adhoc | afternoon all | 06:55 |
---|---|---|
adhoc | has anyone worked on devuan docker containers ? | 06:56 |
gour | morning. usually i have some boot-related problems with my netbook after fresh install andwonder which iso image you recommend for 'rescue' operation? | 08:57 |
lts | gour: any 64-bit glibc livecd should work. I usually use ubuntu because it has easy ZFS, which sometimes is needed | 09:00 |
gour | lts: so, devuan's desktop-live is also ok? | 09:01 |
lts | Should work? You just need to be able to mount the drives you need fix, and potentially chroot in them | 09:03 |
lts | Remember to bindmount /dev, /proc, /sys first if you chroot | 09:04 |
lts | The instructions are pretty much the same for most (all?) distros. Ubuntu has some ok ones at https://help.ubuntu.com/community/LiveCdRecovery | 09:05 |
gour | in recent times i'm experiencing problem with adding new uefi entry "no space on device" and that machine (lenovo e125) some problem bootin from bios/gpt combo, so usually some grub-related work is expected...but i hope that after install we'll stick with devuan on it...as well as on my main desktop machine :-) | 09:07 |
gour | tried to stay with gnome for my other family members, but it always "surprise" me when least expected :-( | 09:07 |
gour | *gnome/debian | 09:08 |
xrogaan | I think I'll need to setup a local deb repository for various software not included. | 09:20 |
xrogaan | I'm starting to get annoyed to have to manually download deb files. | 09:20 |
lunario | any ideas on how to start a program in a wm (dwm) right after the exec dwm command in .xinitrc? | 09:30 |
gnarface | you have to put them before the wm | 09:31 |
gnarface | the wm has to be last | 09:31 |
lunario | ahh, i will try that, thanks :) | 09:32 |
APic | Heya. After the latest full-upgrade of Devuan ceres, i get strange graphical Stripe-Artifacts in _some_ X11-Applications, for Example wine and VirtualBox. | 11:43 |
APic | What could cause these? My Graphics-Chip is: 00:02.0 VGA compatible controller: Intel Corporation CometLake-S GT2 [UHD Graphics 630] (rev 05) | 11:43 |
APic | Kernel is currently 5.10.0-8-amd64 | 11:44 |
* APic will try booting a slightly older one now | 11:44 | |
APic | They still appear with 5.10.0-7-amd64 which worked fine before | 11:46 |
APic | urxvt, Gimp and Chrome work fine without those Graphic-Glitches… | 11:46 |
gnarface | maybe just stuff that uses mesa? | 11:49 |
APic | It could be this: https://bugzilla.kernel.org/show_bug.cgi?id=27572 | 11:49 |
gnarface | that's from 10 years ago though | 11:50 |
gnarface | but i note that a commenter also speculates it could be related to mesa use... | 11:51 |
* APic will try whether a Screenshot will still show them | 11:51 | |
gnarface | that would be a good thing to test... also try to find any common dependencies of the programs that have issues | 11:52 |
APic | When i make a Screenshot with Gimp, the Artifacts are not visible there | 11:52 |
gnarface | could be in the compositing layer then | 11:53 |
gnarface | i assume the artifacts move with the windows so you know it's not display failure? | 11:54 |
APic | No, when i move my VirtualBox-Window, the Artifacts disappear | 11:55 |
APic | But when i switch to another OpenBox-Desktop and back, they reappear | 11:55 |
gnarface | yea, see if it goes away if you disable the compositor | 11:55 |
gnarface | or if your window manager has an option to switch to software compositing it might be worth it as a test | 11:55 |
APic | Also when i move _another_ Window, in this Case urxvt, the Stripes in the VirtualBox-Window are reducing | 11:56 |
gnarface | oh that's definitely weird | 11:56 |
gnarface | that actually sounds like an important clue about what is going wrong that the driver developer might understand | 11:56 |
APic | What is Compositing? GL-Stuff like Compiz? OpenBox does not do such Things anyhow | 11:57 |
gnarface | yes, exactly like compiz | 11:57 |
* APic can try another WM | 11:57 | |
gnarface | to be clear, they stay visible on screen still when you screenshot? | 11:58 |
gnarface | they just don't show up in the screenshot itself? | 11:58 |
gnarface | if screenshotting makes them dissappear on screen too, it could be power management related | 11:58 |
gnarface | otherwise it's very likely a problem in either mesa or the video driver | 11:59 |
gnarface | but i don't know a lot of types of rendering errors that can hide from screenshots without compositing involved | 11:59 |
gnarface | even if it was bad video ram i think it'd show up in a screenshot | 12:00 |
gnarface | and if it was display failure it wouldn't be able to move with the windows | 12:00 |
gnarface | a hamfisted attempt at a HDMI virus could be suspected i suppose but that's an extreme case... | 12:01 |
* APic just tried Plasma (KDE), which _has_ Compositing, and VirtualBox has _no_ Artifacts there | 12:01 | |
gnarface | fascinating | 12:01 |
gnarface | i think that means it's not mesa? not sure though | 12:01 |
APic | Screenshotting does not make them disappear, but on the finished Screenshot they are gone | 12:01 |
APic | Well, Chrome uses GL-Stuff too and works fine without Artifacts | 12:02 |
gnarface | yea, so that doesn't rule out the possibility of it being just some minor mesa function that is not commonly used | 12:03 |
gnarface | but what's weird to me in that case would be all the same programs working fine with compositing | 12:04 |
gnarface | since i'd assume they'd be using all the same function calls or a superset of them | 12:04 |
gnarface | but that could be naive | 12:04 |
gnarface | i wonder if it is something that compositing avoids that you could also avoid in the other window manager by changing some driver flag in the xorg.conf | 12:06 |
gnarface | like picking dri3 over dri2 or something | 12:06 |
gnarface | or even just picking different drivers maybe? could the WM change how auto-detect behaves there? alot of the intel integrated video devices are supported by two different drivers these days, one may work better than the other in certain cases | 12:07 |
* APic is just installing Gnome now, just to try out that one also | 12:08 | |
APic | But i will be afK for w While soon now | 12:08 |
APic | Will play around with it later | 12:08 |
gnarface | yea, worth checking, and i would compare the Xorg.0.log from startup of all these test runs, to see if anything changes related to the driver initialization | 12:08 |
gnarface | there's a smoking gun around there somewhere | 12:09 |
gnarface | i can't imagine it would be the VM guest extensions could it be? | 12:11 |
gnarface | you can recreate it without any VM running, right? or is the VM running always a common factor? | 12:11 |
APic | Gnome is also artifactless | 12:12 |
APic | It probably already uses Wayland | 12:12 |
APic | But i tried out Gnome Classic on Xorg too | 12:12 |
APic | Also Compositing like KDE, and also no Artifacts | 12:13 |
gnarface | definitely onto something here | 12:13 |
APic | Just for the Heck of it tried out FluxBox too: Also Artifacts like with OpenBox | 12:13 |
APic | Nah, it has nothing to do with the VM, because i get those Stripes with Wine-Applications too | 12:14 |
APic | Ok, afK | 12:14 |
APic | cul8r | 12:14 |
APic | Re | 13:43 |
gour | just installed devuan on my old netbook ceres/runit...everything looks and works very nicely...one problem - "switch user" option is greyed out, any hint? | 13:44 |
APic | Even Firefox has those Stripe-Artifacts | 13:46 |
* APic just tried out Pidgin, because it is mentioned in https://bugzilla.kernel.org/show_bug.cgi?id=27572 | 13:51 | |
APic | But it has no Stripes here | 13:51 |
APic | This was yesterday's Full-Upgrade after which the Problem appeared: https://nopaste.linux-dev.org/?1335656 | 13:58 |
APic | Any Hints which Package i shall try downgrading? | 13:58 |
APic | Probably everything with „intel“: intel-opencl-icd:amd64 libdrm-intel1:amd64 libdrm-intel1:i386 intel-media-va-driver:amd64 intel-media-va-driver:i386 | 13:59 |
APic | How do i downgrade them please? | 14:00 |
GyrosGeier | there is also beignet if you have an older GPU | 14:00 |
GyrosGeier | but that's CL only | 14:01 |
APic | The Graphics-Chip is quite new. It is on a MSI Z490-A PRO Motherboard which i bought quite recently | 14:02 |
* APic will try downgrading intel-opencl-icd:amd64 from 20.44.18297-1 to 21.32.20609-1 first | 14:06 | |
APic | Uh | 14:08 |
APic | The other Way round 😉 | 14:08 |
APic | apt install intel-opencl-icd=20.44.18297-1 → E: Version '20.44.18297-1' for 'intel-opencl-icd' was not found | 14:09 |
APic | Is it no longer in the Repo http://sledjhamr.org/devuan/merged or what? | 14:11 |
fsmithred | gour, what display manager are you using? slim, lightdm or other? | 14:14 |
gour | fsmithred: slim...i also noticed that icons are missing in 'Applications' menu? | 14:16 |
gour | fsmithred: if i want to enable autologin for our kid, maybe it's better to move to lightdm? | 14:16 |
fsmithred | the latter problem is a desktop-base issue | 14:17 |
fsmithred | autologin works in slim. switch-user does not (pretty sure about that) | 14:17 |
gour | fsmithred: i should reinstall it? | 14:17 |
fsmithred | no | 14:17 |
gour | i mean, desktop-base | 14:17 |
fsmithred | hang on and I'll find an xml file that works right. | 14:17 |
APic | Why is the old Package no longer there? Do i need to add a Debian-Repo to /etc/apt/sources.list maybe? | 14:19 |
fsmithred | don't add debian repos. What package is gone? | 14:19 |
APic | I am trying to downgrade intel-opencl-icd:amd64 from 21.32.20609-1 to 20.44.18297-1 | 14:20 |
fsmithred | gour, here's a replacement /etc/xdg/xfce4/xfconf/xfce-perchannel-xml/xsettings.xml-devuan-deepsea https://termbin.com/krcb | 14:22 |
fsmithred | you probably didn't get the right desktop-background, either | 14:23 |
APic | (Probably this Package has nothing to do with my Graphics Glitches, but i fail to downgrade other Packages too) | 14:23 |
fsmithred | that's in /usr/share/images/desktop-base/ if you don't find it with the xfce backgrounds | 14:24 |
fsmithred | APic, it's in ceres and chimaera main repo. 'apt download intel-opencl-icd' as user should do it. | 14:25 |
fsmithred | oh, you want the chimaera version | 14:25 |
APic | apt download gets me 21.32.20609-1 | 14:26 |
fsmithred | you're running ceres? | 14:26 |
APic | Yes | 14:27 |
fsmithred | add a line for chimaera, update, then download or install the package. | 14:27 |
fsmithred | if you want to download it with a web browser or wget, you'll have to pull it directly from debian. | 14:27 |
gour | fsmithred: thanks | 14:28 |
fsmithred | gour, it worked ok? | 14:28 |
APic | That works, thanks! | 14:28 |
gour | fsmithred: have to transfer from the desktop to netbook 1st :-) | 14:28 |
gour | fsmithred: it works!! | 14:34 |
fsmithred | :) | 14:34 |
sadoon_albader[m | Are there devuan builds for ppc similar to debian's sid netinstallers? | 14:34 |
fsmithred | gour, I have switch-user working in chimaera. I think it's just a matter of using lightdm. | 14:35 |
fsmithred | sadoon_albader[m, yeah, I think so. There should be mini.isos (business card isos) with the installer. I'll find a link. | 14:35 |
sadoon_albader[m | I'd really appreciate it because I can't find it anywhere | 14:36 |
APic | Ok, i downgraded intel-media-va-driver intel-opencl-icd libdrm-intel1 and rebooted, but the Glitch-Problem is still there | 14:37 |
APic | Any other Hints what Packages i could try downgrading? | 14:37 |
* APic will try everything with „mesa“ in it now | 14:38 | |
fsmithred | sadoon_albader[m, looks like we only have it for beowulf right now. But I think there's a screen that asks you which suite you want to install, so it might work for chimaera. Otherwise, install a minimal beowulf and upgrade it to chimaera. https://pkgmaster.devuan.org/devuan/dists/beowulf/main/installer-ppc64el/current/images/netboot/ | 14:39 |
sadoon_albader[m | Ah no that's ppc64le I already have that :) | 14:39 |
fsmithred | might need to select Expert install | 14:39 |
sadoon_albader[m | I was looking for ppc | 14:39 |
fsmithred | oh | 14:39 |
gour | fsmithred: ok, will switch to lightdm...not a big deal...btw, i'm getting xfce-background with a mouse, should deepsea provide a different one? | 14:39 |
fsmithred | I don't see any other plain ppc | 14:39 |
sadoon_albader[m | I guess there isn't one | 14:40 |
fsmithred | yeah gour, desktop-base is still broken. Set the bg manually. /usr/share/images/desktop-base... | 14:40 |
sadoon_albader[m | Thanks though | 14:40 |
fsmithred | yw. sorry I couldn't help. | 14:40 |
gour | fsmithred: ok thank you! | 14:41 |
fsmithred | gour, what version of deepsea-icon-theme did you get? | 14:41 |
gour | fsmithred: 1.1 | 14:44 |
fsmithred | good | 14:45 |
fsmithred | that fixes the lower panel icons. The defaults don't match the theme. | 14:46 |
APic | Ok, now i downgraded: libglx-mesa0=20.3.5-1 mesa-va-drivers=20.3.5-1 libgl1-mesa-dri=20.3.5-1 mesa-vulkan-drivers=20.3.5-1 libglapi-mesa=20.3.5-1 libegl-mesa0=20.3.5-1 mesa-vdpau-drivers=20.3.5-1 libgbm1=20.3.5-1 | 14:49 |
APic | Then rebooted, and the Problem is gone | 14:49 |
APic | (Before the Reboot it still persisted) | 14:50 |
APic | Now i just need to manually cherry-pick the Upgrades… | 14:50 |
fsmithred | APic, you could pin those packages to current versions or to chimaera so you don't get the ceres versions again. | 15:05 |
fsmithred | see man apt_preferences for details | 15:05 |
APic | fsmithred: Yes, thanks, but i first want to find out what Package exactly causes the Bug, so i can try a current GIT-Revision and see whether they already fixed it there, and if not maybe report the Problem | 15:09 |
fsmithred | good idea | 15:09 |
APic | Already upgraded mesa-vulkan-drivers mesa-vdpau-drivers mesa-va-drivers and they Problem not yet reappeared | 15:13 |
APic | s/they/the/ | 15:13 |
gour | fsmithred: thanks a lot for your assistance! everything is fine now...i'll continue evaluating devuan/ceres for some days before migrating my main desktop machine to it | 15:17 |
* APic cannot really separately install the Packages: libegl-mesa0 libgbm1 libgl1-mesa-dri libglapi-mesa libglx-mesa0 libgbm1 | 15:23 | |
APic | But downgrading all those from 21.2.1-1 to 20.3.5-1 definitely fixes the Problem | 15:23 |
APic | So i will try to build a GIT Mesa now and hope my Problem is already fixed there ☺ | 15:24 |
APic | Aw, gitlab.freedesktop.org is currently down | 15:39 |
systemdlete | cron did not run this morning as it usually does. I set up beowulf on this machine a few days ago and installed logwatch. cron ran logwatch for 2 days at 7:35 am, but not today. I don't recall changing anything. The 00logwatch file is still in /etc/cron.daily. I looked for a cron logfile, but it seems to just log to syslog. I looked through there but don't see anything running at 7:35 for cron | 18:43 |
systemdlete | (I thought cron had its own log also.) | 18:44 |
mason | systemdlete: If you have atime enabled and you're careful not to pollute your data, you can verify that your various crontabs have or haven't been looked at. You can also run cron under strace to watch what it's trying to do. | 18:49 |
systemdlete | "atime enabled" -- is that a cron option, or a systemwide option? | 18:51 |
systemdlete | btw, I did verify that cron is running (should have said that sooner) | 18:51 |
systemdlete | mason: Are you referring to the "at" command perchance? | 18:51 |
mason | systemdlete: atime (and relatime) are filesystem concepts. When a file is read, that's recorded. Some people turn it off for various reasons, but typically it's left turned on. | 18:56 |
systemdlete | Well, this is interesting. It looks like the syslog was renamed at 7:35, so it became syslog.1, and I can see that the cron job that normally launches at that time is spread across the two files, syslog.1 and then syslog. It looks like anacron ran and finished. Maybe this is related? | 18:56 |
systemdlete | I am thinking somehow logrotate running while cron is trying to run anacron messed things up somehow. | 18:57 |
mason | systemdlete: That suggests cron is running, because logrotate is driven by cron. | 18:57 |
systemdlete | cron says that it updated cron.daily at the time. | 18:57 |
mason | Anyway, if you have atime turned on, you can sometimes do useful forensics like seeing that a particular file has been read or not. | 18:58 |
systemdlete | (and no, I did not disable atime, just fyi) | 18:58 |
systemdlete | ls -la 00logwatch? | 18:58 |
mason | Of course, things like inotify can subvert this, where if something sets up an inotify watch it'll know not to bother looking for changes. | 18:58 |
mason | ls -lau | 18:58 |
systemdlete | thanks | 18:58 |
mason | the -u flag is what shows you atime - you see mtime otherwise | 18:58 |
mason | and of course you can sort and such | 18:59 |
mason | ls -lautr | 18:59 |
ham5urg | I just got help from #grub with some GRUB install onto some USB-SSD. This bug https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=991478 was the cause. I needed to use "grub-install --no-uefi-secure-boot". I just wanted to ask if the fixed package will land into Devuan? | 19:00 |
mason | brb, lunch while I can | 19:00 |
systemdlete | yes, the logwatch file in cron.daily has been updated for 7:35 this morning | 19:00 |
systemdlete | (00logwatch, technically) | 19:00 |
systemdlete | I'll look for logwatch trace files, if there are any) | 19:01 |
buZz | ham5urg: huh, uefi on arm64? | 19:01 |
buZz | ham5urg: either way, afaik buster-updates just gets transferred to beowulf-updates | 19:02 |
ham5urg | I hit me on amd64 | 19:04 |
ham5urg | grub-install --no-uefi-secure-boot --target=x86_64-efi --recheck --removable | 19:04 |
ham5urg | was the solution. | 19:05 |
ham5urg | It* | 19:05 |
buZz | i honestly nearly always avoid using uefi and secure boot | 19:11 |
ukine | there was a "hacker song" that went "i don't trust trusted computing, they don't just want it around to stop looting" | 19:12 |
buZz | they dont want what around? | 19:13 |
ukine | TPM | 19:13 |
buZz | they dont want TPM around so they can stop looting? | 19:13 |
ukine | IME, secure boot, etc | 19:13 |
buZz | i dont get it | 19:14 |
ukine | "they" meaning those behind those technologies | 19:14 |
ukine | user tracking, etc | 19:14 |
ukine | remember there was an initial windows 10 requirement of some of those | 19:14 |
ukine | apple had already been doing it | 19:15 |
ukine | but was trying to be responsible about it | 19:15 |
buZz | then why would 'they' not want TPM around? | 19:15 |
ukine | then, PRISM, quantumsquirrel, xkeyscore, etc | 19:15 |
ukine | they bought the biggest names, the alphabet agencies | 19:16 |
buZz | having TPM on your motherboard only makes it easier for them to identify you | 19:16 |
ukine | ^ | 19:16 |
buZz | so they DO want it around to stop looting | 19:16 |
ukine | i know, that was the basic point i was trying to make | 19:16 |
buZz | instead of DONT | 19:16 |
ukine | just, though | 19:16 |
ukine | it's all about balance | 19:17 |
ukine | like in the movie "snowden" | 19:17 |
ukine | nick cage: "military industrial happiness management" ;) | 19:17 |
ukine | that balance was getting a little out of wack, so, here we are on this channel | 19:18 |
onefang | This is a bit #devuan-offtopic. | 19:22 |
ukine | true i just thought about that, apologies | 19:23 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!