maskerino | but i really don't think that's it. | 00:00 |
---|---|---|
maskerino | and yes, i've been running these commands you've posted in apt | 00:00 |
maskerino | and i still get the warnings in apt | 00:00 |
debdog | nala? | 00:00 |
maskerino | it's a different apt front end. but it's entirely cosmetic | 00:01 |
maskerino | it just adds nice block characters | 00:01 |
rwp | I don't know then. If it were me I would browse /var/lib/apt/lists, delete the files there, and then apt-get update again to get fresh copies of them back. | 00:01 |
maskerino | oh i have a feeling that will work let me try | 00:01 |
maskerino | will that remove any extra ppas i've added? | 00:01 |
rwp | Devuan does not have PPAs. | 00:01 |
rwp | Which means that if you have added them that may be the source of the problem. | 00:02 |
rwp | However the files I mentioned above are local cached copies of files that are downloaded. | 00:02 |
maskerino | i can't use add-apt-repository? | 00:02 |
rwp | It's your system! You can do anything you want to do to it. But as the joke goes, if you break it then you get to keep both halves of it. :-) | 00:03 |
maskerino | i did sudo rm /var/lib/apt/lists/* and that didn't work either | 00:03 |
rwp | https://wiki.debian.org/DontBreakDebian#Don.27t_make_a_FrankenDebian | 00:03 |
debdog | are there files in /etc/apt/sources.list.d/ ? | 00:04 |
debdog | other than devuan-baseconf | 00:04 |
maskerino | i see files there and none of them are in devuan-baseconf lol | 00:04 |
debdog | sorry, my bad *other than devuan.list | 00:05 |
rwp | Any files in /etc/apt/preferences.d/* ?? | 00:05 |
maskerino | none in preferences.d, and there isn't devuan.list there either | 00:06 |
rwp | And no /etc/apt/preferences file either? | 00:06 |
maskerino | nope the file doesn't even exist | 00:06 |
rwp | Could you pastebin the output of "apt-cache policy"? | 00:07 |
maskerino | yeah one second | 00:08 |
maskerino | could the problem be that i added a debian source accidentally? | 00:09 |
rwp | Possibly. | 00:09 |
maskerino | some of the files in /etc/apt/sources.list.d had things labelled as debian | 00:10 |
rwp | What files are in /etc/apt/sources.list.d/* ?? That was a previous question asked above. | 00:11 |
maskerino | http://paste.debian.net/1290620/ | 00:11 |
maskerino | oh that's it | 00:11 |
maskerino | let me try something | 00:11 |
rwp | Yes. There are a lot of sources outside of Devuan listed there! | 00:12 |
maskerino | where lol? You're just saying ones that aren't from devuan.org? | 00:13 |
rwp | If it were me I would move all of those files to a saved location not there. Start with just the basic devuan source.list file. | 00:13 |
rwp | Then review each of those files you saved off one by one and verify that they are still useful. | 00:14 |
rwp | It looks like you have multiple mirrors of devuan listed for example. You only need one mirror listed. | 00:14 |
maskerino | so how am i supposed to safely add sources? | 00:16 |
rwp | Also also a VScode repository, which _should_ be compatible but I am always suspicious of 3rd party repositories. | 00:16 |
maskerino | yeah removing those files did the trick | 00:16 |
maskerino | but i use those softwares regularly | 00:16 |
rwp | How to safely add sources? I use my favorite editor. (Which for me is emacs. But vim is just the same too.) | 00:17 |
maskerino | use it to do what | 00:17 |
rwp | Yay! On solving the problem. After having only the devuan sources available and everything happy I would add only the VScode source in and see if things continue to be happy. | 00:17 |
maskerino | yup | 00:18 |
maskerino | vscodium doesn't break it | 00:18 |
rwp | I did not see any other 3rd party repositores in your apt-cache policy output. So at that point you would be done. | 00:18 |
maskerino | wait so i just added all the .list files back and it worked | 00:19 |
maskerino | wth | 00:19 |
maskerino | nothing changed? | 00:19 |
rwp | Why would you put all of those extra mirrors back? After having everything cleaned up? | 00:20 |
maskerino | i didn't clean anything up that's my point lol | 00:20 |
maskerino | i just move the files temporarily to ~, ran apt update, and moved them back and no more errors | 00:20 |
rwp | As to your specific question it is probably a race condition between the different mirror entries. Order matters. | 00:20 |
debdog | apt-magic | 00:20 |
rwp | If there is only the one then there ordering is easy. | 00:21 |
maskerino | gotchaa | 00:21 |
rwp | If there are a dozen then any of the dozen might happen first. | 00:21 |
maskerino | so how can i safetly add more next time | 00:21 |
maskerino | so this doesn't happen again | 00:21 |
rwp | Why do you want to add more? Let's start there. Because I don't understand it. | 00:21 |
maskerino | well i mean i had to add sources for sbt for example | 00:22 |
rwp | SBT? I am not familiar. But the specific ones I am commenting upon are the ones indicated by the apt-cache policy output which shows duplicate devuan repositories. | 00:23 |
maskerino | ok so I just have to avoid duplicates? | 00:24 |
maskerino | why can't apt figure out how to remove duplicates itself? | 00:24 |
debdog | prolly because its missing AI | 00:25 |
debdog | SCNR | 00:25 |
rwp | maskerino, These tools do exactly what you tell them to do. Because that allows people to do clever things. | 00:26 |
maskerino | i've been programming for a long time, i understand that | 00:27 |
maskerino | why isn't there an `apt --clean-duplicates update` option | 00:28 |
maskerino | or a more descriptive error | 00:29 |
rwp | I suppose in seriousness the reason no option exists is that the authors did not think editing the files with an editor and doing it manually was difficult. | 00:29 |
maskerino | can i combine all my sources into one file that way I can check for duplicates more easily? | 00:29 |
rwp | Yes. | 00:30 |
maskerino | cool | 00:30 |
rwp | For example I see mirrors.dotsrc.org/devuan/devuan and mirror.koddos.net/devuan/devuan and mirror.checkdomain.de/devuan/devuan and on and on... | 00:31 |
rwp | For example I see sledjhamr.org/devuan/devuan which is the primary mirror along with the others. | 00:31 |
maskerino | unrelated, but if I close this terminal with irssi (not ctrl-z or disown or whatever), will I be able to read chat history when I come back | 00:31 |
rwp | If sledjhamr is there then no other ones are needed. | 00:31 |
maskerino | you see those in my file? | 00:31 |
rwp | If irssi exits due to terminal close then it will not save history. | 00:32 |
maskerino | ok | 00:32 |
rwp | But the /topic command will show the topic and the topic will show the irc log site. | 00:32 |
maskerino | yeah i figured that's how you're intended to view message history | 00:33 |
rwp | To keep a persistent irssi running it's typical to use screen or tmux to leave it running in the background when one disconnects. | 00:33 |
rwp | Then one can connect back up to the tmux or screen and resume the previous terminal still running irssi. | 00:33 |
maskerino | yeah tmux i've used before | 00:34 |
maskerino | but what about messages that happen when you're offline, can you just not view those easily in terminal? | 00:34 |
rwp | If one is offline from IRC then there is nothing locally to collect those messages. (Other than the independent 3rd party logging site mentioned above.) | 00:34 |
maskerino | yeah but you can't have a script that downloads them when you start up? | 00:35 |
rwp | The IRC servers do not themselves log anything. | 00:35 |
rwp | I am not aware of anything that does that, but something could do that, and I probably would not know about it. | 00:35 |
maskerino | gotcha | 00:36 |
rwp | Also there are irc bouncer proxies which many people use that remain persistently connected and will dump out the last 50 lines (or other configured number) of lines upon reconnection. | 00:36 |
rwp | znc for example does 50 lines by default. I don't know about bip or quassel. | 00:36 |
rwp | All of this discussion about IRC would be excellent discussion in #devuan-offtopic where we chat about things not directly Devuan related. | 00:37 |
debdog | quassel is great (if you have a server which is connected 24/7) | 00:37 |
maskerino | np i have to go anyway. Thank you for your help | 00:37 |
onefang | rwp: sledjhamr isn't "the primary mirror", it was just the first external package mirror, but seems to be one of the better ones since lots of people keep recommending it. Also my mirror. | 02:25 |
onefang | pkgmaster is the primary package server that the others sync to, so not actually a "mirror". But we try to keep the load on pkgmaster low by suggesting people NOT use it directly. | 02:26 |
onefang | Which is why I moved pkgmaster to the end of the mirror list, so it was no longer at the top and getting used a lot. Then sledjhamr ended up at the top, coz it was second. B-) | 02:27 |
al1r4d | > If irssi exits due to terminal close then it will not save history. | 02:40 |
al1r4d | rwp, use xmpp <-> irc bridge, like me :) | 02:40 |
rwp | al1r4d, I solve the problem by never leaving. I just continue to exist always. But both on and away from the keyboard. | 04:14 |
rwp | onefang, The problem I was commenting upon though was that the poster had like a dozen of the mirrors all specified all at the same time in multiple sources.list.d/* files. | 04:15 |
rwp | Which also included sledjhamr, and if sledjhamr was there then there was no need or all of the rest. | 04:15 |
ted-ious | Can devuan handle root on zfs without breaking when zfs updates come out? | 05:21 |
ted-ious | Or do they have to be upgraded carefully by hand or something? | 05:21 |
rwp | ted-ious, AFAIK ZFS is not natively supported by Devuan but by OpenZFS using DKMS (Dynamic Kernel Module Support). | 06:42 |
rwp | And when I say Devuan there I mean Debian since that's the way it is on Debian. It's supported as a DKMS build. | 06:43 |
johndo100 | ZFS is amazing | 06:43 |
rwp | I agree! ZFS is awesome. | 06:44 |
ted-ious | I guess I'm asking how well that works since the packages are not native and the kernel support isn't normal for zfs. | 06:50 |
rwp | I am only running root on ZFS on FreeBSD. I can't say about it on Devuan. But you could try it and then you would know. | 06:55 |
johndo100 | I have ZFS on Devuan before, not on root but it was fine. | 07:02 |
systemdlete | I'm now getting messages from apt update saying that the label for the repo has been changed from "devuan security" to "devuan-security" (the dash is the important part here, I think) | 09:34 |
systemdlete | I've been tooling with apt for days (daedalus) and have not seen this until just now. | 09:34 |
rrq | happened the 25th Aug ... do "apt update" and answer yes | 09:42 |
systemdlete | I did. But I have been doing repeated installs for a week now, and never got this message before. If I had, it would have held up my scripts, but I have not had that. | 09:42 |
systemdlete | I've already fixed my scripts by adding "-y" to all calls to apt. | 09:43 |
rrq | it would have come with the first "apt-get update" | 09:43 |
systemdlete | I would have expected that to be the case, yes. | 09:43 |
systemdlete | Like I said, I've been doing a lot of installs, so I would think I'd seen this by now. | 09:44 |
rrq | unless you have a cacher in the way | 09:44 |
systemdlete | In a VM? | 09:44 |
rrq | I mean some repo cacher (apt-cacher or something) .. | 09:45 |
systemdlete | nah. When I say re-install, I mean, totally from scratch. | 09:45 |
systemdlete | empty root file system | 09:45 |
rrq | right.. only "update", which updates you meta files would be relevant for this | 09:45 |
rrq | upgrades and installs uses the meta files you have | 09:46 |
systemdlete | what does that mean? | 09:46 |
systemdlete | You mean the ones on the CDROM/ISO image I am using? | 09:46 |
systemdlete | I don't think those change! LOL | 09:46 |
systemdlete | (unless I misunderstand you) | 09:46 |
rrq | your installed system has packages and package meta files (or vice versa) | 09:47 |
systemdlete | "installed?" | 09:47 |
systemdlete | sorry | 09:47 |
rrq | all apt commands refers to the meta files you have on your system, except "apt-get update" which updates those meta files | 09:48 |
systemdlete | rrq: Since I am starting over and over again, each time, I should have a fresh copy of whatever comes with the devuan install ISO image. | 09:48 |
systemdlete | So every time I have installed (or re-installed), from scratch, I should have had a re-set | 09:49 |
systemdlete | and then apt update should have shown me those messages, at least for the installs I've done since the 26th when, as you say, those chages took place | 09:49 |
systemdlete | but they didn't. | 09:49 |
systemdlete | It is ONLY when I ran the install this most recent time that I saw the messages. | 09:49 |
systemdlete | No? | 09:50 |
rrq | yes. mmm right.. any time you would do manual apt update after installation it would have mentioned that change | 09:50 |
systemdlete | but... it didn't. | 09:51 |
systemdlete | Like I said, I have carefully taken the precaution of inserting a "-y" wherever I call apt so that it won't hang my scripts. | 09:51 |
systemdlete | but that is only since a few minutes ago, since this last install attempt. | 09:52 |
systemdlete | Look, I'm not really blaming anyone, or claiming there's an awful bug somewhere. But it is bizarre to me. | 09:53 |
systemdlete | And I admit I don't fully undestand all of apt and the debian package system and the repos (though I am slowly catching on) | 09:54 |
rrq | yes that sounds odd. whay's the sources.list line concerned? is it a particular repo that is late updating? | 09:54 |
systemdlete | devuan-security | 09:55 |
systemdlete | not sure which repo | 09:55 |
systemdlete | I saw the message a couple times I think | 09:55 |
rrq | the host part of the line | 09:55 |
systemdlete | http://deb.devuan.org/merged that part? | 09:56 |
systemdlete | should I be using pkgmaster instead? | 09:56 |
rrq | well... some people prefer pkgmaster for devuan-security, yes | 09:57 |
onefang | But I try to discourage everyone using pkgmaster, so the mirrors don't get bogged down. | 09:58 |
rrq | just to avoid possibilityof such updating delays, if that's what it was | 09:58 |
systemdlete | ahhhhh | 09:58 |
rrq | only for daedalus-security | 09:58 |
systemdlete | that WOULD definitely explain it, IFF that is really what happened. | 09:58 |
systemdlete | right, and that's what the message was from | 09:59 |
systemdlete | security | 09:59 |
systemdlete | so pkgmaster gets security updates before the deb ones? | 09:59 |
onefang | I'm also trying to get the package mirrors updating every 30 minutes. Some are still 1 to 3 hours, but I'm making progress. | 09:59 |
rrq | pkgmaster is the master repo that all repos mirror | 09:59 |
systemdlete | ah... so maybe I just ran into a laggard mirror then? | 10:00 |
systemdlete | and that's fine. | 10:00 |
onefang | Possibly. | 10:00 |
systemdlete | In fact, I am smiling now. | 10:00 |
systemdlete | I am RELIEVED. At least there is some logical explanation (at least likely, anyway) | 10:00 |
systemdlete | I didn't know what the heck was going on. | 10:00 |
systemdlete | getting into the guts of installation is interesting, but a bit scary at times. I always wonder if I overlooked something important, probably something I did not know anything about. | 10:01 |
systemdlete | So I really appreciate rrq and onefang taking a moment to clear it up. | 10:02 |
rrq | yeah the correction was something that was discovered when daedalus became stable, or maybe a little bit later | 10:02 |
systemdlete | Well, that's OK. It sounds like the mirrors are just trailing behind a bit. I might have hit one just now, in this most recent install attmempt, that was seriously behind (5 days?) | 10:03 |
rrq | your ISO is a daedalus live I suppose? | 10:03 |
systemdlete | it's the minimal install I've been using. | 10:03 |
onefang | I keep an eye on these things with my apt-panopticon. Any package mirror that isn't updating every 30 minutes I temporarily remove from the deb.devuan.org DNS-RR. | 10:04 |
* systemdlete looks up into the sky to see where onefang is pointing | 10:04 | |
onefang | lol | 10:04 |
systemdlete | that must be the management for the repo servers or somthing like that? | 10:04 |
onefang | https://sledjhamr.org/apt-panopticon/results/Report-web.html http://veritas.devuan.org/apt-panopticon/results/Report-web.html https://ap.in.devuan.org/apt-panopticon/results/Report-web.html https://mishka.snork.ca/apt-panopticon/results/Report-web.html | 10:05 |
* systemdlete adds that link to his "reading list" for someday when they have copious time on their hands... | 10:05 | |
onefang | Stuff I wrote in Lua to track the health of the Devuan package mirrors. | 10:05 |
onefang | Well four links, I track the health from various places around the world. | 10:06 |
systemdlete | The issue here, though, is... 5 days? | 10:06 |
systemdlete | is that even possible? | 10:06 |
onefang | Yes sounds odd. Shouldn't be, due to what I said above. | 10:07 |
systemdlete | I mean, did I get a server that is located in a DC in one of the US's many war zones? | 10:07 |
systemdlete | Or somewhere in the Tampa area? | 10:07 |
systemdlete | who knows, right? | 10:08 |
rrq | you said you started with the "minimal install" whihc was made before 25th I suppose | 10:08 |
systemdlete | right. | 10:08 |
onefang | Mirrors have to tell us which country they are in, and which they are happy serving. | 10:08 |
onefang | And those countries are listed in https://pkgmaster.devuan.org/mirror_list.txt | 10:09 |
systemdlete | I forget exactly when I switched from the netinstall. But it was a few days back. | 10:09 |
rrq | so doing "apt update" on that particular system anytime after 25th would get the message | 10:09 |
systemdlete | Days and nights just all kind of melt into one another when I get deep into one of these projects... | 10:09 |
systemdlete | rrq: Yes. We agree. That is the expected outcome. | 10:09 |
onefang | So what rrq is saying is that the 5 day old repo is your install medium. I think. | 10:10 |
systemdlete | onefang: Could there be a rogue server somewhere? | 10:10 |
systemdlete | supply-chain attack? | 10:10 |
onefang | That's one of the reasons for apt-panopticon. It downloads packages and checks hashes and stuff. | 10:11 |
systemdlete | onefang, you are late to my little "party" here. Ive been doing repeated installs while trying to work out the kinks in my installatin method | 10:11 |
systemdlete | I'm using a combo of debootstrap and some hand-crafted sh scripts | 10:12 |
onefang | I've been busy trying to find anew home, now I'll be busy moving for the second time this year. I hope to settle down so I can work more on things like apt-panopticon. It needs to check CC mirrors now, and other things. | 10:12 |
systemdlete | It's part of a checklist to ensure I don't forget what I did last time... | 10:12 |
systemdlete | "CC"--command and control, Communist China, can't contact, what? | 10:13 |
onefang | I switched from debootstrap to mmdebstrap for chimaera, and write my own scripts for installing using that. | 10:13 |
systemdlete | mmdebstrap? There is such a thing? | 10:14 |
systemdlete | I will HAVE to look at that. | 10:14 |
onefang | Yep. A little faster, with some extra features. | 10:14 |
systemdlete | Is mmdebstrap on the devuan ISO's? | 10:15 |
onefang | Country Codes. CC.deb.devuan.org, where the CC is replaced by the country code you want. | 10:15 |
systemdlete | or do I need to spin one myself? | 10:15 |
onefang | mmdebstrap is in the repos, dunno if it's on the ISOs. | 10:15 |
systemdlete | it is clearly NOT on the minimal ISO | 10:16 |
systemdlete | but that's OK. I am thinking that I want to start creating my own boot media anyway. But that has to wait. I want to get through daedalus first. Maybe for excalibur | 10:17 |
onefang | I just found out about bdebstrap. "bdebstrap is an alternative to debootstrap and a wrapper around mmdebstrap to support YAML based configuration files." | 10:18 |
systemdlete | thank goodness someone had the good sense to create yet another markup language for us to learn. | 10:21 |
systemdlete | (j/k) | 10:21 |
onefang | lol | 10:21 |
systemdlete | Actually, what really helps me the most is when there are 18 variations of the same language, because that helps create confusion as well as lots of incompatibilities to deal with. | 10:22 |
systemdlete | but I am getting waaaay OT and a certain hall monitor is going to send me to the principal's office shortly if I don't quit. | 10:23 |
systemdlete | I also noticed bdebstrap when I was looking up mmdebstrap | 10:23 |
onefang | To right mate, that's a royal pain in the bloody arse, variations in language. Think yourself lucky I didn't use proper 'strain. | 10:24 |
onefang | OOps, that's "proper 'strine". | 10:24 |
systemdlete | actually, I rather like the whole British accent thing | 10:24 |
systemdlete | Sounds odd to this American, but then again, so do a lot of Americans | 10:25 |
onefang | OK, waaay OT now. lol | 10:25 |
systemdlete | lol | 10:25 |
systemdlete | yeah | 10:25 |
Guest4359 | since linux is multiuser why we need seat management software ? | 15:50 |
gnarface | Guest4359: i agree with the sentiment, but it's not really a Devuan support issue. please join #devuan-offtopic for such conversations. | 15:52 |
gnarface | it's by far not the first or the stupidest thing they're doing to us upstream | 15:53 |
Guest4359 | ok sorry | 15:53 |
gnarface | no big deal | 15:53 |
gnarface | we just like to keep this channel clear for actual support issues. if everyone was allowed to bitch about Debian's vandalism of Linux in here, there'd be no room for anything else. | 15:53 |
Guest4359 | i apologize. | 15:54 |
gnarface | we have #devuan-offtopic for all sorts of other philosophical discussions though, feel free to join it and stay connected to both | 15:54 |
Guest4359 | gnarface, i've posted my question there. | 15:55 |
gnarface | were you the person who asked about salix then left 2 minutes later? | 15:56 |
gnarface | that's also a #devuan-offtopic type question | 15:56 |
gnarface | also, in general it's a good idea to wait longer for responses, they can be slow but people do read the scrollbacks | 15:56 |
bgstack15 | I'll start here, because I am indeed using Devuan. Does anyone here have experience getting Nvidia legacy drivers to get your X11 display to get to the 1920x1080 native resolution of the laptop display? I cannot figure out why it doesn't think LVDS-1 can get to that resolution. | 16:27 |
bgstack15 | (Amusingly, btw, X11 happily squishes higher resolutions into its perceived max if you forcefully add with --newmode and --addmode) | 16:28 |
gnarface | bgstack15: not specifically on a laptop but i'm sure i've made it work before. ymmv depending on nvidia card model though... | 16:28 |
bgstack15 | GK107M (GeForce GTX 660M), on an Alienware M71x R4 | 16:29 |
bgstack15 | nvidia-detect choked, but installing nvidia-* whatever warned me that it needed nvidia-tesla-470-driver so I installed that. | 16:29 |
gnarface | one generation newer than the last one i had before switching to AMD. should work... | 16:30 |
bgstack15 | even nouveau got to 1600x900; but contemporary reviews of this laptop called it "native 1920x1080 resolution" | 16:30 |
bgstack15 | And in the graphical nvidia-settings program, it shows "Display devices: None" yet clearly I have a display device... | 16:30 |
gnarface | right, should work. first thing i'd check is to make sure all your packages with nvidia in their names are of the same version | 16:31 |
gnarface | there's more than one legacy, and i see you found the tesla drivers too, so "nvidia-*" definitely is not what you want, that's a mess | 16:31 |
bgstack15 | yes, sorry, I didn't run "apt-get install nvidia-*" I am not sure which "nvidia-driver" or similar package I tried to start with | 16:31 |
gnarface | actually, just paste the output of "dpkg -l |grep -i nvidia" to paste.debian.net and i'll sanity check it for you. | 16:31 |
bgstack15 | I was trying merely to summarize here | 16:32 |
gnarface | now, this is a fully devuan distro of one version, right? you didn't mix other distros or repos or releases in with it, and you didn't use the nvidia.com shell script driver installer, right? | 16:32 |
gnarface | follow-up to that... any backports packages? | 16:33 |
bgstack15 | I'm on Ceres | 16:33 |
gnarface | i see | 16:33 |
bgstack15 | I have not run any nvidia.com shell script | 16:33 |
bgstack15 | I didn't think of that... | 16:33 |
gnarface | ceres could be half the problem right there, but let's proceed anyway | 16:33 |
gnarface | dpkg -l |grep -i nvidia | 16:33 |
gnarface | we make sure you have the right driver packages, then we'll try to poke at the settings with xrandr | 16:34 |
gnarface | you don't have any xorg.conf yet, right? you might need it but start without. | 16:34 |
gnarface | (i'm assuming you're not using wayland) | 16:35 |
bgstack15 | This is X11 | 16:38 |
bgstack15 | I had no xorg.conf, but then I tried the nvidia-xsettings with sudo to save to /etc/X11/xorg.conf | 16:38 |
bgstack15 | that prevented X11 from displaying anything! I'm shocked lightdm didn't fail out. It merely got stuck | 16:39 |
bgstack15 | i'm struggling with dpaste right now... hold on | 16:39 |
gnarface | i don't think it actually saves to xorg.conf, it saves to something like ~/.stupidnvidiasettingsrc and it doesn't work unless you then have nvidia-settings in your WM's startup items (and obviously if you need something for Xorg to even launch first this isn't a viable alternative to actually having /etc/X11/xorg.conf) | 16:40 |
gnarface | bgstack15: please, use paste.debian.net or just /msg it to me | 16:40 |
bgstack15 | https://bgstack15.ddns.net/stackbin/1ea2dcac-2df5-4e6c-a4b6-537cc61336bb | 16:40 |
gnarface | yea, i'm not clicking on that | 16:40 |
bgstack15 | Did curl change?! Now I need --data-binary '@-' instead of before (according to my docs) --data-raw '@-' | 16:41 |
gnarface | dunno, seems plausible though | 16:41 |
bgstack15 | well, that's my pastebin so I don't know how to help you | 16:41 |
gnarface | nobody will click on links to my servers either. don't take it personally. | 16:41 |
bgstack15 | And I bothered to get a let's encrypt cert.... nobody would ever visit when I was using my own CA | 16:42 |
bgstack15 | when I run dpaste, I always get "free(): double free detected in tcache 2" | 16:42 |
gnarface | hmm, i don't know what that's about either. | 16:42 |
bgstack15 | let me get a browser up on the affected system and use the front-end to dpaste | 16:42 |
gnarface | just /msg the paste to me, i'll wait | 16:43 |
bgstack15 | https://paste.debian.net/1290674/ | 16:43 |
bgstack15 | I don't have an irc client on the affected system. | 16:43 |
gnarface | ok that works | 16:43 |
gnarface | ok, yea this is a mess | 16:43 |
gnarface | you need to get all that tesla stuff outta there | 16:44 |
bgstack15 | I believe I removed all nvidia stuff, and then ran "install nvidia-tesla-470-driver firmware-misc-nonfree" | 16:44 |
bgstack15 | The nvidia-driver thing told me I need the tesla thing | 16:44 |
bgstack15 | but OK | 16:44 |
gnarface | well wait, this is a x86 laptop, right? not some nvidia shield or tegra development board right? | 16:44 |
bgstack15 | I will try anything, lol. It's not an important system yet. | 16:44 |
bgstack15 | This is a i7-3610QM | 16:45 |
bgstack15 | so guessing 64-bit | 16:45 |
gnarface | yea, you definitely don't want tesla then. that's their driver for their ARM hardware | 16:45 |
gnarface | why the hell did it even let you install that? sigh | 16:45 |
bgstack15 | hm, nothing on the debian nvidia page made it sound like Tesla was for ARM | 16:45 |
gnarface | maybe i'm confused, but this is definitely not the legacy driver you said you were trying to install | 16:45 |
gnarface | the legacy drivers all have "nvidia-legacy" in their name | 16:45 |
bgstack15 | https://wiki.debian.org/NvidiaGraphicsDrivers#Tesla_Drivers | 16:46 |
bgstack15 | dpkg -l | grep nvid now shows nothing at all! | 16:46 |
bgstack15 | lemme reboot... | 16:46 |
bgstack15 | I actually like nouveau just fine, but for this one game I play, I need Nvidia max performance, lol | 16:46 |
gnarface | i understand. start with this package search: apt-cache search ^nvidia-legacy | 16:47 |
bgstack15 | wilco after reboot | 16:47 |
gnarface | whatever tesla is for, it's not for laptops | 16:48 |
gnarface | and it's definitely not for 600 series geforce cards | 16:48 |
gnarface | maybe it's for their bitcoin mining rigs or something, that's what the description on the link you pasted suggests | 16:48 |
gnarface | i might have got "tegra" and "tesla" conflated in my head but they're definitely not what you want | 16:50 |
bgstack15 | oh, I know what happened. the freaking-annoying dpkg INTERACTIVE frontend said "the above card requires either the non-free legacy NVIDIA driver (package nvidia-tesla-470-driver) or the free Nouveau driver. | 16:50 |
bgstack15 | so that's why I installed that package by name. | 16:50 |
bgstack15 | but now I'm trying nvidia-legacy-340xx-driver | 16:51 |
gnarface | that set should work | 16:52 |
gnarface | do you see any later than 340 in there? | 16:52 |
bgstack15 | Thankfully I'm not interested in chasing the latest drivers (and cannot even, with this 2013 laptop) for anything. I want things to work one way, and then never ever change. So if we can get this going with any random 340xx or 304xx or whatever, then I'll do that. | 16:52 |
bgstack15 | there's a nvidia-legacy-390xx-driver | 16:53 |
bgstack15 | oops, should I have picked that? | 16:53 |
gnarface | yes, use that one. i assume the tesla thing must be some sort of mis-detect, because those are also 470.* drivers, and nvidia.com lists the last supported driver version for your card is 418.113 | 16:53 |
gnarface | so what you want is the latest you can get up to 418.113 | 16:53 |
gnarface | i didn't see any nvidia-legacy-4* packages in pkginfo.devuan.org so 390xx is probably what you want | 16:53 |
bgstack15 | https://us.download.nvidia.com/XFree86/Linux-x86_64/525.105.17/README/supportedchips.html (linked from that Debian nvidia page) indicates that my GTX 660M is supported on the 470.xx driver | 16:54 |
gnarface | hmm, weird because that's not the one their driver search gives me | 16:54 |
bgstack15 | oh, so that tesla business had the 470 in it... | 16:54 |
gnarface | their driver search gives me this https://www.nvidia.com/Download/driverResults.aspx/153717/en-us/ | 16:54 |
gnarface | not sure if the search results will carry through with just that url though... | 16:55 |
gnarface | hmm, seems to though | 16:55 |
gnarface | trust me though, that 660M is definitely in the category of "legacy" | 16:56 |
bgstack15 | I think I used to once use the nvidia.com stuff (in my Fedora days) but I prefer the Devuan packages when possible. I could try the nvidia.com stuff if this legacy-340 stuff or legacy-390 doesn't work | 16:56 |
bgstack15 | Oh, I know. This laptop is from 2013 | 16:56 |
gnarface | only use the nvidia.com shell script as a last resort. it's almost certain to "work" in the sense that it'll probably give you a one-time working setup, but it will run roughshod over your package dependency tree and the next time you try to update your graphical stack it'll break | 16:57 |
gnarface | (it does have an uninstall option now but it usually doesn't actually clean up after itself very well) | 16:58 |
bgstack15 | ah, the link for the 390.144 driver indicates that the GTX 660M is still supported by it. | 17:00 |
gnarface | yea, i'm certain it should go up to 418.113 but it looks like 390 is the latest in debian | 17:01 |
gnarface | and there's probably a good reason why | 17:02 |
gnarface | (like the later versions are all crashy) | 17:02 |
bgstack15 | hm, after the reboot with 340 installed, lsmod shows "nvidia" but fewer than normal nvidia things, and nvidia-settings says I'm not using the nvidia driver... | 17:05 |
bgstack15 | I'm removing, and will install the 390 | 17:05 |
gnarface | after you get the 390 driver in, paste me the "dpkg -l |grep -i nvidia" output so i can sanity check it again | 17:05 |
bgstack15 | yep | 17:05 |
gnarface | and we may still have to do a couple config changes to make it work | 17:06 |
bgstack15 | Is rebooting necessary? Can I just rmmod nvidia and restart lightdm? | 17:07 |
gnarface | specifically, you may need to add "nomodeset" to your kernel command-line, blacklist the nouveau driver, and create a minimal xorg.conf to give it some non-auto-detected operating boundaries | 17:07 |
gnarface | i would recommend a reboot with this hardware. | 17:07 |
gnarface | ymmv but nvidia cards historically have often required a full reboot, and sometimes actually a full shutdown + cold boot | 17:07 |
gnarface | (though the cold boot thing i haven't noticed being strictly necessary since the AGP era except in a couple weird crash cases) | 17:08 |
bgstack15 | https://paste.debian.net/1290678/ | 17:14 |
bgstack15 | my dpkg -l | grep nvid | 17:14 |
bgstack15 | nvidia-settings says it appears that I'm not using the NVIDIA X driver.... | 17:15 |
bgstack15 | should I run `nvidia-xconfig` as root now? | 17:15 |
gnarface | no | 17:15 |
gnarface | blacklist nouveau, add nomodeset to your kernel command-line, then reboot once more, but i think you need to remove these 525.* packages too | 17:16 |
gnarface | to the best of my understanding, these packages are from non-legacy versions and not even the same version at that, they all have to go: nvidia-modprobe, nvidia-vulkan-common, nvidia-xconfig, nvidia-egl-common | 17:17 |
gnarface | but nvidia-settings probably just says that becuase nouveau beat nvidia to the punch. blacklisting it should fix that part. | 17:18 |
gnarface | (let me know if you need help with specifics on those tasks) | 17:18 |
bgstack15 | nouveau is already blacklisted. I updated grub with nomodeset. | 17:18 |
bgstack15 | Unfortunately nvidia-vulkan-common, nvidia-egl-common, and nvidia-xconfig want to remove the *390* drivers.... | 17:18 |
gnarface | damnit | 17:18 |
gnarface | alright, leave them then, but that might be a problem | 17:19 |
gnarface | there's no nvidia-legacy versions of those in sight in the repos you could install isntead? | 17:19 |
gnarface | instead* | 17:19 |
bgstack15 | should I use nvidia-legacy-390xx-vulkan-icd | 17:19 |
bgstack15 | ? | 17:19 |
gnarface | yea, try it. see if that will kick the 525 version out | 17:20 |
bgstack15 | I don't know exactly what yet, the vulkan-common package pulls in | 17:20 |
gnarface | in general you don't want a heterogeneous mix of nvidia driver package versions. it's not smart enough to know it won't work. | 17:20 |
bgstack15 | well I already have the nvidia-legacy-390xx-vulkan-icd and nvidia-legacy-390xx-egl-icd packages | 17:21 |
bgstack15 | but that common business is still here and wants to remove the legacy drivers... | 17:21 |
gnarface | damn, ok so this is ceres and this may just be simply broken but let's ignore those and move forward for the moment | 17:21 |
gnarface | let's get you a decent xorg.conf | 17:22 |
gnarface | uh, hang on, fabricating one | 17:23 |
bgstack15 | oh, I was able to remove the nvidia-xconfig package though. | 17:23 |
bgstack15 | (That probably would have helped with the xorg.conf...) | 17:23 |
gnarface | eh, maybe | 17:23 |
gnarface | this part isn't so hard for me though, just gimme a minute | 17:23 |
gnarface | consulting some notes | 17:23 |
gnarface | ok, we need one set of figures from your laptop's manufacturer manual: we need the horizontal and vertical refresh ranges of the panel | 17:25 |
gnarface | of the display panel | 17:25 |
gnarface | i'm typing this xorg.conf up while you find that | 17:26 |
gnarface | we might need to consult the output of "xrandr" while Xorg is running, but at the moment that's a bit of a catch-22 problem | 17:27 |
bgstack15 | well, the owner's manual and reference guide do not have "horizontal" or "refresh" anywhere in the docs. | 17:28 |
bgstack15 | er, refresh rate, 60Hz | 17:28 |
gnarface | wait, it says 60Hz for which? | 17:30 |
bgstack15 | Maximum resolution 1920 x 1080 | 17:31 |
bgstack15 | Dimensions | 17:31 |
bgstack15 | Height 215 mm (8.46 inches) | 17:31 |
bgstack15 | Width 382 mm (15.04 inches) | 17:31 |
bgstack15 | Diagonal 439 mm (17.28 inches) | 17:31 |
bgstack15 | Refresh rate 60 Hz | 17:31 |
bgstack15 | that is all | 17:31 |
gnarface | stand by | 17:32 |
bgstack15 | now, lshw shows that the display is using driver=i915. Do you think I should blacklist that? | 17:33 |
gnarface | no, not yet. hmm, but is this one of those hybrid laptops with two gpu vendors? | 17:33 |
gnarface | it's possible we're going about this all wrong if the nvidia card isn't the only video device | 17:34 |
gnarface | i just finished the xorg.conf for you though, minus the horizsync and vertrefresh values | 17:34 |
bgstack15 | well, i915 is the built-in Intel thingy | 17:34 |
bgstack15 | I think every i5/i7 always has a native one as part of the chip | 17:35 |
gnarface | well, depending on how the bios is setup the intel one may have to be the primary | 17:35 |
bgstack15 | I don't recall seeing an option for "primary graphics card" but if I had seen one, I would have set it to nvidia. | 17:35 |
gnarface | https://paste.debian.net/1290683/ | 17:36 |
gnarface | well, here's the example xorg.conf i just created under the assumption that this laptop only had a nvidia card | 17:36 |
gnarface | the only thing wrong with it is i had to fudge the values for the Monitor section's HorizSync and VertRefresh fields | 17:37 |
gnarface | maybe nvidia-xconfig can get those | 17:37 |
gnarface | maybe xrandr can too | 17:37 |
gnarface | but the intel/nvidia hybrid gpu situation might explain a lot about why it thought you had a tegra mining rig | 17:38 |
bgstack15 | ok, rebooting with that xorg.conf in place. | 17:38 |
bgstack15 | (Tesla) | 17:38 |
gnarface | tesla, right. | 17:38 |
bgstack15 | ooh, Fn+F7 is "I/D GFX" | 17:40 |
gnarface | so, the way the hybrid laptops typically work, is you actually configure xorg to the intel video card, then you use some command-line program or some environment variable (DRI_PRIME? something like that) as a pass through to offload games and other 3d tasks to the nvidia card on a per-command-execution basis | 17:40 |
bgstack15 | I bet it takes software in the OS to react to it, but maybe I should have tried pressing that... | 17:40 |
gnarface | but, the fancier laptops, have a way to just change the nvidia card to the default one and you don't have to mess with dri prime nonsense | 17:40 |
gnarface | unfortunately i don't have a lot of first-hand experience with this generation of hardware so you'll kinda have to figure it out from there | 17:40 |
bgstack15 | lightdm is not showing or running; haven't figured which. that xorg.conf must be incomplete | 17:41 |
bgstack15 | I did just modprobe -r i915; maybe I should just move that xorg.conf out of the way and see if it loads up the nvidia card now? | 17:41 |
gnarface | if the nvidia card isn't active, it would have no way to use this. it's possible you need a xorg.conf that loads the intel driver | 17:41 |
gnarface | well | 17:42 |
gnarface | try it | 17:42 |
gnarface | but if that doesn't work, try swtiching "intel" for "nvidia" on line 68 of that xorg.conf and reboot | 17:42 |
gnarface | anyone here have one of these alienware hybrid gpu laptops or know if they can have the nvidia card be the default or at least know how to use dri_prime right with it? | 17:44 |
gnarface | bgstack15: see if you can locate a binary called "prime-run" or something like that, in case we have to use the intel card as a passthrough | 17:45 |
gnarface | bgstack15: i think this is the relevant debian wiki page: https://wiki.debian.org/NVIDIA%20Optimus | 17:46 |
gnarface | the thing in their example ServerLayout section might help, i've never seen that; Inactive "intel" | 17:46 |
gnarface | we just need to get xorg up at all, then you can poke at it with xrandr | 17:47 |
gnarface | (nvidia-settings might start working right too then) | 17:47 |
gnarface | sorry we went so far down the road of assuming there was only a nvidia card, that was costly and confusing | 17:49 |
bgstack15 | lol, i hadn't said anything about it having hybrid graphics | 17:53 |
bgstack15 | I suppose it qualifies, merely because i7s always have their own | 17:53 |
gnarface | i think this is the definition of hybrid graphics, but they come in a few different forms | 17:53 |
gnarface | it would explain why the nvidia driver packages and the nvidia-settings tool misdetected you | 17:54 |
gnarface | unfortunately i'm not up on the latest developments for working with them | 17:54 |
bgstack15 | this is a 2013 laptop, so we don't need the latest. | 17:55 |
gnarface | well, when last i looked into it, if you didn't have a bios setting (or bios-enabled hotkey) to switch the nvidia one to the primary, then the only option was to use the intel one as the default for xorg, and as a passthrough for the nvidia card on certain tasks | 17:56 |
gnarface | this wiki page seems to suggest they've moved the ball forward further than that, but it's not clear those developments will help you with the legacy driver | 17:56 |
djph | bgstack15: laptops "only(tm)" offer "Hybrid Graphics" since like ehhh 2009 / 10 or something, where basically the discrete GPU is "only(tm)" a co-processor, rather than actually driving the displays | 17:56 |
gnarface | you're kinda stuck in a weird position here | 17:56 |
djph | Yeah, it's only relatively recent that they started getting less bad (albeit it's still worse than a desktop with truely discrete graphics) | 17:57 |
gnarface | djph: do you happen to know how to fix this? we can't get even xorg running at the native display resolution | 17:57 |
gnarface | *can't even get | 17:57 |
djph | With nouveau, right? | 17:58 |
* djph reads scrollback | 17:58 | |
gnarface | tried with nouveau (presumably, or modesetting) and with nvidia-legacy-390 | 17:58 |
gnarface | working on nvidia-legacy-390 and i think it would have worked at this point if the intel gpu wasn't in the way | 17:59 |
debdog | I have no clue about the technical details of my laptop but in my case this helped: http://us.download.nvidia.com/XFree86/Linux-x86_64/390.42/README/randr14.html | 18:00 |
debdog | a) the xorg.conf and b) the "xrandr --setprovideroutputsource modesetting NVIDIA-0" which I put in fluxbox' startup script | 18:00 |
gnarface | bgstack15: ^^ | 18:02 |
djph | I remember *SOMETIMES* nVidia cards are slow at getting EDID information before xorg wants to timeout and just do something "safe(tm)" (like 640x480)... with ubuntu, restarting X (ctrl+alt+bksp) usually helped. But it's been 5 (6?) years since that 2012 era machine died on me :( | 18:03 |
gnarface | yea, i see that Inactive "intel" thing here too. that might be key | 18:03 |
gnarface | you can actually dump the EDID info to a binary blob and just tell the xorg.conf to load it directly from harddrive. that might help with the latency race condition issue | 18:04 |
djph | yeah, I don't remember what the "permanent(tm)" fix was (I mean, other than replacing it with some no-nVidia laptop because I stopped playing "heavy" games as much) | 18:05 |
bgstack15 | er, well, now, whenever I switch virtual terminals it just goes all blue and does nothing. | 18:08 |
bgstack15 | oh and capslock isn't working either, so maybe it's entirely hung | 18:08 |
gnarface | bgstack15: is that with the "Inactive "intel"" thing in the ServerLayout section of your xorg.conf? | 18:09 |
bgstack15 | I think that's with no xorg.conf at all, so whatever it autogenerates.... | 18:10 |
bgstack15 | although I'm rebooting atm and will try to grab it before lightdm starts up | 18:11 |
bgstack15 | lol, it's a brand new install. Maybe I'll just install Devuan again. It's not the worst thing in the world. | 18:11 |
gnarface | bgstack15: try merging the xorg.conf i gave you with the relevant parts from this one: http://us.download.nvidia.com/XFree86/Linux-x86_64/390.42/README/randr14.html | 18:11 |
bgstack15 | yeah, i did. lightdm wouldn't display anything with any tweaks I tried after merging those. | 18:11 |
rrq | note that the daedalus Xorg with seatd don't like VT switching (I missed the beginning of this though) | 18:11 |
gnarface | bgstack15: what did the Xorg log say? oh, yea, maybe disable seatd too | 18:12 |
bgstack15 | oh, I did disable seatd once. I might have already autopurged it too, though | 18:12 |
gnarface | bgstack15: i'd like to see the Xorg.0.log after a failed run | 18:12 |
rrq | yeah, that Xorg version is not good with VT switching... in fact you might want to roll back xserver-xorg-core to the chimaera version, until it's fixed for daedalus | 18:17 |
bgstack15 | i cannot get it to respond at all, and my sysvinit foos failing. I added to my kernel command "init=bin/bash" but is that the right way to get to single-user mode with sysvinit? | 18:18 |
bgstack15 | er, /bin/bash | 18:19 |
gnarface | bgstack15: "single" | 18:19 |
gnarface | iirc | 18:19 |
gnarface | in the kernel command-line | 18:19 |
rrq | or boot with "init=bin/bash" and use "exec /sbin/init -s" (I couldn't find "single" in the initrd scripts) | 18:28 |
rrq | eh /bin/bash :) | 18:29 |
mason | bgstack15: /bin/sh (dash) | 20:30 |
DrHyde | hello gang, what's up with netatalk? It was in Chimaera, it is in Ceres, but is missing from Daedalus! | 21:06 |
mason | DrHyde: Not in Debian Bookworm, hence not in Daedalus. | 21:06 |
DrHyde | <sadface> | 21:07 |
golinux | Try backports? | 21:07 |
DrHyde | not in bookworm-backports either | 21:07 |
golinux | It may be eventually or works some magic yourself . . . | 21:10 |
golinux | -s | 21:11 |
DrHyde | yeah, i can always build it from source in extremis | 21:11 |
mason | DrHyde: I think most folks are using Samba in that space nowadays anyway. | 21:14 |
mason | Macs can all use SMB, and I think the Netatalk code hasn't seen a lot of use. | 21:15 |
DrHyde | they are, i've been procrastinating about switching for years because i'm very lazy :-) | 21:15 |
mason | https://www.bleepingcomputer.com/news/security/synology-warns-of-critical-netatalk-bugs-in-multiple-products/ | 21:15 |
DrHyde | i suppose now's a good time to switch | 21:15 |
mason | So, that's not as out of date as I thought, if they were updating it that recently. | 21:16 |
mason | Hm, very recent updates: https://en.wikipedia.org/wiki/Netatalk | 21:18 |
mason | So I stand corrected. | 21:18 |
mason | (I used to use it for Time Machine back-ups at a place I worked.) | 21:19 |
gnarface | bgstack15: uh, this probably should have been asked before, but you're in the "video" group, right? | 22:19 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!