc stuff to see if it even exists detectedly 2022-09-06 03:02:48 Did the ls* song and dance, it provides power to connected devices, no dmesg complaints about missing firmware 2022-09-06 03:03:49 I'll try to diff modules between Alpine and a LiveUSB to see what's missing. Should auto-detect be working? Yeah, right? 2022-09-06 03:04:44 most likely yeah 2022-09-06 03:09:39 Sweet. Hopefully I'll be able to track this down and open an MR soon 2022-09-06 05:45:24 Hello! 2022-09-06 05:49:08 what's up 2022-09-06 11:57:19 given the following simple posix shell program: 2022-09-06 11:57:22 unset X ; echo ${X-~} ; echo ${X-~ } ; echo ${X- ~} 2022-09-06 11:57:35 the first echo behaves differently -- why? 2022-09-06 11:58:16 I just can't figure out based on the specification 2022-09-06 11:58:27 (dash, ash, bash do the same) 2022-09-06 11:59:00 you mean the first one expands ~? 2022-09-06 11:59:09 y 2022-09-06 12:04:34 interaction between the stages of shell processing of the line, I bet 2022-09-06 12:04:38 aron: I assume it is because the tilde "may need to be quoted under certain circumstances" 2022-09-06 12:05:30 the tilde is listed a character that "may be special depending on conditions described elsewhere in this volume of POSIX.1-2017" 2022-09-06 12:05:41 well, I don't have to fix this, I'm just interested in the reason behind the odd behavior 2022-09-06 12:06:01 aron: if you quote the ${X} expansions, you'll note that the space is included in the expansion 2022-09-06 12:06:20 tilde expansion takes place before the space is stripped off 2022-09-06 12:06:33 ' ~' fails because ~ is not first 2022-09-06 12:06:49 so in line with the above from POSIX docs, echo ${X-'~'} works like the other 2 2022-09-06 12:06:53 '~ ' fails because there's no user named ' ' 2022-09-06 12:07:13 dwfreed: yes, and no, since tilde is at the beginning of a word 2022-09-06 12:07:30 the shell definition of 'word' and the human definition of 'word' differ 2022-09-06 12:09:26 parameter/variable expansion does not create more words, only brace expansion, word splitting, and pathname expansion can do that 2022-09-06 12:09:36 so ' ~' is still a single word 2022-09-06 12:10:38 hmm 2022-09-06 12:11:34 this should be 2022-09-06 12:11:39 thanks 2022-09-06 12:12:13 this is documented in the bash manpage, but you have to do a lot of reading of it to piece it all together 2022-09-06 12:12:37 I already read the whole stuff a couple of times 2022-09-06 12:12:52 not bash's manpage, but the posix spec 2022-09-06 12:13:27 there are a lot of tiny pieces everywhere :P 2022-09-06 15:27:28 Hi, it looks like xrdp is busted; wanted to run this by to see if someone could double-check to make sure. I'm using 3.16; the package doesn't come with xrdp/xorg.conf (/etc/X11/xrdp/xorg.conf), and that appears to be required when launching a session. Furthermore, xrdp's Xorg drivers aren't available (xrdpkeyb, xrdpmouse), so taking a working xrdp/xorg.conf from another 2022-09-06 15:27:30 environment is insufficient. 2022-09-06 15:28:30 (I've got xrdp-0.9.18.1-r0 installed) 2022-09-06 16:14:14 Noted failure to start xrdp services on Edge a while back, no idea how to solve it. Gave up and used the quite miserable VNC protocol instead 2022-09-06 16:14:29 Do you get any logs or anything when you try to run it? 2022-09-06 17:10:23 Saijin_Naib: Yes. Logs in /var/log/xrdp-* and in ~/.xrdp*. 2022-09-06 17:11:20 The prior lead me to discovering that Xorg wasn't starting correctly -- it was trying to load regular Xorg config and drivers, which results in it trying to access VT8, which it doesn't have access to. 2022-09-06 17:11:49 (VT7 is currently in use by my Xorg instance for the real graphics console) 2022-09-06 17:13:16 If you mkdir /etc/X11/xrdp/ and then copy xorg.conf into that directory from a distribution with a working xrdp (such as raspbian), that fixes the problem with xrdp trying to use VT8 and trying to load KMS drivers and such for the real hardware 2022-09-06 17:13:39 (as it is no longer trying to auto-detect the hardware config and trying to use the xorg.conf for xrdp) 2022-09-06 17:14:42 however, the Xorg plugins (modules? sorry, it's been a while since I've messed with Xorg and linuxy things) that xrdp uses to drive IO on Xorg are missing in the xrdp package. 2022-09-06 17:15:09 i.e. xrdpkeyb and xrdpmouse are missing 2022-09-06 17:15:46 I did sudo find / -iname '*xrdpkeyb*' and sudo find / -iname '*xrdpmouse*', and both yielded nothing on my alpine box 2022-09-06 17:17:01 on Raspbian, I get this: 2022-09-06 17:17:03 /usr/lib/xorg/modules/input/xrdpkeyb_drv.so 2022-09-06 17:18:05 so there should be something similar, presumably, on alpine; but I've reinstalled the package and checked the package manifest for xrdp, and nope, no *_drv.so's in there 2022-09-06 17:18:10 jdestroy: have you installed xorgxrdp ? 2022-09-06 17:18:39 kwadroke[m]: i'll try that right now, thanks 2022-09-06 17:19:05 I just spun up a 3.16.2 VM and I have the xrdpkeyb.so file 2022-09-06 17:19:22 kwadroke[m]: no such file 2022-09-06 17:19:27 *no such package 2022-09-06 17:19:34 Are you on Alpine edge? 2022-09-06 17:19:36 I installed xrdp & xorgxrdp 2022-09-06 17:19:47 no just 3.16.2 2022-09-06 17:19:53 hmmm 2022-09-06 17:20:12 macmini:~$ sudo apk add xorgxrdp 2022-09-06 17:20:14 I do have the edge repos enabled, but didn't have to specify them 2022-09-06 17:20:14 ERROR: unable to select packages: 2022-09-06 17:20:16 xorgxrdp (no such package): 2022-09-06 17:20:18 required by: world[xorgxrdp] 2022-09-06 17:20:24 did you enable community repo 2022-09-06 17:20:30 yes 2022-09-06 17:20:35 I did enable community 2022-09-06 17:20:36 not asking you 2022-09-06 17:21:11 might be why I have it and jdstroy doesn't 2022-09-06 17:21:24 macmini:~$ cat /etc/apk/repositories 2022-09-06 17:21:26 http://dl-cdn.alpinelinux.org/alpine/v3.16/main 2022-09-06 17:21:28 http://dl-cdn.alpinelinux.org/alpine/v3.16/community 2022-09-06 17:21:32 I have main and community 2022-09-06 17:21:36 did you update the index 2022-09-06 17:21:38 yes 2022-09-06 17:23:09 did sudo apk update && sudo apk add xorgxrdp 2022-09-06 17:23:23 first was successful, second gave 2022-09-06 17:23:25 xorgxrdp (no such package): 2022-09-06 17:24:36 it does come from community per pkgs.alpinelinux.org 2022-09-06 17:25:27 what `apk policy xorgxrdp` shows 2022-09-06 17:25:41 are you running x86 (32bit) or x86_64 (64 bit), or some other CPU arch? 2022-09-06 17:26:02 x86 (32-bit) 2022-09-06 17:26:04 aha 2022-09-06 17:26:06 that's why 2022-09-06 17:26:12 that's why 2022-09-06 17:26:12 no 32-bit builds of it 2022-09-06 17:26:51 thanks. that clears up that mystery 2022-09-06 17:27:07 well, guess I'll uninstall xrdp from this machine 2022-09-06 17:27:19 and resort to using janky ol' vnc 2022-09-06 17:27:48 yeah, it's the only cpu that is disabled in the APKBUILD for it 2022-09-06 17:28:20 not quite sure why though 2022-09-06 17:28:50 hmm 2022-09-06 17:29:06 maybe it doesn't build correctly on x86? seems... odd. 2022-09-06 17:29:29 seems odd to me too 2022-09-06 17:29:43 anyway, if it doesn't work on x86, probably there should be a similar disabling of xrdp builds on x86, too, because xrdp isn't functional without it. 2022-09-06 17:29:55 you can use XRDP with VNC 2022-09-06 17:30:02 what? 2022-09-06 17:30:03 instead of xorg 2022-09-06 17:30:22 Oh... that's interesting. 2022-09-06 17:30:34 Okay, I guess I'll play with that 2022-09-06 17:30:55 Thanks 2022-09-06 17:30:58 Learned something new today 2022-09-06 17:31:04 I usually don't. I prefer to use XORG, but I've gottent it to work in the past 2022-09-06 17:31:36 usually VNC is the default unless you install XORGXRDP 2022-09-06 17:32:24 there's an entry in the Alpine wiki for XRDP 2022-09-06 17:32:27 and VNC 2022-09-06 17:32:35 haven't tried it though 2022-09-06 17:32:52 https://wiki.alpinelinux.org/wiki/Remote_Desktop_Server 2022-09-06 17:35:16 jdstroy: would you be willing to open an issue on the aports for xrdp describing what you found to be missing in the Alpine build. 2022-09-06 17:59:54 looks like xorgxrdp builds fine on x86 (unless I missed something) 2022-09-06 18:00:52 Saijin_Naib: Sure. Will be happy to do that a little later after I get off work. 2022-09-06 18:01:09 I would like to be able to use xrdp on x86 :) 2022-09-06 18:02:07 12:32:52 https://wiki.alpinelinux.org/wiki/Remote_Desktop_Server 2022-09-06 18:02:29 Hah. I was actually trying to use that to see if I could enable xrdp/vnc, but it didn't do anything useful. 2022-09-06 18:02:40 I haven't looked for the logs for that yet 2022-09-06 18:03:06 but if xorgxrdp works, that'd be my preference and I won't bother with vnc 2022-09-06 18:58:02 I don't find why it was disabled on x86, any clue? 2022-09-06 19:17:18 it builded https://gitlab.alpinelinux.org/donoban/aports/-/merge_requests/new?merge_request%5Bsource_branch%5D=xorgxrdp-x86 2022-09-06 19:18:12 It was disabled 5 years ago https://gitlab.alpinelinux.org/alpine/aports/-/commit/d68a09e48ce7aead53315d740582a5edef1c37f7 2022-09-06 19:19:29 uhM, interesting but I found this 1 year ago 2022-09-06 19:19:51 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/18554 2022-09-06 19:42:18 I tested my build using the modified xorgxrdp on 32bit. Works fine 2022-09-06 19:43:08 well, it did when I modifed sesman.ini to change param=Xorg to param=/usr/libexec/Xorg 2022-09-06 19:47:36 kwadroke[m]: did you test on edge or 3.16? I suppose that it could be backported but maybe there is some policy against it 2022-09-06 19:47:46 3.16.2 2022-09-06 19:48:03 oh great 2022-09-06 19:48:19 there's another issue open for the /usr/libexec/Xorg entry 2022-09-06 19:48:45 it solved an issue I was having with x86_64 2022-09-06 19:49:21 Here's the issue: https://gitlab.alpinelinux.org/alpine/aports/-/issues/13021 2022-09-06 19:50:56 does the fix breaks other things' 2022-09-06 19:50:57 ? 2022-09-06 19:51:04 not that I can tell 2022-09-06 19:51:15 being it's for XRDP & Xorg 2022-09-06 19:51:37 change made in /etc/xrdp/sesman.ini 2022-09-06 19:52:05 uhM, I will try to add to the opened MR 2022-09-06 19:52:35 there's even instructions for other distros with simalar entries above that line 2022-09-06 19:56:14 ok 2022-09-06 19:58:15 that file is actually contents of the xrdp package, and not the xorgxrdp 2022-09-06 19:58:31 I saw, I'm adding a sed -i... to the APKBUILD 2022-09-06 19:58:49 I'm trying to push to my aports fork on gitlab but it just hangs. Is there some known ongoing issue? 2022-09-06 20:02:08 billiob: I pushed few minutes ago without problem 2022-09-06 20:02:19 kwadroke[m]: is this the right path? sed -i 's/param=Xorg/param=\/usr\/libexec\/Xorg/' sesman/sesman.ini 2022-09-06 20:03:24 ther's not s sesman folder. Full path is /etc/xrdp/sesman.ini 2022-09-06 20:03:34 well, it's the source tarball 2022-09-06 20:03:39 ok 2022-09-06 20:03:59 other than that's looks right so far 2022-09-06 20:04:41 might want to add a "^" to the front though 2022-09-06 20:04:57 ok 2022-09-06 20:04:58 so it won't change other lines such as comments 2022-09-06 20:05:05 wops, psykose already merged the MR 2022-09-06 20:06:09 you can open one for the configuration too 2022-09-06 20:06:14 ok 2022-09-06 20:06:23 should I point the gitlab issue on the commit? 2022-09-06 20:08:05 not a packager myself, but might not hurt 2022-09-06 20:08:48 !38706 2022-09-06 20:09:17 billiob: not aware of any issue. ssh or https? 2022-09-06 20:10:04 You probably want to change the sed line to this: sed -i 's/^param=Xorg/param=/usr/libexec/Xorg/' sesman/sesman.ini 2022-09-06 20:10:20 see the first entry for param= 2022-09-06 20:10:42 uhM 2022-09-06 20:10:53 it stripped the backslashes 2022-09-06 20:10:59 ouch 2022-09-06 20:11:15 when I posted 2022-09-06 20:11:30 I tested it directly and worked 2022-09-06 20:11:46 where did you see the wrong replacement? 2022-09-06 20:11:47 it will work, but it will change comments too 2022-09-06 20:12:05 ah, do you refer to the ^, I've already added it 2022-09-06 20:12:11 yeah 2022-09-06 20:12:25 I'm gonna check the result https://gitlab.alpinelinux.org/donoban/aports/-/jobs/830771/artifacts/file/packages/community/x86_64/xrdp-0.9.18.1-r1.apk 2022-09-06 20:12:30 make it look at the start of the line so comments will be ignored 2022-09-06 20:13:02 ouch, no sesman.ini there 2022-09-06 20:14:56 oh no, it's fine 2022-09-06 20:16:32 ikke: it just finished and was very slow 2022-09-06 20:16:34 yeah, the path looks good 2022-09-06 20:16:55 within the tarball anyway 2022-09-06 20:19:08 I'm gonna use the psykose suggestions with doesn't need scaped / 2022-09-06 20:19:32 definately cleans it up 2022-09-06 20:22:14 uhM 2022-09-06 20:22:40 I was --ammending nothing 2022-09-06 20:23:54 uhM, I push but the MR doesnt' change 2022-09-06 20:24:35 https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/38706/diffs now! 2022-09-06 20:24:46 I have to go, see you guys, thanks psykose ! 2022-09-06 20:24:51 ^>^ 2022-09-06 20:44:30 psykose: nice emoji 2022-09-06 20:44:36 :) 2022-09-06 20:44:56 blech 2022-09-06 20:45:01 this llvm release is even more fucked than the others 2022-09-06 20:45:20 time to bash my brains against the wall for 45 hours and rewrite everything to work 2022-09-06 21:04:16 Whats changed with it 2022-09-06 21:12:28 doesn't seem to be possible to built it split anymore 2022-09-06 21:12:37 which is ok, but then there's 50 things to fix to build it at once 2022-09-06 21:14:30 that, or the main llvm cmake folder is not where it should be again, they forgot that before 2022-09-06 21:14:30 meh 2022-09-06 21:14:38 i'll throw some hours at it when i feel like it 2022-09-06 21:21:08 Good luck 2022-09-06 21:24:39 ^ 2022-09-06 23:19:27 Thanks everyone for getting that sorted with xrdp! 2022-09-06 23:49:53 np 2022-09-07 11:40:17 Hi there. 2022-09-07 11:41:57 I have odroid xu4 which is armv7 cpu type. I would like to install zfs kernel modules and switch from current ext4 -> zfs for my backup hard drives on odroid. How can I eanble zfs module for standard image? Extended image (which has zfs support) only supports x86 and x86_64 2022-09-07 11:42:36 Or should I download generic arm image? armv7 ? 2022-09-07 11:42:44 And enable zfs for this type of image? 2022-09-07 11:43:01 Would like to have zfs support at boot if possible? 2022-09-07 11:48:23 what /etc/pam.d/ file is used on login? there doesn't seem to be an /etc/pam.d/login file as I'm used to on other systems 2022-09-07 11:49:15 I've tried adding my pam module to system-login but it doesn't seem to get run 2022-09-07 11:54:51 ok, I guess one needs to install util-linux-login 2022-09-07 11:55:04 I wonder where I was supposed to find that in the docs 2022-09-07 12:48:29 Hello =) What firewall do you recommend for alpine? I need only basic simple stuff to allow specific traffic/ports. I do not need routing or NAT. What do you think of awall or ufw? 2022-09-07 12:50:44 I use ufw personally, works just fine 2022-09-07 14:04:25 does anyone know why '/dev/cdrom' won't mount when i boot from my usb? when i pass the ISO directly to QEMU, it works. when I pass the USB drive with the iso burned into it, it says "mount: mounting /dev/cdrom on /media/cdrom failed: No medium found" 2022-09-07 14:05:14 i'm running: sudo qemu-system-x86_64 -enable-kvm -hdb /dev/sdb 2022-09-07 14:05:21 and this same problem happens on my real machine 2022-09-07 14:05:54 mount command: mount -t iso9660 /dev/cdrom /media/cdrom 2022-09-07 14:07:50 dd command to burn the iso: sudo dd if=out/sigma-linux-x86_64.iso of=/dev/sdb bs=4M && sync 2022-09-07 14:08:43 is ssh enabled on the standard alpine image? 2022-09-07 14:09:16 this is a custom build, this part happens during the initramfs 2022-09-07 14:09:36 Guest2195: I think if you add it as a USB storage, it is not a CD anymore, so /dev/cdrom won't work 2022-09-07 14:09:43 bl4ckb0ne: no 2022-09-07 14:10:30 Guest2195: qemu-system-x86_64 can do "-cd path/to/your/cd.iso" 2022-09-07 14:10:55 yes, but i use this command to simulate my real machine, which doesn't work 2022-09-07 14:11:35 also, i think that alpine standard's initramfs also mount /dev/cdrom 2022-09-07 14:13:12 it is absolutely necessary to access it as a CD? 2022-09-07 14:13:59 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/fstab 2022-09-07 14:14:06 the /dev/cdrom is on the fstab at least 2022-09-07 14:14:12 if yes, you need more specific qemu options, i.e.: -device file=/dev/sdb,if=ide,index=1,media=cdrom 2022-09-07 14:14:27 hah! if something is on the fstab, it does not mean it is mounted 2022-09-07 14:14:33 aron: it's where the iso contents are, but i don't really mind how it's accessed 2022-09-07 14:15:00 I start my VMs with "-cdrom alpine.iso", tho 2022-09-07 14:15:07 works every time 2022-09-07 14:15:28 so no /dev/sdb magic 2022-09-07 14:16:24 yes, it works if i pass the iso, but i'm simulating my real machine. this error also happens on my real machine, that's why i'm passing the USB to QEMU, so i can debug without restarting all the time 2022-09-07 14:16:44 Guest2195: do you see the "noauto" in that fstab? it means "do not mount it automatically" 2022-09-07 14:17:14 panekj: so there's no easy way to do a headless install 2022-09-07 14:17:21 "USB"? USB what? pendrive? 2022-09-07 14:18:03 yes 2022-09-07 14:18:19 and their initramfs also creates /media/cdrom, so it must be mounted right? 2022-09-07 14:18:45 and how that pendrive will turn into /dev/cdrom? it is a USB mass storage device, not a CD ROM 2022-09-07 14:19:02 the alpine installer ISO will do everything for you 2022-09-07 14:19:10 bl4ckb0ne: you should add apkovl file in your media for headless install 2022-09-07 14:19:34 well, since i burned the iso into the USB drive, i thought linux would make that connection 2022-09-07 14:19:41 Guest2195: no, it doesn't have to be mounted 2022-09-07 14:19:46 nice ill look it up 2022-09-07 14:19:47 thanks 2022-09-07 14:20:01 bl4ckb0ne: https://github.com/macmpi/alpine-linux-headless-bootstrap 2022-09-07 14:20:12 panekj: then where is initramfs pulling all the files from, such as the apk overlay 2022-09-07 14:20:16 if not from /dev/cdrom 2022-09-07 14:20:25 having headless img on the website would be nice 2022-09-07 14:21:03 Guest2195: no, you can write anything on a USB storage, it will always remain a USB storage -- until it stops working ;) 2022-09-07 18:52:48 Did i figure out this stuff right? 2022-09-07 18:52:52 seems so 2022-09-07 18:54:40 how can i make ash look prettier? rn its shows just the host name 2022-09-07 19:01:07 hello? 2022-09-07 19:21:08 PureTryOut: thanks 2022-09-07 20:56:10 how does Alpine deal with external glibc binaries? 2022-09-07 20:56:23 it doesn't 2022-09-08 01:37:25 Quick question about the community branches - if a library has a security hole and package X version a.b-1 depends on it statically 2022-09-08 01:37:36 I assume that'll require a rebuild as a.b-2 2022-09-08 01:38:12 but what does one do if edge already has a a.b-2 without the fix? 2022-09-08 01:39:32 basically would that e.g. rather become a a.b-2+3.3.0-1 or something like that? 2022-09-08 01:39:50 *a.b-1+3.3.0-1 2022-09-08 01:42:15 or does one then first create a a.b-4 on edge, and then create a a.b-3 on the release branch that is a.b-1 + the fix cherrypicked/patched? 2022-09-08 01:49:08 they have no connection 2022-09-08 01:49:35 doesn't matter what the pkgrel is in edge vs stable 2022-09-08 01:49:55 ah, that's how this is resolved 2022-09-08 01:50:26 does mean it's technically slightly dangerous to upgrade a system from one to the other (but usually fine) 2022-09-08 01:50:58 in what sense 2022-09-08 01:51:25 what if for some weird reason (rebuilds) stable has a.b-4 and edge has a.b-3 which actually is newer (yes, weird) 2022-09-08 01:51:45 then wouldn't the package stay at a.b-4 on the system after changing from release to edge? 2022-09-08 01:51:57 or does apk downgrade then 2022-09-08 01:52:00 no, after you change 3.16 to edge and do apk upgrade -a it will downgrade to -3 2022-09-08 01:52:21 if you have both 3.16 and edge repositories (lol) it will stay the same 2022-09-08 01:52:35 yeah 2022-09-08 01:52:51 arguably that's unlikely anyway, usually edge will be ahead by pkgver anyway 2022-09-08 01:53:20 unlikely and also 'unsupported' :p 2022-09-08 01:53:34 probably would require a package that hasn't chanegd upstream for a while AND for some reason more rebuilds on stable than on edge 2022-09-08 01:53:41 which I'd find rather odd 2022-09-08 01:55:53 quite peculiar indeed 2022-09-08 01:57:19 well, I guess it's also acceptable if a community package breaks directly after upgraing from community to edge - it's not needed for booting and is an easy fi 2022-09-08 01:57:21 x 2022-09-08 01:57:31 and also very theoretical 2022-09-08 01:58:31 but anyway, I understand pkgrel now :) 2022-09-08 01:58:50 thanks 2022-09-08 01:59:02 :) 2022-09-08 09:51:03 starting xorg with startx on alpine gives me the error 'libEGL.so.1: No such file or directory'. how do i fix this? 2022-09-08 09:51:35 apk add mesa-egl 2022-09-08 09:51:59 is this required by xorg to start in the first place? 2022-09-08 09:52:37 asking since i used setup-xorg-base 2022-09-08 09:52:38 easy way to find out by seeing if adding it fixes it 2022-09-08 09:52:53 it does fix it, yup 2022-09-08 09:52:59 ah, then you also want mesa-dri-gallium probably 2022-09-08 09:54:07 got it, thanks! 2022-09-08 10:01:19 and then either xf86-video-modesetting of xf86-video-yourgpuorsomethingidk, i forget, and i also forgot if gallium is even used at the same time 2022-09-08 10:01:23 you'll figure it out 2022-09-08 10:01:30 of->or 2022-09-08 12:15:01 Is alpine down? 2022-09-08 12:15:21 in what sense 2022-09-08 12:15:31 Trying to setup alpine under qemu, but keep getting bad address mirrors.alpinelinux.org 2022-09-08 12:15:48 Odd 2022-09-08 12:15:58 When checking manually I can access it just fine 2022-09-08 12:16:16 dns 2022-09-08 12:16:58 So I tried to add random mirror... 2022-09-08 12:17:15 What happens is: /sbin/setup-apkrepos: line 31: divide by zero 2022-09-08 12:17:21 Room for improvement 2022-09-08 12:17:42 That happens when it didn't get a repo list 2022-09-08 12:17:58 ikke: yeah not sure why it can't get a repo list 2022-09-08 12:18:17 Some network issue I suppose 2022-09-08 12:19:20 ur network is broke 2022-09-08 12:20:48 Yeah dns issue.. can't ping mirrors.alpinelinux.org from qemu for whatever reason ... and I remembered :s 2022-09-08 12:21:13 It probably tried to go through my lan card which is currently on some other network 2022-09-08 12:21:20 OK thank you guys! 2022-09-08 12:21:42 add more lan cards ;) 2022-09-08 13:14:30 hi, i have noticed that `swi-prolog` package has been introduced into the testing repository and it reports bsd clause 2 licence. SWI Prolog can use that licence if `gmp`(that is licenced as lgpl) is disabled, but looking at the build script, it is compiled with default args and `gmp` is included (if i read the cmakefiles correctly) 2022-09-08 13:15:32 (i'm just a user that happen to have Alpine and deal with `swi-prolog` recently, i'm not involved into any of those projects 🙃) 2022-09-08 13:17:08 this aspect of `swi-prolog` is described in their website: https://www.swi-prolog.org/license.html 2022-09-08 13:17:15 it is, but the final result is identical 2022-09-08 13:17:58 e.g. it says mit (even though its "not"), but then you install it and it pulls in gmp which has lgpl so to use anything you have to follow that licence anyway.. 2022-09-08 13:17:59 or something 2022-09-08 13:18:45 there's probably thousands of packages missing something like that by proxy via linkage 2022-09-08 13:19:06 nevertheless i amended it 2022-09-08 13:28:38 any environment variables set in `/etc/conf.d/` should be used by the service started with `/etc/init.d/` no? 2022-09-08 13:28:54 not without export 2022-09-08 13:29:29 i see, thanks a lot for the quick fix! 2022-09-08 13:29:51 x=1; ./thing doesn't put x in the env; in the same fashion sourcing a file with x=1 won't do that, but if you put export x=1 then sourcing the file exports it 2022-09-08 13:30:04 the conf.d is merely exported 2022-09-08 13:30:06 Uniformbuffer[m]1: np 2022-09-08 13:30:09 psykose: interesting. Can I just do `export VARIABLE=` in the conf? 2022-09-08 13:30:12 sourced* 2022-09-08 13:30:15 yeah 2022-09-08 13:30:31 it's just a sourced shell script, you can break whatever you wish in there too :) 2022-09-08 13:30:46 i forget if there is any special handling of it at all 2022-09-08 13:30:54 ah ok interesting 2022-09-08 13:36:51 it does seem to ignore `supervisor=supervise-daemon` still though 2022-09-08 13:37:30 what's the init.d 2022-09-08 13:37:56 that doesn't ever have to be exported 2022-09-08 13:39:45 ah custom one 2022-09-08 13:40:10 basically I'm getting pipewire to run as a system service. I'll upstream the init scripts when done 2022-09-08 13:40:38 show it 2022-09-08 13:41:16 uh sure 2022-09-08 13:41:19 ACTION sent a code block: https://matrix.org/_matrix/media/r0/download/matrix.org/QEeozvcgJPJHpVgLNcHNFAnG 2022-09-08 13:41:22 👆️ 2022-09-08 13:41:35 although line 8 was moved to conf.d 2022-09-08 13:41:41 (which just broke it) 2022-09-08 13:42:50 that would be identical, unless the conf.d is not named the same 2022-09-08 13:42:55 and so never gets loaded 2022-09-08 13:42:58 it is named the same 2022-09-08 13:43:05 can also fail for a bunch of other reasons 2022-09-08 13:43:12 what makes you think supervise-daemon is unused 2022-09-08 13:43:29 also this init script for some reason redirects the log/output to /dev/null. I have no clue where to get the service logs from 2022-09-08 13:44:01 because with supervise-daemon in conf.d rather than init.d, the services just fail to launch (they need to be sent to the background which without supervise-daemon doesn't happen in the current script) 2022-09-08 13:44:04 ah, you're missing pidfile="/run/$RC_SVCNAME.pid" and command_background=true 2022-09-08 13:44:16 those are only needed when not using a supervisor ime 2022-09-08 13:44:20 no 2022-09-08 13:44:47 no? 2022-09-08 13:45:51 nope even with that, just crashing 2022-09-08 13:46:04 or well not crashing, but OpenRC thinks it crashed 2022-09-08 13:46:18 according to ps the services are running fine still, just nothing is tracking it anymore 2022-09-08 13:46:47 i copied the file, added the two vars, added the export for runtime_dir, and it works 2022-09-08 13:47:15 with supervisor=supervise-daemon in `/etc/conf.d/pipewire`? 2022-09-08 13:47:18 yea 2022-09-08 13:47:23 wtf 2022-09-08 13:48:22 even more wtf, now it works here as well. And I changed basically nothing. Uh... debugging further 2022-09-08 13:48:25 it's possible something is already running to it crashes due to conflict 2022-09-08 13:48:54 ah I see what happens. OpenRC is unable to take the service down again. It says they are stopped, but ps reveals they are still running 2022-09-08 13:49:32 does pipewire not die on sigterm or something 2022-09-08 13:49:38 weir 2022-09-08 13:49:39 d 2022-09-08 13:49:40 I... Don't know 2022-09-08 13:50:03 although tbf it might be trying to kill dbus-run-session only 2022-09-08 13:51:13 ah 2022-09-08 13:51:17 yeah, it's that thing all over again 2022-09-08 13:51:18 hah 2022-09-08 13:51:26 i don't think you can use dbus-run-session there to begin with 2022-09-08 13:51:39 the other pipewire-pulse/wireplumber won't be in the same session 2022-09-08 13:52:02 dbus-run-session runs as a parent and not as a child so when you kill it it doesn't transfer as expected 2022-09-08 13:52:07 I can assure you without it, those other services won't connect to pipewire at all 2022-09-08 13:52:17 I used dbus-launch before, let's try that again 2022-09-08 13:53:42 hmm that does take pipewire down, but it keeps a dbus-daemon instance around 2022-09-08 13:55:30 although it does seem pipewire itself doesn't need dbus-launch, only services connecting to it do 2022-09-08 14:02:50 i'm curious, is there a reason for system-pipewire 2022-09-08 14:02:58 i don't think it does anything special 2022-09-08 14:03:19 PureTryOut: I am running pipewire as a system service for some time now, my config: http://okturing.com/src/14169/body 2022-09-08 14:04:28 (maybe helpful) 2022-09-08 14:06:47 Thanks. That's close to what I have, although I don't seperate runtime and state dir currently 2022-09-08 14:07:02 psykose: not particularly, just simpler to have it start with other services + multiple users can use same daemon 2022-09-08 14:13:09 https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image - alpine-make-vm-image doesn't seem like third party. Its repo is under umbrella of alpine linux organisation 2022-09-08 14:13:53 if it's on github, it's pretty much 3rd party since only jirutka maintains stuff on GH 2022-09-08 14:15:57 well ok 2022-09-08 14:18:36 psykose: in my case the only thing this machine does is be a speaker, so it being a system service makes more sense imo. It's also easier to keep track of the status of it 2022-09-08 14:18:44 yeah makes sense 2022-09-08 14:20:31 there is the problem that the speaker is capable of 5.1 surround, but over the network PipeWire only exposes stereo channels :( 2022-09-08 14:21:23 and although it does upmix it so sound comes from all the individual speaker modules, it's hard to fully understand/hear speech in 5.1 surround movies 2022-09-08 16:54:58 Hi! I wonder why the postgrey package is not in 3.16 - it was present in 3.3... 2022-09-08 16:55:53 It was removed due to licencing issues 2022-09-08 17:04:48 ah. 2022-09-09 04:02:27 hello 2022-09-09 04:02:32 can y'all see this? 2022-09-09 04:04:04 assuming yall can read my messages 2022-09-09 04:04:04 alright, so I'm on postmarketOS (downstream of Alpine, already asked in their channel but figured it made sense to ask here too) and installing PipeWire according to the instructions on the alpine linux wiki. That mentions an `rtkit` group. Is that group created when the package is installed or do I have to create it? 2022-09-09 04:04:23 that is when the rtkit package is installed 2022-09-09 04:04:49 i ran `groups` but that did not show an rtkit group. I added my user to the rtkit group, will that just work? do I have to also create the group? 2022-09-09 04:20:30 Can edge be compared with rolling release? 2022-09-09 04:23:02 capybara_[m]: yes and no 2022-09-09 04:23:22 Why no? 2022-09-09 04:24:36 packages / updates are not tested before being pushed to edge 2022-09-09 04:25:41 Are they pushed from “stable” newest version of that package that is provided by developers of that package? 2022-09-09 04:26:46 generally, yes, but sometimes also rc's 2022-09-09 04:27:36 So package may be broken 2022-09-09 04:27:53 yes 2022-09-09 04:28:02 Ok thank you 2022-09-09 04:30:50 ikke: wait but what is the point in edge testing repository then? 2022-09-09 04:31:06 that's for comletely new packages 2022-09-09 04:31:32 whatever is in testing does not end up in the next release 2022-09-09 04:46:53 my issue has been resolved 2022-09-09 04:47:02 well, partially. I still need to get pipewire working 2022-09-09 04:47:45 chexo4[m]: btw, if you add a user to a group, you need to login again before it's in effect 2022-09-09 04:47:56 noted 2022-09-09 04:47:58 if you were able to add your user, it means it exists 2022-09-09 04:47:59 thanks 2022-09-09 04:48:16 yeah I figured that out from being a dumbass but now I know how that works 2022-09-09 04:48:22 so that part is good I think 2022-09-09 12:49:38 Hi. I'm trying to make sway work with elogind (not seatd) under alpine linux, but with no success 2022-09-09 12:50:44 i have installed polkit, pam-rundir, elogind, dbus, sway and ran sway with "dbus-run-session sway", all to be greeted with "no sessions" 2022-09-09 12:51:25 I've tried enabling all these services manually (instead of them being triggered via dbus), but still doesn't budge 2022-09-09 12:51:53 any idea how to make elogind work for session management in alpine? 2022-09-09 13:07:42 :( 2022-09-09 13:07:47 chexo4: i wouldn't follow the guide after installing pipewire and setting up rtkit. Use wireplumber as your media session manager and it should just work. I never got the manual way from the guide to reliably work, especially when sources need to switch (plug headphones, connecting Bluetooth, etc) 2022-09-09 16:34:37 i try to mount restic with fuse... if i `modprobe fuse` i get this: "modprobe: FATAL: Module fuse not found in directory /lib/modules/5.15.51-0-lts" 2022-09-09 16:34:40 any ideas about this? 2022-09-09 16:35:31 fuse and fuse3 is installed 2022-09-09 16:39:32 oh a new kernel is installed but not rebootet 2022-09-09 16:50:37 :) 2022-09-09 17:08:14 (: 2022-09-09 19:33:20 I am not too familiar with X11 installs but I was trying an old laptop with neomagic graphics and noticed that there is no xf86-video-neomagic package? Wasn't sure if something else was expected to work instead? 2022-09-09 19:40:09 isn't that like GPU from 2000 2022-09-09 19:56:07 probably ;) 2022-09-09 19:56:35 thinkpad 240: https://en.wikipedia.org/wiki/IBM_ThinkPad_240 2022-09-09 19:56:57 introduced 1999 :p 2022-09-09 19:57:44 this is only 7 months old though: https://gitlab.freedesktop.org/xorg/driver/xf86-video-neomagic 2022-09-09 19:57:54 3 years* 2022-09-09 19:57:59 I suppose I can make a package for it and build it. I know how to do that. 2022-09-09 19:58:35 7 months ago: typo fixes, different tarball compression, add CI 2022-09-09 19:59:10 oh I know, I am fooling. I'll try building a package. 2022-09-09 22:48:50 just use xf86-video-vesa 2022-09-09 23:19:27 Bringing me back! Compaq Presario 1240 and struggling to run Quake2 at 160x120 on that awful chipset 2022-09-09 23:36:58 160x120 sounds -very- low :) 2022-09-10 00:12:04 "just use xf86-video-vesa" <- yeah, I think the problem probably is getting the external display to be the output, I'll dig around things again. I had it working pretty nicely before by building espeakup, worked great, just wanted to try out a GUI for a change ;) 2022-09-10 04:45:12 Hi! I was attempting to open a merge request on aports (first time), using the email address given by the window "Email a new merge request to this project" link, and get an email in response that says "g Unfortunately, your email message to GitLab could not be processed. You are not allowed to perform this action. If you believe this is in error, contact a staff member." I didn't see anything 2022-09-10 04:45:18 under the contributing documentation that indicates there is something else I need to do. 2022-09-10 04:47:21 do you have an account on the alpine linux gitlab? 2022-09-10 04:48:03 I do 2022-09-10 04:52:30 When I look under the profile settings, it shows in green my email is verified, the primary emai, and blue as commit email and default notification email. 2022-09-10 04:54:40 as crazy as this sounds, try making the MR by email against your fork of aports 2022-09-10 04:54:44 the mailing list integration is broken currently and doesn't work 2022-09-10 04:54:49 the emails don't do anything 2022-09-10 04:55:22 this isn't mailing list integration, though? and clearly the message *is* getting to the GitLab server 2022-09-10 04:56:16 is this not opening merge requests by email? 2022-09-10 04:56:24 I'm using the generated email ending in merge-request@gitlab.alpinelinux.org. 2022-09-10 04:56:37 yes, it's been broken for over a month 2022-09-10 04:57:00 How should I be doing it? 2022-09-10 04:57:14 like any other merge request on gitlab 2022-09-10 04:57:52 fork public -> checkout -> make branch (not master) -> commit changes -> push branch to fork -> open merge request into alpine/aports 2022-09-10 05:41:02 Ok, that works, thanks! I didn't see the merge request link in aports, but I didn't have the repo cloned on my account 2022-09-10 14:06:27 how to install not busybox dd to alpine? 2022-09-10 14:06:33 I cannot find package name... 2022-09-10 14:07:31 coreutils 2022-09-10 14:08:02 apk search cmd:dd 2022-09-10 14:08:09 thanks! 2022-09-10 14:08:38 or https://pkgs.alpinelinux.org/contents?file=dd 2022-09-10 14:10:34 I've forget cmd: everytime... 2022-09-10 14:10:36 thanks! 2022-09-10 14:10:53 busybox's dd doesnt provide progress bar... 2022-09-10 14:21:32 boombim: "The ‘status=’ operand is a GNU extension to POSIX." https://www.gnu.org/software/coreutils/manual/html_node/dd-invocation.html 2022-09-10 14:22:38 yes, so they install coreutils dd to be able to use it 2022-09-10 14:24:09 boombim: or use "pv" together with BB dd instead? 2022-09-10 14:58:07 minimal: Thanks! I'll try to figure out wit pv 2022-09-10 20:23:18 hello. quick question regarding package triggers: https://wiki.alpinelinux.org/wiki/APKBUILD_Reference#triggers 2022-09-10 20:23:30 "The triggers variable must include the triggers for subpackages too if they have any" 2022-09-10 20:24:10 when a subpackage needs to use triggering, the trigger keyword should be added on the main package, on subpackage, or both? 2022-09-10 20:26:55 pretty sure it should work on sub 2022-09-10 20:33:37 i believe so too. at least build works fine. 2022-09-10 22:18:34 Please help me, I'm becoming stupid by proof.... (full message at https://matrix.org/_matrix/media/r0/download/matrix.org/LnIqnkHJZyKZhdEOxwKrbWmh) 2022-09-10 23:33:46 good for you 2022-09-10 23:33:55 "how does it feel?" to take a sport question 2022-09-10 23:35:48 ? 2022-09-10 23:36:03 helmen[m]: you seem to have a typo in apline_dev= 2022-09-10 23:36:14 Apachez: what is that supposed to mean 2022-09-10 23:37:41 psykose: thanks, typo is only here, that parameter works well in my syslinux.cfg where is properly written 2022-09-10 23:37:51 ah, okay 2022-09-10 23:37:55 the problem is the other 2022-09-10 23:38:44 i cant understand how to fill it with the correct directory 2022-09-11 13:30:53 morning. quick question: how would i configure a static tun-interface with network/interfaces? i seem to be unable to find complete information on what it can and can not do. 2022-09-11 14:04:54 nevermind, apparently i need to add the necessary up/down scripts by hand. 2022-09-11 14:47:09 I have a couple of typos to report with 'lbu package|pkg -h', bold for emphasis: "If is specified, and is not a direcotry, a package with the specified name willbe created."." 2022-09-11 15:53:49 anyone successfully serving files to windows 10 via samba? 2022-09-11 15:54:05 file sharing works but the share does not show up under Network 2022-09-11 15:54:07 yes 2022-09-11 15:54:31 I've enabled smb1 on windows but still doesn't show up, I have to manually type in \\hostname\ 2022-09-11 15:55:09 bleb: you can try using wsdd 2022-09-11 15:55:15 panekj: is discovery working for you? 2022-09-11 15:55:19 available in the testing repo, or https://github.com/christgau/wsdd 2022-09-11 15:58:32 I don't use SMBv1 (because ew) so no, I do not get discovery 2022-09-11 15:58:43 nor do I need it, since I map all drives anyway 2022-09-11 15:59:23 so u have a windows drive like Z: which is always mapped to \\yourhostname\yourshare ? 2022-09-11 16:00:34 Yes https://usercontent.irccloud-cdn.com/file/DLpZaW5p/drives.png 2022-09-11 16:05:24 cool 2022-09-11 16:05:28 why do they all say NTFS 2022-09-11 16:45:22 Network neighborhood has never worked well, in any incarnation of windows 2022-09-11 16:45:31 It's always a crapshoot 2022-09-11 16:46:14 worked fine for me whenever samba was either the WINS server or the master browser. 2022-09-11 16:46:35 Absolutely terrible whenever there's no WINS server or any MS OS is the master browser. hah 2022-09-11 18:29:43 Yeah, because the way the master browser is determined is essentially a yelling contest 2022-09-11 18:30:18 Throw wifi and buggy hardware into the mix and forget about it 2022-09-11 18:51:53 morning. quick question for abuild: i'm trying to create a package with a -openrc subpackage and i seem to not understand how this works. i just need it to package a rather simple openrc init.d script but 'abuild -r' errors out with "$pkgdir/$pkgname-openrc" not found. what am i doing wrong? 2022-09-11 18:56:11 apparently i didn't RTFM correctly. nvm. fix'd. 2022-09-11 19:04:33 :) 2022-09-11 19:04:52 for the record, they should be installed in /etc/{init.d,conf.d}/* 2022-09-11 19:58:29 is it possible to build an alpine .iso with alpine-conf git version? 2022-09-11 20:01:09 I guess I need to make alpine-conf-git package and feed it to mkimg 2022-09-11 20:02:05 yes 2022-09-11 20:18:31 ikke: well, since i figured how to do it, it's now an 'install -Dm755 $startdir/$pkgname.init $pkgdir/etc/init.d/$pkgname' ;) 2022-09-11 21:40:48 What can be used to monitor services and be notified when they go down and back up (like Uptime Robot)? I'm trying to find something that works for OpenRC and/or runit 2022-09-11 21:42:15 And is there something nicer to navigate system logs? 2022-09-11 22:32:05 i just managed to make my custom iso fully work with squashfs, now it barely uses any memory (even managed to work with 512M). the problem is that, since i'm booting from a USB, i apparently can't mount /dev/cdrom as the root-ro, instead i have to mount the raw device (e.g /dev/sdb). is there a way to automatically detect the device i want to mount from the initramfs? right now 2022-09-11 22:32:07 i'm mounting it manually with a debug shell (losetup -r -f /dev/sdX ; mount -t iso9660 /dev/loop0 /media/cdrom). 2022-09-12 00:13:04 i think it has something to do with nlplug-findfs 2022-09-12 01:16:37 found a hacky way to detect the ISO device, still open for sugestions though: BOOTDEV="$(blkid | grep -E '(.*):.*\s*LABEL="sigma-linux".*\s*TYPE="iso9660".*$' | head -n 1 | cut -d ':' -f 1)" 2022-09-12 02:29:56 what is the keys combination to scroll tty? 2022-09-12 02:32:23 shift+pgup/pgdn, but note that linux recently removed that functionality, I believe 2022-09-12 02:34:55 https://www.phoronix.com/news/Linux-5.9-Drops-Soft-Scrollback 2022-09-12 02:35:17 dwfreed: but they speak about framebuffer 2022-09-12 02:37:15 that is all console 2022-09-12 02:38:48 sure that dropping scroll support because everything is done graphically is absurd ... isn't it full of headless servers? and linux is not mainly used in servers?!? 2022-09-12 02:39:10 you can use screen or tmux if you need scrollback 2022-09-12 02:39:53 or you're likely using ssh or a serial console instead of a framebuffer 2022-09-12 02:40:13 wanted to look at the output of the alpine booting 2022-09-12 02:40:52 i dont think that i can load tmux in time 2022-09-12 02:41:56 i vaguely recall that pre-login text often couldn't be scrolled with shift-pgup/pgdn anyways, possibly because it's sent to /dev/console 2022-09-12 02:45:14 Hello71: in fact, I also remember that on alpine I never managed to scroll 2022-09-12 03:35:28 dmesg should be saved to a file somewhere for all the kernel booting output 2022-09-12 03:35:32 but i forget if it worked or not :p 2022-09-12 03:35:44 it should, normally, that's how you're "supposed" to read it, not by scrolling backward 2022-09-12 03:36:13 and for every other use the userspace variants (tmux et al) are at least 500x times better for having a working framebuffer with scrollback than whatever that ancient code was 2022-09-12 03:36:51 the openrc log can be enabled to output to a file too, but that usually doesn't have any info anyway 2022-09-12 03:37:26 ah, it's /var/log/rc.log by default for that 2022-09-12 03:37:51 and /var/log/dmesg has the startup dmesg 2022-09-12 03:38:04 and running `dmesg` has the rest 2022-09-12 03:38:23 all the other output would be after you log in 2022-09-12 04:17:34 thanks for the info 2022-09-12 13:35:14 psykose: you need to set rc_logger="YES" in /etc/rc.conf to have /var/log/rc.log written to 2022-09-12 13:36:49 psykose: also I've found that as "dmesg" command doesn't show the full set of kernel messages as its a fixed size buffer (defined in kernel config) that I've needed to this to the cmdline options to get the full output: "log_buf_len=32768" 2022-09-12 13:36:58 aha 2022-09-12 13:37:02 what's the default 2022-09-12 13:37:21 psykose: don't remember, 16k I think 2022-09-12 13:37:24 smol 2022-09-12 13:39:09 psykose: looking at lts.x86_64.config in aports, CONFIG_LOG_BUF_SHIFT=14, which means to mean 16KB 2022-09-12 13:40:13 but whatever value required is somewhat arbitary, if you've got a large server with lots of devices, etc then typically more output would be written than on an embedded system 2022-09-12 13:40:45 but I've found the current setting to be small even for VMs 2022-09-13 03:54:32 Are proprietary media codecs available on Alpine Linux? 2022-09-13 04:27:13 if they are propriatory, most likely not directly 2022-09-13 05:56:33 hello, does anyone here use a custom kernel with alpine linux? If so, do you do it manually or is there something similar to genkernel on alpine linux? 2022-09-13 06:01:10 define "custom kernel" 2022-09-13 06:24:23 a kernel that you build yourself for use FOR yourself 2022-09-13 06:24:27 by* 2022-09-13 07:27:00 there is no genkernel 2022-09-13 07:27:39 you can build your own kernel if you want, that's pretty easy (with or without initramfs) 2022-09-13 09:06:38 Is there a way to automate that process tho 2022-09-13 09:06:40 just like you can automate that process using genkernel on gentoo 2022-09-13 09:11:49 Duck[m]: modify an existing kernel aports and build that locally? 2022-09-13 09:12:14 sounds like itt 2022-09-13 11:43:12 I have a working APKBUILD for x2goclient, anyone interested? 2022-09-13 13:11:06 I have a strange problem, I try to configure NetworkManager with XFCE, I can backup with LBU, restart ... stop working. If I recover the old apkovl happen the same thing, it only works the first time that I simply configure and after the reboot the WiFi does not go. It happens if I add NetworkManager to the runlevels or if I don't do it and continues so even if I remove networking from the services. 2022-09-13 13:11:06 now i have to go away but if you have an idea, please let me know it 2022-09-13 14:33:05 helmen[m]: I suppose that the NetworkManager is saving the conf in some place not backed by the lbu 2022-09-13 14:34:16 LBU is not very convenient for a desktop use case, why did choose it? a bootable pendrive? 2022-09-13 14:45:50 "helmen: I suppose that the..." <- Oh, then I'll try to see even if it won't be easy to understand. 2022-09-13 14:47:40 "LBU is not very convenient for a..." <- Yes, I've always liked the distro that run in RAM. Because they push me to do experiments 2022-09-13 14:47:45 add new paths to lbu is easy, just try to discover exactly where NetorkManager sotres it 2022-09-13 14:48:03 stores* 2022-09-13 14:49:25 donoban: You can be anywhere and I must be able to understand that they are NetworkManager changes. 2022-09-13 14:50:07 do you back your home with lbu? 2022-09-13 14:50:29 however, having no other idea I would say that it is the best track to follow 2022-09-13 14:50:45 donoban: no 2022-09-13 14:51:27 I suppose that if you configure it inside a user xfce session it will use some user directory 2022-09-13 14:52:13 donoban: can be but it is strange, I never had this problem 2022-09-13 14:53:21 I followed a guide on the Alpine wiki 2022-09-13 14:53:59 uhM, can you check if it's on /etc/NetworkManager/ ? 2022-09-13 14:54:44 I can't help you very much since I don't use both them 2022-09-13 14:54:54 https://wiki.alpinelinux.org/wiki/NetworkManager 2022-09-13 14:56:27 can you check if /etc/NetworkManager is in your lbu? 2022-09-13 14:56:48 donoban: Now I am not in front of the PC but obviously all or changes in etc are tracked by LBU 2022-09-13 14:57:09 yeah I suppose that 2022-09-13 15:02:43 uhM, if you don't backup the home dir 2022-09-13 15:02:53 do you run xfce as some user right? 2022-09-13 15:03:34 could you try to backup for testing? 2022-09-13 15:07:23 I was reinstalling, the next step was to choose what backup in the home 2022-09-13 15:07:51 donoban: yes this can be my next attempt 2022-09-13 15:11:52 well, try to backup whole home :P 2022-09-13 15:18:59 yes I can try, when I will return in front of my pc. I'm resting a little now, I'm exhausted :p 2022-09-13 15:20:08 hehe ok 2022-09-13 15:29:29 "hehe ok" <- but really thank for the help! 2022-09-13 15:30:11 :) 2022-09-13 23:45:20 Hi there, I'm pretty new to Alpine, but has anyone else had struggle at installing sudo? 2022-09-13 23:46:01 I'm able to find other packages not `sudo`, for some reason 2022-09-13 23:48:40 blursedpotato: alpine uses doas instead. If you really want to use sudo, you must use the community repo. 2022-09-13 23:51:34 Thanks for the help. It seems like there are community repos listed in /etc/apk/repositories, but community packages don't show up anyway. 2022-09-13 23:54:53 any chance they have been commented out? 2022-09-13 23:54:59 beyond that idk 2022-09-13 23:55:19 They are :facepalm: 2022-09-13 23:55:32 I feel so embarassed right now 2022-09-13 23:55:44 Sorry for bothering yo guys 2022-09-13 23:55:49 *you 2022-09-13 23:55:51 no problem, happens to everybody! 2022-09-13 23:55:53 \o/ 2022-09-13 23:57:04 I believed they weren't when I checked, but I it seems like didn't check that well ;-; 2022-09-14 04:18:02 I gather that there's a bit of a debate with busybox at the moment, is there a mailing list thread I can read about it? 2022-09-14 04:21:53 there's this: https://gitlab.alpinelinux.org/alpine/tsc/-/issues/55 2022-09-14 04:38:59 not really 2022-09-14 05:01:51 ptrc: Thanks. :) 2022-09-14 07:40:22 How can I use ngrock on alpine? 2022-09-14 07:40:41 ngrok* 2022-09-14 07:45:17 just `./ngrok`? It's a static binary, it runs just fine out of the box 2022-09-14 07:59:03 no package for it? 2022-09-14 08:03:21 boombim: doesn't seem like it. feel free to make one 2022-09-14 08:04:46 no, as it's proprietary 2022-09-14 08:05:23 The original 1.x version is still open source 2022-09-14 08:05:33 is that still usable with their service though? 2022-09-14 08:05:56 good question 2022-09-14 08:33:54 ":)" <- Ok, I tried every combination, is not a question of not backuped file :( 2022-09-14 08:48:55 Maybe some better open source solution for my task? I want to create demo service to test frontend side of site until backend will done. 2022-09-14 08:49:30 I suppose people use ngrok for this but maybe not only ngrok or whatever 2022-09-14 08:55:40 so exactly how is one supposed to use the 'generic armv7 image'? 2022-09-14 09:00:53 boombim: I don't know of any. The fact that you need a service for it makes it an easy target for proprietary software sadly. You could write your own I suppose 😜 2022-09-14 09:27:04 what can I do if lrelease-qt5 just exits with 1, but no error message is produced? 2022-09-14 11:07:11 helmen[m]: it's strange 2022-09-14 11:16:23 "helmen: it's strange" <- i know and the base netwirking work good 2022-09-14 11:16:48 i have problem only with networkmanager 2022-09-14 11:17:07 * i know and the base networking works good 2022-09-14 11:31:10 when you reboot, networkmanager service is running? 2022-09-14 11:32:49 yes 2022-09-14 11:33:37 only at the first installation 2022-09-14 11:33:51 so, when you reboot it fails to start? 2022-09-14 11:34:15 my alpine work live in diskless mode 2022-09-14 11:34:32 yes I know 2022-09-14 11:34:48 donoban: yes, do not show wireless connection 2022-09-14 11:34:57 your problem is that you configure some wifi network, it works, you commit your lbu 2022-09-14 11:35:14 and when you reboot, the networkmanager service runs but the wireless connection is lost? 2022-09-14 11:35:27 i have also an exernal wifi card that swich on a led when is wirking, no light 2022-09-14 11:35:39 donoban: yes 2022-09-14 11:36:04 uhm, maybe you miss some kernel module? 2022-09-14 11:36:44 could you compare 'lsmod' on success and failure scenarios? 2022-09-14 11:36:54 i dont think 2022-09-14 11:37:30 donoban: uhm only the output of this command? 2022-09-14 11:37:55 you can discover if some module is not loaded 2022-09-14 11:38:44 donoban: wifi work correctly with the builtin setup-networking and work good at the first run with networkmanager, so I dont think that I miss kernel module 2022-09-14 11:39:06 donoban: I can try 2022-09-14 11:39:22 I'm not sure but maybe the setup scripts load them but they are not configured proplerly for autoloading 2022-09-14 11:39:49 can be but 2022-09-14 11:40:12 uhm I have to try 2022-09-14 11:40:57 when you say that it fails, you mean that it doesn't show any wireless network? 2022-09-14 11:41:55 yes 2022-09-14 11:42:17 also the light of my external wifi cars is off 2022-09-14 11:42:30 s/cars/cards/ 2022-09-14 11:42:40 s/cars/card/ 2022-09-14 11:43:01 so it could be a kernel module problem 2022-09-14 11:43:35 do you know if the network interface exists? (wlan0 ?) 2022-09-14 11:43:56 I'm watching lsmod output with a currently working configuration with setup-imterfaces 2022-09-14 11:44:17 if you want post it with tpaste 2022-09-14 11:44:20 lsmod | tpaste 2022-09-14 11:44:32 how i can undersand if some module is missing? 2022-09-14 11:45:00 just go to the failing scenario and run lsmod again 2022-09-14 11:45:20 I suppose that just configure lbu, commit and reboot 2022-09-14 11:45:31 i have to restart the oc and configure the bad networkingmanager for check 2022-09-14 11:46:10 ok so just make twoa comparisom with a diff between a working and not working state? 2022-09-14 11:46:43 yes, lsmod just shows the loaded modules, if someone is missing you could notice it 2022-09-14 11:47:46 ok, i need some time to make all step 2022-09-14 11:47:56 but i want to try now 2022-09-14 11:48:14 good luck, I have to go in few minutes 2022-09-14 11:48:43 (im sorry for the text, im writhing from a phone, i hate writing on a phone xD) 2022-09-14 11:48:52 if you discover that there are some modules missing you can just add them to /etc/modules 2022-09-14 11:49:01 they will be loaded on startup 2022-09-14 11:49:04 donoban: thanks twice for the support 2022-09-14 11:49:35 donoban: ok thanks 2022-09-14 11:49:43 👍 2022-09-14 11:49:53 np 2022-09-14 19:01:34 I'm running 3.16 with some packages from edge, but I'm forgetful (specially about dependencies). Is there a way to list the installed packages with the tag/repo they came from? 2022-09-14 19:16:16 swinnenb: afaik, apk doesn't keep track of that 2022-09-14 19:16:30 at most, you can run apk policy to see where it would install it from 2022-09-14 19:17:38 I thought that's what the s: tag was for in /lib/apk/db/installed? (although I don't know of an interface to actually deal with it) 2022-09-14 19:18:38 zcrayfish: I don't have s: in there 2022-09-14 19:18:52 only S:, which is size 2022-09-14 19:18:59 policy does show the ../db/installed and the repo for the package, perhaps I can deduce a list from that 2022-09-14 19:20:45 ikke: it seems to be for apks from tagged repos only. 2022-09-14 19:21:24 ahh 2022-09-14 19:22:22 yeah, I do see a few 's:edge' entries in the 'installed' file 2022-09-14 19:24:54 I think I can get what I need with this info. thanks. 2022-09-15 08:11:38 I installed Alpine to a N4500 based BRIX, but i915 crashes after 1-30 seconds after modesetting (kernel dies with it), using nomodeset it runs without crash, but the 800x600 resolution on a 4K display is not optimal 2022-09-15 08:11:47 what can I do? 2022-09-15 08:50:51 aron: configure something like netconsole or serial console to capture the crash? 2022-09-15 09:00:32 dwfreed: it dies exactly the same way as the installer iso does 2022-09-15 09:01:08 so, just booting the iso, which performs modesetting, then it dies 2022-09-15 09:01:34 without touching the keyboard ;) 2022-09-15 09:01:56 sure, but without your hardware it's hard to debug without the crash log 2022-09-15 09:02:16 I tried 2022-09-15 09:02:47 disable modesetting and setup kdump to capture the crash, then boot with modesetting enabled 2022-09-15 09:02:48 it dies so quick, it is really hard to setup quickly anything 2022-09-15 09:03:29 kdump can send it on network? 2022-09-15 09:04:02 Afaik yes if kdump kernel was built with network support 2022-09-15 09:05:40 even just the log sent via netconsole would work 2022-09-15 09:05:48 netconsole can be configured via kernel command line 2022-09-15 09:06:02 sounds good 2022-09-15 09:07:37 looking into its documentation, this should work 2022-09-15 09:08:15 one instance I saw a backtrace on the heavily blinking display :) 2022-09-15 09:08:38 (or a part of it) 2022-09-15 09:09:06 thanks, I'll be back with the details 2022-09-15 09:15:53 > np... (full message at ) 2022-09-15 09:16:57 ops, the command is "ip link" non "if link" 2022-09-15 09:31:54 helmen[m]: what happens if you run 'modprobe iwldvm' 2022-09-15 09:34:26 I'm not sure if using a modloop /etc/modules makes sense 2022-09-15 09:35:17 I think that the module is missing due kernel version missmatch or missing firmware 2022-09-15 09:35:30 did you notice other missing modules? 2022-09-15 09:40:41 donoban: Im sorry, I was using fractal on alpine for matrix but it crashed. Now im using gomuks 2022-09-15 09:41:40 no I dont need moodprobe, before doing the reboot I put the module in /etc/module and it shows correctly loaded with lsmod 2022-09-15 09:46:22 anyway it seems that the module works, I rebooted and tried just using `setup-interfaces` 2022-09-15 09:46:41 3584 1 iwldvm 2022-09-15 09:46:42 cfg80211 1052672 6 ath9k_htc,ath9k_common,iwldvm,mac80211,iwlwifi,ath[201~ 2022-09-15 09:46:42 iwlwifi 323584 1 iwldvm 2022-09-15 09:47:12 cfg80211 1052672 6 ath9k_htc,ath9k_common,iwldvm,mac80211,iwlwifi,ath 2022-09-15 09:48:37 this was an lsmod | grep iw 2022-09-15 09:50:57 \toggle typingnotif 2022-09-15 09:55:44 i advise you not to do moodprobe on me. 2022-09-15 10:03:25 "i advise you not to do moodprobe..." <- 😋 2022-09-15 10:07:18 clandmeter: instead do you have any typ for this problem? https://www.reddit.com/r/AlpineLinux/comments/xb3dqv/apkovl_boot_parameter/ 2022-09-15 10:10:06 I ask you because I had seen that in your github there were references to that parameter 2022-09-15 10:10:08 moodprobe -rfa anxiety 2022-09-15 10:13:15 helmen[m]: so it's working fine now? 2022-09-15 10:16:05 donoban: not with networkmanager 2022-09-15 10:16:35 :( 2022-09-15 10:19:00 In this momet i'm connected with usb0 interface, i want to try one time with wifi phone tethering and setup-interfaces 2022-09-15 10:32:33 but the wireless card looks operative? 2022-09-15 10:35:31 YES, I can see the ssids of wifi networks. 2022-09-15 10:35:31 On first startup I can also do it with networkmanager 2022-09-15 10:40:44 donoban: ok, I can confirm, with setup-interfaces wlan work correctly 2022-09-15 10:40:56 now i'm connecter with tethering 2022-09-15 10:41:18 * connecter with wifi tethering 2022-09-15 10:41:40 and does it works with lbu? 2022-09-15 10:42:21 donoban: ok, one moment, I try 2022-09-15 10:47:28 "and does it works with lbu?" <- yes, also with lbu 2022-09-15 13:16:50 Is there any way to clone the current alpine run? 2022-09-15 13:16:53 a script? 2022-09-15 13:22:57 uhm, maybe involving setup-bootable 2022-09-15 15:02:35 what you mean with "the current alpine run"? 2022-09-15 15:03:02 create another pendrive? 2022-09-15 15:53:29 dwfreed: netconsole does not send a single message, even if I echo stuff into /dev/kmsg 2022-09-15 15:55:36 however, I can boot with nomodeset, and then unload i915, then modprobe again with explicit modeset=1, which cause the same crash 2022-09-15 15:55:53 I thought to ssh to the machine, and "dmesg -w" on it 2022-09-15 15:55:57 did not worked 2022-09-15 15:56:18 kernel died before message reached the ssh client 2022-09-15 15:58:01 Btw do you experience crash on other distros? 2022-09-15 15:58:24 tbh, I did not tried it 2022-09-15 15:59:06 but netconsole should work, tho 2022-09-15 16:00:38 let's try arch linux 2022-09-15 16:00:48 14 minutes left :/ 2022-09-15 16:27:17 Ermine: yes, arch linux 2022-09-03 dies exactly the same way 2022-09-15 16:28:30 So, this is a legit kernel bug 2022-09-15 16:28:53 it seems 2022-09-16 02:52:10 ibus appears to be screwed up on edge. 2022-09-16 02:52:26 Something about missing python module gi. 2022-09-16 02:53:58 Got it somewhat working by editing the ibus-setup file and explicitly defining python3 but still having problems changing IME to say ibus-anthy (does not switch). 2022-09-16 02:54:44 Just going to try downgrading my whole setup to 3.16. 2022-09-16 03:23:36 hey, does anyone know what script automatically mounts partitions in /media? I'd like to disable it 2022-09-16 03:35:32 that's generally a function performed by a desktop environment 2022-09-16 03:49:10 I mean at boot, when I simply start alpine it mounts /dev/mmcblk0p1 to /media/mmcblk0p1 2022-09-16 03:49:54 I mean, you could grep /etc/init.d for /media, but that's most likely an fstab entry 2022-09-16 03:51:36 there's nothing in fstab except /dev/cdrom and /dev/usbdisk 2022-09-16 03:52:30 it seems like something the base ISO does without any customisation 2022-09-16 03:57:22 I'm sure a grep of /etc would be enlightening 2022-09-16 03:59:10 `rg /media /etc/` only shows entries in /etc/fstab and /etc/apk/repositories 2022-09-16 03:59:34 fstab still only mentions cdrom and usbdisk 2022-09-16 04:00:37 oh, searching for "mmc" shows /etc/mdev.conf, I think that's what I'm looking for 2022-09-16 04:01:14 mmcblk.* root:disk 0660 */lib/mdev/persistent-storage 2022-09-16 04:02:47 you'll find that /lib/mdev/persistent-storage does not contain anything for mounting, it just sets up the symlinks people often use from udev 2022-09-16 04:03:16 yeah you're right 2022-09-16 04:16:48 a small question, why "apk info -R" don't show dependaces by package name? 2022-09-16 04:17:29 so:libncursesw.so.6 2022-09-16 04:17:29 so:libc.musl-x86_64.so.1 2022-09-16 04:17:29 nano-6.4-r0 depends on: 2022-09-16 04:18:09 I assume because there can be multiple packages that provide the same dependency 2022-09-16 04:18:12 example, but the package names of the dependencies are: musl and ncurses-lib 2022-09-16 04:18:48 ok but why https://pkgs.alpinelinux.org/package show the package name dependencies? 2022-09-16 04:25:22 I guess the website just lists the providers of the dependencies to make it easier to use 2022-09-16 04:28:11 if you search for "tracing dependencies" in the build log it prints the libraries, not packages, so that's probably where apk gets it from https://build.alpinelinux.org/buildlogs/build-edge-x86_64/main/nano/nano-6.4-r0.log 2022-09-16 04:36:42 * ok mistake104 , thanks for the reply, at least now this behaviour have a reason 2022-09-16 04:39:14 If i want to create a copy of the cache directory (/media/usb/cache) cleaned of all packages that are not used? is there a way? 2022-09-16 04:39:35 not only old version, all'not used packages 2022-09-16 04:46:08 honestly I'd like to know how to do that too 2022-09-16 05:00:02 "honestly I'd like to know how to..." <- I only found one procedure though involves having to re-download everything from the internet and that was what I wanted to avoid 2022-09-16 05:00:58 https://arstechnica.com/civis/viewtopic.php?p=34749529#p34749529 2022-09-16 05:02:26 apk cache clean? 2022-09-16 05:02:27 I almost managed to do it by creating a fake repository from the cache but then I couldn't do it 2022-09-16 05:03:18 ikke: for example I have fractal on my cache but I dont load it, if i run `apk cache clean` fractal stay there 2022-09-16 05:03:34 apk cache clean remove only old package 2022-09-16 05:04:16 s/package/packages/ 2022-09-16 05:14:35 the man page also claims it would remove packages that are no longer needed 2022-09-16 05:16:02 ikke: I just tried 2022-09-16 05:26:39 helmen[m] try `apk cache clean --purge`, it should delete everything that isn't currently installed 2022-09-16 05:28:40 if you want to clean a *copy*, then copy the cache to a new directory and run `apk cache clean --purge --cache-dir /path/to/cache` 2022-09-16 05:38:52 ok one moment 2022-09-16 05:41:35 ok this time seems ok 2022-09-16 05:42:10 "if you want to clean a *copy..." <- thanks for the help, this is really helpful 2022-09-16 06:27:49 anyone got redshift to work on raspberry? 2022-09-16 06:27:53 4b 2022-09-16 06:46:38 "anyone got redshift to work on..." <- No, however I noticed that in a minimal alpine installation with XFCE also the Display Brightness control does not work. What packages will be missing? 2022-09-16 06:49:13 helmen[m]: i wonder if it is the driver tbh 2022-09-16 06:57:48 uhm mesa is nkt enough? 2022-09-16 07:29:17 s/nkt/not/ 2022-09-16 08:05:00 Is this the proper channel to discuss docs? https://wiki.alpinelinux.org/wiki/Upgrading_Alpine#Upgrading_Alpine_Linux_on_CD seems to have a lot of duplicate info which is already present in https://wiki.alpinelinux.org/wiki/Upgrading_Alpine#Upgrading_an_Alpine_Linux_Hard-disk_installation ... Perhaps an idea to refer to the first header for info which is duplicate? 2022-09-16 08:13:33 raspbian os has a "sd copier" that can copy a sd/usb-device to another, copies the UUID and doesnt have to be the same size! seems to even work well, is there a way to do this easily on the command line? 2022-09-16 08:14:06 eg copied a 500GB SDcard (only 6GB used) to a 16GB usbstick.. 2022-09-16 08:20:25 frag: I mean, it's not doing anything special, just making sure it makes the new filesystems with the same UUIDs as the source filesystems 2022-09-16 08:21:44 dwfreed: i guess it has to set boot flags and partition and such too, i dunno even how to choose UUID 2022-09-16 08:22:05 manpages are your friend :) 2022-09-16 08:22:43 but also the thing is open source, and *very* basic: https://github.com/raspberrypi-ui/piclone 2022-09-16 08:46:32 there is a place where I can find all boot parameter for alpine? 2022-09-16 09:07:53 Whoops 2022-09-16 09:09:30 can I force to mount the boot partition everytime at the same mountpoint? for example con fstab ? 2022-09-16 09:24:32 Breezy seems to be broken on edge: "No module named 'fastbencode'" 2022-09-16 09:24:56 And it doesn't seem packaged separately either. 2022-09-16 17:35:37 So I loaded the Alpine Linux image on a USB today and I could not mod probe the speak up module. I know it is included in the Linux kernel maybe somebody could help me out. 2022-09-16 17:37:22 sethmhurst: what does "modinfo speakup" give? 2022-09-16 17:38:12 I'm not sure as I can't see the screen. I could try loading it onto my laptop and see what comes up. Does it automatically login as root or do I have to type that at a prompt. 2022-09-16 17:40:13 from memory there's a login prompt so type "root" and Return, there's no password so should drop to shell then 2022-09-16 17:41:29 How many minutes do I want to wait just to make sure I'm out of prompt. 2022-09-16 17:41:38 At 2022-09-16 17:41:42 should be instantly 2022-09-16 17:42:59 sethmhurst[m]: I used to work with a visually impaired colleague 30 years ago doing research into GUI access for the visually impaired. 2022-09-16 17:54:46 ACTION uploaded an image: (97KiB) < https://matrix.org/_matrix/media/r0/download/matrix.org/GzEyejzMHkuMvYxNklUEVHhI/20220916_135341_3500367918746412897.jpg > 2022-09-16 18:04:51 I did a mod info and it is showing up. But I'm not sure how to load it 2022-09-16 18:05:07 I'm getting no sound. 2022-09-16 18:06:41 sethmhurst: "modprobe speakup" should load it. However I see there are multiple other speakup related kernel modules for various "brands" such as apollo, bns, dectlk etc 2022-09-16 18:06:55 which screen reader are you using? 2022-09-16 18:10:27 Just the built-in software synthesizer you speak up. 2022-09-16 18:10:44 Not using any hardware screener just the output from my laptop speakers. 2022-09-16 18:26:05 sethmhurst[m]: I think to use software synthesizer you also need to do "modprobe speakup_soft" 2022-09-16 18:26:51 sethmhurst[m]: and also have the espeakup package installed and have espeakup running 2022-09-16 18:27:55 sethmhurst[m]: I haven't looked at this stuff before. Let me try installing in a VM and see what is or is not there 2022-09-16 18:29:05 Sounds good. 2022-09-16 18:56:35 sethmhurst[m]: FYI i'm still here, I'll come back to you in a couple of minutes with what I've found 2022-09-16 19:03:34 sethmhurst[m]: ok, here's the situation: if you boot either alpine-standard or alpine-extended the speakup and speakup_soft modules are present. This lets you hook into a software synth. However there is no software synth package present in either the standard or extended ISO. 2022-09-16 19:04:29 sethmhurst[m]: Alpine has packaged one speech synth designed to work with speakup and that is espeakup 2022-09-16 19:05:28 sethmhurst[m]: however that package is only present in the "testing" repository which means it is only available for Alpine Edge, not for any of the Alpine releases like 3.16 2022-09-16 19:05:35 I'm familiar with that one. 2022-09-16 19:06:44 so I could discuss with the Alpine espeakup maintainer to move it from testing to community which, if that happened, would mean that it would present in the next Alpine release, 3.17, due around November timeframe 2022-09-16 19:07:22 I've only ever used it on a linode over SSH and interested to see how it would do on bare metal. 2022-09-16 19:07:54 That would be fine. Is there any way to get the Alpine edge version? 2022-09-16 19:08:32 However, even with espeakup being present in a release there is nothing currently in place to automatically activate it. Either or both the standard and extended ISOs would need to have espeak added to their packages set and then some thought given as to activate it upon boot 2022-09-16 19:09:14 There is the SYS Linux file that allows you to append modules that you want loaded but the file on the ISO that I have is read only. 2022-09-16 19:10:26 to activate it manually would involve running the commands: "modprobe speakup_soft", "apk add espeakup", and "/etc/init.d/espeak start" - that runs it manually and I've tested that it works (on Alpine Edge) but after a reboot you'd have to do that again 2022-09-16 19:13:29 to set it up to be running every time would involve adding the "speakuk_soft" module to either a "modules=X" bootloader config (e.g. Syslinux or Grub) or else to /etc/modules file and also run "rc-update add espeakup" once to configure espeak to automatically run on each boot. This is assuming you do a "Sys" installation of Alpine where it runs from disk, if you intend to just run from Ram the configuration would be similar but I'm not as 2022-09-16 19:13:29 familiar with that 2022-09-16 19:15:26 sethmhurst[m]: to get edge, you change the repositories in /etc/apk/repositories from v3.x to edge, and then run upgrade --update --available 2022-09-16 19:16:39 Sadly the only way I can access the system is through the speaker module. 2022-09-16 19:16:47 ikke: it would be nice to make Alpine more accessible but there's nothing to "autodetect" to determine if a screen reader should be enabled 2022-09-16 19:17:55 Maybe an image that specifically loads that. 2022-09-16 19:18:25 sethmhurst[m]: yeah that was the only mechanism I could think of 2022-09-16 19:19:14 sethmhurst[m]: or a script to tweak a downloaded ISO before it is written to CD / USB 2022-09-16 19:19:57 I have the ISO file but I'm not sure how to edit it. Because I could just write the script and put it on there and see if it works. 2022-09-16 19:21:22 sethmhurst[m]: well I was thinking once espeakup is in a release such a script (i.e. that adds a parameter to be passed by Syslinux) would be somewhat simple 2022-09-16 19:22:40 with the currently situation I guess a script could be used to "insert" another script into the ISO that could be run to do the whole 3.16 to Edge upgrade as well as setting up espeakup 2022-09-16 19:22:42 I've also got a VM and I have it on a USB. So I can test either way. 2022-09-16 19:23:23 I've got espeakup running fine here on Edge in a VM as a screen reader 2022-09-16 19:25:08 sethmhurst[m]: I can have a go at writing a script over the weekend 2022-09-16 19:26:14 That would be very awesome. If only there was a way to load up SSH if it was connected to a wired connection which I can connect my laptop to a wired connection. 2022-09-16 19:30:59 sethmhurst[m]: need to grab some food now but I'll be around on this channel over the weekend (and I check the logs of when I'm not on here) 2022-09-16 19:32:51 Well I've got a pretty busy weekend but I'll be around. 2022-09-16 19:54:12 how buildozer builds armv7 packages? 2022-09-16 19:54:42 my solution was to start qemu-openrc, install armv7 chroot on x86_64, and chroot into it 2022-09-16 19:54:48 this works fine 2022-09-16 19:54:50 until... 2022-09-16 19:54:57 qt5 is involved 2022-09-16 19:55:21 which behaves in this qemu-arm/chroot environment badly 2022-09-16 19:58:17 armv7 container on aarch64 hardware in 32-bits mode 2022-09-16 19:59:14 :( 2022-09-16 19:59:27 this is bad news 2022-09-16 20:00:01 maybe if I set up a complete qemu-system-arm 2022-09-16 20:00:13 it should work, right? 2022-09-16 20:01:25 such as alpine-virt--armv7.iso 2022-09-16 20:03:27 there is no such iso 2022-09-16 20:03:53 I was able to download it 2022-09-16 20:04:00 :) 2022-09-16 20:04:12 from where 2022-09-16 20:04:28 https://alpinelinux.org/downloads/ "VIRTUAL" / "armv7" 2022-09-16 20:04:58 oh, not for edge, for v3.16 2022-09-16 20:05:19 yep, sorry, I didn't mentioned it 2022-09-16 20:05:26 edge has no version number tho :) 2022-09-16 20:06:07 nod, I should've known 2022-09-16 20:06:13 we only do snapshots for edge once in a while 2022-09-16 20:47:55 i guess the rpi tarball doesnt support full disk encryption? 2022-09-16 20:53:53 does anyone have a working zoom from flathub on edge? 2022-09-17 02:55:03 Yes, last I checked about two days ago. What's wrong with yours 2022-09-17 03:45:40 i tried to put a rule in fstab to force alpine diskless partition to mount at a certain mountpoint. It doesn't work, what am I wrong? 2022-09-17 04:25:59 <_Void> Hi, I've been trying to mount a CD on alpine linux by using "mount /dev/cdrom /media/cdrom", but it just gives me "/dev/cdrom on /media/cdrom failed: Invalid argument". I was told online to use -t to force it to use iso9660, but that gave me the same error, what am I doing wrong? 2022-09-17 04:32:21 whoa, optical drives 2022-09-17 04:32:57 try using /dev/sr0 or whatever it really is 2022-09-17 04:34:14 might also want to make sure your mount point exists 2022-09-17 04:43:33 <_Void> Just tried that, I'm now getting a completely different "mounting /dev/cdrom on /media/cdrom failed: I/O error" error. on both sr0 and cdrom. 2022-09-17 04:44:12 <_Void> Also, the mount point does exist, and I know for a fact that this drive works since other operating systems that i've tried use it fine, so it seems like a software problem rather than it just being broken. 2022-09-17 04:44:29 check dmesg 2022-09-17 04:48:17 <_Void> It's full of a lot of "blk_update_request: I/O error, dev sr0, sector 0 op 0x0:(READ) flags 0x1000 phys_seg 1 prio class 0" and "EXT4-fs (sr0): unable to read superblock", but I'm a bit new so I don't understand most of it and might be missing something important, should I send the whole thing? 2022-09-17 04:49:57 pastebinning the whole thing wouldn't be bad, but it sounds like your disc drive doesn't work in Linux 2022-09-17 04:50:23 laptop or desktop optical drive? 2022-09-17 04:52:05 <_Void> laptop, and I've tested the disc drive with a live USB of endeavourOS earlier so it should work 2022-09-17 04:54:17 <_Void> and here's the dmesg: https://pastebin.com/mdHPK7c7 2022-09-17 04:56:25 it's trying to mount as ext4 not iso9660, but it looks like the drive can't see the disk... what kind of media is this, CD-R? DVD+R? 2022-09-17 05:03:18 <_Void> Well I'd assumed it was CD, but it actually doesn't say anywhere on it so I might have to use another OS to check 2022-09-17 05:14:16 it's probably a some junk CD-R vendor and suffering cd rot or faded dye 2022-09-17 05:15:34 all my CMC magnetics cd-r's degraded like crazy, I bought a distributor case of taiyo yudens and they are the only ones that stood the test of time with no failures after 15 years of storage 2022-09-17 05:16:21 <_Void> Thing is, I just loaded up endeavourOS on my laptop and it's reading the disc just fine 2022-09-17 05:17:57 rules out hardware then 2022-09-17 05:19:25 can you open a terminal on that and run mount? 2022-09-17 05:19:30 see what it mounted it as 2022-09-17 05:20:45 <_Void> I'm not even sure it IS mounted, actually 2022-09-17 05:20:59 <_Void> it shows up as "cdda://sr0/" in the file manager 2022-09-17 05:21:09 it's an audio CD 2022-09-17 05:21:51 won't mount unless you have something like fuse-cdfs 2022-09-17 05:22:20 <_Void> oh i see 2022-09-17 05:23:31 <_Void> so in order to mount this on alpine i'd have to install that then? 2022-09-17 05:24:19 i doubt they package it, but it's not a filesystem... that gives you a virtual filesystem, but I guess the better question is 'what do you want to do with it?' 2022-09-17 05:25:01 play it? most media players have cd audio support, mpv or audacious come to mind 2022-09-17 05:27:09 <_Void> the plan is to access the actual files on the disc from another PC that doesn't have its own disc drive, so i need to mount it as a file system 2022-09-17 05:29:49 yeah but it's an audio cd, so it doesn't have files unless there's a filesystem after that on the disc 2022-09-17 05:30:48 <_Void> maybe i could use something like network block device to share the actual disc drive itself and then let the other PC handle actually reading from it? 2022-09-17 05:31:52 if you want to rip it to files you'd want something like cdparanoia to get them to WAV, and then probably encode them to FLAC to save some space.. and ideally with something that would tag it for you 2022-09-17 05:32:52 <_Void> well, the live USB of endeavourOS just let me copy them straight out of the file manager, and my main PC also runs that so it'll probably work fine 2022-09-17 05:33:55 i dunno what it's running, some desktop environments built in ripping support to their file managers, so it might just rip them as wav files for you 2022-09-17 05:46:42 <_Void> Well, thanks for the help, I'm gonna go figure out how to setup NBD or some other solution like that, and possibly find a non-audio CD to test on! 2022-09-17 05:47:09 \reply 2022-09-17 05:48:20 * so the reply seems that this is impossible: https://unix.stackexchange.com/questions/502610/alpine-persistent-disk-mounted-to-media-sda-even-though-fstab-has-media-usb 2022-09-17 05:49:01 "i tried to put a rule in fstab..." <- seem that this is impossible 2022-09-17 05:49:28 https://unix.stackexchange.com/questions/502610/alpine-persistent-disk-mounted-to-media-sda-even-though-fstab-has-media-usb 2022-09-17 08:00:53 Saijin_Naib: when i try to run zoom with "flatpak run --verbose us.zoom.Zoom" it yields https://pastebin.com/9wncBp9m 2022-09-17 08:02:41 not sure what's the problem since i didn't mess with flatpak whatsoever 2022-09-17 08:38:41 a while back we talked about how to create a partial git checkout (or whatever it is called in git) 2022-09-17 08:39:02 I made an APKBUILD for x2goclient, and I want to contribute it 2022-09-17 08:40:11 sadly my qemu/chroot-based build does not work well with qt, so I made a decision to add it to the official repo 2022-09-17 08:42:52 and I forgot how to create the partial git checkout :) 2022-09-17 08:53:05 it was --depth 1 but it does not work, nvm 2022-09-17 09:15:17 I keep getting "remote: HTTP Basic: Access denied. The provided password or token is incorrect" :( 2022-09-17 10:12:51 I can't believe, I made it! 2022-09-17 10:12:57 only 3 hours 2022-09-17 10:58:37 shouldnt devices in /etc/conf.d/dmcrypt be decrypted at boot? 2022-09-17 11:05:08 (the sysmte doesnt have grub btw) 2022-09-17 11:11:47 i also added to /etc/crypttab but that doesnt seem to do anything either 2022-09-17 11:34:47 context? 2022-09-17 11:37:14 Duck[m]: trying to decrypt a luks partition to ~ on rpi at boot 2022-09-17 11:42:46 are you using syslinux 2022-09-17 11:43:07 cat /etc/mkinitfs/mkinitfs.conf 2022-09-17 11:43:11 what does this say 2022-09-17 12:07:46 Duck[m]: oh it had just a 4-5 "features" (its on another computer) 2022-09-17 12:08:00 Duck[m]: should i add cryptosetup to it? 2022-09-17 12:08:41 frag: obviously 2022-09-17 12:08:48 are you not following a guide? 2022-09-17 12:08:51 > <@_oftc_frag:matrix.org> Duck 🏞️: should i add cryptosetup to it? 2022-09-17 12:08:51 * yes 2022-09-17 12:09:10 frag: add cryptsetup 2022-09-17 12:09:12 not cryptosetup 2022-09-17 12:09:20 and are you using lvm 2022-09-17 12:09:31 yes! 2022-09-17 12:09:45 guides told me to add to crypttab 2022-09-17 12:09:55 never looked at mkinitfs 2022-09-17 12:13:11 did you add lvm to the crypttab as well 2022-09-17 12:13:20 add nvme in it as well if you are using an nvme drive 2022-09-17 12:13:52 hmm still doesnt seem to do anything (and no errors, just error from fstab not finding /dev/mapper/home) 2022-09-17 12:14:42 Duck[m]: crypttab has "home /dev/sdb3 none luks" 2022-09-17 12:15:14 try UUID=... instead of "home" 2022-09-17 12:15:34 uhm. scratch that. I'm stupid 2022-09-17 12:15:46 satmd: home is the name it will use for /dev/mapper/home i think 2022-09-17 12:15:53 yeah, scratch that idea 2022-09-17 12:16:33 right? 2022-09-17 12:16:34 frag: you do get asked for the passsword tho 2022-09-17 12:16:45 also cmdline.txt on raspberry doesnt suppoert UUID, so ill just use device name uniformly 2022-09-17 12:16:58 Duck[m]: i dont 2022-09-17 12:17:25 wait 2022-09-17 12:17:28 are you using syslinux or grub 2022-09-17 12:18:01 Duck[m]: im actually not sure, its a raspberry machine, not sure it uses either? not grub at least afaics 2022-09-17 12:19:07 alpine and raspberry pis are special 2022-09-17 12:20:59 indeed 2022-09-17 12:34:35 is there an easy way to just run a command for decrypting and mounting the device _before_ login? (i dont understand well the runlevel scripts btw) 2022-09-17 12:34:50 like a shell command 2022-09-17 12:41:18 i just found something 2022-09-17 12:42:31 "is there an easy way to just run..." <- https://github.com/alpinelinux/mkinitfs/edit/master/mkinitfs-bootparam.7.in 2022-09-17 12:42:57 bootparameters, i'm curious to understand if this is a man page 2022-09-17 12:43:37 helmen[m]: hmm, have to log in on that url? 2022-09-17 12:44:23 uh, one moment, maybe this is "wrong" url 2022-09-17 12:46:00 https://github.com/alpinelinux/mkinitfs/blob/master/mkinitfs-bootparam.7.in 2022-09-17 12:46:22 im sorry the previos was the edit link 2022-09-17 12:47:45 frag: ping 2022-09-17 12:48:41 why is startup not just a shell script :| 2022-09-17 12:51:56 no i really don't know anything about encryption, but i was reading this 2022-09-17 15:19:06 frag: there's is no FDE for Raspberry Pis as the boot partition has to be unencrypted. Now if you want almost-FDE that's another matter ;-) 2022-09-17 15:22:53 frag: also with LUKS you'll probably want to use Adiantum as the cipher as it will work better on the RPI 2022-09-17 15:25:55 minimal: yeah its a bit slow but maybe ok 2022-09-17 15:26:22 frag: adiantum will be faster, it's designed for lower-end CPUs 2022-09-17 15:27:05 minimal: oh... it is slow only to decrypt or slows down the system after that too? 2022-09-17 15:28:04 frag: LUKS is encrypted so any read or writes on a fs on top of LUKS involves encrypt/decrypt all the time 2022-09-17 15:29:28 frag: Adiantum was designed by Google for Android devices which obviously use lower-end CPUs than PCs 2022-09-17 15:29:59 especially as some/many Arm CPUs do not have hardware AES acceleration 2022-09-17 15:32:39 minimal: yes, its already slow enough :p 2022-09-17 15:33:10 frag: you need to pass "cryptroot=UUID= cryptdm= root=/dev/mapper/" entries on cmdline in the RPI's cmdline.txt file 2022-09-17 15:35:09 frag: so switching to adiantum should speed it up a bit 2022-09-17 15:37:14 frag: you might want to look at my script which can create a RPI encrypted disk image which you can then "dd" to a SDcard 2022-09-17 15:38:43 frag: https://github.com/dermotbradley/create-alpine-disk-image 2022-09-17 15:40:57 minimal: of thanks! will look at this later ++ 2022-09-17 17:07:20 => load mmc 0:1 ${fdt_addr_r} armada-385-turris-omnia.dtb 2022-09-17 17:07:20 => setenv bootargs earlyprintk console=ttyS0,115200 2022-09-17 17:07:20 => load usb 0:1 ${kernel_addr_r} boot/vmlinuz-lts 2022-09-17 17:07:20 I am trying to get alpine to boot on a turris omnia (armv7 with onboard u-boot). I am using the following commands but it gets stuck at `Starting kernel ...` and the hardware resets after about a minute: 2022-09-17 17:07:20 19617 bytes read in 21 ms (912.1 KiB/s) 2022-09-17 17:07:20 6251008 bytes read in 480 ms (12.4 MiB/s) 2022-09-17 17:07:20 => load usb 0:1 ${ramdisk_addr_r} boot/initramfs-lts 2022-09-17 17:07:22 6743161 bytes read in 525 ms (12.2 MiB/s) 2022-09-17 17:07:22 => bootz ${kernel_addr_r} ${ramdisk_addr_r}:${filesize} ${fdt_addr_r} 2022-09-17 17:07:24 Booting using the fdt blob at 0x2000000 2022-09-17 17:07:24 ## Flattened Device Tree blob at 02000000 2022-09-17 17:07:26 Loading Device Tree to 0f989000, end 0f990ca0 ... OK 2022-09-17 17:07:26 Loading Ramdisk to 0f991000, end 0ffff479 ... OK 2022-09-17 17:07:28 Starting kernel ... 2022-09-18 02:51:26 grisha, my log looks similar but it runs just fine. 2022-09-18 02:52:05 This is what I have for a Flatpak repair one-liner 2022-09-18 02:52:06 flatpak uninstall --unused --delete-data && flatpak repair && sudo ostree prune --repo=/var/lib/flatpak/repo 2022-09-18 06:47:03 I am trying to get alpine to boot on a turris omnia (armv7, onboard u-boot) but u-boot gets stuck at `Starting kernel ...`. Any pointers on how to debug this? 2022-09-18 10:37:58 pinsl: could you boot without 'quiet' parameter? 2022-09-18 10:38:16 maybe it shows something useful 2022-09-18 10:38:46 donoban: I am booting with: setenv bootargs earlyprintk console=ttyS0,115200 modules=loop,squashfs,sd-mod,usb-storage 2022-09-18 10:38:53 and I get no output... 2022-09-18 10:40:23 uhm, do you know if 'Starting kernel ...' comes from Alpine or its u-boot message? 2022-09-18 10:41:55 it's u-boot 2022-09-18 10:44:34 oh I read your other messages 2022-09-18 10:45:59 have you booted another system and console=ttyS0,115200 worked proplery? 2022-09-18 10:48:47 yes, openwrt and debian boot just fine 2022-09-18 10:49:07 but both of these also ship the dtb file for the board which alpine doesn't 2022-09-18 10:54:25 So is the issue a missing dtb file? 2022-09-18 10:55:31 oh hey, its pinsl. quick question, you know how archlinux has their pacstrap command, and debian and its kids deboostrap? something similar for alpine? 2022-09-18 10:55:59 apk 2022-09-18 10:56:09 with --initdb and --root 2022-09-18 10:56:28 ikke: no, they're providing one of their own, not sure the providence. the implication is that alpine is not shipping that kind of dtb because they don't flip any of that soc's CONFIG_* flipped on, which generally causes that family of soc's dtbs to be built. 2022-09-18 10:56:30 hanetzer2: apk add -X https://dl-cdn.alpinelinux.org/alpine//main --initdb --root /path/to/chroot alpine-base 2022-09-18 10:56:52 ikke: nice. and startup scripts/openrc setup? 2022-09-18 10:57:25 it's included in alpine-base 2022-09-18 10:57:42 including the enabling of them and such? 2022-09-18 10:57:46 no 2022-09-18 10:57:55 If you want a bootable system, I'd use setup-alpine 2022-09-18 10:58:49 is that something I can do after the apk add -X in chroot? 2022-09-18 10:58:59 hanetzer: make sure alpine-conf is installed, then yes 2022-09-18 10:59:25 nice. I'll give that a go when I get back to my hw 2022-09-18 11:00:34 Can anyone offer me advice on an Alpine installation bodge? :) 2022-09-18 11:01:04 What is an 'Ik zal duidelijk aangeven wanneer de afbeeldingen in beeld gebracht mogen worden en wanneer ze weer weg mogen. Maar hierbij nog even een indicatie: Na een minuut of 10 volgen de plaatjes 1/tm 3 elkaar vrij snel op. Een paar minuten later volgen plaatje 4 en 5 vlot achter elkaar. En in de laatste 5 minuten zijn plaatje 6 en 7 aan de beurt. 2022-09-18 11:01:13 sorry 2022-09-18 11:01:20 What's an 'Alpine installation bodge' 2022-09-18 11:03:35 I have tried to copy all dirs/files from .iso to a fresh partition and configured for UEFI boot. But the boot fails upon switch_root. I have a feeling the CD initramfs is expecting a specific drive layout 2022-09-18 11:07:19 See error messages like "mkdir: can't create /sysroot//dev: read-only file system", looks like something missing in the middle 2022-09-18 11:08:06 sysroot would become the new root, so /sysroot/dev looks correct 2022-09-18 11:08:57 It's the double forward slash (//) that looks suspicious 2022-09-18 11:09:20 that can happen if you do foo=root/; $foo/bar 2022-09-18 11:09:40 btw. how does alpine handle, if at all, 'upgrading' packages like u-boot-$foo on devices where u-boot is dd'd into a block device? 2022-09-18 11:09:58 yeh, double slash isn't a problem I've ever seen actually be a problem. 2022-09-18 11:12:19 ok 2022-09-18 11:16:31 My drive layout is two partitions, sda1 and sda2, sda1 is vfat ESP containing Grub and config, sda2 is ext4 containing "apks/", "boot/" and ".alpine-release" 2022-09-18 11:17:28 The ISO's build of Grub doesn't contain an ext4 module, so I had to add that and use insmod in the config 2022-09-18 11:18:39 I also added "root=UUID=" to the kernel params 2022-09-18 11:22:52 I've tried adding the "single" param and peppering the "/init" script with echos to try to determine where exactly it fails, then exit from singlemode to continue, but it never seems to trigger anything that I add 2022-09-18 11:31:42 I basically want diskless (root in RAM) from an ext4 partition source 2022-09-18 11:49:14 SMurf7: / should be alreayd tmpfs, if it says it's read-only, it seems something else is mounted there 2022-09-18 11:53:30 does alpine handle cross-arch apk add -X , generally speaking? 2022-09-18 11:53:51 depends what you mean 2022-09-18 11:53:58 you can create chroot for different arch 2022-09-18 11:54:06 you can't install different arch package 2022-09-18 11:54:20 *generally speaking* 2022-09-18 11:56:27 panekj: I mean, similar to how you can do the debian debootstrap to create a cross-arch chroot 2022-09-18 11:59:03 apk add -X https://dl-cdn.alpinelinux.org/alpine/edge/main --root /submarine/data/alpine-chroot --initdb --arch aarch64 --allow-untrusted alpine-base 2022-09-18 11:59:04 add --arch to apk add 2022-09-18 12:02:42 nice. 2022-09-18 12:11:59 ke it is, which controls the fstab. 2022-09-18 12:11:59 It's too complicated for me to understand the source though.[201~init.in) it looks to me like it is, which controls the fstab. 2022-09-18 12:11:59 It's too complicated for me to understand the source though.[201~init.in) it looks to me like it is, which controls the fstab. 2022-09-18 12:12:50 does anyone know if there is a way to force the boot disk to mount at a specific mountpoint? 2022-09-18 12:15:33 I had read that it's not possible but taking a look at the initfram ( https://github.com/alpinelinux/mkinitfs/blob/master/initramfs-init.in ) it looks to me like it is, which controls the fstab. 2022-09-18 12:15:53 however It's too complicated for me to understand the source 2022-09-18 12:16:32 Not really sure what you want to achieve exactly 2022-09-18 12:16:48 ok i explain better 2022-09-18 12:17:36 I have an alpine diskless that automatically mounts in /media/sda1 2022-09-18 12:18:19 can I force it to mount automatically in another partition (for example /media/usb or /media/alpine) ? 2022-09-18 12:18:36 for what reason do you want that to happen? 2022-09-18 12:18:43 I had tried to put in a fstab rule and it doesn't work 2022-09-18 12:18:47 that's controlled by mdev 2022-09-18 12:19:01 You need to adjust mdev 2022-09-18 12:20:10 because I want to avoid always having to change the apkcache location when for various reasons the memory changes mountpoint 2022-09-18 12:20:24 ikke: what is mdev? 2022-09-18 12:22:33 it seems a service 2022-09-18 12:23:20 https://skarnet.org/software/mdevd/ 2022-09-18 12:24:37 thanks 2022-09-18 12:24:42 i take a look 2022-09-18 12:26:05 note that mdevd != mdev 2022-09-18 12:27:26 ok but in the repo we have only mdevd and not mdev 2022-09-18 12:27:38 mdev is provided by busybox 2022-09-18 12:27:46 busybox-mdev-openrc 2022-09-18 12:27:48 ah is this ? 2022-09-18 12:28:02 those are service units 2022-09-18 12:28:50 no ok, mdev is already installed with busybox 2022-09-18 12:29:24 I figured that skarnet page would be better documentation than bb :^) 2022-09-18 12:30:09 but skarnet speak abaut mdevd not mdev, is it the same? 2022-09-18 12:31:26 or I can simple read their documentation and after apply the readings for mdev of busybox ? 2022-09-18 12:33:53 gentoo documentation quoting: "The mdev application is definitely suitable as long as the system does not use a full-fledged desktop" 2022-09-18 12:34:14 helmen[m]: mdevd is mdev but run as a daemon instead of a program that is executed 2022-09-18 12:35:43 ok but i need that this mountpoint configuration happen before the apk loading 2022-09-18 12:36:54 I really don't know the boot sequence of alpine, rc.service "boot" it is the correct place? 2022-09-18 12:37:09 s/it// 2022-09-18 12:37:26 helmen[m]: You mean before lbu restores packages? 2022-09-18 12:38:30 yes 2022-09-18 12:40:21 Then openrc boot levels are already too late 2022-09-18 12:41:29 right now the pen is not restoring packages because the apk cache (/etc/apk/cache symlink) points to another mountpoint. I can adjust it though, and it bothers me to have to do it every time and I'd rather be able to force a fixed mountpoint to the disk. 2022-09-18 12:48:15 But does mdev continue to be the solution or not even mdev ?But does mdev continue to be the solution or not even mdev ? 2022-09-18 12:49:27 pinsl: hi 2022-09-18 12:49:36 mps: :) 2022-09-18 12:49:58 build not yet finished 2022-09-18 12:50:15 going to make another coffee in meantime 2022-09-18 13:01:45 pinsl: https://dev.alpinelinux.org/~mps/linux-edge-5.19.9-r0.apk 2022-09-18 13:03:24 to install it mount sdcard and run `apk --root=/mnt --allow $pkgname` on another alpine linux arm32 machine 2022-09-18 13:03:41 `apk --root=/mnt --allow add $pkgname` 2022-09-18 13:04:35 Saijin_Naib: this oneliner didn't help and i had to use zoom through chromium, it wasn't that bad afterall. thanks anyway :) 2022-09-18 13:15:26 mps: I don't have another arm32 machine, I just extraced the package onto a usb disk and tried booting it but it gets stuck like before 2022-09-18 13:30:36 pinsl: hmm, will try to find somewhere kernel config which works 2022-09-18 13:42:52 mps: I'd like to try the openwrt config but I am struggling to cross compile with abuild 2022-09-18 13:46:05 pinsl: in kernel tree is the file arch/arm/configs/mvebu_v7_defconfig which probably will work but I don't like idea to have option set =y 2022-09-18 13:46:34 and cross build on alpine is very hard if possible at all 2022-09-18 14:04:10 pinsl: could you download linux-edge again, same file, and test. this is built by `make mvebu_v7_defconfig` 2022-09-18 14:05:16 if you can get console messages from kernel rest will be easier 2022-09-18 14:08:49 bingo :) 2022-09-18 14:08:58 [ 0.000000] Linux version 5.19.9-0-edge (mps@mps-edge-armv7) (gcc (Alpine 12.1.1_git20220630) 12.1.1 20220630, GNU ld (GNU Binutils) 2.39) #1-Alpine SMP Sun, 18 Sep 2022 13:58:39 +0000 2022-09-18 14:09:04 mps: what do you need? 2022-09-18 14:09:50 https://pastebin.com/mBrju1SM 2022-09-18 14:21:56 pinsl: ok, now we know it can boot 2022-09-18 14:22:52 and please if you can use ix.io or tpaste.us for pasting text and not this stupid pastebin 2022-09-18 14:23:15 mps: haha ok I will 2022-09-18 14:23:33 now we can make diff with this kernel and previous one which didn't booted 2022-09-18 14:24:01 bu we made first step to make normal linux-edge to work on turris 2022-09-18 14:26:12 s/bu/but/ 2022-09-18 14:29:31 pinsl: kernel couldn't find root partition `Please append a correct "root=" boot option; here are the available partitions:` 2022-09-18 14:31:18 could you paste u-boot config 2022-09-18 14:32:05 well I only had the kernel on disk 2022-09-18 14:33:14 I now combined your kernel with the alpine standard iso and loaded it with the lts ramdisk: https://tpaste.us/WQgL 2022-09-18 14:33:39 now idea how much sense that makes but it reaches a login prompt... 2022-09-18 14:34:18 oh, it boot now? 2022-09-18 14:35:13 With your kernel yes 2022-09-18 14:36:34 good, now you have to install linux-edge package to get kernel modules and initramfs-edge 2022-09-18 14:38:10 which dirs in / do i need to copy the contetnt of to copy the system to another partition? (and which are ok to only mkdir) 2022-09-18 14:38:18 I have no idea how boot process works on this device so I'm 'shooting in the dark' 2022-09-18 14:39:12 what is your machine on which you prepare sdcard 2022-09-18 14:39:19 and what OS 2022-09-18 14:39:24 arch x86 2022-09-18 14:39:28 and its a usb stick not a sd card 2022-09-18 14:39:43 the board does not have a sdcard reader only soldered on mmc 2022-09-18 14:40:11 ah 2022-09-18 14:40:44 so you copy files from your machine to usb stick and boot from it? 2022-09-18 14:40:51 correct 2022-09-18 14:40:57 ok 2022-09-18 14:42:49 look here to get idea what is needed on your workstation https://arvanta.net/alpine/alpine-on-visionfivev1/ 2022-09-18 14:43:41 that is `for other linux distros apk-tool-static could be downloaded from https://dl-cdn.alpinelinux.org/alpine/edge/main/x86_64/ and unpacked with tar` 2022-09-18 14:44:15 then you can use `apk` on other distros and archs 2022-09-18 14:44:26 I have a laptop with alpine set up 2022-09-18 14:44:33 hah 2022-09-18 14:44:45 then it will be easier 2022-09-18 14:45:49 but anyway mount usb stick and run `apk --arch armv7 --root /mnt --allow-untrusted add linux-edge-....` 2022-09-18 14:46:25 linux-edge is full path and name of kernel apk you downloaded 2022-09-18 14:47:27 and replace boot/initramfs-lts with boot/initramfs-edge in u-boot config 2022-09-18 14:58:37 mps: hmm I am getting an `unable to lock database: no such file` error; I am pointing --root to the root of the usb disk that has the alpine standard iso conents on it, is that correct? 2022-09-18 14:59:53 pinsl: this will not work on iso image 2022-09-18 15:00:18 yeh, disk with dd'd iso is ro 2022-09-18 15:00:33 oh no i am unpacking the iso contents on a fat32 patition 2022-09-18 15:03:59 ah, I see turris have email on contact page, I think I will write mail to them and ask more infos 2022-09-18 15:04:54 pinsl: do you know by chance does u-boot on board supports booting with extelinux.conf 2022-09-18 15:05:12 yes it does, I tried that 2022-09-18 15:05:19 extlinux.conf 2022-09-18 15:05:52 do you know does it need to be on separate partition 2022-09-18 15:06:29 so I could try to make bootable usb/mmc image which could be dd-ed 2022-09-18 15:06:34 to skip iso 2022-09-18 15:06:37 Well I once tried booting the generic arm image and it found the extlinux.conf file 2022-09-18 15:07:55 nice, I will try later (maybe at night) to make image 2022-09-18 15:07:56 u-boot also loads grub on the standard 2022-09-18 15:08:04 iso 2022-09-18 15:08:18 aha, good to know 2022-09-18 15:08:51 I tested grub on armv7 only in qemu VM 2022-09-18 15:09:58 yeah grub works fine on the omnia, but of course when it loads the kernel it fails 2022-09-18 15:10:04 I have now to try to make linux 6.0-rc and test does it boots on M1 2022-09-18 15:10:50 is the M1 stuff mainline? 2022-09-18 15:12:15 no, not yet 2022-09-18 15:13:00 some things are but now when Linus uses M1 I hope things will go faster 2022-09-18 15:13:30 if you have M1 you can use alpine on quite fine 2022-09-18 15:13:42 on it* 2022-09-18 15:14:19 there are even small guide ;) https://arvanta.net/alpine/install-alpine-m1/ 2022-09-18 15:14:54 alpine was the first distro used as daily workstation on apple silicon, btw 2022-09-18 15:15:09 oh nice 2022-09-18 15:15:33 and first user was 'your truly' :) 2022-09-18 15:17:06 sorry for self promotion (but you will forgive me because I promote alpine at the same time ;) ) 2022-09-18 16:29:07 is there a way of getting the real package name of installed packages without guessing by parsing `apk info -I`? i ultimately want to see which packages i have installed that aren't in world 2022-09-18 16:30:54 lopid: try something like.......... 2022-09-18 16:30:55 grep ^'o:' /lib/apk/db/installed | less 2022-09-18 16:31:30 maybe o: is the wrong letter. but you get the idea. 2022-09-18 16:32:06 that's a new one to me, thanks 2022-09-18 16:32:24 I think it's actually P: 2022-09-18 16:37:56 looks like this will do it: while IFS=: read -r _ p; do grep -Fxq "$p" /etc/apk/world || echo "$p"; done < <( grep '^P:' /lib/apk/db/installed ) 2022-09-18 16:43:20 ACTION adds a "| sort | less" on for good measure 2022-09-18 16:43:24 that's pretty cool 2022-09-18 16:44:34 me too :) 2022-09-18 17:34:23 this is a lot faster: comm -13 <( sort /etc/apk/world ) <( grep '^P:' /lib/apk/db/installed | while IFS=: read -r _ p; do echo "$p"; done | sort ) 2022-09-18 17:51:11 and this even faster: comm -13 <( sort /etc/apk/world ) <( grep "^P:" /lib/apk/db/installed | cut -d: -f2 | sort ) 2022-09-18 17:53:50 pinsl: could you try to download https://dev.alpinelinux.org/~mps/armv7/armv7-mmc.img.gz, gunzip it and 'dd' it to your usb stick and try to boot on omnia 2022-09-18 23:57:11 sethmhurst[m]: if you're around I've testing a script for modify stock ISO for enabling espeak. I'd need a little info from yourself regarding some VM/machine specifics to tweak it for you. 2022-09-19 00:47:19 does there exist a 'linux-edge' variant of the generic-arm tarball? 2022-09-19 00:47:35 lts doesn't have various rockchip configs enabled :P 2022-09-19 07:38:05 is there any location where I can find the current (supported) versions of alpine linux? e.g. in json/xml/yml or any other format? 2022-09-19 07:38:28 https://alpinelinux.org/releases.json 2022-09-19 07:39:04 https://alpinelinux.org/releases/ 2022-09-19 07:39:53 and check for eol_date key 2022-09-19 07:40:14 oh, wonderful, thanks! 2022-09-19 07:43:32 Note that only the main repo is supported until the eol date 2022-09-19 09:11:33 I vaguely remember there was a command to concatenate two ongoing streams 2022-09-19 09:16:48 hanetzer1: no, official alpine releases user linux-lts 2022-09-19 09:20:14 hanetzer1: and tarballs are intended for containers (docker, lxc ...) and not for bare metal install, though could be installed bare metal with some additional 'tweaks' 2022-09-19 09:21:37 if tarballs are intended for containers then why raspberry pi and generic arm are released with tarballs 2022-09-19 09:25:02 panekj: side effect 2022-09-19 09:25:35 for installation on bare metal there are isos 2022-09-19 09:26:33 OS on raspberry pi is installed via extracting rootfs 2022-09-19 09:27:05 rpi is in better position because for them bootloader could be set and added to tarball 2022-09-19 09:28:10 but generic-arm is complicated, it need different u-boot for different SBCs, different boot configs etc 2022-09-19 09:29:15 panekj: I wrote earlier about RPi install https://arvanta.net/alpine/flash-rpi-zero/ 2022-09-19 09:30:21 and I wrote about better install method for different SBCs https://arvanta.net/alpine/install-alpine-armv7-sbc/ 2022-09-19 09:31:24 not perfect I know, but more as a hints/starting point for newcomers to alpine who want to install alpine on arm boards 2022-09-19 09:45:16 panekj, you don't install an OS on a pi like you do on a laptop/desktop 2022-09-19 09:45:31 I know 2022-09-19 09:45:36 and with qemu-user it's so convenient 2022-09-19 09:45:52 that's why I'm correcting mps, since "tarballs are for containers" is just untrue 2022-09-19 09:45:58 basically mount; tar xf; chroot apk add; chroot rc-update; umount -> plug into your pi 2022-09-19 09:46:16 I love that alpine images are provided via tarballs 2022-09-19 09:46:45 it takes somewhat like 2-3 minutes to install our appliances while our old setups based on raspbian takes half hour 2022-09-19 09:49:30 panekj: you don't need to correct me, I know well what is for what, but I'm just lazy to repeat all these for years 2022-09-19 09:49:46 then don't mislead people? (: 2022-09-19 09:50:10 meh, people have to learn to think and read 2022-09-19 11:12:39 ikke: https://gitlab.alpinelinux.org/alpine/mkinitfs/-/issues/5 2022-09-19 11:12:45 the proble with fstab is a bug 2022-09-19 11:20:27 Hey guys, I'm trying to get alpine to run on a NanoPi R2S (based on Rockchip 3328). I made it through u-boot and am now in the recovery shell, but it seems like I can't get further because no block devices are being detected (neither sd card nor usb drive). The `dw_mmc_rockchip` kernel module is being loaded, as are `sd_mod` and `usb_storage`. Do you have any ideas? 2022-09-19 11:27:29 I see armbian for the same board has some kernel modules built in instead of used as modules, like `CONFIG_ROCKCHIP_PHY`, however I'm not sure if that would help and if so, which one I'd need :/ 2022-09-19 12:52:26 dmesg might have more information 2022-09-19 12:52:30 like e.g. missing firmware 2022-09-19 13:03:29 I've went through `dmesg` pretty extensively and haven't found anything 2022-09-19 13:57:23 Hi. Anyone familiar with diskless setup? It seems partition is no longer mounted RO on 3.16 :-/ 2022-09-19 15:47:12 mps: image works :) https://tpaste.us/6EdZ 2022-09-19 15:48:28 pinsl looks like you're doing something similar to what I want to do 2022-09-19 15:51:52 minimal: got raspberry to prompt for enc passphrase :] BUT there was an error about midding dm_mod or something, do you know what could be the problem? kinda using a unenc sys disk now, and mounting an enc partition manually.. 2022-09-19 15:54:57 frag: did you add "cryptsetup" to the features list in /etc/mkinitfs/mkinitfs.conf and rebuild the initramfs? 2022-09-19 15:55:51 minimal: eh no! 2022-09-19 15:57:24 frag: that is need to add the crypto kernel modules to the initramfs - otherwise how can the initramfs mount the encrypted rootfs? 2022-09-19 15:58:54 minimal: i basically dont understand anything about booting :p, how do you rebuilt initramfs? 2022-09-19 15:59:22 mkinitfs 2022-09-19 15:59:30 ty! 2022-09-19 16:19:03 pinsl: good, now we have to find what from this kernel should be enabled on real linux-edge (this one is custom configured for your device) 2022-09-19 16:20:01 pinsl: are you ready to help by installing different kernels and test them 2022-09-19 16:20:41 we can do this in few steps, no hurry 2022-09-19 17:01:20 minimal: hmm still getting "Cannot initialize device-mapper. Is dm_mod kernel module loaded?" 2022-09-19 17:02:20 frag: are you using the correct cmdline options? (in grub.cfg or extlinux.conf) 2022-09-19 17:04:04 minimal: alpine rpi uses those?? 2022-09-19 17:05:08 i use cmdline.txt, fstab and now mkinitfs.conf 2022-09-19 17:06:38 frag: forgot you're using RPI, ok so in cmdline.txt do you have the right values for encryption? 2022-09-19 17:08:17 minimal: i used what u gave me: cryptroot=UUID= cryptdm= root=/dev/mapper/ 2022-09-19 17:09:15 frag: and you put the correct UUID in there for the LUKS partition? 2022-09-19 17:09:25 im still using luks tho, added lvm too 2022-09-19 17:10:05 frag: LVM-on-LUKS or LUKS-on-LVM? 2022-09-19 17:10:41 minimal: ill double-check, it doesnt need to be quoted does it? the uuid 2022-09-19 17:10:59 minimal: not sure lvm-on-luks or the other way around :S 2022-09-19 17:11:07 u 2022-09-19 17:11:39 frag: well it is kinda important to know which way around it is... 2022-09-19 17:14:23 minimal: i most likely used this: doas cryptsetup luksFormat --type luks1 $1 2022-09-19 17:18:55 freq: that don't give any indication of whether its LVM-on-LUKS or LUKS-on-LVM. Have you been following any guide on setting this up? 2022-09-19 17:20:11 minimal: no i didnt find a guide 2022-09-19 17:23:06 frag: so you're just "randomly" trying to get it working? did you look at my script? 2022-09-19 17:25:52 frag: there are also generic guides on the Alpine Wiki 2022-09-19 17:30:05 ah i have followed the alpine wiki on my normal computer yes, and just tried to find bits and pieces from the net regarding rpi 2022-09-19 17:31:46 when i ran your script it said arch:x86_64 (running ./create-alpine-disk-image --physical rpi4 --release 3.16 --script-filename create.sh ) 2022-09-19 17:32:52 Hi... (full message at ) 2022-09-19 17:36:25 heh 2022-09-19 17:39:36 frag: perhaps try adding "--lvm" and "--encrypt" options when calling my script and then look at create.sh to see how it does things? 2022-09-19 17:40:33 alright 2022-09-19 18:31:36 mps: ready 2022-09-19 19:22:36 pinsl: I was busy today and I'm already tired. Will try to find some free time to work on it. will inform you when I prepare new kernel or disk image 2022-09-19 19:23:06 and thank you console output, it will help 2022-09-19 19:24:39 mps: no problem 2022-09-19 20:33:32 Sup 2022-09-19 20:39:37 In the name of Elohim the one and only the King of kings the right the sustainer of existance the most powerful 1_ Judaism is a Monotheism religion, it is the belief in Elohim the only one, no God except he, from prophet Moses way, The right belief is: there is no God except one Elohim; and that prophet Moses is the messenger of Elohim 2_ No body asked U to commite crimes and kill innocent believers and poor people, Systemically which you are judged 2022-09-19 20:39:49 ---------- 2022-09-19 20:39:49 3_ By the authority of Elohim the one the greatest and from Prophets Moses Aaron and Joshua way: 4_ People u r now appointed by us to achieve big achievements - its your bad acts and deeds and God the redeemer Big punishments getting back on u till u do whats requested well. 2022-09-19 20:39:59 ---------- 2022-09-19 20:40:00 5_ You are requested to eliminate personalities and positions in power and in the ruling Israeli current regime inplace of what u did for this tyrant regime of crimes until its end, under consequence of big punishments from Elohim the Greatest redeemer if u were late (complicated sinarios are welcomed by our Greatest Elohim). 2022-09-19 20:40:10 ---------- 2022-09-19 20:40:10 Evidences of whats mentioned above from The most famous Ten Commandments of his majesty to Moses:[ I am the Lord your God, who brought you out of Egypt, out of the land of slavery You shall have no other gods before me (1) You shall not make for yourself a statue (2) You shall not murder(6)] Exodus 20 2022-09-19 20:40:22 ---------- 2022-09-19 20:40:22 firsts are first - U can Send 100 of these message [Moses of Elohim. M] 2022-09-19 20:40:22 Which indicates that Monotheism [God is One] is the most important Commandment of All in Judaism, aswell as in islam; as they both worship same God, and indicates the great wrongness of paganism and also indicates the big dangerousness of murder 2022-09-19 20:40:56 ok 2022-09-19 23:37:33 sethmhurst[m]: you around? 2022-09-20 01:07:04 Here 2022-09-20 01:22:43 If you have a Gmail account, a computer or a smartphone and you have 2 hours a day, you can earn 345 EUR a day. Ask me how.... (full message at ) 2022-09-20 01:30:52 But I want 346 EUR a day 2022-09-20 01:39:04 i guess you will need two Gmail accounts, two computers or smartphones, and 4 hours a day 2022-09-20 08:10:02 can i merge .apk(s) into the diskless rootfs, so they don't use RAM? 2022-09-20 09:21:00 for some reason `mount` ignores my swapfile. I can however use `swapon /path/to/swapfile` just fine, but this means I have to do it manually every time I boot my PC. `--verbose` doesn't actually tell me why it ignores it, just that it does. Anything I can do to fix this? 2022-09-20 09:23:58 `swapon -a` (so it reads the swapfile from fstab) also seems to work fine 🤔 2022-09-20 09:24:07 merge .apk(s) into the diskless rootfs: gave up, found this instead: https://wiki.alpinelinux.org/wiki/Raspberry_Pi#Persistent_storage 2022-09-20 09:32:10 hmm there is a swap OpenRC service which works if I start it manually, but not when I add it to the boot runlevel. Does it need to be in some other runlevel? 2022-09-20 09:52:25 PureTryOut: boot runlevel is correct; enable rc_logger in rc.conf and then look at why it's failing at startup ? 2022-09-20 09:52:55 ideas: is the swapfile on a partition that needs to be mounted first? 2022-09-20 09:53:26 it is but the swapfile is mentioned later in fstab 2022-09-20 09:53:42 dwfreed: thanks for the tip, will do next time I reboot 2022-09-20 09:54:20 you need to edit /etc/conf.d/swap 2022-09-20 09:54:26 PureTryOut: 2022-09-20 09:54:33 oh, that would actually explain it; if the swapfile is not on an fs already mounted, then you have to reorder the deps 2022-09-20 09:54:42 because by default swap runs before localmount 2022-09-20 09:55:33 ooohhhh... Yeah `/etc/conf.d/swap` actually mentions it 2022-09-20 09:56:55 thanks that makes sense 2022-09-20 09:57:25 fstab has a column which influences the order 2022-09-20 09:57:35 fstab isn't automatically done from top to bottom 2022-09-20 09:57:48 check the passno field 2022-09-20 09:58:11 (hint: 6th, last field) 2022-09-20 09:58:20 passno is for fsck, not mounting 2022-09-20 09:58:28 https://man7.org/linux/man-pages/man5/fstab.5.html "The order of records in fstab is important because fsck(8), mount(8), and umount(8) sequentially iterate through fstab doing their thing." 2022-09-20 09:58:49 "The sixth field (fs_passno). This field is used by fsck(8) to determine the order in which filesystem checks are done at boot time." 2022-09-20 10:00:11 besides that theory, give it a try. ;) 2022-09-20 10:00:15 and mount (and thus localmount) does not handle swap at all; that is what swapon/swapoff (and the swap initscript) are for 2022-09-20 10:01:15 so, does swap depend on localmount/mount? if not, you can either add rc_need in swap for localmount or move the swapfile into / 2022-09-20 10:04:05 satmd: having just searched the util-linux source, I can definitively say that passno is not used for mounting in util-linux 2022-09-20 10:29:27 Hi I want to install alpine on USB(diskless). I used alpine-setup and at a later.stage I was asked to choose a device to store config files. I chose trueroot but I get the reply: device not found. Pls help 2022-09-20 12:14:42 Hey guys, does anyone know how I can get `kernel/drivers/phy/rockchip/phy-rockchip-emmc` into the initramfs? There's no `phy` in `/etc/mkinitfs/features.d/` 2022-09-20 12:16:59 You can create one yourself 2022-09-20 12:33:41 Right, that's what I thought, but my end goal is to have this done in a `aports/scripts/mkimg.rockchip.sh` kind of way, and I guess it's not clear to me yet where it find those features and if I can even add one there. I'm assuming it takes the `mkinitfs` package for the target arch and there's no way to patch it without editing `mkimg.base.sh` maybe? 2022-09-20 12:55:34 filoozom: is this for 32 or 64 bit machine 2022-09-20 13:07:02 For a 64 bit RK3328 2022-09-20 13:07:22 I'm trying to get Alpine running on it, managed to get through u-boot to the kernel, but now it can't find my SD card 2022-09-20 13:07:34 I recompiled the kernel and got a USB stick to be detected, but still no SD 2022-09-20 13:08:13 And re-compiling takes quite a bit of time even with a 48C EPYC, so I'm checking if I can avoid it by just building a new initramfs with the right modules 2022-09-20 13:11:08 I have kernel for rk3399 and I'm using it fine on gru-kevin for few years. not sure if it can help you but I can paste config if want it to try for your board 2022-09-20 13:11:47 filoozom: i recommend ccache and doing the kernel build as producible as possible (https://www.kernel.org/doc/html/latest/kbuild/reproducible-builds.html) 2022-09-20 13:52:35 Good day everyone. Where in alpine Linux do I set my kernel commandline parameters, please? 2022-09-20 13:54:12 cts: that depends on which bootloader you are using 2022-09-20 14:08:24 pinsl: I pushed linux-edge 5.19.10 and enabled some options for mvebu platform (which is in omnia). could you please test does it boots 2022-09-20 14:09:19 pinsl: also I sent mail to turris asking them for help us to make omnia working with alpine out-of-the-box 2022-09-20 14:24:24 minimal: as I have found /boot/menu.c32 I would assume syslinux. I haven't changed the "alpine-install" standard so far. 2022-09-20 14:26:22 minimal: Ahh, and there's my /boot/extlinux.conf . Thanks ;) 2022-09-20 14:27:40 Now that I've been helped with this, one more question if I may: is there a way to suppress Alpines service-starting-up "[OK]" / "[FAIL]" boot messages? 2022-09-20 14:38:47 cts: the "alpine install" bootloader depends on what you installed it on - if its booting via UEFI then grub is installed, whereas if its installed on a BIOS machine then Syslinux is used 2022-09-20 14:51:20 cts: add quiet to the kernel arguments 2022-09-20 15:03:47 Has anyone tried to package a python package built with scikit-build? I'm new to packaging, and having a headache trying to get this sorted 2022-09-20 15:04:14 I have all the build dependencies working, and the plugins I am trying to package get built, but I can't figure out what should be in the `package()` set 2022-09-20 15:04:16 *step 2022-09-20 15:05:05 ikke: I already had that, that only seems to affect the kernel boot messages. I have found a rather obscure youtube video detailing how to " >/dev/null" all the init scripts. I'll try that. 2022-09-20 15:13:28 I have an unorthodox exec >/dev/null in /etc/rc.conf 2022-09-20 15:13:34 please do not try this at home 2022-09-20 15:14:48 or kernel option console=/dev/null 2022-09-20 15:58:54 mps: awesome, will test in about an hour 2022-09-20 16:07:16 Hi all - is there a way to have an APKBUILD create/update an environment variable on the system it's being installed on? Specifically, I am packaging a plugin for PDAL, and I need to update PDAL_DRIVER_PATH to include the location of the installed plugin 2022-09-20 16:47:14 I've tried making image, standard profile and booting it with qemu uefi. When uefi shell appeared, I tried to launch grub through it, but it says: 'Command Error Status: Access Denied' 2022-09-20 16:52:20 Ermine: using qemu-system-x86_64 directly and passing it OVMF files? 2022-09-20 16:52:44 minimal: I launched it through virt-manager 2022-09-20 16:53:27 Ermine: ok, I assume it makes use of OVMF then. You've created a disk image or an ISO image? 2022-09-20 16:53:41 ISO image 2022-09-20 16:55:44 Upd: I chose OVMF manually and it worked. 2022-09-20 16:56:39 Ermine: the online references I found about that error related to UEFI Secure Boot but I assume you weren't trying that (it uses a different OVMF AFAIK) 2022-09-20 16:58:28 probably. There are 4 options to choose from: "UEFI", "UEFI x86_64: /usr/share/OVMF/OVMF_CODE_4M.fd", "UEFI x86_64: /usr/share/OVMF/OVMF_CODE_4M.secboot.fd" and "UEFI x86_64: /usr/share/OVMF/OVMF_CODE_4M.ms.fd" 2022-09-20 16:58:53 Maybe "UEFI" means the secboot one 2022-09-20 16:58:57 Ermine: the "secboot" stuff is for Secure Boot AFAIK 2022-09-20 17:01:26 Ermine: ah, "ms" in the name means Secure Boot with Microsoft signing keys and is apparently a softlink to the secboot file 2022-09-20 17:01:56 basically you want /usr/share/OVMF/OVMF_CODE_4M.fd 2022-09-20 17:02:08 Yes, this one works 2022-09-20 18:51:26 mps: the new kernel does not boot: https://tpaste.us/nVOM 2022-09-20 18:52:02 mps: I tried starting it directly as I did not manage to install it on the other image you sent me 2022-09-20 18:54:33 pinsl: I see. have to find more time to look what is needed to enable to get at least console to work. then it will be easier for testing 2022-09-20 18:55:22 mps: hopefully the turris guys send an answer 2022-09-20 18:56:40 pinsl: I hope, but usually have to wait few days for companies to answer 2022-09-20 18:58:40 debian kernel config you posted is 'confusing' because they don't follow upstream style of config file 2022-09-20 23:40:42 Now I'm very new to this, but it seems like sdmmc should be enabled (https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/rockchip/rk3328-nanopi-r2s.dts#L348-L361) and use "rockchip,rk3328-dw-mshc" or "rockchip,rk3288-dw-mshc" (https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/rockchip/rk3328.dtsi#L865-L875) 2022-09-20 23:40:46 However, loading these modules clearly is not enough (https://dpaste.org/dkDZz) 2022-09-20 23:41:41 Piraty_: thanks, that makes sense. For now I was just going the `abuild -r` route, but clearly that's super slow 2022-09-20 23:42:04 mps: that would be lovely! Have been bashing my head in for a few days now and really can't get it to work... 2022-09-21 08:27:26 mps: I tried starting it directly as I did not manage to install it on the other image you sent me 2022-09-21 08:27:47 oops sorry 2022-09-21 08:36:27 pinsl: don't understand what you intended to say 2022-09-21 08:45:20 pinsl: I will create new image today with bootable kernel and on with name linux-mvebu for tests. u-boot will have menu to select kernel to boot and you will be able to install testing kernels from omnia, I hope network interface works 2022-09-21 08:56:45 test 2022-09-21 10:03:00 anyone been able to get minikube working on alpine? 2022-09-21 11:12:28 mps: ok sounds good 2022-09-21 11:19:25 got it to work with root, rootless is an issue 2022-09-21 11:42:43 pinsl: I'm uploading new usb/mmc image. it have two options in u-boot to select which kernel to boot: 1 - is defconfig for mvebu (omnia), 2 - standard linux-edge 2022-09-21 11:43:20 you have 5 seconds on boot to select else it will boot mvebu by default 2022-09-21 11:43:49 https://dev.alpinelinux.org/~mps/armv7/armv7-mmc.img.gz 2022-09-21 12:09:58 mps: could I ask you for your RK3399 config? :) I'm guessing it's not going to solve my issue as it's probably botting from eMMC instead of SD but might as well try 2022-09-21 12:12:20 mps: great, will try it later 2022-09-21 12:14:18 filoozom: what's your issue? 2022-09-21 12:15:43 hanetzer: I'm trying to get alpine running on RK3288. Made it through u-boot into the kernel, but then it can't boot because it can't find the SD card. 2022-09-21 12:16:05 ah. what's your device? 2022-09-21 12:16:38 NanoPi R2S 2022-09-21 12:16:50 My latest update: 2022-09-20 23:40:42 Now I'm very new to this, but it seems like sdmmc should be enabled (https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/rockchip/rk3328-nanopi-r2s.dts#L348-L361) and use "rockchip,rk3328-dw-mshc" or "rockchip,rk3288-dw-mshc" (https://github.com/torvalds/linux/blob/master/arch/arm64/boot/dts/rockchip/rk3328.dtsi#L865-L875) 2022-09-20 23:40:46 However, load 2022-09-21 12:17:05 is not enough (https://dpaste.org/dkDZz) 2022-09-21 12:17:10 aight, you have proper serial on that so I can't really help you there. 2022-09-21 12:17:41 Proper serial? 2022-09-21 12:18:08 yeah like using a usb to uart adapter to see boot failure messages. 2022-09-21 12:18:16 I do have that 2022-09-21 12:18:44 filoozom: http://ix.io/4b2C 2022-09-21 12:19:33 filoozom: you probably have to change/add some drivers or options 2022-09-21 12:19:58 filoozom: dmesg? 2022-09-21 12:20:06 mps: I guess so, but I've compiled all rockchip related things as builtins and I'm not sure what next steps to take 2022-09-21 12:20:17 I don't have any rk3288 to test 2022-09-21 12:21:18 my config build kernels which can boot from internal emmc, external mmc and usb ssd 2022-09-21 12:21:46 hanetzer: https://dpaste.org/oW1YX this is a full boot log with debug_init and including dmesg 2022-09-21 12:24:23 filoozom: what alpine img you use to boot machine 2022-09-21 12:25:23 Initially I used generic arm, but now I'm running a custom kernel and mkimg to try to get the SD to work 2022-09-21 12:28:07 filoozom: well, I don't like to give advises but long ago I stopped these methods and created scripts to making disk images or making bootable media directly 2022-09-21 12:28:37 here are some I documented a little https://arvanta.net/alpine/ 2022-09-21 12:28:54 they could be used as hints or starting point 2022-09-21 12:32:14 Awesome, I'll read through it, thanks! 2022-09-21 12:41:11 Could it be mdev related per chance? 2022-09-21 12:46:38 maybe, didn't looked carefully in your boot log 2022-09-21 12:49:02 hmm, it can't find boot media 2022-09-21 12:50:09 `nlplug-findfs: exit due to timeout (5000)` 2022-09-21 12:51:30 filoozom: maybe you could try with linux-edge instead of linux-lts 2022-09-21 12:52:04 iirc I enabled most drivers for rockchip 2022-09-21 13:13:19 mps: yes, the new kernel I'm building is based on edge already, figured it had more chance of working 2022-09-21 13:14:23 At this point I have all `CONFIG_*ROCKCHIP*` set to `y`, still doesn't work though haha 2022-09-21 13:16:23 hmm in you boot log I see this `Linux version 5.15.68-0-lts (filoozom@adaff143f7ec` 2022-09-21 13:18:41 Click on the link and thank me later 👉👉👉https://t.me/+6418gUGZ6qQ2YzE0 2022-09-21 13:21:15 mps: ah you're right, it's `linux-lts` in the `master` branch, not actually `linux-edge`, my bad 2022-09-21 13:50:51 Hi all - I'm packaging some new projects. Is there a way to run the APK pipelines that get run on Merge Requests against my branch in my fork on the Alpine Gitlab? 2022-09-21 13:51:56 Hi all - I'm packaging some new projects. Is there a way to run the APK pipelines that get run on Merge Requests against my branch in my fork on the Alpine Gitlab? 2022-09-21 13:54:29 Hi, I am trying to see if there is a known path for mapping mdev events to a specific usb device ID? The best I have worked out is using the envvar DEVPATH and trying to regex compare against dmesg | tail 2022-09-21 13:57:14 brianvuyk: pipelines are created when you create a merge request 2022-09-21 13:58:15 Ikke - hmm... so I've had several times where that pipeline has failed (for example, the s390x build failed). Then I need to close my MR, fix the issue, squash the commits in my branch, and open a new MR 2022-09-21 13:58:39 It would be ideal to open the MR with the knowledge that it will pass ;) 2022-09-21 13:59:35 brianvuyk: you do not need to close it 2022-09-21 13:59:56 ikke: If I don't, I get failed for having multiple commits 2022-09-21 14:01:35 brianvuyk: you can amend the commit and force push 2022-09-21 14:01:57 Hmm, hadn't considered that. Thanks 2022-09-21 14:22:26 CI needs to be based on merge requests so that it knows what packages to build 2022-09-21 14:33:15 does anyone else have weird visual artifacts in the context menu on latest firefox in edge? 2022-09-21 14:35:07 hi! just made a fresh install of alpine on a nvme drive with the cryptsys option on the installer and Alpine shit itself (drop me in a recover shell) when booting 2022-09-21 14:35:35 it didn't asked me for a password, so I guess that's the problem 2022-09-21 14:51:42 eletrotupi: BIOS or UEFI booting system? 2022-09-21 14:52:37 UEFI 2022-09-21 14:53:50 so it installed fine, just when you rebooted afterwards it wouldn't boot? 2022-09-21 14:54:22 yeah 2022-09-21 14:55:00 it even go further to allow me to select "Alpine 3.16" but afterwards it fails to mount and drop me in a recover shell 2022-09-21 14:55:16 I think I saw someone talking about some issue like that in the past 2022-09-21 14:55:24 regarding cryptsetup and nvme drives 2022-09-21 14:55:30 but I can't find anything on my logs 2022-09-21 14:57:18 there were issues in the past where the nvme kernel module wasn't been loaded by initramfs, I *think* that was fixed but you could try at the grub menu editing the mneu option and adding "modules=nvme" (or editing any existing modules= line) 2022-09-21 14:57:49 though that assumes that nvme was enabled in mkinitfs' config when the initramfs was built 2022-09-21 14:57:51 right, I think thats the issue to be honest 2022-09-21 14:58:04 because when dropped on the recovery shell 2022-09-21 14:58:09 I can't seem to find my nvme drive 2022-09-21 14:58:11 on /dev 2022-09-21 14:58:22 but obviously on the live install it's there 2022-09-21 14:59:05 oh, it's there 2022-09-21 14:59:22 sd-mod,usb-storage,ext4,nvme cryptroot=... 2022-09-21 14:59:26 hm 2022-09-21 14:59:51 right, but is the module actually present in the initramfs? it can't load what isn't there 2022-09-21 15:00:17 how can I check that? 2022-09-21 15:00:42 after install but *before* rebooting you'd have to look at the mkinitramfs.conf file inside where it installed Alpine on the drive 2022-09-21 15:00:44 from the recovery shell or I need to go into the live install, mount everything, etc 2022-09-21 15:00:51 to see if the features list included nvme 2022-09-21 15:01:14 aight, so I'll do a new install 2022-09-21 15:01:54 if its not present in that file then you'd have to add it and then rebuild the initramfs before rebooting 2022-09-21 15:02:22 aight 2022-09-21 15:04:26 electrotupi: you're using 3.16.2? 2022-09-21 15:05:03 yes! 2022-09-21 15:05:52 nvme is certainly present in the features list of mkinitfs Git master, trying to see if that's the case for 3.16.2 2022-09-21 15:08:02 eletrotupi: nvme has been in features list since Oct 2020, so it should already be the case in 3.16.2 2022-09-21 15:08:35 so when you end up in recovery shell try a "lsmod | grep nvme" to verify it is loaded? 2022-09-21 15:09:15 yeah, it was there on /etc/mkinitfs/mkinitfs.conf 2022-09-21 15:09:28 anything I can poke before reboot (I just reinstalled) 2022-09-21 15:09:56 when running the command outlined here: https://wiki.alpinelinux.org/wiki/NVME 2022-09-21 15:10:34 it outputs: /sbin/mkinitfs: line 178: can't create /boot/initramfs-lts: noexistent directory 2022-09-21 15:10:57 and /boot don't exists, dunno if that should be correct tho 2022-09-21 15:11:50 you have to modify paths and your running that from the live CD/USBstick whereas you'd want to do so inside the installed Alpine 2022-09-21 15:12:00 anyway as nvme is in the features list you don't need to 2022-09-21 15:12:43 is "cryptsetup" also in the features list? *NOT* in /etc/mkinitfs/mkinitfs.conf but inside the created disk install (/mnt?) 2022-09-21 15:14:40 dumb question: but right after the installation, does it drops you inside a shell on the fresh install? 2022-09-21 15:14:46 because I don't see anything under /mnt 2022-09-21 15:15:23 and my shell is showing the hostname I've picked (I'm on the live install right after setup alpine has finished) 2022-09-21 15:16:00 not as far as I know, you're still in the "live" Alpine until you reboot 2022-09-21 15:16:23 all right, then I have to mount everything, right 2022-09-21 15:17:18 Is there any way I can reduce the number of 'linux-firmware' packages that are installed? 2022-09-21 15:17:23 yeah, plus as its encrypted you'd have to use cryptsetup as well 2022-09-21 15:17:31 adhawkins: install linux-firmware-none 2022-09-21 15:17:36 and then select the ones you need 2022-09-21 15:17:48 Will installing 'none' remove all the others? 2022-09-21 15:17:51 yes 2022-09-21 15:18:00 Excellent. It's a VM so I doubt I need any of them. 2022-09-21 15:18:56 Great, that seems to have done the trick. That'll reduce the 'noise' in my 'packages needing upgrades' emails! 2022-09-21 15:19:30 minimal: yeah, I can mount what appears to be /boot but can't mount / 2022-09-21 15:19:44 I can open the third partition via cryptsetup which is prolly /home? 2022-09-21 15:20:22 oh no, the third partition is actually / 2022-09-21 15:20:33 for UEFI and crypt I'd assume part1 is ESP, part2 is swap, and part3 is LUKS (with rootfs inside) 2022-09-21 15:20:34 what is that second 4GB partition then? 2022-09-21 15:20:48 hm, right 2022-09-21 15:20:58 fdisk -l /dev/nvme0n1 to check? 2022-09-21 15:22:08 yeah, it doesn't show if is swap or not 2022-09-21 15:22:26 but regardless, I've mounted everything in place 2022-09-21 15:22:42 (btw thanks for the help so far, minimal) 2022-09-21 15:23:43 by poking on /mnt/root/etc/apk/world I see cryptsetup there 2022-09-21 15:24:43 and nvme on /mnt/root/etc/mkinitfs/mkinitfs.conf too 2022-09-21 15:25:11 hm, nvme cryptsetup keymap 2022-09-21 15:25:29 right in that order, should keyboard be before cryptsetup or something? 2022-09-21 15:34:39 electrotupi: order in features list is unimportant, this is just used to decide which kernel modules to place in initramfs 2022-09-21 15:34:49 aight 2022-09-21 15:35:21 on 2022-07-21 ptrc was having the same issue 2022-09-21 15:35:30 the recommendation that worked was: [12:36:28] nm17[m]: okay, so.. could you try adding the grub manually with `efibootmgr --create --disk /dev/nvme0n1 --part 1 --label 'Grub' --loader /EFI/boot/bootx64.efi` 2022-09-21 15:35:46 oh sorry, wasn't ptrc (sorry for the highlight) 2022-09-21 15:36:20 does that makes sense, minimal? I think maybe the path for the loader will have to be different to match where it's mounted, but anyway 2022-09-21 15:36:37 electrotupi: that's setting up Grub on ESP, if that wasn't done right then you wouldn't have seen the Grub menu, which you did 2022-09-21 15:37:07 the problem rather is after the initramfs loads and its init runs 2022-09-21 15:37:52 ok, yeah 2022-09-21 15:37:54 that makes sense 2022-09-21 15:38:23 I'm not sure man, it seems to me that everything is in-place related to nvme 2022-09-21 15:38:38 try it again and assuming it fails, check again if /dev/nvme0n1 etc exist or not? 2022-09-21 15:38:45 ok 2022-09-21 15:39:05 and if that is missing boot again with "nvme" added to the modules list and check if then /dev/nvme0n1 etc exist 2022-09-21 15:40:06 yeah, it failed 2022-09-21 15:40:21 lsmod shows me that nvme and nvme_core is there 2022-09-21 15:40:36 there is nothing on /dev/nvme 2022-09-21 15:40:50 that is weird 2022-09-21 15:41:05 did you check dmesg output to see anything interesting? e.g. nvme mention 2022-09-21 15:42:27 well, dmesg says that there's an unknown kernel command and that will pass that user-space 2022-09-21 15:42:36 that's normal 2022-09-21 15:42:47 but is there any sign of it mentioning nvme? 2022-09-21 15:43:06 the modules arg is sd-mod,usb-storage,ext4,nvme 2022-09-21 15:43:11 nothing else mentioning nvme 2022-09-21 15:43:21 dmesg | grep nvme outputs those three items 2022-09-21 15:43:52 the command line, kernel command line (which is the same as before), then it complains that will pass to user-space 2022-09-21 15:44:15 so it is not detecting the nvme drive for some reason 2022-09-21 15:44:20 why make/model is it? 2022-09-21 15:44:26 s/why/what/ 2022-09-21 15:44:49 bah, can't remember the model, but I know it's a shitty ADATA one 2022-09-21 15:45:00 it was shipped by default on this dell lappy 2022-09-21 15:45:15 does recovery have "lspci"? 2022-09-21 15:45:27 no, it doesn't 2022-09-21 15:45:43 I'll reboot into the live install to check, anything else I can try from the recovery shell first? 2022-09-21 15:46:02 man such dejavu, because I can swear I've did the same thing when I received this laptop 2022-09-21 15:46:09 and stumbled upon the same problems 2022-09-21 15:46:22 and the outcome was not good: ended up with arch 2022-09-21 15:50:27 without lspci you could scan around inside /sys/devices or /sys/bus/ to try and see the full set of devices the kernel knows about but it wouldn't be slow/messy 2022-09-21 15:50:53 bingo! 2022-09-21 15:50:55 that nvme modules are loading but no device found is very weird 2022-09-21 15:50:56 it was the fucking bios 2022-09-21 15:51:06 it was set to Raid mode with the nvme storage 2022-09-21 15:51:12 Intel Quick Start or similar? 2022-09-21 15:51:14 changing to AHCI fixed 2022-09-21 15:51:37 ah, yeah, that's sort of what I was starting to suspect 2022-09-21 15:52:00 with the other mode then the "vmd" module might have helped 2022-09-21 15:52:28 right, hm 2022-09-21 15:52:29 except it is not in the typical initramfs AFAIK. I remember looking at this with someone about 1 month ago 2022-09-21 15:52:31 should I enable that? 2022-09-21 15:52:45 was able to fully login 2022-09-21 15:54:42 is you don't need the raid then best to stay with AHCI 2022-09-21 15:55:06 the previous person's issue was on 28th Aug 2022-09-21 15:55:35 I was intended to raise a MR to add vmd to the initramfs 2022-09-21 16:02:38 eletrotupi: adding "vmd" to the initramfs is straightforward, the question is how to detect the presence of such a device so that its module can be autoloaded 2022-09-21 16:11:19 mps: "alpine omnia" boots: https://tpaste.us/x5Nr 2022-09-21 16:11:23 the other does not 2022-09-21 16:28:20 Cool, I worked out how to use mdev script calling to auto add /dev/zigbee and /dev/zwave as static paths and remove them on unplug. A little crufty with the bash but w/e works fine enough 2022-09-21 16:32:43 eletrotupi: how that you have it booting does lspci show any sign of the vmd device? 2022-09-21 16:47:03 sorry, I was out for lunch 2022-09-21 16:47:20 lemme see 2022-09-21 16:47:55 hm, not it doesn't 2022-09-21 16:48:05 should it tho? 2022-09-21 16:48:52 I've also run lspci | grep vmd after modprobe vmd 2022-09-21 16:51:03 linux-edge clearly isn't working out for me - drivers/gpu/drm/nouveau/nvkm/engine/device/base.c:2630:1: error: redefinition of 'nv173_chipset' 2022-09-21 16:55:35 eletrotupi: I guess the presented pci/pcie "name" (and possibly ids) change depending on the UEFI/BIOS setting. what ids does the NVME device now have? just wondering how to cleanly detect vmd device for autoloading driver 2022-09-21 16:59:45 right! 2022-09-21 16:59:48 "IM2P33F3A NVMe ADATA 512GB" 2022-09-21 17:03:05 this is what lspci output: http://paste.debian.net/plain/1254583 2022-09-21 17:05:32 electrotupi: sorry, its "lspci -nn" output that's useful as that shows the ids at the end of each line 2022-09-21 17:19:05 pinsl: ah yes, I didn't named boot options correctly, but I think it is not important now 2022-09-21 17:20:23 pinsl: anyway you now have bootable media on which you can install 'fixed' linux-edge and test further 2022-09-21 17:22:18 pinsl: does option 2 (linux-edge) shows anything after 'Starting kernel ...' 2022-09-21 17:22:29 mps: no 2022-09-21 17:23:33 filoozom: do you use serial console or monitor and keyboard as primary console 2022-09-21 17:24:29 pinsl: hmm, so have to try to enabme more base drivers and build again 2022-09-21 17:25:21 mps: serial console, the NanoPi R2S doesn't have video input and it's more practical anyways 2022-09-21 17:27:18 filoozom: aha, drivers/gpu/drm/ misslead me to think that it have video out 2022-09-21 17:28:05 filoozom: do you have boot log to paste 2022-09-21 17:28:51 mps: nothing more than previously (https://dpaste.org/oW1YX) as I haven't been able to compile `linux-edge` yet because of that compile error 2022-09-21 17:29:16 Now I'm wondering if I can cross-build from x86 instead of going the binfmt / docker aarch64 route to make it faster 2022-09-21 17:29:26 Because I'm not a fan of waiting 2 hours every time 2022-09-21 17:30:55 filoozom: few years ago I used debian a lot for cross building kernels for different boards, it worked well. but in last 3-4 years I build in 'native' env 2022-09-21 17:33:28 mps: yeah I'm sure that's best. Would've been nice to speed things up a bit though, seing as I don't have a powerful aarch64 and qemu-arm is super slow 2022-09-21 17:34:40 filoozom: cross build is not so slow, but right qemu is slow 2022-09-21 17:36:00 mps: yeah, hence why I'm trying to get cross-compiling to work from alpine x86 to alpine aarch64, but there's not a lot of documentation out there. Read about scripts/bootstraph.sh but not sure that's right 2022-09-21 17:36:30 How can I check the version of a package? Using apk info returns two versions (I assume from two different repos..) 2022-09-21 17:36:37 filoozom: cross build on alpine is not so easy 2022-09-21 17:37:42 filoozom: if you want just to build kernels I think using some other distro is better idea, debian or arch linux 2022-09-21 17:38:14 filoozom: but ofc any distro with which you have experience and have cross tools is ok 2022-09-21 17:39:03 filoozom: I used debian on alpine in lxc container for such tasks 2022-09-21 17:40:44 mps: got it, thanks! On what hardware do you build aarch64 kernels usually? 2022-09-21 17:41:47 filoozom: nowadays on apple silicon M1pro but in previous years I did on arm64 chromebooks 2022-09-21 17:42:32 imagine build complete alpine kernel packkage in lest than 2 minutes :) 2022-09-21 17:42:44 less* 2022-09-21 17:43:32 gentoo is good for cross (eg chromeos uses the gentoo toolchain) 2022-09-21 17:44:05 invoked: I guess yes, though never tried on gentoo 2022-09-21 17:44:26 Never tried gentoo either 2022-09-21 17:44:36 but any distro with needed cross tools is ok 2022-09-21 17:44:58 And if I start with another distro then I have to figure out how to use my built kernel with scripts/mkimage.sh which I don't really want to do haha 2022-09-21 17:45:16 gentoo's defacto use case is the user building everything from source, so it's very good at that 2022-09-21 17:45:23 Have been thinking of buying a M1, do you need to run linux on it to compile a kernel with it? 2022-09-21 17:45:31 I read docker runs in a VM and is thus slower 2022-09-21 17:46:45 filoozom: you can use qemu on macos but that is 'not good' 2022-09-21 17:47:13 filoozom: I run alpine linux on m1 from the previous october/november native 2022-09-21 17:47:43 mps: ah that's awesome, didn't know it could run on mac :D 2022-09-21 17:47:52 actually alpine was the first distro used as daily driver on m1 :) 2022-09-21 17:48:24 I love alpine so much, bit sad that there's not a "huge" community around it 2022-09-21 17:48:31 Run it on all my servers 2022-09-21 17:48:37 ACTION was first who used m1 with asahi kernel for daily workstation 2022-09-21 17:49:17 the way i look at it is, alpine doesn't have all the noise that other distros have. better signal to noise ratio 2022-09-21 17:49:31 filoozom: even wrote short guide https://arvanta.net/alpine/install-alpine-m1/ 2022-09-21 17:49:47 Yeah for sure, but documentation is clearly lacking in a bunch of subjects 2022-09-21 17:50:14 mps: I'll be sure to bookmark that! 2022-09-21 17:50:44 there are users who use M1 with alpine as production servers 2022-09-21 17:51:07 and workstations ofc 2022-09-21 17:51:10 Apparently I ventured into cross-compiling before, but can't really remember anything haha https://gitlab.alpinelinux.org/alpine/aports/-/merge_requests/1774 2022-09-21 17:51:56 Don't tempt me, I might end up a single-use M1 just to get alpine running on my nanopi haha 2022-09-21 17:52:10 end up buying* 2022-09-21 17:52:28 :) 2022-09-21 17:53:22 I'm still running different arm32 SBCs for different things, servers, routers, dedicated tasks ... 2022-09-21 17:53:36 and arm64 ofc 2022-09-21 17:54:23 Same. Recently moved some services from my epyc server to a RPi so I could save some power and not have the server run constantly 2022-09-21 17:54:29 Thanks energy crysis 2022-09-21 17:55:54 my reasons to stop using intel was well, ugly cpus 2022-09-21 17:56:09 ugly? 2022-09-21 17:56:17 last intel machine I bought in 2010 2022-09-21 17:56:33 ugly design 2022-09-21 17:57:14 Yeah I do like arm quite a bit better, but there's just a lack of hardware that makes it a bit sad 2022-09-21 17:57:18 mps: did you try any riscv sbcs? 2022-09-21 17:57:39 filoozom: ugly architecture design I mean to say 2022-09-21 17:57:43 Either you get (well, there's no recent alternative I believe?) an Ampere eMag or ThunderX (which normal people can't find and costs a shit ton) 2022-09-21 17:57:50 Or you get an SBC which can't do anything 2022-09-21 17:58:11 pinsl: yes, I've got one riscv SBC to make alpine working on it 2022-09-21 17:58:37 pinsl: here is guide https://arvanta.net/alpine/alpine-on-visionfivev1/ 2022-09-21 17:59:46 actually I'm not only alpine dev who got this board, we even have channel #alpine-riscv64 for developing alpine on riscv 2022-09-21 17:59:52 mps: ah interesting 2022-09-21 18:01:00 pinsl: there is page on alpine gitlab somewhere also 2022-09-21 18:13:36 hey everyone! 2022-09-21 18:14:03 just wanted to ask whether any of you have experienced the same issue as me: 2022-09-21 18:14:43 when I work with diskless mode, my boot media (USB stick) gets randomly mounted either as /media/usb or as /media/sdb1 2022-09-21 18:15:30 LBU was set up to use /media/sdb1, so when it [randomly] gets mounted under /media/usb, my customizations won't be loaded... 2022-09-21 18:16:26 usbmounter: there were changes made to mdev over the last (year?) or so that affected /media/usb 2022-09-21 18:17:58 hmm, what is the solution then? I noticed this behavior on various architectures if that matters (x86_64, armv7 and armv8 too) 2022-09-21 18:18:15 which Alpine version are you using? 2022-09-21 18:18:23 3.16.2 2022-09-21 18:19:03 I ran the setup-alpine script and chose (the default offered) sdb1 as LBU storage 2022-09-21 18:19:37 now my boot USB stick is randomly mounted at either /media/usb or /media/sdb1 at every boot 2022-09-21 18:19:45 and the former scenario breaks everything of course 2022-09-21 18:20:36 but I also see that fstab has the entry to always mount /dev/usbdisk to /media/usb 2022-09-21 18:20:49 and /dev/usbdisk is a symlink to sdb1 2022-09-21 18:20:57 in 3.16.2 /dev/usbdisk should be a softlink to the actual usb device name 2022-09-21 18:20:59 so that seems to be logical 2022-09-21 18:21:08 yepp, but it doesn't work :( 2022-09-21 18:21:26 I'm in front of a sample setup right now 2022-09-21 18:21:47 maybe I start from scratch and choose /media/usb as LBU storage instead of sdb1 2022-09-21 18:22:00 I just chose sdb1 because it offered it by default 2022-09-21 18:22:05 chose /dev/usbdisk? 2022-09-21 18:22:24 yepp will try that give me 5 min 2022-09-21 18:22:41 Can't you just uncomment the extra fstab entry? 2022-09-21 18:23:21 that breaks it because it is supposed to be on sdb1 (which is in the lbu.conf as well) 2022-09-21 18:23:34 Setting up cross-compiling for the kernel on Ubuntu was actually way more painless than I anticipated 2022-09-21 18:23:38 let me start clean and go the right path immediately at start when choosing LBU storage 2022-09-21 18:23:40 the problem with relying on a specific sdX device, especially with USB, is that naming can be performed based on the order of devices, so if you plug in 2 usb devices then which one is sda depends on things like which usb socket you plug them into 2022-09-21 18:23:41 The thing is flying now 2022-09-21 18:23:52 Hopefully the result will be the same if I use the same .config? 2022-09-21 18:25:12 filoozom: could you please give me the link to the tutorial you followed to make a custom kernel build? wanted to do that for a long time just didn't have the power to do so! If you say it was straightforward I'm super interested 2022-09-21 18:25:13 thank you! 2022-09-21 18:27:55 Aaaaand it's done, why did I bother with qemu haha 2022-09-21 18:28:17 usbmounter: are we talking with cross-compile or without? 2022-09-21 18:28:29 with would be nice 2022-09-21 18:28:41 actually I only care about x86_64 armv7l and armv8 2022-09-21 18:29:37 update: started from scratch, chose "usb" instead of "sdb1" to store configs (LBU) 2022-09-21 18:29:52 installed my stuff and did an lbu ci 2022-09-21 18:30:10 usbmounter: honestly I mostly just looked at this: https://github.com/alpinelinux/abuild/blob/master/abuild.in 2022-09-21 18:30:22 rebooted, now I see something I never experienced before: the USB stick is mounted at BOTH /media/usb and /media/sdb1 2022-09-21 18:30:41 You can either recreate than manually, or git clone aports and then change your config file and run abuild checksum && abuild -r 2022-09-21 18:30:44 which *should* be fine. but my preinstalled APKs are not available :( 2022-09-21 18:32:26 no deps to be installed? or the Makefile takes care of it? 2022-09-21 18:32:46 I wanted to dockerize the whole thing to make sure no unneeded deps will be lurking around on my host after the build 2022-09-21 18:32:55 but that seems really easy what you wrote 2022-09-21 18:33:31 Sorry, wrong link 2022-09-21 18:33:37 I meant this: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/community/linux-edge/APKBUILD 2022-09-21 18:37:52 wow, looks great! I guess I just need to edit the arch related config and run abuild checksum && abuild -r as you said? 2022-09-21 18:45:20 Hi, does the default linux-lts kernel support sound? I looked inside /proc/asound/oss/sndstat and it says "Audio devices: NOT ENABLED IN CONFIG" and same for Synth and Midi devices. I am trying to set up ALSA, but no sound cards are detected (/proc/asound/cards is empty). 2022-09-21 18:46:22 glt: do you see any relevant output in dmesg? 2022-09-21 18:47:56 minimal: https://dpaste.org/vPALZ/raw 2022-09-21 18:47:56 I'm not sure what I should search for. No results searching for the strings "alsa", "card", or "sound" 2022-09-21 18:49:58 and you don't see any cards when running aplay -l ? 2022-09-21 18:50:21 My understanding is that ALSA uses the sound cards it finds in /proc/asound/cards, and that information is provided by the kernel. So I am shifting my focus to the kernel for that reason. 2022-09-21 18:50:32 aplay -L gives only the null sink and nothing else 2022-09-21 18:51:52 interesting... had my troubles with ALSA but it always could list the cards at least... 2022-09-21 18:53:27 Now why do I end up with a 25MB kernel instead of the 7MB in linux-edge 2022-09-21 18:53:48 sounds like it is uncompressed vmlinux? 2022-09-21 18:54:08 or you baked windows10 drivers in :DDDD 2022-09-21 18:54:29 My laptop is pretty new, could it be that linux-lts doesn't have new enough modules for it? 2022-09-21 18:55:23 it should *at least* recognize the NVIDIA GPU (MX450 as I've seen) at least IMO 2022-09-21 18:56:59 filoozom: make zImage 2022-09-21 18:57:06 it might see the device but it doesn't mean there are drivers for it 2022-09-21 18:57:22 Lol my initramfs is 89M compressed 2022-09-21 18:57:36 I know that, but I don't think the MX450 is THAT new 2022-09-21 18:58:18 just depends on what nouveau supports 2022-09-21 18:58:26 enabling drivers in alpine kernels is usually done when someone request 2022-09-21 18:59:05 or you compile your custom kernel, filoozom just gave the link a couple messages before 2022-09-21 19:00:25 mps: but there's no make zImage here: https://gitlab.alpinelinux.org/alpine/aports/-/blob/master/community/linux-edge/APKBUILD 2022-09-21 19:02:38 Does anyone else get "Audio Devices: NOT ENABLED IN CONFIG" if you cat /proc/asound/oss/sndstat ? 2022-09-21 19:02:54 Idk if it's just me 2022-09-21 19:03:20 yeah i get it too 2022-09-21 19:03:29 But your audio is working? 2022-09-21 19:03:46 If you use alpine as a desktop 2022-09-21 19:04:10 this is kinda new system, but yeah it works through HDMI, have to edit asoundrc to get pcm to work 2022-09-21 19:04:27 filoozom: this line arm*|aarch64) _install="zinstall dtbs_install"; 2022-09-21 19:04:41 Do you remember /proc/asound/cards ever being empty? 2022-09-21 19:05:27 Because I'm pretty sure that's what I need to fix, the alsa-openrc service just loops through all the cards and does something to them but problem is there are no cards I guess 2022-09-21 19:05:40 filoozom: and later in _package() 'make -j1 modules_install $_install' 2022-09-21 19:06:55 glt: nah 2022-09-21 19:07:38 filoozom: https://tpaste.us/YrR4 is my make kernel shell script which builds kernel from upstream tarball out of any distro build 'system' 2022-09-21 19:07:44 I'll just try linux-edge kernel I guess 2022-09-21 19:08:13 Thanks all for answering my questions 2022-09-21 19:08:38 filoozom: you only need to add CROSS_BUILD options there and you will got ready to install kernel tarball 2022-09-21 19:23:23 glt: did you load the relevant kernel modules for your sound device? 2022-09-21 19:34:28 mps: ah, I did run that modules_install but must have fucked up somewhere else. I now have the right size! 2022-09-21 19:34:34 Added some scripts here: https://github.com/filoozom/kernel-cross-build 2022-09-21 19:34:40 That kind of emulate how abuild works 2022-09-21 19:35:24 Now for some reason I end up with `vmlinuz-5.19.10-0-edge` instead of `vmlinuz-edge`, but that's not too bad 2022-09-21 19:36:27 Now if I can package this in an apk I could use it in my other build scripts, or I'll have to do the initramfs stuff manually 2022-09-21 19:36:55 filoozom: if you followed my script then you get version in kernel name and tarball, this is intentional when testing kernels to have every one separate 2022-09-21 20:01:32 OK so about my previous soundcard issue, I fixed it by changing my initramfs. I have no idea why this affected the ability to detect the soundcards :) 2022-09-21 20:02:29 When I installed Alpine I was having trouble with the initramfs loading the necessary modules so I just build an initramfs that loaded EVERY module. I just gutted all the sound modules from it and now my sound cards get detected. 2022-09-21 20:09:21 if you want to share your experience you could compare the is it modlist who lists all loaded modules before and after 2022-09-21 20:09:50 probably some "incorrect" module hooked on your soundcard so the correct module couldnt since your soundcard was already loaded (even if it didnt worked) 2022-09-21 20:10:04 like back in the days with soundblaster compatability vs gravis soundcards etc 2022-09-21 20:13:41 mps: do you happen to have a script that builds the initramfs with your custom kernel as well? 2022-09-21 20:13:59 Gonna test this config: https://github.com/filoozom/kernel-cross-build/blob/main/configs/nanopi-r2s 2022-09-21 20:14:17 Do the modules loaded in the initramfs carry over to the actual kernel? 2022-09-21 20:14:55 there is only one kernel 2022-09-21 20:15:05 so yes 2022-09-21 20:15:18 Oh 2022-09-21 20:15:32 I guess I don't understand what initramfs is 2022-09-21 20:15:45 filoozom: no, I don't use initram with testing kernels, build all essential drivers in kernel and when finish with tests then I merge in linux-edge 2022-09-21 20:16:17 glt: it's basically a small filesystem that gets loaded first to setup the kernel to boot the actual system 2022-09-21 20:17:02 I found concept of initrd as not so good idea 2022-09-21 20:17:32 Less than 4 minutes to build the kernel :D 2022-09-21 20:18:08 'kay, but then how do I boot a diskless install without initramfs 2022-09-21 20:18:20 I don't 2022-09-21 20:18:21 mps: any particular reason why not? I suppose it's required if you want to have a kernel that is generally usable without building every module in 2022-09-21 20:19:01 So how is the initrd different from something like the bootloader? Because syslinux/grub "modules" aren't for linux or whatever other system you boot into. And I've seen the grub "grubx64.efi" referred to as an image or a kernel. 2022-09-21 20:19:29 ikke: it is distros invention, to have a lot of modules and load them from initrd, but this leads to more problems usually than 'solution' 2022-09-21 20:19:54 I prefer tailored kernels for particular SBC 2022-09-21 20:20:01 Sure, but you build your own kernels 2022-09-21 20:20:05 not everyone can do that 2022-09-21 20:20:13 ikke: right 2022-09-21 20:20:28 but I think distros should do same 2022-09-21 20:20:44 how do you avoid having to build and maintain 100 kernels? 2022-09-21 20:20:46 or more 2022-09-21 20:20:57 I think there is some option with the linux kernel "make allconfig" then you certainly don't get issues with missing modules :) 2022-09-21 20:20:57 100 maintainers 2022-09-21 20:21:05 mps: that's unatainable 2022-09-21 20:21:43 ikke: well, if the users wants specific kernel they should be maintainers 2022-09-21 20:21:51 I think "make allconfig;make;make modules" took my old computer like 4 hours to build 2022-09-21 20:22:07 mps: users don't want specific kernels, they want generic kernels :) 2022-09-21 20:22:12 glt: heh, I think you are right 2022-09-21 20:22:33 ikke: who cares what users wants ;p 2022-09-21 20:22:56 I'm a user :P 2022-09-21 20:23:28 Well, I just want a kernel that works on RK3328, and `linux-lts` ain't it... ;p 2022-09-21 20:23:47 ikke: you know that I'm trying long time to make good kernel for different arm SBCs but I get very little feedback from users 2022-09-21 20:24:16 mps: in general, you don't hear from users when everything is working alright 2022-09-21 20:24:34 people only show up when things are broken 2022-09-21 20:24:37 that's why i make issues 2022-09-21 20:24:37 so I stoped to care for general one and only work with users who are ready to help and give feedback 2022-09-21 20:24:38 like filoozom :P 2022-09-21 20:24:44 if no one tells me it's broken i assume no one uses it 2022-09-21 20:24:45 so i break it 2022-09-21 20:24:47 :p 2022-09-21 20:24:58 psykose: universe brain 2022-09-21 20:25:07 or galaxy 2022-09-21 20:25:12 Haha 2022-09-21 20:26:05 ikke: you know well I think that I try to help all users who are willing to report problems and test 2022-09-21 20:26:46 mps: yes, as you've shown here 2022-09-21 20:27:12 but it is just waste of time and effort to fix problem not described correctly 2022-09-21 20:27:28 and no feedback 2022-09-21 20:27:42 you could add CONFIG_SND_SOC_AMD_RENOIR_MACH=m for sound on some amd ryzen laptops 2022-09-21 20:27:46 one a meter away from me 2022-09-21 20:27:50 For 99% of the users, linux-lts + initramfs works 2022-09-21 20:28:02 psykose: you are telling to me 2022-09-21 20:28:07 yea 2022-09-21 20:28:22 Hey, is there an Alpine Linxu developer handbook being made? That part of the website is empty last I checked 2022-09-21 20:28:37 did you told me this earlier, I can't remember, sorry if you did 2022-09-21 20:28:38 glt: No one is actively working on it 2022-09-21 20:28:48 realistically the best generic distro kernel is almostallmodconfig with all the modules that don't really increase size anyway, and some module subpackaging 2022-09-21 20:28:49 Is there a way to create an apk from a folder without going through the full abuild process? 2022-09-21 20:28:51 not sure why we don't do that 2022-09-21 20:29:00 filoozom: nfpm can do that 2022-09-21 20:29:00 mps: haven't 2022-09-21 20:29:30 filoozom: you can also create an apkbuild that just packages whatever is in a specific fir 2022-09-21 20:29:42 ikke: oh ok 2022-09-21 20:29:57 psykose: ok, granted, I can be stubborn sometimes but at the end I usually do what users request 2022-09-21 20:30:20 nfpm has a yaml file that describes the metadata and maps files 2022-09-21 20:31:58 actually, now that linux-edge is used more and more I would appreciate help from other devs (hello psykose) and users 2022-09-21 20:31:58 ikke: Yeah I saw, wonder if it would be easier to insert my cross-compiler kernel somewhere in `linux-edge` to make sure that apkbuild does the right thing 2022-09-21 20:32:12 Or if I should go down that route, learn yet another tool and risk further issues 2022-09-21 20:33:20 don't think there's much to do aside from enable things people want to use 2022-09-21 20:33:50 I have never worked on kernels 2022-09-21 20:34:16 yes, I think I finally made it very nice, separate and cleaner -dev, separate -doc 2022-09-21 20:34:45 and no more linux-edge4virt 2022-09-21 20:34:46 the dev thing doesn't make much sense as the size doesn't matter as it's just pulled for akms sometimes 2022-09-21 20:34:53 and instead it's just broken for some configs as you've seen 2022-09-21 20:35:06 then again, akms is also awful, so hard to pick there 2022-09-21 20:35:17 well, I don't see bug reports 2022-09-21 20:35:40 psykose: what's wrong with akms? 2022-09-21 20:35:51 nobody tested it with anything 2022-09-21 20:35:59 use any non-alpine localversion? broken 2022-09-21 20:36:03 _rcX? broken 2022-09-21 20:36:54 not in aports? https://github.com/jirutka/akms/issues/10 broken again 2022-09-21 20:37:04 it was made and tested with linux-lts-dev and forgotten about i guess 2022-09-21 20:37:45 then again, i don't even remember the issues dkms would have instead as it's been so long, haha 2022-09-21 20:38:45 Before we had nothing 2022-09-21 20:38:51 so it is nice that jirutka made akms 2022-09-21 20:38:56 even though it's not complete 2022-09-21 20:39:25 Should there be a xtables-addons-edge? 2022-09-21 20:40:20 the addons are per-kernel out of tree modules, it's a lot of maintenance to keep them per extra kernel aside from lts 2022-09-21 20:40:28 but there should probably be a -src of it for akms 2022-09-21 20:40:32 that would be ok 2022-09-21 20:40:37 not sure why one isn't provided for all of them 2022-09-21 20:40:53 Running an alpine docker container, I'm having trouble with apk, if I `apk update` it just hangs. If I `wget https://dl-cdn.alpinelinux.org/alpine/v3.16/main/x86_64/APKINDEX.tar.gz` from the host it works fine but running the same from the container just hangs. I'm all out of ideas to try. 2022-09-21 20:40:53 because no one made the effort to build packages for it? 2022-09-21 20:40:56 someone made a zfs-src for instance, there should be one from every modules package as a subpackage 2022-09-21 20:41:01 filoozom: linux-edge is not official kernel 2022-09-21 20:41:16 psykose: do you volunteer? :P 2022-09-21 20:41:41 i would but i can't even install -src packages to test them because i use my own kernel 2022-09-21 20:41:45 it is as name suggest 'edge' (of knife) 2022-09-21 20:42:15 I made one for rtl8821ce, as I maintain that package 2022-09-21 20:42:30 yeah, they're not too hard 2022-09-21 20:43:19 Honestly I'm not even sure what I'd need those addons for, I just encountered that issue while trying to setup my custom kernel 2022-09-21 20:43:33 I guess I'll just remove that for now 2022-09-21 20:54:43 hm, I followed the pipewire wiki page, but still haven't had much luck so far, when running wpctl status, it only displays a "Dummy Output" 2022-09-21 20:55:26 I've also already installed linux-firmware 2022-09-21 20:55:44 because I initially thought that was the issue, but seems that there's something else 2022-09-21 20:57:43 What's the pkgdir variable set to in APKBUILD? 2022-09-21 20:58:23 $STARTDIR/pkg/$pkgname 2022-09-21 21:08:36 `nfpm` stuff done, now I guess I have to create an APKINDEX somehow 2022-09-21 21:08:50 apk index 2022-09-21 21:09:12 eletrotupi: assuming that's after rebooting, maybe your sound card is missing the kernel module being enabled entirely 2022-09-21 21:15:04 psykose: according to inxi, I need the sof-audio-pci-intel-tgl driver 2022-09-21 21:15:26 there's the sof-firmware package for those i think 2022-09-21 21:15:29 separate 2022-09-21 21:15:36 as well as sof-tools etc 2022-09-21 21:23:45 yep, that was it 2022-09-21 21:23:47 thanks psykose 2022-09-21 21:23:53 :) 2022-09-21 21:44:30 Mr501: in your container is there a /etc/apk/repositories file? 2022-09-21 21:48:04 minimal: yeah. I'm using docker image `node:16-alpine`. I realised it's not an alpine issue... I can only reproduce the issue on one host machine, seems to work find on other hosts. Weird AF because it's only that one url that doesn't work. Like the docker container can ping dl-cdn.alpinelinux.org but just can't download that file. 2022-09-21 21:57:17 Mr501: does it retrieve any other files? you didn't indicate that 2022-09-21 21:58:35 minimal: well when I `apk update` it tries that file first and just hangs. it never gets to try the other files. 2022-09-21 21:59:30 I can wget that url from the host, no problem. but I can't wget that url from the docker container when it's on that host. When the docker container is on another host everything works as expected. 2022-09-21 22:03:32 Mr5o1: sounds like the resolv.conf inside the container isn't right? or something else DNS related 2022-09-21 22:07:49 How does `apk index` sign the index? 2022-09-21 22:11:38 Or maybe it's just my nfpm package not being signed 2022-09-21 22:13:44 hmm I can ping that server though so I don't think it's DNS 2022-09-21 22:14:42 Mr5o1: is either end dual-stack? 2022-09-21 22:15:05 IDK what that means LOL so I guess not ? 2022-09-21 22:15:18 i.e. IPv4 + IPv6 2022-09-21 22:15:53 I've seen apk fail in weird ways if, for example, both ends have an IPv6 address, but the actual connection is unusable 2022-09-21 22:21:01 Never mind, found the apkindex signing thing (https://github.com/alpinelinux/abuild/blob/master/abuild.in#L1838) 2022-09-21 22:26:42 I FIGURED IT OUT. A rogue wireguard config I'd forgotten about. 2022-09-21 22:26:57 cool 2022-09-21 22:50:58 Well, got my kernel to run after a lot of work 2022-09-21 22:51:04 But the SD card still doesn't work 2022-09-21 22:51:07 Sad sad life 2022-09-21 22:51:09 Off to bed now 2022-09-21 22:53:37 Now `modprobe dw_mmc-rockchip` doesn't even show the module in `lsmod` anymore 2022-09-21 23:00:44 https://dpaste.org/iEfyy 2022-09-22 11:39:09 https://89d1-185-213-154-164.eu.ngrok.io 2022-09-22 14:06:36 Hello. 2022-09-22 14:09:32 I'm trying to decrypt the root partiotion on a raspberry pi remotly. I found this git repo: . Is it possible to get something like this in an apk? I didn't find anything. 2022-09-22 15:24:38 rommi: I submitted a MR 1 year ago to add a remote decryption option to Alpine but it was not merged: https://gitlab.alpinelinux.org/alpine/mkinitfs/-/merge_requests/86 2022-09-22 15:24:56 soon ™️ 2022-09-22 15:42:56 minimal: Pretty nice, thank you. 2022-09-22 15:48:20 rommi: I haven't kept the MR up-to-date since I submitted it. I still would like to get it merged but mkinitfs still has no testing framework in place 2022-09-22 15:48:52 yes, i'm reading the comments right now. 2022-09-22 15:53:37 So lazy me could git clone your repo, maybe git pull master from mkinitfs, and apkbuild everything? 2022-09-22 15:54:41 rommi: you'd like have to adapt my diff as mkinitfs will have changed in the meantime 2022-09-22 15:56:01 rommi: there's also another change for dropbear to create a new subpackage. Have a look at https://github.com/dermotbradley/alpine-remote-LUKS-unlock/ for the full set of changes 2022-09-22 15:56:37 Yes, but having an apk would be nice. I'm playing around with the pi, so I'd have to edit the files some times probably. 2022-09-22 15:58:04 rommi: yes that's the fundamental issue/problem with the changes not being merged, if you set it and then later update/upgrade your Alpine packages it *will* stop working as the updated Alpine mkinitfs/dropbear packages won't contain the functionality 2022-09-22 15:58:47 rommi: that's why I have not kept the changes up-to-date, as there was no likelyhood of it being merged (until after a testing framework was in place) 2022-09-22 16:08:22 minimal: To sumarize what I understood: Until there is a framework, there is no merge. No merge means manual intervention after updating mkinitfs or dropbear (which I will forget sometime). Thanks for the explaination. Maybe I'm better of plugin the keyboard in on reboot. 2022-09-22 16:08:37 *test framework 2022-09-22 16:10:29 Hey guys, I have a few new clues by booting armbian on my nanopi: boot (https://dpaste.org/iPA17), udevadmin (https://dpaste.org/Dq2Xj) and dmesg (https://dpaste.org/EBZ1j) 2022-09-22 16:10:59 DRIVERS=="dwmmc_rockchip" I have in my kernel so that can't be it 2022-09-22 16:11:10 Could DRIVERS=="mmcblk" be missing? 2022-09-22 16:13:45 I have a feeling that that has to be `CONFIG_MMC_BLOCK` which I have on (https://github.com/filoozom/kernel-cross-build/blob/main/configs/nanopi-r2s#L6018) 2022-09-22 16:22:33 filoozom: so what does dmesg show when you try the kernel you built? which of the MMC-related kernel messages that occur with the Arbian boot do you/do you NOT see with Alpine? 2022-09-22 16:29:34 mps: well, actually there's no MMC logs in my custom kernel 2022-09-22 16:29:52 Even though the modules are clearly builtin 2022-09-22 16:31:47 This is the last output from yesterday night: https://dpaste.org/iEfyy 2022-09-22 16:36:21 minimal: whoops I thought you were mps for some reason haha 2022-09-22 16:36:45 minimal: I'm his stunt double ;-) 2022-09-22 16:38:08 I feel like dwmmc_rockchip doesn't start in my kernel even though it's builtin? Is that even possible? 2022-09-22 16:41:38 kernel/drivers/mmc/host/dw_mmc-rockchip.ko seems to be missing from my kernel but is in armbian's 5.15.63 2022-09-22 16:45:16 filoozom: you have CONFIG_MMC_DW_ROCKCHIP=y so its builtin 2022-09-22 16:46:39 minimal: yes, but it doesn't appear in `cat /lib/modules/5.19.10-nanopi/modules.builtin | grep rockchip`?! 2022-09-22 16:46:51 I feel like I'm going cray-cray 2022-09-22 16:51:00 filoozom: are you using the same DTB for both kernels? 2022-09-22 16:51:16 Good question 2022-09-22 16:51:32 that could be the problem - compare them 2022-09-22 16:53:14 They are different, but to be fair it's a 5.15 vs 5.19 2022-09-22 16:55:18 (Well, the hash is different, I'm not too sure how to "decompile" them 2022-09-22 16:56:08 filoozom: there's a "dtc" command to convert dtb files into dts files 2022-09-22 16:56:37 on Alpine it's in the "dtc" package strangely enough ;-) 2022-09-22 16:57:27 dtc -I dtb -O dts -o 2022-09-22 17:02:54 Hello all 2022-09-22 17:04:18 minimal: http://ix.io/4b94 2022-09-22 17:04:26 I'm in for another long night aren't I 2022-09-22 17:04:28 I would like to make qemu VM to make mini labs on GNS3, what do you guys suggest me to start with? Standard, extended (as I want to setup tcpserver on one of the VM for instance), or Virtual maybe? 2022-09-22 17:14:17 virtual 2022-09-22 17:15:05 @panekj If you answer me, thanks. I don't really know Alpine enough to decide alone 2022-09-22 17:17:09 Disregard the kernel/drivers/mmc/host/dw_mmc-rockchip.ko thing, it is indeed in `modules.builtin`... 2022-09-22 17:26:44 From armbian: [ 3.335961] vcc_sd: supplied by vcc_io [ 3.343392] rk3328-dmc ff780000.dmc: current ATF version 0x101 2022-09-22 17:26:47 Might be missing that 2022-09-22 17:29:54 filoozom: that's a DEVFREQ driver, don't think it affects MMC 2022-09-22 17:30:47 minimal: it's just that the SD card shows up right after that in armbian: https://dpaste.org/EBZ1j#L32,316 2022-09-22 17:31:02 In my head it kinda goes "apply voltage to the SD card slot so you can read it" 2022-09-22 17:31:16 But I might be inventing stuff now haha 2022-09-22 17:31:16 filozom: doesn't necessarily mean it is related 2022-09-22 17:32:35 true 2022-09-22 17:34:00 filoozom: from the DTB diffs such as "vmmc-supply", "dwmmc", "vcc_io_sdio", "vcc_sd", and a lot of "sdmmc" stuff 2022-09-22 17:34:26 so I'd suggest that perhaps the DTB file you're using with Alpine may not be correct 2022-09-22 17:34:37 minimal: I tried to boot with armbian's dtb but it didn't change anything 2022-09-22 17:34:54 the DTB is used to "announce" devices to the kernel, it doesn't matter what drivers the kernel has if the DTB is wrong/missing stuff 2022-09-22 17:34:56 I'm using the standard one used in the kernel 2022-09-22 17:35:18 And the card does work in u-boot 2022-09-22 17:37:00 Is vcc_sd a module I can enable somewhere? Might as well try 2022-09-22 17:40:49 filoozom: REGULATOR_RK808 perhaps? which you don't have set because you also don't have CONFIG_MFD_RK808 set 2022-09-22 17:41:19 though that might not be used for your particular Rockchip chip 2022-09-22 17:48:24 filoozom: what is your u-boot config file 2022-09-22 17:50:13 filoozom: I also notice you don't have CONFIG_I2C_RK3X set and you have CONFIG_PL330_DMA as a module and in the Armbian kernel dwmmc_rockchip's kernel messages say "Using internal DMA controller" 2022-09-22 17:52:31 filoozom: also, can you tell me your device model 2022-09-22 17:52:45 to find kernel config on the net for it 2022-09-22 17:53:15 mps: nanopi r2s 2022-09-22 17:53:35 filoozom: thanks 2022-09-22 17:53:45 My u-boot config is the generic one found here: https://u-boot.readthedocs.io/en/latest/board/rockchip/rockchip.html 2022-09-22 17:54:36 minimal: about the dtb: true, but the armbian dtb works for a fact and is the one I'm using now with no success 2022-09-22 17:55:36 filoozom: as different kernel version change things then DTB files may need to be revised accordingly 2022-09-22 17:56:02 minimal: regarding the DMA modules... I hoped dependencies would be enabled automatically? :sweat_smile: 2022-09-22 17:57:17 filoozom: only if it is defined as a dep in the Kconfig file. I've seen at least one general case where that is not the case (Realtek network driver which does not have a dep on the PHY module but needs it) 2022-09-22 17:57:45 minimal: damn, this doesn't make things easier. Gonna read up on CONFIG_I2C_RK3X and stuff 2022-09-22 17:58:05 rockchip,rk3228-i2c seems to be used by my sbc 2022-09-22 17:58:35 filoozom: I asked on #linux-rockchip IRC on libera 2022-09-22 17:59:13 mps: thanks! 2022-09-22 17:59:56 u-boot defconfig have CONFIG_MMC_DW_ROCKCHIP=y for this board 2022-09-22 18:00:38 default u-boot config is here https://tpaste.us/jPWl 2022-09-22 18:00:38 mps: yes, I have that enabled 2022-09-22 18:01:09 I've enabled everything they have enabled, that was basically my first debug step :p 2022-09-22 18:02:16 Well, all the things I found. For example RK8XX is nowhere to be sean in the config 2022-09-22 18:02:47 filoozom: they have REGULATOR_RK8XX enabled, I mentioned REGULATOR_RK808 earlier to you was not set 2022-09-22 18:06:38 filoozom: could you find config file for armbian? ls -l /proc/configs.gz ? 2022-09-22 18:07:06 maybe have to first 'modprobe configs' 2022-09-22 18:07:36 Apprently all those RK808 things need to be modules and can't be set to "y" 2022-09-22 18:07:44 mps: I'll quickly boot that, one sec 2022-09-22 18:07:46 filoozom: both Armbian and Openwrt appear to have CONFIG_PL330_DMA built into the kernel, my suspicion is this is needed for the SD controller to work 2022-09-22 18:08:23 pinsl: I've got mail from turris but it is not of big help. they just sent their kernel config and for outdated kernel 5.15 2022-09-22 18:08:46 minimal: yes, probably this is needed 2022-09-22 18:09:46 Alright, I'm recompiling with that 2022-09-22 18:09:54 And at the same time looking at the armbian config 2022-09-22 18:10:14 mps: from the Arbian dmesg the sd card mentions DMA so that's why I suspect it is needed. Does the Turris kernel config also built that into the kernel? 2022-09-22 18:10:47 minimal: turris is marvel 2022-09-22 18:11:20 I mean turris omnia is marvel mvebu soc 2022-09-22 18:11:44 mps: ah, got mixed up with the other conversation :-) 2022-09-22 18:12:27 armbian config: https://file.io/69QQ1KIRYbYm 2022-09-22 18:13:50 do you use swap? partition or file? 2022-09-22 18:13:55 http://ix.io/4b9x 2022-09-22 18:13:58 This is easier 2022-09-22 18:14:32 Does have CONFIG_PL330_DMA as well 2022-09-22 18:15:08 CONFIG_I2C_RK3X=y 2022-09-22 18:15:22 filoozom: yes, that's visible early in the dmesg output you showed 2022-09-22 18:15:27 CONFIG_PINCTRL_RK805=y 2022-09-22 18:15:37 mps: already suggested the I2C setting to him 2022-09-22 18:15:48 CONFIG_MFD_RK808=y 2022-09-22 18:17:18 CONFIG_PL330_DMA=y, yes minimal is right 2022-09-22 18:17:34 But when I set the RK808 configs it changes them back to modules 2022-09-22 18:18:11 sorry I have some eyes strain in last two days, I need like tree times more to read something 2022-09-22 18:19:38 filoozom: you could use this armbian config file to build kernel instead of one I posted 2022-09-22 18:20:53 Yeah, I'll try that next, because the DMA is apparently not enough... :'( 2022-09-22 18:21:28 [ 1.022852] dma-pl330 ff1f0000.dma-controller: Loaded driver for PL330 DMAC-241330 2022-09-22 18:24:57 Ah but it's a 5.15 config and it's all messed up 2022-09-22 18:26:49 I'll try this one: https://github.com/armbian/build/blob/master/config/kernel/linux-rockchip64-edge.config 2022-09-22 18:28:35 filoozom: yes, this sounds better 2022-09-22 18:29:28 (filozof in serbian means philosopher ;) ) 2022-09-22 18:29:56 Thank god you told me about cross-compiling from something other than alpine 2022-09-22 18:29:57 My god 2022-09-22 18:32:36 anyone know does pmOS have support for nanopi-r2s 2022-09-22 18:36:09 Well 2022-09-22 18:36:12 Still nothing 2022-09-22 18:37:42 did you set proper dtb to u-boot 2022-09-22 18:48:25 Tried the standard 5.19.10 one and the 5.15 one from armbian 2022-09-22 18:52:22 Also just tried the one from the armbian-edge image 2022-09-22 18:53:41 Could my u-boot be the issue? 2022-09-22 18:53:57 Maybe it doesn't hand the sd card over or something? 2022-09-22 19:02:17 filoozom: I guess it is u-boot doesn't load proper dtb 2022-09-22 19:03:17 last option is to go to kernel tree and do 'make defconfig' and build kernel and try with it. though kernel will biiiig 2022-09-22 19:03:42 be* 2022-09-22 19:04:19 If that kernel is known to work on armbian, is there a reason it could not work on alpine? 2022-09-22 19:04:24 If there's a chance I'll try it 2022-09-22 19:04:55 wrong dtb could be problem 2022-09-22 19:04:59 And I'm sure u-boot loads the right dtb because there's only one on my card 2022-09-22 19:05:17 Retrieving file: /boot/dtbs-nanopi/rockchip/rk3328-nanopi-r2s.dtb 2022-09-22 19:05:22 And it doesn't print any errors 2022-09-22 19:05:23 dtb must be from same kernel build 2022-09-22 19:06:40 if you build kernel copy proper dtb from that build to mmc 2022-09-22 19:07:12 That's what I initially did 2022-09-22 19:07:19 I build the entire kernel like the linux-edge package 2022-09-22 19:07:26 package it and then build an image with mkimage.sh 2022-09-22 19:07:30 make script I posted create in tarball all needed things 2022-09-22 19:07:34 So it takes the dtb from there, right? 2022-09-22 19:08:01 idk what mkimage.sh does, it is suspicious always to me 2022-09-22 19:08:33 that is why I prefer to build boot images with scripts 2022-09-22 19:08:34 I'll shasum both files but I'm fairly sure they're right 2022-09-22 19:09:09 with scripts I control everything 2022-09-22 19:09:31 Wonder if my dts needs patching of some sort 2022-09-22 19:09:41 Because the armbian-edge is completely different than mine as well 2022-09-22 19:09:48 Despite being built for the same kernel 2022-09-22 19:10:24 kernel dev told me on #linux-rockchip that R2S works out of the box with mainline kernels 2022-09-22 19:11:04 Hmm 2022-09-22 19:11:35 filoozom: you can join #linux-rockchip on libera and ask people competent with rockchip SOCs 2022-09-22 19:13:10 but because I don't have this board I mostly shooting in the dark 2022-09-22 19:13:53 seems like an interesting SBC but the "local" prices I see for it are crazy 2022-09-22 19:21:59 minimal: well, honestly it's basically the only thing I can get here 2022-09-22 19:22:05 I'd be happy for any suggestion 2022-09-22 19:22:27 It's 85€ on Amazon here (at least it provides fast shipping) but there's nothing in bigger retailers 2022-09-22 19:22:39 The beaglebone is quite available but a bit weak 2022-09-22 19:23:36 And I was looking into the Khadas VIM1 / VIM1S maybe instead of the Nanopi, but the latter is fairly "beautiful and compact" and comes with the heatsink case by default for the price 2022-09-22 19:24:51 ARM jungle 2022-09-22 19:25:48 Quite sad the RPi thing 2022-09-22 19:26:26 RPi is worst in arm jungle 2022-09-22 19:27:53 mps why? 2022-09-22 19:28:17 everyone talking about rpi 2022-09-22 19:29:31 It's just not possible to get one at this point 2022-09-22 19:29:55 mps: how big should a `make defconfig` kernel be? 2022-09-22 19:30:00 rly? 2022-09-22 19:30:14 They've been sold out for 2 years haha 2022-09-22 19:32:00 frag: rpi is broadcom SOC and broadcom is not free software friendly, actually opposite 2022-09-22 19:32:07 seems to still sell em at my local computer shop 2022-09-22 19:32:38 Oh that's what you're talking about 2022-09-22 19:32:55 filoozom: really big 2022-09-22 19:33:03 My `make defconfig` kernel is only 12 MB big 2022-09-22 19:33:07 That can't be right 2022-09-22 19:33:10 mps ah yeah .. kinda cool tho 2022-09-22 19:34:22 mps: which boards would you recommend? Gotta throw this nanopi out :p 2022-09-22 19:34:28 rockchip, allwinner, exynos are more free software friendly 2022-09-22 19:35:02 filoozom: I think rockchip is fine, only you hit some problems 2022-09-22 19:36:05 apparently there's a quite powerful RK3588 coming out soon 2022-09-22 19:36:20 if I have to buy new SBC I would first look at rockchip 2022-09-22 19:37:49 As long as it's not rk3288 2022-09-22 19:38:37 :) 2022-09-22 19:40:01 this is problem with using not popular linux distro, I always had some problems to get alpine on different boards 2022-09-22 19:41:01 How can I load the i2c-dev kernel modul? 2022-09-22 19:42:03 clius[m]: modprobe i2c_dev 2022-09-22 19:45:42 mps: rebuilt the entire kernel and it's still only 12M 2022-09-22 19:45:51 Then comes the message: 2022-09-22 19:45:51 modprobe: FATAL: Module i2c_dev not found in directory /lib/modules/5.15.68-0-lts 2022-09-22 19:46:31 clius[m]: so it is not enabled in -lts 2022-09-22 19:47:57 filoozom: kernel image is not big but look at modules 2022-09-22 19:48:09 total file is big 2022-09-22 19:48:27 Ah, it's 98MB then 2022-09-22 19:48:39 That's not too bad 2022-09-22 19:53:06 Win up to $1000 in crypto trading when you invest with just the minimum of $50... (full message at ) 2022-09-22 19:57:12 OH MY GOD 2022-09-22 19:57:16 make deps actually works!!! :D :D 2022-09-22 19:57:23 <3 <3 <3 2022-09-22 20:04:49 filoozom: ? 2022-09-22 20:13:07 Well, the Nanopi R2S boots on alpine :) 2022-09-22 20:15:16 Now is there a way to see which drivers are actually being used and build a slimmer kernel based on that? 2022-09-22 20:16:20 how things goes I guess alpine kernels will be built with 'make allyesconfig' soon :) 2022-09-22 20:22:22 mps: can't you split out extra kernel modules to a separate package if you don't like them? 2022-09-22 20:23:29 Newbyte: heh, I don't care if the kernel pkg is big 2022-09-22 20:23:51 mps: what do you care about then? kernel image? 2022-09-22 20:25:13 Newbyte: about nothing. If users want big kernel than we will make it 2022-09-22 20:26:06 mps: what's with the snarky comments then? 2022-09-22 20:27:16 my style 2022-09-22 20:28:55 For what it's worth I didn't think they were snarky, I thought they were information about big -> will take time to compile and stuff 2022-09-22 20:29:36 not sure what allyesconfig[=y] has to do with some extra modules needed for hardware support 2022-09-22 20:30:14 (people lost sense of humor) 2022-09-22 20:30:20 filoozom: if this was the first time I wouldn't have batted an eyelid 2022-09-22 20:30:39 mps: maybe get new jokes then 2022-09-22 20:35:03 pinsl: sorry, I will left this channel (tired of it), if you want to continue getting alpine on omnia with me please connect me on private chat 2022-09-22 20:49:10 I only need CONFIG_SQUASHFS_XZ for Alpine, right? 2022-09-22 21:13:49 * Starting busybox mdev .../lib/rc/sh/openrc-run.sh: line 15: can't create /proc/sys/kernel/hotplug: nonexistent directory 2022-09-22 21:14:02 Just have this issue left then I think I'm good to go! 2022-09-23 01:58:11 hello 2022-09-23 01:59:08 hi 2022-09-23 02:00:18 i am running alpine linux on my server, I have two SAS drives connected to my server via an HBA. I want them to mount as storage on boot, I have tried putting them in my fstab file to mount at a certain directory but they will not mount. 2022-09-23 02:00:30 there are no errors from mount -av 2022-09-23 02:01:32 can you pastebin your fstab? 2022-09-23 02:03:29 https://pastebin.com/jQBpTPCr 2022-09-23 02:03:40 that is just currently 2022-09-23 02:04:07 i have been troubleshooting, I had the UUIDs of the disks where LABEL is currently in the paste 2022-09-23 02:04:18 still never mounted on boot 2022-09-23 02:05:02 and yes I know that the pound symbol comments them out 2022-09-23 02:05:59 but the previous configuration, of course, had the pounds removed and the blkid (UUID) in place... 2022-09-23 02:06:09 yes 2022-09-23 02:06:21 anyway, the disks mount just fine after boot is complete? 2022-09-23 02:06:27 yes they do 2022-09-23 02:09:08 well since nobody else has spoken up, I'm just going to give my best guesses based on my previous (but admittedly not recent) experience with SAS drives... I am wondering if the controller is just incredibly slow to come up (ran into an HP/LSI controller that took a full 30 seconds to come online after a fresh boot)... 2022-09-23 02:09:29 or if it's not mounting because of the kernel module for the controller not being in the initrd? 2022-09-23 02:09:32 that is what i was thinking as well 2022-09-23 02:09:46 i couldnt find any info on delaying the auto mount 2022-09-23 02:10:06 so i am just going to write a script and run it with cron 2022-09-23 02:10:14 or something equivalent 2022-09-23 02:10:19 haha. that's what I was thinking too 2022-09-23 02:10:28 an @reboot line in /etc/crontabs/root 2022-09-23 02:10:39 yes, i wanted to avoid doing it 2022-09-23 02:11:21 before i axed my install and went to alpine i had it running debian and it was basically held together with duct tape so to speak 2022-09-23 02:11:27 very hacky workarounds 2022-09-23 02:13:19 anyways thanks for the two sense, have a good day brother 2022-09-23 02:13:54 oh darn, he left before I could ask for his mkinitfs.conf :x 2022-09-23 09:13:54 is there a tool similar to xlocate (from void's xtools) that searches for (possibly not installed) files in packages? xlocate keeps its own file index which is rebuilt by xlocate author and needs to be synced to local. i could scrape https://pkgs.alpinelinux.org/contents sure, but is there anything already? 2022-09-23 09:19:36 or maybe the file index backing pkgs.alpine* is publicly available? 2022-09-23 09:20:33 Piraty_, if needed I've coded a small utility that takes advantage of pkgs.alpinelinux.org 2022-09-23 09:20:33 https://github.com/fcolista/apkfile 2022-09-23 09:21:37 exactly this, thanks 2022-09-23 09:22:05 the distinction of file and path in the form is weird 2022-09-23 09:23:45 sometimes you want to get a filename at any path 2022-09-23 09:23:52 a mere *name does not give you that 2022-09-23 09:24:24 and sometimes you want to get all files in a path, but /path/* gives you subpaths too without fancy options 2022-09-23 09:24:35 realistically name/path split is the most versatile one at a glance 2022-09-23 09:24:37 maybe i'm wrong 2022-09-23 09:25:21 allow regex and it's fine ;) 2022-09-23 09:25:55 yeah i sure want to type /path/[^/]* instead of just the path 2022-09-23 10:26:39 what was the printf way of telling the termianl to not echo back keys pressed? 2022-09-23 10:27:30 whatever tput cinvis does :P 2022-09-23 10:28:12 tsk tsk, such bloat! 2022-09-23 10:30:13 `tput civis`* 2022-09-23 10:31:18 frag: printf '\e[?25l' 2022-09-23 10:31:58 found with: tput civis | xxd 2022-09-23 10:32:33 ikke: hmm that seems to be for hiding the cursor 2022-09-23 10:32:43 oh, right 2022-09-23 10:40:36 frag: the only thing I can find is `stty -echo`, but that does not work for me 2022-09-23 10:42:50 ikke: that does not work for me either, hmm, thought this would be something basic haha 2022-09-23 10:43:07 frag: read -s does work (with non-posix versions of read) 2022-09-23 10:46:17 ikke, yeah i have a script that worked on a faster computer, but now it prints pressed chars on the screen, i guess between 'read' commands 2022-09-23 10:46:36 i wonder why stty -echo doesnt work, i have that in this script too .. 2022-09-23 10:58:59 does anyone have the asoundrc for the rpi 4b, using headphones and one HDMI? 2022-09-23 13:21:13 finally, i got here. 2022-09-23 13:26:26 yay 2022-09-23 13:27:15 frag: I think stty works on ttys and in scripts, you don't have a tty 2022-09-23 13:27:21 it's a "interactive console" thing 2022-09-23 13:29:11 i am trying to boot a ARM board with extlinux, looks like the 'modloop=' option was not get loaded. 2022-09-23 13:29:36 everyone using ARMs whats going on 2022-09-23 13:29:48 satmd: it is a (zsh) script 2022-09-23 13:31:06 so i downloaded th Generic tar.gz for aarch64, everything is extracted to a fat32 usb stick. 2022-09-23 13:31:34 frag: because ARM is no noise and low power and fast enough 2022-09-23 13:32:12 uuidNuniq: mhm :] 2022-09-23 13:32:20 which arm u use? 2022-09-23 13:32:37 and i slietly modifyed the extlinux.conf to add an option 'modloop=modloop-lts' 2022-09-23 13:33:05 frag: lowend A53 4core 1.5Ghz with 2GB ram and 8GB emmc 2022-09-23 13:36:40 frag: home firewall server and router. 2022-09-23 13:37:10 the hardware is way good enough for the purpose... i don't need a huge x86 to do that. 2022-09-23 13:37:14 uuidNuniq: hmm, but its a phone? 2022-09-23 13:37:53 a phone is just a small laptop : ) 2022-09-23 13:37:53 not a phone, just a Single Board Computer. palm size. 2022-09-23 13:38:30 okoko, yeah i switched to rpi 4b recently 2022-09-23 13:39:09 rpi is more powerful than my board. but rpi does not have hardware AES 2022-09-23 13:39:50 here is my CPU flags : fp asimd evtstrm aes pmull sha1 sha2 crc32 cpuid 2022-09-23 13:40:31 havent used any tools to test, but seems to me its not much difference between encrypted and not encrypted root partition 2022-09-23 13:42:27 when it comes to https tls1.2 or tls1.3 that kind of stuff, there is a difference 2022-09-23 13:43:25 ah 2022-09-23 13:43:41 http can slow 2022-09-23 13:44:50 it's not related to 'http slow' at all 2022-09-23 13:50:02 for ssl issues: check your random number generator first 2022-09-23 13:50:10 s/ssl/ssl performance/ 2022-09-23 13:51:22 i got a hardware rng 2022-09-23 13:51:39 so, there is hardware rng and aes on my board. 2022-09-23 13:52:36 and it is being used, I assume? 2022-09-23 13:53:51 aes is being used. not sure about the rng... i don't see a service on rc-update related to rng 2022-09-23 14:15:17 you would want to add rng-tools and add rngd 2022-09-23 14:16:24 psykose: but arch wiki says it is not needed anymore. 2022-09-23 14:16:42 since kernel 5.6 2022-09-23 14:17:32 and then rngd -l can list the sources, hwrng should show up 2022-09-23 14:18:30 if it does, then it works 2022-09-23 14:18:48 i assume from there the kernel can use it directly and rng-tools themselves can be deleted 2022-09-23 14:18:55 psykose: i used rng-tools before, and yes, it works on my hardware. but since the wiki says it is not needed i disabled it. 2022-09-23 14:20:03 this wiki : https://wiki.archlinux.org/title/Random_number_generation 2022-09-23 16:05:23 which ARM works best as desktop? 2022-09-23 16:06:06 frag: maybe this: https://www.solid-run.com/arm-servers-networking-platforms/honeycomb-lx2/ 2022-09-23 16:07:29 the desktop socketed 80-core altra, obviously 2022-09-23 16:07:33 you have $10k, right 2022-09-23 16:10:43 psykose: thats just half a btc! cheap! 2022-09-23 16:11:11 maybe, i have no idea what it is in funny money 2022-09-23 16:12:50 ok boomer! 2022-09-23 16:13:52 you are probably older than me 2022-09-23 16:50:14 Win up to $1000 in crypto trading when you invest with just the minimum of $50... (full message at ) 2022-09-23 16:52:54 Hello, I am having trouble getting iio-sensor-proxy to run. I can run 'rc-service iio-sensor-proxy start' or 'rc-update enable iio-sensor-proxy default' but when checking the status is says 'stopped'. Does anyone have any idea what I could be doing wrong or try next? 🤔 2022-09-23 17:07:03 it means it merely failed to start for some reason 2022-09-23 17:07:33 you can try invoke it with iio-sensor-proxy -v as root by hand i think to see what it prints if anything 2022-09-23 17:07:37 i'd guess it's just broken 2022-09-23 17:09:17 tail -f /var/log/message? 2022-09-23 17:16:32 that only works for things that log to syslog 2022-09-23 17:23:08 psykose: at least the SSD or SD reports some of issues by default 2022-09-23 17:26:25 '$thing exited with return code 0; respawned $thing too many times, exiting' 2022-09-23 17:31:38 uuidNuniq: Hi, Alpine rng-tools package maintainer here. The kernel support and use for hwrng's has changed quite a bit in the past 6 months or so 2022-09-23 17:33:07 recent kernels so use hwrngs (as well as bootloader seeds if provided and hypervisor seeds). However it is not as simple as if kernel > 5.6 2022-09-23 17:37:51 minimal: does this matter? cat /proc/sys/kernel/random/entropy_avail 2022-09-23 17:37:57 uuidNuniq: plus Alpine's kernel is still on 5.15.x (unless you use linux-edge) and so it doesn't contain many of the recent rng-related changes 2022-09-23 17:38:02 it reports 256 2022-09-23 17:38:59 uuidNuniq: the value of the that /sys entry changed in recent kernels, that sounds like a typical value for a recent kernel 2022-09-23 17:39:25 i recall when i used rng-tools, it reports around 4000 2022-09-23 17:40:50 uuidNuiq: on older kernels "poolsizde" was 4096 and a *good* value for "entropy_avail" would be around 3072 2022-09-23 17:41:14 i see, this is a big change 2022-09-23 17:42:18 there have been quite a few changes and there are more in kernels after the 5.15.x that Alpine uses (e.g. I noticed another change submitted last week that I assume will be in 6.0 kernel) 2022-09-23 17:43:21 for a long time the linux kernel did use hwrngs but did not "credit" their entropy (and so "entropy_avail" did not increase) whereas rngd did credit entropy 2022-09-23 17:44:32 so yes with recent kernels there is the question of whether it is worth using rng-tools (and haveged) 2022-09-23 17:45:20 there are still some use cases for rng-tools, e.g. if you want to add additional entropy sources 2022-09-23 17:49:55 i see 256 with and without rngd 2022-09-23 17:50:02 6.0_rc6 2022-09-23 17:50:46 psykose: yes, that is the new behaviour 2022-09-23 17:50:51 neat 2022-09-23 17:51:43 from memory they changed the hash algo (Blake2?) and this counter changed as a result 2022-09-23 17:52:52 some of the recent changes are mentioned here: https://www.zx2c4.com/projects/linux-rng-5.17-5.18/ 2022-09-23 17:54:59 mhm 2022-09-23 17:55:46 psykose: there's a link in that page to random-related kernel commits, you'll see there have been quite a lot the past 18months or so 2022-09-23 18:09:20 eletrotupi: are you around? wondering if you could check something for me re: vmd 2022-09-23 18:14:23 does it ever make sense to install the binutils- package on the same architecture it's for? like, install binutils-aarch64 on an aarch64 device? 2022-09-23 18:14:23 In my head I'm thinking it could be a simple fix for programs that assume aarch64 is not the native architecture, and don't let you change the name it uses to invoke the tools. 2022-09-23 18:15:01 no 2022-09-23 18:15:24 fair 2022-09-23 18:16:11 well, sure, fair enough- they are named with the full triple 2022-09-23 18:16:36 but if they had hardcoded triples- why would they be invoking $arch-alpine-linux-musl-as ? would be weird 2022-09-23 18:17:03 would need to be an automatic-detection type of build system that is modified to weirdly call something that way, i'd guess 2022-09-23 18:17:13 never seen it (compared to just hardcoding the usual -none-linux) 2022-09-23 18:17:21 er, unknown 2022-09-23 18:17:53 usually fixing that is just a 1-2 line patch or so anyway 2022-09-23 18:45:37 heya minimal, I'm here now 2022-09-23 18:45:46 yeah, for sure man 2022-09-23 18:45:49 what do you need? 2022-09-23 18:52:40 eletrotupi: was just wondering if you could try re-enabling VMD in the BIOS/UEFI and booting the Alpine ISO/USBstick to collect some info for me, looking for info to detect presence of VMD 2022-09-23 18:57:43 right, can we do this in 1h +-? 2022-09-23 18:58:21 I'm at worktime right now, and I'm kind of stuck at the moment 2022-09-23 18:58:21 sure, no rush. Let me DM you with details 2022-09-23 18:58:29 ah sweet! 2022-09-23 19:35:22 Win up to $1000 in crypto trading when you invest with just the minimum of $50... (full message at ) 2022-09-23 19:49:59 an echo with redirection in a temporary file but nothing gets written.[201~started when I do an lbu commit. 2022-09-23 19:50:29 I put a script in /etc/lbu/post-package.d/ but it doesn't seem to get started when I do an lbu commit. 2022-09-23 19:50:34 To test I only put an echo with redirection in a temporary file but nothing gets written. 2022-09-23 19:52:20 https://wiki.alpinelinux.org/wiki/Alpine_local_backup#Execute_a_script_as_part_of_a_backup 2022-09-23 19:53:25 cat /etc/lbu/post-package.d/00_apkovl_hardlink 2022-09-23 19:53:25 echo "works" > /tmp/output 2022-09-23 19:53:34 where I'm wrong? 2022-09-23 19:54:02 is it executable? 2022-09-23 19:54:15 ye 2022-09-23 19:54:16 yep 2022-09-23 19:57:34 does executing `run-parts /etc/lbu/post-package.d` output something? 2022-09-23 19:58:22 run-parts: can't execute '/etc/lbu/post-package.d/00_apkovl_hardlink': Exec format error 2022-09-23 19:58:50 head -n1 /etc/lbu/post-package.d/00_apkovl_hardlink 2022-09-23 19:59:20 there is no shebang line 2022-09-23 19:59:26 I need to put one? 2022-09-23 19:59:28 Yes 2022-09-23 19:59:39 At least, the error indicates it wants one 2022-09-23 20:00:18 ok, if is this, we need to update the wiki 2022-09-23 20:00:29 one moment, I try 2022-09-23 20:01:47 uh 2022-09-23 20:01:57 but this script is runned also with lbu status 2022-09-23 20:02:02 :| 2022-09-23 20:03:17 apparently cmd_status runs cmd_package 2022-09-23 20:03:34 which runs the scripts 2022-09-23 20:04:31 It creates a temporary archive to compare 2022-09-23 20:04:47 ok 2022-09-23 20:05:13 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/lbu.in 2022-09-23 20:05:17 It's a shell script 2022-09-23 20:05:52 I know, recently i had patched this file for my previous problef (if you remember) 2022-09-23 20:06:29 Directory permissions? 2022-09-23 20:06:42 no, one moment i can send to you the link 2022-09-23 20:07:17 https://gitlab.alpinelinux.org/alpine/alpine-conf/-/issues/10534 2022-09-23 20:07:56 I don't think I was involved 2022-09-23 20:08:21 or maybe I was, but don't remember 2022-09-23 20:09:26 * ok, it's not important now but in this case I wouldn't want to edit lbu.in because what I want to do shouldn't be inside that script. 2022-09-23 20:09:51 maybe it' s better if I ask you what I want to do. 2022-09-23 20:10:04 I mentioned it because it's quite readable 2022-09-23 20:10:17 After I do a commit I want to save an hardlink of the backup in another dir. 2022-09-23 20:10:17 I'm not all that familiar with LBU 2022-09-23 20:13:12 is there another way that you recommend? 2022-09-23 20:13:50 i was thinking of creating a special script that would wrap lbu ci. 2022-09-23 20:13:58 the package hooks are more for changing things that end up in the apkovl I gather 2022-09-23 20:14:22 helmen[m]: yeah, I was thinking about suggesting that as well 2022-09-23 20:15:15 OK then just one more question. 2022-09-23 20:16:44 when lbu ci is launched the script mounts in write, saves and mounts again in read. 2022-09-23 20:17:05 Should I rewrite this behavior in my script as well, or is there a way to avoid it? 2022-09-23 20:18:20 And is there an environment variable that tells where the booting disk is mounted? 2022-09-23 20:18:46 only for generalize the script (i have more than one memory stick with alpine) 2022-09-23 20:19:10 There is no such environment variable 2022-09-23 20:19:47 ok, I think I can read it in /etc/lbu/lbu.conf 2022-09-23 20:22:29 thanks for the hel ikke 2022-09-23 20:22:34 no problem 2022-09-23 20:41:14 psykose: uuidNuniq: I'm the alt for BitHermit. I tried all your suggestions, but came up with nothing. I guess I have to submit a bug =( 2022-09-23 20:55:33 Bug is filed: https://gitlab.alpinelinux.org/alpine/aports/-/issues/14209 2022-09-23 21:00:10 hmm, does not help that the package maintainer does not have time anymore for Alpine Linux 2022-09-23 21:01:21 ikke: I found another solution, make the copy by first checking if the file has changed. 2022-09-23 21:01:56 that way lbu st should not give me any problems 2022-09-23 23:16:42 getting some weird errors, think with the latest dbus update, not sure tho 2022-09-23 23:17:07 starting i3 with dbus-launch gives "error relocating /usr/bin/dbus-launch" "symbol not found" 2022-09-23 23:17:20 also saw weechat having a similar issue with loading the perl plugin 2022-09-23 23:17:52 what kind of symbol? 2022-09-23 23:23:03 not sure, the dbus one is somekind of dbus related symbol _dbus... 2022-09-23 23:28:36 _dbus_lookup_user_dbus: symbol not found 2022-09-23 23:29:04 when launching weechat, it's something about the perl_newsv_type symbol 2022-09-23 23:29:08 not sure what's going on with that 2022-09-23 23:29:53 everything else seems to be fine. i can launch i3 without dbus-launch, firefox works, etc. 2022-09-23 23:46:02 $5 says you're mixing release and edge 2022-09-23 23:48:00 i see, did have a non edge repo enabled 2022-09-23 23:48:09 how do i resolve this? 2022-09-23 23:52:13 just using /apk fix.../ ? 2022-09-23 23:52:28 pick either a stable release or edge, keep repositories only for that one and do `apk upgrade -Ua` 2022-09-23 23:56:02 alright, gonna reboot really quick 2022-09-23 23:58:54 still seeing that symbol not found error 2022-09-23 23:59:21 should i try to just reinstall everything? 2022-09-23 23:59:23 could you copy-paste the whole message as you see it on screen? 2022-09-24 00:01:25 can't from tty, but dumped the outpout to a log file 2022-09-24 00:01:27 one sec 2022-09-24 00:04:08 X.Org X Server 1.21.1.4 X Protocol Version 11, Revision 0 Current Operating System: Linux TSUKI 5.19.10-0-edge #1-Alpine SMP PREEMPT_DYNAMIC Tue, 20 Sep 2022 13:36:04 +0000 x86_64 Kernel command line: BOOT_IMAGE=vmlinuz-edge root=/dev/alpinevg/rootlv modules=sd-mod,usb-storage,ext4 cryptroot=UUID=531a4a1c-a540-46e8-a8fe-ba143389e7cd cryptdm=root quiet rootfstype=ext4 2022-09-24 00:04:10 initrd=intel-ucode.img,initramfs-edge Current version of pixman: 0.40.0 Before reporting problems, check http://wiki.x.org to make sure that you have the latest version. Markers: (--) probed, (**) from config file, (==) default setting, (++) from command line, (!!) notice, (II) informational, (WW) warning, (EE) error, (NI) not implemented, (??) unknown. (==) Log file: 2022-09-24 00:04:12 "/home/tsuki/.local/share/xorg/Xorg.0.log", Time: Fri Sep 23 19:02:28 2022 (==) Using system config directory "/usr/share/X11/xorg.conf.d" xf86EnableIO: failed to enable I/O ports 0000-03ff (Operation not permitted) Error relocating /usr/bin/dbus-launch: _dbus_lookup_user_bus: symbol not found xinit: connection to X server lost waiting for X server to shut down (II) Server 2022-09-24 00:04:14 terminated successfully (0). Closing log file. 2022-09-24 00:05:48 i've been using edge for quite a while, so edge is what i need and want 2022-09-24 02:29:00 i tried wiping apk cache and re-installing everything, but still getting that symbol not found error 2022-09-24 02:42:21 apk policy dbus dbus-libs dbus-x11 2022-09-24 02:42:58 but i'm heading out now 2022-09-24 07:28:45 Did anybody try setting up dm-verity in Alpine? 2022-09-24 10:16:51 when will alpine use s6 2022-09-24 10:20:07 uuidNuniq: when it's done 2022-09-24 10:56:41 try to configure ossec, install package successful but no bin folder with ossec-control (needed for /etc/init.d/ossec-hids 2022-09-24 10:57:33 https://pkgs.alpinelinux.org/contents?file=ossec-control&path=&name=&branch=edge&arch=x86_64 2022-09-24 11:00:29 apk search ossec, I only get ossec-hids-3.6.0-r1,ossec-hids-openrc-3.6.0-r1, ossec-hids-contrib-3.6.0-r1,ossec-hids-doc-3.6.0-r1 2022-09-24 11:00:59 md_: what arch are you on? 2022-09-24 11:01:16 Linux smdcs007 5.15.55-0-rpi #1-Alpine Mon Jul 18 10:04:18 UTC 2022 armv6l Linux 2022-09-24 11:01:28 aah, arhmf 2022-09-24 11:01:31 armhf* 2022-09-24 11:01:49 so this is apparently not built for armhf 2022-09-24 11:02:25 nor any other arm based arch 2022-09-24 11:02:29 sad.ok.thx 2022-09-24 11:54:42 Ermine: will the there be such a day? 2022-09-24 11:55:04 is there already a plan or discusstion? 2022-09-24 11:55:19 uuidNuniq: s6 is not ready yet to replace openrc 2022-09-24 11:56:13 s6 is quite interesting according to the dcos on their homepage. 2022-09-24 11:59:26 sorry, too many typos 2022-09-24 14:21:59 Hello there. I need some clarifications about installing Alpine on a single partition (possibly 2 with /boot) on a virtual machine. I would like to make tiny empty VMs to use with GNS3 for quick exportable labs to do with students. So, I created an ext4 (^has_journal) filesystem on one partition mounted on /mnt. Setup-alpine with "none" on everything related to apk and disk. Then `setup-disk -m sys -s 0 /mnt`, which seems to install e 2022-09-24 14:21:59 verything correctly. On boot, I get "no default configuration found" and "no ui" on next line. Am I missing something please? 2022-09-24 14:23:55 Seems like syslinux was not installed properly 2022-09-24 14:24:15 Can you verify the contents of /boot ? 2022-09-24 14:25:38 uuidNuniq: this day will definitely happen. Be patient 2022-09-24 14:30:09 ikke, /boot has files and even immutable files 2022-09-24 14:31:13 setup-disk also installs the MBR right? 2022-09-24 14:31:35 Yes 2022-09-24 14:32:19 And the process to install the system on a single ext4 partition doesn't have any tricks, does it ? 2022-09-24 14:35:00 You have something like /dev/sda[12]? 2022-09-24 14:36:11 I didn't when I tried installing on a single partition, I did mkfs.ext4 -O ^has_journal /dev/sda, and I only got /dev/sda mountable on /lnt 2022-09-24 14:36:18 /mnt * 2022-09-24 14:38:32 Hmm, I have no experience with such a setup, so it can indeed be that you need to do something special, if even supported. 2022-09-24 14:39:20 So better to have 2 partitions then? If I'm not mistaken, the wiki advise to have at least /boot on FAT16 and the rest can stay ext4, is it right ? 2022-09-24 14:48:45 axel50397: so you have not actually installed Alpine on a partition, you have installed it directly on a disk with no partitions 2022-09-24 14:50:33 axel50397: so you'd have to install syslinux differently than usual as it expects a partitioned disk 2022-09-24 14:50:33 Ok, let me try again. 1 partition, ext4, mounted on /mnt, setup-disk with -m sys -s 0 /mnt. Is it right ? 2022-09-24 14:52:34 if you want to install on a partitioned device then why not just run setup-alpine and let it set things up for you? 2022-09-24 14:53:48 ah, you don't want a swap partition, right 2022-09-24 14:53:52 Yes 2022-09-24 14:54:17 so the above command should do it then 2022-09-24 14:54:24 Light as hell, to put in GNS3 2022-09-24 14:54:31 On thanks, I try it now 2022-09-24 14:57:43 axel50397: setup-alpine -c myconf ; edit myconf ; setup-apline -f myconf 2022-09-24 14:58:08 axel50397: ah, I think setup-disk creates a separate /boot partition, let me check 2022-09-24 14:59:00 uuidNuniq, thanks for the suggestion but I'm creating this vm with proxmox and have no access to the fs. I could do a conf file but won't have access to it after next reboot 2022-09-24 14:59:39 minimal, what I did was creating a /boot folder manually in /mnt, then /mnt/boot was installed with extlinux/syslinux files 2022-09-24 15:00:22 proxmox? mount an Alpine ISO... 2022-09-24 15:00:32 axel5-397: but was a boot partition created? 2022-09-24 15:01:31 with a valid conf file, alpine can be installed with 60 seconds :P that is the fastest on the earth 2022-09-24 15:01:33 minimal, no because it seems setup-disk can accept a mounted root as last argument 2022-09-24 15:02:30 uuidNuniq, I'm using alpine-virt but as I'm trying to install this since 2 days, everytime I reboot, I'm either on my non working install or iso 2022-09-24 15:02:31 axel50397: that has nothing to do with it - setup-disk for Sys appears to create a 100MB boot partition by default 2022-09-24 15:02:59 DISKOPTS="-k virt -s 0 -m sys /dev/vda" 2022-09-24 15:03:01 you cna override this by setting BOOT_SIZE environment variable to something other than 100 2022-09-24 15:03:03 here is what i do ... 2022-09-24 15:03:13 same virt in QEMU 2022-09-24 15:03:35 minimal, so it doesn't create a boot partition in my case as boot files are simply copied in /boot currently 2022-09-24 15:03:57 uuidNuniq, this is exactly my command line using setup-disk (minus -k) 2022-09-24 15:04:08 -m sys creates a ESP and a ROOT on UEFI 2022-09-24 15:04:47 axel50397: you're not listening to what I said - whether or not boto files are put inside /mnt is a different matter than whether a separate partition is created for boot use 2022-09-24 15:05:04 axel5-0397: do a "fdisk -l /dev/sda" to see how the disk is partitioned 2022-09-24 15:05:29 minimal, no I understood, I was answering your question. It doesn't create a separate boot partition so I may have done something wrong 2022-09-24 15:05:36 Ok 2022-09-24 15:06:10 I see no option available to tell setup-disk not to create a boot partition for Sys install, unless perhaps setting BOOT_SIZE=0 can prevent this, not sure 2022-09-24 15:06:48 Well, now I have started over with a blank .raw file. So I'm using cfdisk to create an msdos partition table first 2022-09-24 15:07:40 So right now I have /dev/sda1 starting at 2048, with a type 83 (linux) 2022-09-24 15:08:10 fdisk -l shows the same 2022-09-24 15:08:13 axel50397: https://gitlab.alpinelinux.org/alpine/alpine-conf/-/blob/master/setup-disk.in#L1274 2022-09-24 15:08:58 minimal, so 0 might do the trick as swap size does ? 2022-09-24 15:11:25 Now ls /dev/sda* shows only /dev/sda T_T 2022-09-24 15:11:41 Even though fdisk shows /dev/sda1 2022-09-24 15:11:57 axel50397 I think so but you'd need to run it as "BOOT_SIZE=0 setup-disk ...." 2022-09-24 15:12:13 I'll try. I reboot and see if the partition appears 2022-09-24 15:16:29 if [ -n "$USE_EFI" ] && [ -z "$BOOT_SIZE" ]; then 2022-09-24 15:16:41 so boot_size is only for the ESP 2022-09-24 15:16:55 there is no boot_size for mbr 2022-09-24 15:18:05 uuidNuniq: there is setting BOOT_SIZE for UEFI *if* no BOOT_SIZE is already specified 2022-09-24 15:18:57 I partitioned the disk with gparted. Now I have /dev/sda1 2022-09-24 15:19:28 Now classic install without FS first 2022-09-24 15:19:37 uuidNuniq: line 1163 defaults to 100M, so if BOOT_SIZE is not passed in then the code you pointed to sets a value for UEFI systems, otherwise when native_disk_install is called the 100M is used for MBR systems 2022-09-24 15:20:24 ah, i see 2022-09-24 15:20:47 axel50397: why not let sys-disk do the partitioning for you? 2022-09-24 15:21:43 Because… The disk is 80MB in size and I didn't get any "no space left" yet. As boot is 100M by default, I didn't risk it 2022-09-24 15:22:30 axel50397: but by running "BOOT_SIZE=0 setup-disk -m sys -s 0" you avoid both boot and swap partitions... 2022-09-24 15:22:38 Indeed 2022-09-24 15:23:20 axel50397:> also you expect to successfully install Sys mode in 80MB??? 2022-09-24 15:24:00 Well… That's why I'm trying to do, as I didn't hit any "no space left", I had hope ^^ 2022-09-24 15:24:11 I don't remember how small I've done VM Alpine installs in but don't think I've done so in 80MB 2022-09-24 15:24:46 Also, this article let me understand it's possible to set your own partition layout 2022-09-24 15:24:47 linux-virt-5.15.69-r0 installed size: 32 MiB 2022-09-24 15:24:53 minimal, https://wiki.alpinelinux.org/wiki/Setting_up_disks_manually#For_.22sys.22_disk_mode_installs_2 2022-09-24 15:25:10 uuidNuniq, default installer?? 2022-09-24 15:25:15 32MiB ? 2022-09-24 15:25:26 apk info linux-virt 2022-09-24 15:26:04 I mean, you simply install it with setup-alpine and it's 32MiB ?! 2022-09-24 15:26:23 axel50397: i think setup-disk allow a mount point, so, it is possible to setup a layout 2022-09-24 15:26:42 kernel itself is 32M 2022-09-24 15:26:43 axel50397: no, he is saying that the virt kernel package alone takes 32MiB of disk space 2022-09-24 15:26:53 Ah yes ok 2022-09-24 15:27:41 I also don't need openssh as It's intended to be used on GNS3 via terminal 2022-09-24 15:30:57 BOOT_SIZE=0 setup-disk -m sys -s 0 /mnt, got me "/mnt/boot is device /dev/sda1, /boot is device /dev/sda1, you might need fix the MBR to be able to boot" 2022-09-24 15:32:25 minima-l, according to this (https://wiki.alpinelinux.org/wiki/Dualbooting#In_Alpine_2.2.3_or_newer) setup-disk with a mounted partition won't create a /boot, and the fix is to change the LINUX and INITRD to /boot/*** 2022-09-24 15:32:56 Logically I just have to edit /mnt/boot/extlinux.conf 2022-09-24 15:33:30 install_mounted_root "${1%/}" \ && echo "You might need fix the MBR to be able to boot" >&2 2022-09-24 15:33:38 that is a success :) 2022-09-24 15:33:45 \o/ 2022-09-24 15:34:34 axel50397: so looks like while setup-disk can use BOOT_SIZE=0 it doesn't correctly handle it with regard to the Syslinux setup I guess 2022-09-24 15:34:48 for Sys mode 2022-09-24 15:34:54 there is a setup-bootable 2022-09-24 15:35:10 The .raw disk size is currently 51M (ls -lsh) 2022-09-24 15:35:33 ls is not correct 2022-09-24 15:35:38 use du -h 2022-09-24 15:36:29 du outputs the same size 2022-09-24 15:36:44 I can send it to you guys if you want 2022-09-24 15:36:44 and do a fstrim -v / 2022-09-24 15:37:09 fstrim -v /mnt i mean i this case 2022-09-24 15:37:36 Wait, the du/ls -lsh was done on the host machine, not the guest 2022-09-24 15:37:47 fstrim should be done on the guest alpine right? 2022-09-24 15:37:52 yeah 2022-09-24 15:38:32 is there any change after fstrim? 2022-09-24 15:38:44 Ok so. Now I'm booting the guest on the HDD, it shows "Booting from Hard Disk" and that's it. It's stuck there, no "SYSLINUX ..." shown :( 2022-09-24 15:40:07 minimal, actually I don't really care about sys or data mode as I would like to simply change settings or write some scripts (sometimes) on volatile machines. The single partition is for me to avoid choosing to put scripts in the "biggest" partition as, as a GNS3 user, I may put it anywhere and destroy it few hours later 2022-09-24 15:41:16 i might have an idea? 2022-09-24 15:41:24 Shooooot 2022-09-24 15:41:33 setup-alpine -c myconf ; edit myconf 2022-09-24 15:41:56 now put : DISKOPTS="-k virt -m sys /mnt" 2022-09-24 15:42:15 Your suggestion is to add the -k option basically ? 2022-09-24 15:42:15 and change other things, and setup-alpine -f muconf 2022-09-24 15:42:19 *myconf 2022-09-24 15:42:26 I'm trying right now 2022-09-24 15:42:46 -k and /mnt 2022-09-24 15:42:58 -k is to set kernel flavor 2022-09-24 15:43:41 ss? 2022-09-24 15:43:46 /mnt instead of a device name, meant to be installing to a mount point 2022-09-24 15:44:10 well you have to format and mount before setup-alpine -f 2022-09-24 15:44:22 This is already what I do with setup-disk 2022-09-24 15:44:34 I put a mount point instead of a device 2022-09-24 15:44:43 i know,but setup-disk does not take care of other things. 2022-09-24 15:44:53 ok 2022-09-24 15:45:08 you can config myconf to let setup-alpine handle everything for you 2022-09-24 15:45:35 Well, while the idea isn't bad, when I reboot to check the installation, I will have to redo again the script 😂 2022-09-24 15:46:59 what script? 2022-09-24 15:47:15 The conf sorry 2022-09-24 15:47:39 copy it to somewhere 2022-09-24 15:48:00 It's in a qemu guest on a distant machine 2022-09-24 15:49:04 I'm trying anyway 2022-09-24 15:49:20 the conf is very simple.. not too much to edit 2022-09-24 15:52:09 the whole thing are shell scripts... oh by the way 2022-09-24 15:52:16 axel50397: correcting what I said earlier, I do have a VM using approx 70MiB of disk space 2022-09-24 15:52:22 add a 'hwclock -s' on top of the 'myconf' 2022-09-24 15:52:32 so you get the correct date and time 2022-09-24 15:53:30 uuidNuniq, what would be the interfacesopts for setup-alpine to ignore interface settings? I already set them using ip a and ip r, but the end machine should stay dhcp 2022-09-24 15:53:31 when the official ISO boots up, the date was always wrong 2022-09-24 15:53:40 Thanks for the tips 2022-09-24 15:53:46 minimal, Good :D 2022-09-24 15:54:12 uuidNuniq: clock wrong on a VM? it should have a RTC that is provided by the hypervisor 2022-09-24 15:55:20 axel50397: you definately what "-k virt" to use linux-virt, if you use linux-lts it will drag in lots of unnecessary firmware files 2022-09-24 15:55:23 axel50397: the interface config is needed to install basic stuff i guess? 2022-09-24 15:56:18 uuidNuniq, I setup the interface manually already, I answer "done" when setup-alpine asks me about it. I deleted interfacesopts to check 2022-09-24 15:57:11 It's installing 2022-09-24 15:57:20 With -k virt :) 2022-09-24 15:57:20 if there is no such option it willdo : setup-interfaces ${quick:+-a} ${rst_if:+-r} 2022-09-24 15:57:37 oh no 2022-09-24 15:57:39 sorry 2022-09-24 15:58:19 if you comment out INTERFACESOPTS, setup-alpine will do nothing for the interface 2022-09-24 15:58:37 can i delete my words? 2022-09-24 15:58:45 shit... i am sleepy 2022-09-24 15:59:08 if [ -n "$INTERFACESOPTS" ]; then ..... else setup-interfaces ${quick:+-a} ${rst_if:+-r} 2022-09-24 16:00:28 I forgot to format the partition and hit "no space left on device", now starting again 2022-09-24 16:00:42 .... 2022-09-24 16:00:57 Without restart fortunately :) 2022-09-24 16:02:02 that means a mount point instead of a dev name is working for setup-alpine 2022-09-24 16:04:23 minimal: i use QEUM i don't know if it is due to my qemu parameters 2022-09-24 16:04:57 hey everybody 2022-09-24 16:04:59 but anyway, i run hwclock -s first everytime 2022-09-24 16:05:34 apkhelp: let me guess, you need help with apk 2022-09-24 16:05:34 just wanted to ask if someone could help me out a bit with how APK works 2022-09-24 16:05:44 how'd you know? :D 2022-09-24 16:05:53 apkhelp: what part? 2022-09-24 16:06:06 I'll try to summarize my goal and everything I know so far 2022-09-24 16:06:43 I want to build a custom image with preinstalled APK packages. I know there is specific documentation for that, but I want to learn the internals and why I can't do that manually by hand 2022-09-24 16:07:30 So my idea was to: 1) download the standard ISO, extract the contents of it using bsdtar to a FAT formatted USB drive, then boot that using UEFI (this works of course) 2022-09-24 16:08:03 2) preload the APKs I need using "apk fetch --recursive" 2022-09-24 16:08:28 minimal, uuidNuniq, same. Doesn't boot but doesn't loop either. Maybe that's where setup-bootable is useful? -_-' 2022-09-24 16:08:35 3) Place the fetched files in the cache directory of the prepared USB stick (this includes the APKINDEX tar files as well) 2022-09-24 16:09:27 4) I expect this to work as - the way I see it - is the same as running setup-interfaces, setup-apkrepos, setup-apkcache and then installing said packages 2022-09-24 16:09:28 axel50397: yea, try that, you don't need to reinstall 2022-09-24 16:09:54 But for some reason it does not work :( 2022-09-24 16:10:04 the RSA pubkeys are in place as well 2022-09-24 16:10:31 axel50397: oh no Copy the contents of SOURCE to DEST and make DEST bootable. 2022-09-24 16:10:31 I did a diff of the filesystem now and I can't seem to be able to find a difference 2022-09-24 16:10:59 between my prepared USB stick and going through the relevant setup-* scripts and using an apkovl 2022-09-24 16:11:32 so basically the question is how can I preload APK packages onto a prepared bootable USB stick manually? 2022-09-24 16:11:55 I know this exists: https://wiki.alpinelinux.org/wiki/How_to_make_a_custom_ISO_image_with_mkimage 2022-09-24 16:12:04 just want to do it manually and understand why I failed so far... 2022-09-24 16:12:53 apkhelp: use setup-alpine to install to your USB, the usb you used to boot up your computer 2022-09-24 16:13:23 I can go through the process that way. I mean if I setup LBU it works fine 2022-09-24 16:13:50 the thing is that I want to prepare images declaratively (which can be built in CI) rather than imperatively preparing them 2022-09-24 16:14:05 after that, use the USB to boot again, now your USB is like a read hard disk 2022-09-24 16:14:31 but then it also will have a lot of wear on it, would prefer to run from RAM 2022-09-24 16:14:37 you can install whatever on it ,and take away with it to boot any computer? 2022-09-24 16:15:56 The image I am preparing will be built using docker. Must not have manual steps involved either 2022-09-24 16:16:06 Just fetching the right files and putting them in the proper directories 2022-09-24 16:16:59 I thought I'd need: 1) the relevant APKINDEX from the repository the package is located in, 2) the .apk file(s) 3) the RSA pubkeys to verify said .apk files 2022-09-24 16:17:13 but no luck that way... what else might be required? 2022-09-24 17:22:00 uuidNuniq, minimal, so far I was only able to produce a bootable system with BOOT_SIZE=30 (20 isn't enough) setup-disk -k virt -m sys -s 0 /dev/sda, letting the script handle the partitioning with a separate /boot partition. 2022-09-24 17:23:00 I'm still tinkering 2022-09-24 17:41:37 The final .raw file is 45M according to du. Installed and fstrim'med 2022-09-24 17:44:47 still seeing that dbus symbol not found issue from yesterday. one user mentioned checking the apk policy, but everything's all edge. i just accidentally had a release repo enabled, the issue is not stemming from apk downgrading packages to release or something 2022-09-24 17:45:26 then not to mention i wiped apk cache in /etc and /var, and re-added everything after making sure the release repo was actually disabled 2022-09-24 17:45:30 so i am just not sure what's going on 2022-09-24 18:21:42 could you post `apk policy dbus-x11 dbus-libs dbus` and `ldd /usr/bin/dbus-launch` 2022-09-24 20:16:43 I am getting errors when trying to access usb drives: 2022-09-24 20:16:45 fdisk: cannot open /dev/sdb: Invalid argument 2022-09-24 20:17:29 or 2022-09-24 20:17:31 dd: error writing '/dev/sdb': No space left on device 2022-09-24 20:17:43 any thoughts? 2022-09-24 20:40:43 dmesg 2022-09-24 20:42:12 when plugging in the device? 2022-09-24 20:42:36 https://tpaste.us/yvxx 2022-09-24 20:48:31 oh missed something: https://tpaste.us/5k8N 2022-09-24 20:50:42 just send the entire buffer 2022-09-24 20:52:05 https://tpaste.us/QNwl 2022-09-25 00:47:58 Out of curiosity, does anyone know why LBU doesnt backup /etc/init.d by default? Whats the reason behind it? 2022-09-25 01:33:38 can't scroll back through weechat log, not sure why it's not logging 2022-09-25 01:33:57 someone wanted me to post the /apk policy/ and ldd of what? 2022-09-25 01:40:57 dbus policy: 1.15.0-r0: lib/apk/db/installed http://dl-cdn.alpinelinux.org/alpine/edge/main dbus-libs policy: 1.15.0-r0: lib/apk/db/installed http://dl-cdn.alpinelinux.org/alpine/edge/main dbus-glib policy: 0.112-r1: lib/apk/db/installed http://dl-cdn.alpinelinux.org/alpine/edge/main 2022-09-25 01:41:16 /lib/ld-musl-x86_64.so.1 (0x7f806c072000) libdbus-1.so.3 => /usr/lib/libdbus-1.so.3 (0x7f806c015000) libX11.so.6 => /usr/lib/libX11.so.6 (0x7f806bef2000) libc.musl-x86_64.so.1 => /lib/ld-musl-x86_64.so.1 (0x7f806c072000) libxcb.so.1 => /usr/lib/libxcb.so.1 (0x7f806becb000) libXau.so.6 => /usr/lib/libXau.so.6 (0x7f806bec6000) libXdmcp.so.6 => /usr/lib/libXdmcp.so.6 2022-09-25 01:41:18 (0x7f806bebe000) libbsd.so.0 => /usr/lib/libbsd.so.0 (0x7f806beaa000) libmd.so.0 => /usr/lib/libmd.so.0 (0x7f806be9e000)Error relocating /usr/bin/dbus-launch: _dbus_lookup_user_bus: symbol not found 2022-09-25 01:41:23 i think that was it 2022-09-25 01:41:52 having some issues with weechat tho, so it's a sloppy paste 2022-09-25 03:06:50 use a pastebin 2022-09-25 03:10:26 and policy of dbus-x11 2022-09-25 03:10:56 dbus-launch comes from that, the specific symbol doesn't even exist anymore 2022-09-25 03:43:21 http://ix.io/4blk 2022-09-25 03:45:53 much confusion 2022-09-25 03:46:15 how come? 2022-09-25 03:46:52 it looks like everythings should be fine, right? 2022-09-25 03:47:18 ye 2022-09-25 03:48:06 can you sha512sum dbus-launch 2022-09-25 03:48:30 86b8742ca20ebaa87abe252153a0b31cd50d3e562d6abc64a24ac99217aa31815f9e9307fd0fee92e5f09d4bc137f2fb3f4f0c17680b0fca97854231abc78f56 2022-09-25 03:48:37 not the correct one 2022-09-25 03:48:38 strange 2022-09-25 03:49:20 i wiped /etc and /var apk cache's, and reinstalled everything the other night 2022-09-25 03:49:32 oh 2022-09-25 03:49:38 dbus-x11 policy doesn't say installed 2022-09-25 03:49:54 it's some random other dbus-launch 2022-09-25 03:50:19 so i bet apk info -W /usr/bin/dbus-launch will give you nothing 2022-09-25 03:50:33 either way if you apk add it it should give the correct one 2022-09-25 03:51:46 507cf3895f2ea7dc3e25d19fc3167ed5d97387e8da73577f61ba5d7598a1f4ad03b9123f7b25051db5a2adc3d427795658fc358af44ad2ffa2e7c2c87b153b73 2022-09-25 03:51:56 yep 2022-09-25 03:52:43 alright, let me just start i3 with it, brb 2022-09-25 03:54:16 w00t w00t 2022-09-25 03:54:27 it works 2022-09-25 03:55:34 thank you :) 2022-09-25 03:56:14 weechat is complaining about a symbol missing perl plugin issue, so i wonder if it's a similar issue. but i will look at that another time 2022-09-25 03:56:53 i'm assuming something weird just happened with accidentally leaving release main enabled 2022-09-25 03:57:52 it's weird cuz i had it enabled for quite a while without issue 2022-09-25 04:00:07 alright, w.e., happy weekend 2022-09-25 04:00:08 o/ 2022-09-25 04:17:08 have fun :) 2022-09-25 10:27:55 Hi. Morning =) 2022-09-25 10:28:23 I have just installed Alpine with Gnome. I wonder if there is an extension that can bring back the old system tray? 2022-09-25 10:36:38 This is a plain install so only alpine and gnome is installed with nothing else so far. I am getting some segfaults in various gnome apps: 2022-09-25 10:36:40 [ 1934.940845] gnome-calendar[7416]: segfault at 18 ip 00007f79c047aa89 sp 00007ffc62957b98 error 4 in libglib-2.0.so.0.7200.1[7f79c0431000+66000] 2022-09-25 10:36:56 WebKitWebProces[4518]: segfault at 14a8 ip 00007f610455e343 sp 00007fff14ca3110 error 4 in libGLESv2.so.2.0.0[7f610455e000+5000] 2022-09-25 10:37:08 glib was updated, so that might be causing issues 2022-09-25 10:37:45 uname -a : Linux alpine01 5.15.69-0-virt #1-Alpine 2022-09-25 10:37:56 Alpine 3.16.2 2022-09-25 10:38:14 Maybe its a wayland thing too? 2022-09-25 10:38:22 oh, not edge 2022-09-25 10:38:29 edge? 2022-09-25 10:38:59 The alpine branch where work on for the next release 2022-09-25 10:39:02 http://dl-cdn.alpinelinux.org/alpine/v3.16/main 2022-09-25 10:39:02 repos: 2022-09-25 10:39:02 http://dl-cdn.alpinelinux.org/alpine/v3.16/community 2022-09-25 10:39:27 sorry. I know what edge is, but I am not using edge 2022-09-25 10:40:21 ok 2022-09-25 10:40:52 glib-2.72.1-r0 x86_64 {glib} (LGPL-2.1-or-later) [installed] 2022-09-25 10:41:20 can you pastebin a backtrace with webkit2gtk-dbg mesa-dbg glib-dbg etc 2022-09-25 10:42:23 how do I do a backtrace? 2022-09-25 10:42:29 run it with gdb 2022-09-25 10:42:34 apk add gdb 2022-09-25 10:42:36 run what ? 2022-09-25 10:42:40 sorry :D 2022-09-25 10:42:53 the apps that give those errors 2022-09-25 10:43:00 gnome-callendar 2022-09-25 10:43:49 Ah. I can try. I don't actually run those myself. The errors show in dmesg when I login on the desktop 2022-09-25 10:48:33 something like this? https://dpaste.com/GGCD7FHGL 2022-09-25 10:50:19 type `bt full` 2022-09-25 10:50:25 ah, that error 2022-09-25 10:50:30 i ran into that earlier on edge.. 2022-09-25 10:50:34 don't think it's library related 2022-09-25 10:50:43 strange 2022-09-25 10:51:16 if i had to guess it's a missing dep or env var, since it's identical for me on edge running it outside of gnome 2022-09-25 10:53:26 https://dpaste.com/6ZSVF7Z6Q 2022-09-25 10:53:48 Yea, just what is missing.. 2022-09-25 10:56:27 " g_time_zone_get_identifier: assertion 'tz != NULL' failed" Seems to indicate a timezone issue? So I did setup-timezone and restarted. Lets see if it helps? 2022-09-25 10:56:43 worth trying 2022-09-25 10:58:05 so far so good 2022-09-25 10:58:25 I don't remember if I did setup-timezone during the install 2022-09-25 10:58:33 /etc/localtime is different now 2022-09-25 10:59:46 It is now "/etc/localtime -> /etc/zoneinfo/Europe/Stockholm" and before "/mnt/rootvol/snapshots/root/etc/localtime -> /etc/zoneinfo/UTC". So it should probably have worked anyway? 2022-09-25 10:59:59 mine is correct but it still fails 2022-09-25 11:00:14 ahh /etc/zoneinfo/UTC is missing 2022-09-25 11:00:18 hm ok. 2022-09-25 11:00:28 also you said various- are there others that fail the same 2022-09-25 11:00:42 yes. gnome-calendar and webkit-gtk 2022-09-25 11:01:01 but I have tried both the browser epiphany and gnome-calendar 2022-09-25 11:01:24 you mean when you start epiphany it segfaults or 2022-09-25 11:01:24 oh. no. I take it back. 2022-09-25 11:01:27 webkit-gtk is a library 2022-09-25 11:01:36 gnome-calendar segfaults 2022-09-25 11:01:48 i installed gnome-calendar and now it works 2022-09-25 11:01:49 mfw 2022-09-25 11:02:02 yea. I don't know what app uses webkit-gtk. I thought it was epihany, but maybe something else? 2022-09-25 11:02:27 gnome-calendar for one 2022-09-25 11:02:40 quite a few gnome apps use it 2022-09-25 11:02:46 Ah ok. 2022-09-25 11:03:01 hm 2022-09-25 11:03:24 what's weird is if i remove tzdata then i get the same crash 2022-09-25 11:03:27 but i know it was always there 2022-09-25 11:04:15 https://paste.tnonline.net/files/eTf33m3wRhMe_alpine_segfault_1.png 2022-09-25 11:05:17 That's when you have an invalid timezone? 2022-09-25 11:06:01 No. now when it is correct to Europe/Stockholm 2022-09-25 11:07:11 if the /etc/localtime symlink is invalid it crashes 2022-09-25 11:07:16 glmark2 works nice though :) https://paste.tnonline.net/files/dldOJ5JddXyV_alpine-virt-glmark2.png 2022-09-25 11:07:19 but that shouldn't be possible..? it depends on tzdata 2022-09-25 11:08:13 -rw-r--r-- 1 root root 2298 Sep 25 12:50 /etc/zoneinfo/Europe/Stockholm 2022-09-25 11:08:55 can't tell if the contents are correct though. How do I check? Looks like binary stuff to me 2022-09-25 11:09:40 "file /etc/zoneinfo/Europe/Stockholm" gives "/etc/zoneinfo/Europe/Stockholm: timezone data (fat), version 2, 9 gmt time flags, 9 std time flags, no leap seconds, 143 transition times, 9 local time types, 18 abbreviation chars" 2022-09-25 11:09:53 that looks completely wrong? it should be /etc/localtime -> /usr/share/zoneinfo/Europe/Stockholm 2022-09-25 11:10:02 where did you even get them into /etc from 2022-09-25 11:10:11 setup-timezone 2022-09-25 11:11:38 ..why does it copy a time zone at all??? 2022-09-25 11:11:43 anyway, if i use it still works 2022-09-25 11:11:54 as long as the /etc/localtime symlink points to something 2022-09-25 11:14:22 https://paste.tnonline.net/files/QQ6dacoeMvoj_alpine-timezone.png 2022-09-25 11:14:29 Seems it recreates the symlink to etc 2022-09-25 11:16:12 gnome-calendar still fails the same way if I make a symlink /etc/localtime to /usr/share/zoneinfo/Europe/Stockholm 2022-09-25 11:28:24 I guess maybe we should wait for a new glib version? 2022-09-25 11:28:29 Or downgrade? 2022-09-25 11:29:32 Or are you saying the version in edge is working? 2022-09-25 11:30:47 if you link it to a better city like Europe/Oslo does it change at all 2022-09-25 11:31:11 ll 2022-09-25 11:31:13 lol 2022-09-25 12:03:22 psykose: better, hah. I'll try later. Have gone out for a bit 2022-09-25 12:07:41 take care!@ 2022-09-25 12:28:29 hello 2022-09-25 12:28:33 anone here 2022-09-25 12:28:39 anyone* 2022-09-25 12:28:47 uhh 2022-09-25 12:28:56 am i speaking into a void? 2022-09-25 12:30:02 it's not the emergency line, nobody responds in 20 seconds 2022-09-25 12:30:06 goodsamaritan: hi 2022-09-25 12:30:22 ohh 2022-09-25 12:30:40 hey @psykose sorry i'm new to this irc thing 2022-09-25 12:30:47 hello @Newbyte 2022-09-25 13:49:38 hi, what happened to qt on wayland? 2022-09-25 13:51:53 I got this today after some updates when launching telegram-desktop: `qt.qpa.plugin: Could not find the Qt platform plugin "wayland" in "". This application failed to start because no Qt platform plugin could be initialized. Reinstalling the application may fix this problem.` 2022-09-25 14:02:54 I read the exact same question some days ago but forgot the exact solution 2022-09-25 14:03:04 it was something along setting an environment variable 2022-09-25 14:03:45 export ...QPA...=wayland or something 2022-09-25 14:04:09 you need qt6-qtwayland installed 2022-09-25 14:06:26 don't need QT_QPA_PLATFORM for qt6 unless the compositor doesn't set wayland_display 2022-09-25 14:08:27 in usual qt fashion it used to explicitly link qt6-qtwayland and now it doesn't, nothing special 2022-09-25 18:34:33 psykose: unfortunately, Oslo doesn't have magic tzdata, so gnome-calendar segfaults just the same 2022-09-25 18:41:17 I'd guess we have to debug the program to find out where / how tz is normally set. 2022-09-25 18:41:29 Does it help to set the TZ env variable? 2022-09-25 20:15:03 ikke: unfortunately our didn't help. Also tried UTC and others too 2022-09-25 21:21:57 I suppose it's this assertion that's failing? https://github.com/GNOME/glib/blob/main/glib/gdatetime.c#L2658 2022-09-26 05:32:13 ACTION https://t.ly/BTD6 2022-09-26 06:03:40 Ikke yes. Looks like it. 2022-09-26 07:01:46 hello. I want to setup a joystick in alpine/linux, but... bad luck ATM 2022-09-26 07:06:07 hello. I want to setup a joystick in alpine/linux, but... bad luck ATM. I'm searching the web for help, but, i'm not getting the desired results 2022-09-26 07:07:32 could it be that someone here will give me clues? 2022-09-26 07:07:45 usuario: a refernece : https://wiki.archlinux.org/title/Gamepad 2022-09-26 07:08:12 but when you do , use alpine commands... 2022-09-26 07:08:45 i think the wiki may help you understand how it works 2022-09-26 07:09:06 I took a look at that page, but ... 2022-09-26 07:10:43 arch works with systemd and standard c libraries, isn't? 2022-09-26 07:11:20 What joystick? 2022-09-26 07:27:37 usuario: yes, arch linux is systemd + glibc 2022-09-26 07:28:07 alpine is so small and so efficient, love it so much. 2022-09-26 07:29:08 the best distro so far after my years experience with gentoo debian arch centos. 2022-09-26 07:44:54 ikke it's an Activbb X6-34U USB 2022-09-26 07:48:23 uuidNuniq: yes, me too. I started with debian in the 2002, twenty years later still use debian, but tested a lot of other distros, and yes, alpine, alpine... 2022-09-26 07:48:38 :) 2022-09-26 07:50:44 what do you see in dmesg when you connect it 2022-09-26 07:50:55 I'm an average user. no programming knowledge, a lot of things I don't know 2022-09-26 07:53:21 usuario: i only do shell script. that is all i can understand. 2022-09-26 08:04:50 psycose: I hope this help 2022-09-26 08:04:53 https://hastebin.com/raw/nudiwiqoha 2022-09-26 08:08:00 and... what is nv50cal ? 2022-09-26 08:14:09 looks like it's detected, if you cat that somewhere in the /dev tree (look for those id's, input30, etc) and hit some buttons does it print random garbage to the console 2022-09-26 08:16:06 usuario: what do you mean by that you're not "getting the desired results"? 2022-09-26 08:40:05 I am wanting to use 1 package thats in edge/testing, I am running 3.16. Is the best way to do this just downloading the apk file directly and installing locally? 2022-09-26 08:40:16 The package has no deps (fdm) 2022-09-26 08:41:20 is it possible to install a 64-bit (x86_64) kernel on an otherwise 32-bit system? 2022-09-26 08:41:42 I guess I can't have kernels for both architectures because of /lib/modules though... 2022-09-26 08:49:55 jay: use @ 2022-09-26 08:51:00 jay: edit /etc/apk/repositories and add a @edge in front of the edge mirror URL : "@edge https://....../edge/main" 2022-09-26 08:51:02 uuidNuniq: Nice one thanks, I did read about repo pinning but was unsure if that was the best way. Will give that a try 2022-09-26 08:51:17 and then apk add pkgname@edge 2022-09-26 08:51:35 uuidNuniq: Thanks v much 2022-09-26 08:52:08 do the same with testing : "@test https://..../testing/main" , apk add pkgname@test 2022-09-26 08:53:08 that is the best way. if one day you want the pkg go back to stable, just do this : apk add pkgname ; apk fix 2022-09-26 08:53:47 that is controlled by /etc/apk/world file. 2022-09-26 08:54:57 when you do "apk add something" with or without @tag, it will be seen in /etc/apk/world , and "apk add" will update the world file. so, in the end, just do : apk fix 2022-09-26 08:56:52 besides the @tag, you can also control the version of package : "apk add linux-lts=5.15.69-r0 ; apk fix " , you will be stay at this version for ever unless you update the world file like this : "apk add linux-lts" 2022-09-26 08:58:30 just love apk. 2022-09-26 08:58:43 uuidNuniq: fantastic, just gave it a try and it worked a charm! thanks for all the info, going to be a great help. 2022-09-26 09:01:34 jay: one thing i havn't tried yet is this : apk add linux-edge=5.19.10-r0@edge 2022-09-26 09:02:08 can we use the = > < ~= together with @? i did not test yet :P 2022-09-26 09:04:59 grawity: i think you can run 32bit programs in a 64bit system. but not be able to install a 64bit system to a pure 32bit hardware 2022-09-26 09:06:48 uuidNuniq: Just gave it a try, for science, this seems to work apk add fdm@test=2.1-r0 2022-09-26 09:07:23 uuidNuniq: format is name(@tag)([<>~=]version) 2022-09-26 09:09:27 cool 2022-09-26 09:09:48 that is so nice. 2022-09-26 09:15:18 apk fix : OK: 191 MiB in 200 packages 2022-09-26 09:15:36 the smallest full feature system i have ever used 2022-09-26 09:15:50 OK: 9547 MiB in 1304 packages 2022-09-26 09:15:51 close enough 2022-09-26 09:16:06 well, this is a server here. 2022-09-26 09:16:28 also the part i love a lot is that i can just install a package doc without installing the package 2022-09-26 09:17:32 mandoc seems do not have a database for "man -k" to search 2022-09-26 09:17:48 instead i installed man-db which works pretty well with `man -k` 2022-09-26 09:18:08 you probably wanted mandoc-apropos 2022-09-26 09:18:55 psykose: thanks! 2022-09-26 09:18:59 note that will run on any transaction at all 2022-09-26 09:19:05 so basically any adding/removing gets quite a bit slower 2022-09-26 09:19:12 whenever man stuff gets touched 2022-09-26 09:19:19 but it should keep a full db for man -k 2022-09-26 09:20:02 man-db is pretty fast on transactions 2022-09-26 09:20:17 i'll take a note on mandoc. 2022-09-26 09:20:45 yeah, it's more tuned to that usage 2022-09-26 09:22:57 it would be nice to have 'apk info -L' also list the not-installed-package content. 2022-09-26 09:29:50 hello. is there a way to list the contents of an apk file? `apk info -L foo.apk` doesn't do it. 2022-09-26 09:30:50 Terminus: in that case , foo.apk has to be installed. 2022-09-26 09:31:16 if it is installed, then just do `apk info -L foo` 2022-09-26 09:31:53 if it is not installed, you have to go https://pkgs.alpinelinux.org/contents 2022-09-26 09:32:58 oh, sorry, you mean a local .apk file 2022-09-26 09:33:27 yep. local apk file. 2022-09-26 09:33:28 did abuild -r and saw package files appear and i'm trying to inspect them. 2022-09-26 09:38:01 Terminus: guess you can unpack and pipe to '| less' 2022-09-26 09:40:29 i'll give that a try. 2022-09-26 09:42:10 uuidNuniq: lol i had no idea it was just a tarball. `tar tf foo.apk` works fine. thanks. =) 2022-09-26 09:42:48 Terminus: tar -t -f foo.apk | less 2022-09-26 09:43:23 it is a gzip compress file 2022-09-26 09:43:43 acct-6.6.4-r1.apk: gzip compressed data, max compression, from Unix, original size modulo 2^32 194560 2022-09-26 09:43:56 uuidNuniq: yup. saw the contents. 2022-09-26 09:44:20 funny that gunzip -l does not show content, has to use tar -t 2022-09-26 09:44:25 file only showed me gzip. nothing about being a tarball so i wasn't sure. 2022-09-26 09:44:45 uuidNuniq: i tried gunzip on it. just gave me a plain tarball. that's why gunzip alone won't work. 2022-09-26 09:48:08 Terminus: yeah, tar is the king. i guess 'tar -t -f ' can be used on any linux package 2022-09-26 09:49:08 uuidNuniq: not any i think. rpm needs cpio, not tar. 2022-09-26 09:49:31 oh right, initramfs also needs cpio 2022-09-26 10:01:37 unless it's the libarchive-based bsdtar, which will read cpio (and zip, rar, iso9660...) 2022-09-26 13:52:20 didn't expect to file an issue and see a fix released in 4 hours. it didn't even need to take that long. the delays were because it took me at least 30 minutes to notice a new message. O_o 2022-09-26 13:53:25 also, i thought learning how to use abuild would take a lot of effort. it didn't. heh. 2022-09-26 14:28:50 i've set up greetd to run agreety, which seems to work but only registers input one in every ~3 keystrokes, seemingly nondeterministically 2022-09-26 14:29:10 can anyone think of anything that might be conflicting re input, or is this likely an agreety bug? 2022-09-26 14:30:12 don't know much about input, but if i had to harbor a guess nondeterministic stealing of input wouldn't make sense- maybe it's some weird timeout on poll for input? 2022-09-26 14:30:46 if it works in regular desktop session (via libinput or whatever other one) and perhaps also in the console, then i would say it's an agreety bug 2022-09-26 14:38:46 Cool, I'll report it on the issue tracker there 2022-09-26 14:42:06 well, i'm not too sure of course :) but the default agreety is just on the tty almost the same as the regular login prompt.. so if that one is fine.. 2022-09-26 14:43:10 if you know how to read through code, it's all in one file https://git.sr.ht/~kennylevinsen/greetd/tree/master/item/agreety/src/main.rs 2022-09-26 14:44:19 considering there is nothing fancy there (it's just rust std::io::stdin reading), who knows 2022-09-26 20:11:24 there is a way to comment lbu commit? 2022-09-26 20:11:47 There is no log being kept or a history of all commits 2022-09-26 20:12:50 ok, so the only solution is to create and track a commebt file bymyself. as usual, tanks ikke🙂 2022-09-26 20:13:07 * ok, so the only solution is to create and track a comment file file bymyself. as usual, tanks ikke🙂 2022-09-26 20:48:10 i installed cronie as a cron daemon and want to set some standard commandline options. i saw /etc/init.d/cronie has the $CRON_OPTS env var. how do i set it? 2022-09-26 20:48:33 /etc/conf.d/cronie 2022-09-26 20:48:46 also i'm assuming the command will be called with `$command $command_args` with openrc? 2022-09-26 20:49:46 ikke: thanks! 2022-09-26 20:50:24 openrc uses start-stop-daemon or supervisord to actually start the service 2022-09-26 21:39:32 ikke: Logically the directory contained in /etc/apk/repositories is the boot partition? 2022-09-26 21:40:47 since the apks directory would like to be in the boot disk, right? 2022-09-27 00:51:17 nft ipsets not living up to needs, can i get a review on my piece of shot script now ment to work for openwrt 22.03? https://github.com/bldur/bones/blob/main/wrt-badhost.sh Or I mean, you can use this with its quirks, but the facilities should work better. 2022-09-27 00:58:28 I'm curious why you're asking in #alpine-linux and not #openwrt 2022-09-27 00:59:36 whole host of things that would be 10x easier if you didn't do this in shell, not to mention all the useless uses of cat 2022-09-27 01:11:27 dwfreed: fairly certain this channel cares more that nft ipsets converts 224.0.0.0/3 to 224.0.0.0-255.255.255.255 and the 567M addressees involved, than openwrt community as it involves a vastly different target community. 2022-09-27 01:12:35 ive poked it with a banana, and complained to #netfilter, but it is also tremendously annoying. 2022-09-27 01:14:56 zpiro: can you provide a small, reproducible example of this issue? 2022-09-27 01:17:54 https://pastebin.com/YQX9Ghpj 2022-09-27 01:18:30 and this is pulling the carpet under thiis POS script. https://github.com/bldur/bones/blob/main/wrt-badhost.sh 2022-09-27 01:19:56 And I was thinking, "off you go" not my lawn. 2022-09-27 01:22:35 aaaand what's the issue? 2022-09-27 01:22:51 224.0.0.0/3 literally is 224.0.0.0-255.255.255.255 2022-09-27 01:23:32 there's a difference in displaying the addresses, the /4 one is just being displayed as-is, the /3 one is being converted to this range syntax 2022-09-27 01:23:44 but it's behaving correctly 2022-09-27 01:28:59 it's actually 536.9M addresses 2022-09-27 01:29:48 but yes, the behaviour is correct, what do you expect /3 to mean? 2022-09-27 01:31:26 ptrc: and among 37k nft ipset entries, more than 2000 /cidr addresses, 224.0.0.0/3 just happens to be converted, as the only entry among close to 40k, to the ip range to that happens to include the broadcast address of the entire planet of the world. 2022-09-27 01:33:10 what? 2022-09-27 01:33:41 from a binary perspective, 224.0.0.0.-255.255.255.255 is the same as 224.0.0.0/3, but it is properly weird and untrustwworthy. 2022-09-27 01:34:04 no, that's exactly what that cidr notation means, same as the other 'more than 2000'. the others just don't have /3 2022-09-27 01:34:33 if you put /3 you get /3, what do you want it to do? tell you 'oops too big'? print a warning? reverse it to /29? 2022-09-27 01:35:29 what do you mean by "close to 40k"? this is not an issue, this is just one subnet being printed in a different way 2022-09-27 01:35:40 psykose: yes i know, but it doesnt explain why no other netmask regions including 0.0.0.0/0 to 255.255.255.255/32 is excluded from the majesty of xxx.xxx.xxx.xxx-xxx.xxx.xxx.xxx notation. 2022-09-27 01:36:05 oh, i see what you mean 2022-09-27 01:36:15 only this one specific format is printed as / instead of x-y 2022-09-27 01:36:23 and you are parsing this data with grep and it breaks your scripts 2022-09-27 01:36:30 ok, just don't use grep or shell scripts. fixed it 2022-09-27 01:36:32 backwards, only this one is printed as x-y, the rest are / 2022-09-27 01:36:37 right 2022-09-27 01:36:46 assuming it is, but this is not my area of expertise. 2022-09-27 01:36:50 and yeah, that's what I've said for a while 2022-09-27 01:37:05 (they were in #openwrt in the past) 2022-09-27 01:37:14 if it is merely nft being politely and a helpful user infterface, i am in error. 2022-09-27 01:39:02 you are 2022-09-27 01:39:09 use nft --json and whatever else at least 2022-09-27 01:39:20 then it i guess should be consistent 2022-09-27 01:39:26 unless they print the same thing into json :') 2022-09-27 01:40:05 I'm not born in 2010, at the very least i brought the liqour and cofee talk from #netfliter here. 2022-09-27 01:42:03 https://git.netfilter.org/nftables/commit/?id=7e6be917987c3ab0261bf543eb307cbb2679294f 2022-09-27 01:44:26 wow, just fixed huh 2022-09-27 01:44:42 probably from his ranting and raving in #netfilter 2022-09-27 01:45:13 I mean, the output is technically correct; it's certainly not ideal, though, so the patch is appropriate 2022-09-27 01:56:27 My work here then seem done, and no useful comment was goven outside of fact. 2022-09-27 02:27:05 https://ptrc.gay/cDxvtcLH, can we get a +b for this nice guy 2022-09-27 02:27:27 ( posted on libera/#alpine-linux ) 2022-09-27 02:29:34 who cares 2022-09-27 02:30:21 libera alpine-linux is basically a quarantine 2022-09-27 02:30:40 they can say anything there if they want, as long as it's not here :) 2022-09-27 02:30:44 lmao 2022-09-27 02:31:32 i mean, sure 2022-09-27 02:32:21 if they come back for some truly unproductive discussion, sure 2022-09-27 11:25:30 Hi, I'm trying to build a package directly with abuild but I get permission errors on abuild-apk 2022-09-27 11:25:49 ERROR: https://mirror.leaseweb.com/alpine/edge/main: Permission denied 2022-09-27 11:25:54 uhM 2022-09-27 11:26:20 obv I'm on abuild group and abuild-sudo seems properly suid 2022-09-27 11:31:35 4990 root 0:00 /sbin/apk update 2022-09-27 11:35:22 openat(3, "lib/apk/db/lock", O_RDWR|O_CREAT|O_LARGEFILE|O_CLOEXEC, 0600) = -1 EACCES (Permission denied) 2022-09-27 11:36:54 it's interesting, if I remove lib/apk/db/lock it appears again when trying 'abuild-apk update' 2022-09-27 14:00:14 uhM the problem seems here 2022-09-27 14:00:18 setuid(0) = -1 EPERM (Operation not permitted) 2022-09-27 14:03:57 Apparently it dropped privileges 2022-09-27 14:04:24 can you reproduce it? 2022-09-27 14:04:37 I suppose that someone would noticed it already 2022-09-27 14:07:03 I tried abuild-apk update and it worked successfully. abuild-apk upgrade works too 2022-09-27 14:07:19 and are you running edge? 2022-09-27 14:07:26 Yes 2022-09-27 14:08:22 it fails even I run it as root :D 2022-09-27 14:10:16 does it setuid to something else? 2022-09-27 14:16:45 uhM 2022-09-27 14:17:46 if I do, sudo -s, it works fine 2022-09-27 14:18:12 but I do, 'su', it fails the same way that as my user 2022-09-27 14:19:38 if I do, sudo -s, then 'su donoban', it works 2022-09-27 14:19:40 lol 2022-09-27 14:19:50 looks that the problem is something in my env 2022-09-27 14:23:44 Hi. I try to setup an USB joystick unsuccessfully. OS: Alpine Linux edge x86_64. Kernel: 5.15.69-0-lts. Seems like I need to load kernel modules that are not present (after lsmod). Modules I need to load: gamepad, analog and joydev (last one is present). Joystick is detected by the system, dmesg returns this : https://hastebin.com/raw/nudiwiqoha . 2022-09-27 14:25:09 I'm stuck, please, help :) 2022-09-27 14:25:32 I need to build those modules or what? 2022-09-27 14:26:17 what shows lsusb? 2022-09-27 14:28:50 if you are sure that you need that modules you probably have to rebuild with them enabled, you could open an issue asking for enable them 2022-09-27 14:30:22 lsusb show 8 lines with info like this 2022-09-27 14:30:24 Bus 002 Device 003: ID 046d:c52b 2022-09-27 14:38:13 just one more thing... Can I follow this instructions, or don't apply to kernel 3.15? 2022-09-27 14:38:34 https://unix.stackexchange.com/questions/606073/how-to-build-kernel-modules-in-alpine-3-12 2022-09-27 14:39:30 kernel... 2022-09-27 14:40:06 kernel... alpine version.. I mean 2022-09-27 14:47:07 I think that it should work 2022-09-27 14:51:13 @minimal yo, thanks 2022-09-27 14:58:36 @minimal I think it is working, but I am going to do one more test before I call it fixed. Thank you so much for you help! 2022-09-27 15:07:43 @minimal WOOOO! Yeah hotplugging works, reboot works, thank you so much. 2022-09-27 15:10:57 you're welcome 2022-09-27 15:14:40 Next step is to figure out how to pass those devices into my k8s cluster =) 2022-09-27 17:00:41 whats a good way to run 'btattach -B /dev/ttyAMA0 -P bcm -S 115200 -N' at boot? (rpi seem to need it to find the bluetooth) 2022-09-27 17:36:25 frag, I'm not on Alpine computer atm, but on Gentoo there is OpenRC service called "local" that runs scripts that are placed in /etc/local.d (https://wiki.gentoo.org/wiki//etc/local.d) 2022-09-27 17:44:29 townsfolkPravda: ah thanks, didnt know about this 2022-09-27 19:31:54 does anyone have a good reference for writing one-shot init scripts for openrc? 2022-09-27 19:32:29 something that doesn't need supervision (no daemon) and has an empty stop() 2022-09-27 19:33:24 just do what you need to do in start() like normal 2022-09-27 19:33:30 and make an empty stop 2022-09-27 20:08:39 start () { your commands ; } 2022-09-27 20:08:53 stop () { : ; } 2022-09-27 20:09:05 no need to set any of the other variables? 2022-09-27 20:09:17 maybe a name? 2022-09-27 20:09:24 not needed. 2022-09-27 20:09:26 sure, I meant to influence behavior 2022-09-27 20:09:32 ok, thanks 2022-09-27 20:09:52 you can also enable local.service and ... do stuff in /etc/local.d? 2022-09-27 20:11:33 ahills: less /etc/local.d/README 2022-09-27 20:11:35 this one is getting promoted from local.d to a proper service 2022-09-27 20:11:36 read. 2022-09-27 20:12:04 rc-update add local and then add stuff in /etc/local.d/ 2022-09-27 20:12:08 read the readme 2022-09-27 22:19:39 ei hi hello! after do this: apk add alpine-sdk linux-lts-dev 2022-09-27 22:19:55 i get his: depmod: WARNING: could not open modules.order at /lib/modules/5.15.70-0-lts: No such file or directory 2022-09-27 22:20:57 the same for modules.builtin and modules.builtin.modinfo 2022-09-27 22:22:15 should I worry? o_O 2022-09-28 08:56:46 This is odd. Anyone know why the kernel is listed twice in a list of upgradeable packages? https://paste.debian.net/1255265/ 2022-09-28 08:59:46 hm 2022-09-28 08:59:56 maybe it's an edgecase with origin packages and having linux-virt 2022-09-28 09:00:06 since linux-virt comes from linux-lts, and it resolves origin 2022-09-28 09:00:13 so having both installed gives you the origin twice 2022-09-28 09:00:18 do you have linux-lts+linux-virt? 2022-09-28 09:00:20 if not, then no idea 2022-09-28 09:00:25 it's a bug either way it seems 2022-09-28 09:10:25 usuario: apk list -vvu 2022-09-28 09:12:34 adhawkins: psykose is right : https://0x0.st/o48V.txt 2022-09-28 09:13:19 the stuff within { } is a package basename ... 2022-09-28 09:14:29 it would still be a bug for --upgradeable 2022-09-28 09:14:46 it should show the real packagename and not an origin or a resolved virtual or whatever 2022-09-28 09:15:20 usuario: you need to do : apk updat && apk upgrade -a 2022-09-28 09:15:24 if you run --upgradeable now, do you see update like the above 2022-09-28 09:15:25 then, do apk add. 2022-09-28 09:16:00 psykose: i don't have outdated packages atm 2022-09-28 09:16:01 well, you'd have to make sure they're actually upgradable 2022-09-28 09:16:03 xD 2022-09-28 09:16:04 yeah 2022-09-28 09:31:58 in order to run alpine package in a GNU/linux distro, i have to compile with -static 2022-09-28 09:35:19 uuidNuniq: psykose: Thanks. My apk skills are failing me. How do I tell if a package is installed? 2022-09-28 09:35:34 apk list -I | grep linux 2022-09-28 09:36:38 https://paste.debian.net/1255269/ No sign of linux-virt 2022-09-28 09:37:01 hm, ok 2022-09-28 09:37:06 not related to anything i said then 2022-09-28 09:37:23 no idea 2022-09-28 09:38:20 you can also skip grep and just do apk list -I 'linux' 2022-09-28 09:38:48 s//\´*/, s//\*/ 2022-09-28 09:38:52 s//\*/, s//\*/ 2022-09-28 09:38:54 I give up 2022-09-28 09:39:55 xD 2022-09-28 09:40:13 -I '*linux*' 2022-09-28 09:40:47 my \ key has gotten quite the workout 2022-09-28 09:40:57 i have to use it a thousand times a day just to stop getting italics in markdown 2022-09-28 09:42:20 i almost instinctively type \*thing\* for everything nowadays, it sucks 2022-09-28 10:11:19 adhawkins: can you apk list -vvu 2022-09-28 10:11:42 you may find out why there are 2 linux-lts 2022-09-28 10:12:52 https://paste.debian.net/1255273/ 2022-09-28 10:14:24 !bug 2022-09-28 10:15:37 that is a bug for sure 2022-09-28 10:16:06 adhawkins: try apk fix 2022-09-28 10:16:24 any report from apk fix? 2022-09-28 10:17:23 adhawkins: maybe you configured more than 2 mirrors? like a edge mirror with a 3.16 mirror? 2022-09-28 10:17:46 cat /etc/apk/repositories 2022-09-28 10:17:53 Nothing from fix 2022-09-28 10:17:56 hah 2022-09-28 10:18:09 Ah yes, I have both 3.15 and edge. Will fix that. Thanks uuidNuniq 2022-09-28 10:18:21 so not bug 2022-09-28 10:19:00 adhawkins: i use only one mirror : that is latest-stable 2022-09-28 10:19:00 yeah, in this case those are distinct packages 2022-09-28 10:19:10 there is a linux-lts-5.15.70-r0 in 3.15 2022-09-28 10:19:11 and in edge 2022-09-28 10:19:25 so there are two packages it could upgrade to.. and only because they have exactly the same version 2022-09-28 10:19:31 adhawkins: you can prefix a '@edge ' in front of the edge mirror 2022-09-28 10:19:35 if one of the repositories had a lower version, then only the bigger would be there 2022-09-28 10:19:51 but since they're the same, and named the same, it's just two identical packages it could upgrade to 2022-09-28 10:20:09 not sure how it determines which to really use- they're not necessarily actually identical 2022-09-28 10:20:28 apk policy would tell you which repo it really comes from (after upgrade) 2022-09-28 10:20:34 but i don't know how it picks it :) 2022-09-28 10:21:51 it depends on linux-firmware so it should pick it fine unless linux-firmware has same versions as well 2022-09-28 10:22:26 it also depends on how you upgrade , with -a or -i 2022-09-28 10:22:34 it also depends on how you upgrade , with -a or -l 2022-09-28 10:22:47 -i is interactive so I'm curious what that would change 2022-09-28 10:22:57 -a is latest 2022-09-28 10:23:09 -a or -l 2022-09-28 10:24:21 in his/her case, -a -l is not the same i guess? 2022-09-28 10:25:15 Ok, so only one listed now, thanks guys! 2022-09-28 10:25:56 so if i use latest-stable together with edge, i would add a prefix on the edge URL: @edge https://....edge/main 2022-09-28 10:26:23 generally yes, but also, not recommended to mix repos 2022-09-28 10:27:25 -a isn't latest 2022-09-28 10:27:33 i'm also not sure why any of those would change anything at all 2022-09-28 10:27:58 pretty sure it makes no difference 2022-09-28 10:28:02 adhawkins: man apk-world and man apk-repositorie, learn about the 'tag' 2022-09-28 10:28:16 man apk-repositories 2022-09-28 10:28:24 psykose: brain damage 2022-09-28 10:28:32 me 2022-09-28 10:28:37 >:( 2022-09-28 10:30:44 apk sometimes confusing 2022-09-28 10:31:02 to search, there is apk search name , apk list "*name*" 2022-09-28 10:31:22 to update : there is apk add -u ; apk upgrade -a or -i ; apk fix -u 2022-09-28 10:32:29 uuidNuniq: I'm not sure what the point of having both a 'stable' and edge repo is. Just removed the stable one. 2022-09-28 10:32:30 and the man says you don't need to run 'apk update' 2022-09-28 10:33:02 because generally, you don't always have to run apk update 2022-09-28 10:33:21 adhawkins: to have newer or lower version of the package... with the help of @tag feature 2022-09-28 10:33:25 apk will update indexes from time to time when they are detected as stale 2022-09-28 10:34:35 apk search looks into more than just name 2022-09-28 10:35:12 adhawkins: e.g. i want to use nginx stable but with everything else edge... 2022-09-28 10:35:21 something something apk search -x cmd:ss 2022-09-28 10:35:42 apk list -P cmd::ss 2022-09-28 10:35:43 uuidNuniq: why..... 2022-09-28 10:35:45 apk list -P cmd:ss 2022-09-28 10:36:50 apk list gives too much info package name 2022-09-28 10:37:09 yeah, but they have similar feature 2022-09-28 10:37:21 there is cmd so pc ? what else 2022-09-28 10:37:22 s/apk list gives too much info package name/apk list gives too much info, more than package name 2022-09-28 10:37:40 they do not have similar feature? 2022-09-28 10:37:42 apk list package name info more much than too 2022-09-28 10:37:56 I will scream 2022-09-28 10:38:01 record it 2022-09-28 10:38:12 upload to youtube 2022-09-28 10:38:17 directly to my inbox 2022-09-28 10:38:40 name it "Shouting in the datacenter" 2022-09-28 10:40:07 apk search -x so:*libc* ; apk list -P so:*libc* 2022-09-28 10:40:13 i like the latter 2022-09-28 10:41:07 the former you don't know what actual .so it provides ... you only have the package name 2022-09-28 10:41:46 the latter gives too much output which would have to be |cut to use in scripts 2022-09-28 10:42:27 yeah, depends on use case. 2022-09-28 10:44:58 wow there is a gcc6 package 2022-09-28 10:45:31 used by nothing except the openjdk bootstrap process 2022-09-28 10:49:30 (since on every release we build all the repos from scratch, as little as possible is copied to start the process) 2022-09-28 10:49:43 so since openjdk8 needs 7 to build, and so on in the other direction 2022-09-28 10:49:52 7 needs java-gcj-compat, and that only builds with gcc6 2022-09-28 10:50:21 in theory it could by ported to build with new gcc and gcc6 can be dropped, but that's a lot of work for what is solved by just one old compiler somewhere 2022-09-28 10:50:27 at some point alpine should probably switch to cross-compiling openjdk 2022-09-28 10:50:31 in comparison, making gcc6 buildable with new gcc is not terribly hard 2022-09-28 10:50:36 cross in what sense 2022-09-28 10:50:38 i replaced one ubuntu server and one debian server with alpine. feels so good not more waiting on updating and no more obsolete packages. 2022-09-28 10:50:50 compile openjdk with openjdk instead of compiling gcc with gcc to compile openjdk 2022-09-28 10:51:06 that's not possible without copying a jdk to start with 2022-09-28 10:51:09 which is a pain in the ass 2022-09-28 10:51:31 why harder than copying gcc 2022-09-28 10:51:55 because currently nothing is copied for it? 2022-09-28 10:52:34 the list of things that get copied every release is gcc (duh), the revdeps, a few more things from main/, go, rust, maybe 1-2 other things i forgot 2022-09-28 10:52:39 the point is to not expand that list 2022-09-28 10:52:51 for openjdk there is nothing explicitly copied, it just needs the gcc everything else does 2022-09-28 10:53:04 changing to that means copying even more things (and all the revdeps) just for openjdk 2022-09-28 10:53:46 the cost is having to maintain the bootstrap path instead, as gcc6/gcj-compat have to remain buildable with new gcc. but that's a lot less effort in my experience 2022-09-28 10:53:52 the end result is the same, you build openjdk 2022-09-28 10:55:03 in theory it would allow dropping gcc6, gcj, openjdk7, and then copying openjdk8 instead 2022-09-28 10:55:11 i guess nothing else, all the other deps can be built normally 2022-09-28 14:19:05 Hi, I'm trying to install Alpine Linux as desktop, but I keep receiving the error: "unable to select packages: linux-firmware-b43" 2022-09-28 14:20:11 I asked for help in the Telegram group and after telling me to try with the extended iso (didn't work) I was redirected here 2022-09-28 14:20:13 That's a bug / oversight in the installer 2022-09-28 14:21:05 So how am I supposed to deal with it? 2022-09-28 14:21:48 b43 2022-09-28 14:21:58 if you don't mind editing the files, apply https://gitlab.alpinelinux.org/alpine/alpine-conf/-/merge_requests/104/diffs to /sbin/setup-disk 2022-09-28 14:22:06 just adds a line 2022-09-28 14:22:47 You stok 2022-09-28 14:22:56 no you 2022-09-28 14:23:25 Still need to install it manually later if you want wireless 2022-09-28 14:24:12 for b43 you have to get the firmware yourself from the back of a truck, but yes 2022-09-28 14:24:36 b43-fwcutter can extract it for you 2022-09-28 14:24:42 haven't tried as i don't own any b43 wireless 2022-09-28 14:28:17 So it it enough to edit the script as in the link you sent me, or do I also need to smth else? 2022-09-28 14:30:08 just that and rerun whatever you were doing i guess 2022-09-28 14:30:23 but then to actually use b43 wifi.. bunch more steps that i don't know about :) 2022-09-28 14:47:18 psykose: isn't that for b43legacy? 2022-09-28 14:47:46 maybe 2022-09-28 14:49:55 a bunch of firmware files were included in linux-firmware and I've not tested any device since that required splitting it's firmware from the proprietary driver 2022-09-28 14:50:58 hmm, though most seem to be for brcmfmac 2022-09-28 14:54:47 then i guess it might be ok :) i would have guessed there's some overlap between firmware-exists-anyway and b43 kernel module 2022-09-28 14:56:00 hmm, no, I must have installed it from somewhere else because I don't have it anymore... :/ 2022-09-28 15:48:41 there is a 'modules=' in kernel cmdline under alpine, but i don't see docs about that parameter on kernel.org, what does this do? 2022-09-28 15:49:19 https://gitlab.alpinelinux.org/alpine/mkinitfs/-/blob/master/initramfs-init.in#L461 2022-09-28 15:54:10 i see. so i have modules=sd-mod,usb-storage,ext4 rootfstype=ext4 , both get passed to initramfs scripts. but rootfstype= is also used by kernel? so don't mind there are 2 ext4 ? 2022-09-28 16:05:05 apparently not :) 2022-09-28 16:06:37 what is the point of telling the vmlinuz that rootfstype is ext4 2022-09-28 16:07:33 mkinitramfs needs to know the type to be able to mount it 2022-09-28 16:07:44 and some other things 2022-09-28 16:10:19 to force the kernel to mount root as that type only 2022-09-28 16:12:45 funnily the only documented use i can find is to "mount ext3 as ext2" or something 2022-09-28 16:12:52 i don't think it's very useful for most other cases 2022-09-28 16:17:40 maybe related to busybox, mount without -t will fail 2022-09-28 16:18:33 only if the module has not been loaded yet 2022-09-28 16:20:45 that's outside the kernel 2022-09-28 16:20:55 you asked for vmlinuz :) 2022-09-28 16:21:17 as an actual kernel argument, outside all the initfs stuff 2022-09-28 16:53:52 psykose: thanks. why is there a System.map in /boot, can i get rid of it somehow 2022-09-28 16:58:15 shrug 2022-09-28 17:02:18 uuidNuniq: get rid of it somehow? there's a command called "rm" ;-) The file does have its uses though: https://en.wikipedia.org/wiki/System.map 2022-09-28 17:04:51 minimal: lol, rm is not going to work... you know it will come back 2022-09-28 17:05:23 does it work with kaslr? 2022-09-28 17:06:14 uuidNuniq: you only ask about getting rid of it, you didn't specify for how long ;-) 2022-09-28 17:08:40 i think a pickaxe to the storage of the computer also works for getting rid of it? 2022-09-28 17:09:23 might as well torch the whole thing 2022-09-28 17:09:25 but no, we'd have to make a -dbg subpackage with the stuff in it, etc 2022-09-28 17:12:09 maybe a hook or someting, kernel post install 2022-09-28 17:18:51 so if i set linux-lts=5.15.70-r0, will zfs-lts get upgraded to e.g. 5.15.71 when apk upgrade? 2022-09-28 17:19:56 zfs-lts-5.15.70-r0 depends on linux-lts=5.15.70-r0 , oh i see. 2022-09-28 19:56:06 system.map is pretty crap, it doesn't work without aslr and it's also not compressed on the iso 2022-09-29 12:56:14 hi! what would be the easiest way to have a Docker image with a custom Alpine package? I'd like to have `nsd` but with a single function removed 2022-09-29 12:59:17 eloy_: multi-stage Dockerfile 2022-09-29 15:45:53 is there a discord application in repo? 2022-09-29 15:46:02 (chatting application) 2022-09-29 15:46:36 Aren't you only supposed to use the official client? 2022-09-29 15:47:39 ikke: i have no idea, just been using the web-interface .. you think the official one works on alpine? 2022-09-29 15:47:51 not directly 2022-09-29 15:47:58 mhmh 2022-09-29 15:55:54 official discord client is non-free, so it won't go to alpine repos. And it is built against glibc. 2022-09-29 16:06:10 why cant everyone use irc! /k 2022-09-29 16:06:15 *j/k 2022-09-29 16:37:30 frag 6cord is an unoffcial client for discord but works okay in my experience! As ikke mentioned, using an unofficial client can get you banned so beware! 2022-09-29 16:40:01 thanksbud: hmmm! doesnt sound like a nice platform at all eh 2022-09-29 16:58:34 frag: yeah, totally agree on that one it's not very good! 2022-09-29 17:33:12 whats a situation when swap space is used? i only have 8GM ram, but never caught it running out... 2022-09-29 17:36:22 eight gigamegs 2022-09-29 17:39:19 frag: if you don't want swap, you can comment out or remove respective entry from /etc/fstab 2022-09-29 17:40:15 lopid: X) 2022-09-29 17:40:46 Ermine: yes, i dont have it, was considering swap-file if that is still a thing, but dunno if it will help or not... 2022-09-29 17:41:38 help with what? 2022-09-29 17:44:24 i guess speed up _something_ 2022-09-29 17:48:55 fetching data from disk is much slower than from RAM, so if you have enough memory for your tasks, you're better off without swap 2022-09-29 17:50:17 Ermine: nice! came to think of i have so little hdd space too, dunno if i want to use 8GB+ for swap file.. 2022-09-29 17:59:49 that's not how it works 2022-09-29 18:29:36 swap was an obsolete thing. it was created during the era RAM was small and expensive 2022-09-29 18:46:45 Is the aports mailing list out of action indefinitely or still in the works to resolve the issues with it? (couldn't see anything on the wiki/ML about it) 2022-09-29 18:55:42 To answer my own question, I found an issue about it on gitlab. infra/10766 2022-09-29 19:57:27 how can I download an alpine package with wget? 2022-09-29 19:58:12 wget https://dl-cdn.alpinelinux.org/alpine/v3.16/main/x86_64/aaudit-0.7.2-r2.apk 2022-09-29 19:58:15 for example 2022-09-29 19:58:43 Mirrors have dirlisting enabled, so you can grab the link from there 2022-09-29 20:04:17 ah ye sthanks 2022-09-29 20:11:22 ikke: like curl? 2022-09-29 20:11:49 what do you mean? 2022-09-29 20:11:54 curl -L https://dl-cdn.alpinelinux.org/alpine/v3.16/main/x86_64 2022-09-29 20:12:22 get the list of packages in a HTML form 2022-09-29 20:12:38 and parse it with tool 2022-09-29 22:03:00 I installed alpine and MATE desktop environment, and firefox-esr . But firefox do not launch ! I try to start firefox from terminal /usr/bin/firefox-esr And I see message Segmentation fault . How can I fixed it ? 2022-09-30 10:07:11 Does current firefox work? 2022-09-30 10:10:08 Does anyone know if it should be possible i Alpine 3.16 to boot with Grub into an ext4 ilesystem located in a zfs volume? I get error: no such and error: unknown filesystem 2022-09-30 10:24:08 current worked when i tried it last week 2022-09-30 11:51:46 you might be able to load the kernel but i doubt mkinitfs can mount it 2022-09-30 12:03:46 hello there! anyone around who is familiar with specifying apkovl=URL as a kernel parameter? can't get it working :( 2022-09-30 16:43:12 network-apkovl: what does that do? 2022-09-30 16:43:59 network-apkovl: https://wiki.alpinelinux.org/wiki/Alpine_local_backup 2022-09-30 22:57:03 lopid: did you use zfs as the filesystem as well or ext4 in a zfs volume? Also did you use grub-efi or grub legacy?