[00:00] well poop... can't boot at all now on any kernel [00:01] 21.04 install iso can't boot either [00:06] Maybe a BIOS update... [00:10] BIOS update seems to have improved things [00:20] OKay now I'm getting somewhere... 5.12.4 is running but need to upgrade all the things.. doing that now, but feeling hopeful [00:21] * rud0lf holds his thumbs [00:23] Ubuntu has come a long way.. I remember when do-release-upgrade was all but guaranteed to brick your machine [00:39] I usually only use LTS releases [00:39] But I'm not using all the latest hardware [00:41] Yeah this stuff is kinda bleeding edge I think [00:56] How to install caps2esc on ubuntu? [00:57] Intelo: hopefully whoever wrote it put an INSTALL file next to it [00:59] sarnold: i wanted to/ expected some package. Just checking it there isnt [01:00] Intelo: 'locate caps2esc' doesn't show anything on my unpacked ubuntu mirror [01:00] name resolution is broken on my system, resolv.conf is nameserver 127.0.0.53, and it is listening on 53 as seen by the output of netstat -nl, however dig returns SERVFAIL. I can dig @4.2.2.2 just fine, so there is no fundamental networking issue. Any idea? [01:04] Entelin: is there anything in journalctl -u systemd-resolved output that looks related? [01:06] sarnold: nothing in two weeks [01:11] Entelin: static IP or dynamic/auto IP? [01:12] Entelin: static right? [01:14] static [01:15] ec2 server with two ips I think it may have broken when I added the second [01:19] Entelin: heh, strange [01:32] Entelin: it'll work just fine after you undo the layers (of the onion) that you added before it "broke" [01:33] yeah I think i'm done with ubuntu honestly. I've used it for at least 10 years as my primary server os, but between this proprietary snap nonsense and them changing how the network gets configured every couple years I think it's time to move on. [01:33] going to move my ec2 instance to alpine or arch perhaps [01:34] Entelin: ...well..good luck in your endeavors [01:34] there shouldn't be an onion to peel [01:34] o/ [01:38] hmm, I only remember the one change, /etc/network/interfaces -> netplan to generate configs for systemd-network.. what else did I miss? [01:41] i must have missed it also, nothing that actually gets in the way or prevents production/productivity === Roey is now known as lock-in-a-sock === gabkdlly_ is now known as gabkdlly === lock-in-a-sock is now known as Roey [02:36] Why wouldn't it work? I can still press use caps lock https://imgur.com/HeM24zY.png [02:52] solved, nevermind === Roey is now known as rock-in-a-frock === rock-in-a-frock is now known as Roey [04:25] something messed up with caps2esc and now caps is always onn. any quick tips? [04:25] sudo /usr/bin/nice -n -20 /usr/bin/udevmon -c /etc/interception/udevmon.d/deafmute-ppa-caps2esc.yaml [04:25] this doesnt help either [04:58] Okay I disabled systemd-resolved service altogether, restarted dnsmasq, ran ifdown eth0 && ifup eth0 [04:59] dhclient's lease show DNS servers but nothing is reflected in resolv.conf [04:59] BTW using 20.04 Ubuntu Server [05:16] I found this and editing dhclient's hook file fixed the issue: https://bugs.launchpad.net/ubuntu/+source/systemd/+bug/1745463 [05:16] Launchpad bug 1745463 in systemd (Ubuntu) "Disabling systemd-resolved breaks dhclient resolvconf integration" [Undecided,Confirmed] [05:36] Hello :-) I am new to Linux and Ubuntu. I installed Ubuntu for the first time ever last week. I really love the look and feel of Ubuntu coming from windows 10. [05:39]  I have a problem when I try to move any open window they look like they are glitching or something. I also have a similar problem with youtube, scrolling webpages or playing some games. I am not sure what this problem is but after looking on google I think it might be something called screen tearing ? [05:41] There is plenty of fixes listed for Ubuntu but I am new and learning so I do not want to try anything that I do not understand. I hope someone here may be able to help me understand and fix this problem :-) === denningsrogue682 is now known as denningsrogue68 [08:16] hey guys [08:16] how do you set supervisor to create logs under running users? [08:16] supervisord runs at root, but all child processes also create logs for root, even if child process is non-root [08:16] causes log issues [08:26] heller_: most process supervisors are designed to work like that, because they are meant to be run and used by system administrators. [08:27] heller_: you can run multiple instances of 'supervisord' as different users if you wish though. [08:38] Issue here being, that because the logs are owned by root, the child process cant write to them [08:38] Hi , can you advice ? I'm trying to install ppa kernel 5.12, with it's dependencies. But got the following error. [08:38] linux-headers-5.12.0-051200-generic : Depends: libc6 (>= 2.33) but 2.27-3ubuntu1.4 is to be installed [08:39] Is it possible to find out how to install libc6-2.33 on 18.04.5 LTS [08:49] Walex: any ideas? [08:55] heller_: as I was saying you can run 'supervisord' as another user. [08:56] heller_: regardless the logs that *supervisord' are creating are meant to be the 'stdout' and 'stderr' logs, so certainly the process can write to them even if 'supervisord' creates them as 'root'. [08:57] heller_: this is described quite explicitly here: http://supervisord.org/logging.html#child-process-logs [09:06] XATRIX: not a good idea. [09:08] Walex: damn, so i can't install recent kernel on a 18.04 ? [09:14] XATRIX: of course you can, e.g. you can build a kernel package yourself for example. [09:14] I'd like to use ppa kernel built [09:14] https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.12/ [09:14] XATRIX: then find a PPA that builds recent kernels for 18.04 :-) [09:15] I downloaded a kernel from the link i provided. is it supported by ? [09:16] XATRIX: "supported" is a difficult word, especially for a PPA, especially for a PPA with "Test" in its name :-) [09:17] XATRIX: there is a convenient 'ubuntu-support-status' tool for checking. [09:18] XATRIX: perhaps the 'edge' kernels from the 18.04 "official" repositories has already advanced to 5.12 [09:19] You have 2032 packages (65.4%) supported until April 2023 (Canonical - 5y) [09:19] You have 187 packages (6.0%) that can not/no-longer be downloaded [09:19] You have 886 packages (28.5%) that are unsupported [09:20] XATRIX: look at the options and check for yourself :-) [09:20] copy. thanks [09:22] XATRIX: but PPAs are unlikely to be "supported", as a ruly "main" and "restricted" get some support. [09:23] XATRIX: https://ubuntu.com/about/release-cycle "Maintenance and security status" section [09:23] Openvpn adds routes when it starts. Can I prevent or alter that? [09:28] XATRIX: 'apt-cache policy linux-image-generic-hwe-18.04-edge' might also be useful in general [09:29] Walex: it's a bit dated kernel [09:29] :) [09:30] XATRIX: note that Debian and Ubuntu often backport significant new features to "official" kernels, and as a rule backport security fixes. [09:30] yea, got it [09:30] XATRIX: LTS releases are meant to be "stable". [09:31] XATRIX: so it depends a lot on why you need a newer kernel. [09:32] XATRIX: but if you *really* need a latest, unsupported kernel, downloading the source package from the PPA and rebuilding it should not give problems. [09:32] Ok, i'll try to [09:32] thanks [09:34] disregard my earlier question, found options [09:34] XATRIX: for example I am trying out 'bcachefs' so I am occasionally using a patched mainline kernel, and rebuilding it with something like " nice make -j 8 KERNELRELEASE=5.11.0-20210509 bindeb-pkg" [09:35] you mean, i have to compile kernel myself ? [09:35] XATRIX, the mainline kernels are explicitly not supported and lack a lot of ubuntu patches and settings ... they are for verifing kernel bugfixes only and should not be used on production devices [09:36] ogra: i'm trying to run it on my own laptop [09:36] newer supported kernels are in the linux-hwe stack ... [09:36] XATRIX: if you want the very latest... otherwise rebuild the PPA source package. [09:36] well, unless you want to verify a bug, dont do that ... [09:37] just re-building the PPA package wont give you the ubuntu patchset either ... [09:38] yeah please dont advice compile and rebuilding kernels to users Walex, not really the ubuntu way right [09:38] XATRIX: depending on why you want a newer kernel, it might help to have a look at the "linux-modules-extra" packages [09:38] Walex: just like to be running a latest one. nothing special [09:39] well, the latest supported one is 5.8 ... [09:39] XATRIX: on 18.04? :-) [09:40] Walex: yes, i'm planning to upgrade to 20., but it's not about in a few days :) [09:40] (though not sure if 5.8 made it into the xenial hwe stack already) [09:41] for low values of "supported" there is the "-oem" line of kernel packages, that on 20.04 reaches 5.10 [09:41] with this kernel 5.4.0-73-generic #82-Ubuntu the BlueDevil-Assistent never identifies my headphones [09:41] *especially* if you want to ugrade soon, do not switch to an unsupported deb ... [09:41] that will cause headdaches during the upgrade [09:41] :( [09:42] yea, i can rollback to a supported one, during upgrade. but i'd like to use the recent one after [09:43] well, it is your system after all ... and you can do what you want ... but note that these kernels do not get updated, do not get any security fixes at all and do lack a lot of patches and defaults .... and will likely cause problems for release-to-release upgrades [09:43] and it is most likely pointless too... [09:43] some userspace things will not work with them as well [09:43] since the patches are expected [09:50] ogra: i usually leave one official kernel to boot in case of. and one recent to use all the modern features [09:50] what "modern features" ? [09:50] you will be lacking features ... [09:51] articulary security ones that come from ubuntu patches and integrate with userspace bits [09:51] *particulary [09:52] if there is some HW not working it probably makes sense to run such a kernel temporary and report the missing bit in launchpad so it can be added to the next supported kernel, thats about the only reason why you should run such a kernel [09:52] hi guys, whats the network and channel for ubuntu 21.10? [09:52] (or if a kernel team member asks you in a bug to use it to verify the bugs is gone with a mainline build) [09:56] CryptoSiD: #ubuntu+1 [09:58] Hi, after upgrade to 21.04 I lost resume from swap [09:59] Hi, after upgrade to 21.04 I lost resume from swapduring boot I have : gave up waiting for suspend/resume device [09:59] If I disable encypt swap and use normal swap it works [10:05] nomad_fr: do you encrypt the swap with a randomly generated key? :-) [10:08] Walex: like this cryptsetup luksFormat --cipher aes-xts-plain64 --verify-passphrase --key-size 256 /dev/nvme0n1p2 [10:14] Walex: so not a generated one, I set up it in the prompt [10:21] nomad_fr: have you checkd the steps in https://rephlex.de/blog/2019/12/27/how-to-hibernate-and-resume-from-swap-file-in-ubuntu-20-04-using-full-disk-encryption/ it might help.But your error message "gave up waiting for suspend/resume device" seems to come from 'systemd' and perhaps indicates that you have not configured 'systemd' to decrypt the swap. But I don't have 21.04 so cannot check. [10:29] Walex: but i want to run supervisor as root, but have the logs for child processes [10:29] since at the moment, they are not able to write to the logs [10:37] heller_: hos is "the logs that *supervisord' are creating are meant to be the 'stdout' and 'stderr' logs" not clear? [10:38] heller_: how is "the logs that *supervisord' are creating are meant to be the 'stdout' and 'stderr' logs" not clear? === westor{off} is now known as westor === denningsrogue683 is now known as denningsrogue68 [11:11] Walex: i believe it would be configurable ? but i believe i used stdout. why u ask? [11:14] Walex: a quick look at the man page reveals what u want to know: The server piece of supervisor is named supervisord. It is responsible for starting child programs at its own invocation, responding to commands from clients, restarting crashed or exited [11:14] subprocesseses, logging its subprocess stdout and stderr output, and generating and handling "events" corresponding to points in subprocess lifetimes. [11:14] TheBigK: paste.ubuntu.com /topic [11:15] thats 2 lines. thats fine i think? [11:19] c/lose [11:19] hahahah === ghoti_ is now known as ghoti === westor{off} is now known as westor [11:23] TheBigK: I was referring to my advice to "heller_": [11:23] [08:57] Walex | heller_: this is described quite explicitly here: http://supervisord.org/logging.html#child-process-logs [11:39] hey losers [11:39] steamcmd:i386 : Depends: libc6:i386 (>= 2.12) but it is not going to be installed [11:39] Depends: libstdc++6:i386 but it is not going to be installed [11:39] Idk how to fix this, I have 'multiverse' distribution component is already enabled for all sources. === LabMonkey is now known as Mechanismus [11:40] I didn't mean the losers comment, just tired of Ubuntu [11:40] being all ... stupid [11:40] horribleprogram: works for me :-) [11:40] Walex: I'm the biggest loser here === Walex is now known as loser_walex [11:41] horribleprogram: works for me :-) Pls let us know the version of Ubuntu. [11:42] horribleprogram: have you enabled "multiarch"? [11:42] uname? [11:43] uname -o? [11:43] horribleprogram: as in, do you have any other "i386" packages installed? [11:43] how I get the thingy [11:43] 5.4.0-73-generic [11:43] but that's the kernel [11:43] horribleprogram: do, "multiarch" is a feature for DPKG+APT [11:43] horribleprogram: 64b amd64 systems can also run 32b packages in "compatibility" mode [11:44] sudo dpkg --add-architecture i386 ? [11:44] sudo add-apt-repository multiverse [11:44] horribleprogram: look good the "dpkg --add-architecture i386" [11:44] loser_walex: "horribleprogram> Idk how to fix this, I have 'multiverse' distribution component is already enabled for all sources." [11:44] horribleprogram: you will have found this I guess: https://help.ubuntu.com/community/MultiArch [11:45] horribleprogram: note this is a guess. [11:45] of course [11:45] but I'll read it again [11:46] horribleprogram: if you cannot install the i386 'libc' package it must be either because you don't have multiarch enabled, or because you already have it installed but it is the wrong version and you cannot install the right version on top/along [11:46] there is no multiarch file [11:46] in /etc/dpkg/dpkg.cfg.d/ [11:46] horribleprogram: so also check: dpkg -l libc6 [11:46] ii libc6:amd64 2.31-0ubuntu9.3 amd64 GNU C Library: Shared libraries [11:47] do I create a multiarch file [11:47] horribleprogram: so enable multiarch and then 'apt install libc6:i386' to dobulke check [11:47] Hi folks [11:47] horribleprogram: you got it right above. [11:47] sudo dpkg --add-architecture i386 [11:47] doesn't add the multiverse thing [11:47] multiarch file [11:47] horribleprogram: don#t worry about that. [11:47] i'm trying to automate a ubuntu 20.04 install, but can't found in the documentation any option to allow halting not rebooting at the end. Is currently possible? I know that with halt on late_commands works, but at least the logs and not saved in the final system. I found this question https://discourse.ubuntu.com/t/how-to-poweroff-autoinstall-instead-of-reboot/17449 but no answer [11:47] horribleprogram: you also have to run 'apt update' "of course". [11:48] of course [11:48] still doesn't work [11:48] steamcmd:i386 : Depends: libc6:i386 (>= 2.12) but it is not going to be installed [11:48] same output [11:48] E: Unable to correct problems, you have held broken packages. [11:48] horribleprogram: ahhh that is a different thing. [11:49] horribleprogram: your package set is in an iffy state so DPKG+APT are hilding back [11:49] horribleprogram: that's what I have: https://paste.ubuntu.com/p/GzFPn5xtyQ/ [11:50] ueah ur output is way better [11:50] it shows the i386 installed as well [11:50] mine just has the amd64 [11:51] horribleprogram: pls install 'aptitude' too then we can look at which packages are broken more easily [11:51] sounds good ty, installing as we speak [11:52] wait wait [11:52] is sudo apt update [11:52] the same as sudo apt-get update [11:52] cuz I always use apt [11:52] yes same [11:52] kk sweet [11:53] Anyone know why the nvidia driver might be stuck at 640x480 after upgrading to 21.04? [11:53] honestly it seems like an ez fix [11:53] add the i386 architecture, update, and install [11:53] but I have broken packages [11:53] horribleprogram: aptitude search '~b' [11:53] to not it's about resolving broken packages [11:53] no output [11:54] actually it didn't work [11:54] echo $? returned 1 [11:54] there was some prompts that flashed really quickly [100%] building dependency tree [11:54] was one [11:55] horribleprogram: what does 'aptitude install steam-launcher |& pastebinit' say? [11:56] https://paste.ubuntu.com/p/b9kqmhgctd/ [11:56] that's [11:56] embarassing [11:56] horribleprogram: what does 'sudo aptitude install steam-launcher |& pastebinit' say? [11:56] https://paste.ubuntu.com/p/Jns2wjX7Qn/ [11:56] loser_walex: don't do that... [11:57] horribleprogram: did you add the "steampowered.com" repo? [11:58] no :'( [11:58] https://developer.valvesoftware.com/wiki/SteamCMD#Linux [11:58] I'm following this guide [11:58] horribleprogram: https://repo.steampowered.com/steam/ [11:58] ty :D [11:59] horribleprogram: you can boostrap by doing "wget steam_latest.deb" and then "sudo dpkg --install steam_latest.deb", "sudo apt update", "sudo apt upgrade" [12:00] horribleprogram: or just follow the manual steps in "Installing via apt" in that page. [12:01] yeah it's really ez [12:01] tysm [12:01] just out of curiosity [12:01] do u even know how to install a .gpg [12:02] horribleprogram: in that case it just means "just copy it there" [12:03] just testing u :D [12:05] loser_walex: omfg it worked ur the best lol [12:06] I wanna host a sweet GMod Server [12:06] cool! [12:06] Steam for Ubuntu LTS works really well here. [12:07] :D yeah apart from some driver issues Linux is damn near on par with Windows [12:08] I found that with Valve's version of WINE even latest DOOM works pretty decently. [12:08] horribleprogram: BTW make sure that you got the vulkan drivers installed, they seem to be fine. [12:08] horribleprogram: ahhh but you want to run a server, not the graphical game. [12:09] loser_walex: server on the Linux box [12:09] but yeah I have the nvidia ppa and did all that :D [12:09] vulkan-utils and those packages [12:12] horribleprogram: BTW 'aptitude' in full-screen mode is really good for looking at available packages, package status, dependencies etc. and I wrote a nice guide here: http://www.sabi.co.uk/blog/13-one.html?130414#130414 [12:12] loser_walex: tysm reading it now :D [12:13] loser_walex: uk some pretty deep syntax lol [12:13] ~i(!~M!~E)(~R~i) [12:13] jesus... [12:13] lol that's deep af [12:13] horribleprogram: that's computer engineering stuff [12:14] boolean algebra [12:14] horribleprogram: the syntax look bad, but each ~thing has a simple meaning and the expression is pretty clear actually [12:14] ab means a AND b? [12:15] like ~i(M) for example would mean NOT i AND M [12:15] ? [12:15] horribleprogram: yes concatenation means AND [12:15] that's cool [12:15] I think they do that for the .deb clauses too [12:15] horribleprogram: but "NOT" is "!", "~" means "match" more or less [12:15] how to let adjusting brightness works? or check where go wrong, thanks [12:15] like Recommends: Depends: etc [12:16] ahhh [12:16] horribleprogram: the latest versions of 'apt' have an 'aptitude' like query language too [12:16] horribleprogram: so "NOT i AND M" is: "(!~i)~M" [12:17] describing dependencies requires a couple of things right [12:17] but ultimately, NOT, AND, OR, a way to express package names, and then package version of that name [12:17] horribleprogram: ywes indeed [12:17] so like a POM syntax with package:version, but also expressing stuff like... libgcc >= 2.4 [12:18] horribleprogram: that is indeed quite useful in manay cases [12:18] yeah, it's actually something I'm going to dive deeper into soon [12:18] package management [12:18] I hate not knowing why stuff doesn't work [12:19] horribleprogram: with the fullscreen interface you can look directly at package dependencies and see what happens, and then 'aptitude' stages installations in a list so you can see exactly what is going to happen or cancel it. [12:19] me too [12:19] I hate not knowing why stuff doesn't work [12:19] jvava: ikr [12:19] also yeah [12:20] visualization for these sorts of complex dependency graphs is probably the best bet [12:20] brb 20 mins anyways gotta restart [12:20] tysm tc [12:20] what is ikr, horribleprogram [12:20] horribleprogram: jvava: while DPKG is a bit rubbish, APT works really well. [12:20] I dare not say that, loser_walex [12:21] basically, I think dpkg can do more === ghoti_ is now known as ghoti [12:31] I've always thought apt was built on the top of dpkg [12:37] it is === mateen1 is now known as mateen [13:13] what is better with kvm vm's Spice ot VNC? [13:14] *or [13:15] assuming viwing vm on the same host [13:19] aiena: i would connect via ssh and use the kvm virtmanager? [13:20] my vm's are on the same host. So I can run the virtmanager as a GUI app. [13:21] thats fine as well. why then using vnc? [13:21] do u need specific features ? [13:21] TheBigK: because kvm offers a choice between SPICE and VNC [13:21] and my VM is a desktop OS [13:22] I was unfamiliar with spice which it defaults to and wasn't sure about the differences [13:23] looking at the wikipedia page... it has some advantages over vnc [13:23] u can forward usb over that ... which u cant do with vnc [13:23] copy paste things are also possible over that... so the default choice seems fine to me [13:24] yes looks like spice is superior to vnc [13:24] im using virtualbox for that matter... but i dont like oracle... and i thought switching to kvm as well ;D [13:25] virtualbox was my favourite too but I think kvm is more secure and smoother [13:25] wouldnt be suprising considering its from oracle :D [13:25] I found that I need to passthrough pulse for guest to avoid latency [13:25] with kvm [13:25] but that's about it not bothered with GPU passthrough at this point [13:26] in some regard virtualbox is better [13:26] if if you setup right kvm is incredible [13:26] virtualbox is very noob friendly [13:26] im a kvm user already... i know :) [13:26] but i didnt use it on desktop yet [13:27] Ah. Actually not sure about the difference between a desktop and server OS in ubuntu [13:27] as in either can be made the other and all desktops are half servers [13:28] are the desktop kernel's different from the server ones [13:30] doesn't look like === thinkofher1 is now known as thinkofher === grimple_ is now known as grimple === jcnmark6_ is now known as jcnmark6 === westor{off} is now known as westor [14:02] hrrm, ubuntu 21.04 on ausus x201e netbook -- probelm seems to be in kernel 5.11 with blackscreening unless 'nomodeset' used... === jrm2 is now known as jrm [14:02] manually installing debian LTS kernel 5.10.0-6 debian package ... booting into that, 'solves' issue [14:02] wondering if 21.04 users of various versions seen this sort of thing, I wonder what erms to best look for bug === blackest_mamba_ is now known as blackest_mamba [14:05] enyc, probably it totally depends on which graphics adapter you have .... specifically nvidia is known to get problems now and then because they are not very supportive about opensource [14:15] bluesceada: this case something like 3rd gen intel integrated graphics [14:17] Hello, say I'm logged in through ssh to a laptop and someone else is also using the computer "normally" (as in, logged in through the gnome session and not remotely). Can I launch app on the user's screen and session? [14:17] Like, can I launch firefox on the user's screen? [14:20] Ah I got it `DISPLAY=:0 gnome-terminal` === ghoti_ is now known as ghoti === bildramer1 is now known as bildramer [15:30] Antoine_: indeed, also often better 'env DISPLAY=:0 ...' === loser_walex is now known as Walex === sauvin_ is now known as Sauvin [15:33] jvava: DPKG does (mostly) packages, APT does dependencies among packages, and then uses DPKG. But there is for example a version of APT that uses RPM as the package manager instead of DPKG. === lotuspsychje_ is now known as lotuspsychje === Scotty_Trees2 is now known as Scotty_Trees [16:19] Walex: There's a version of APT that uses RPM? :) I knew there was a version for Microsoft Windows, but RPM?! :) [16:32] jo-erlend: yeah, apt-rpm https://en.wikipedia.org/wiki/APT-RPM -- I used it on yellowdog linux once upon a time [16:32] Yellowdog doesn't come up often. :) [16:33] jo-erlend: apt-cache search and so on where nice to have, but the entire process of apt-get update && apt-get dist-upgrade was pretty slow compared to using apt+dpkg on debian .. [16:33] indeed :) hehe [16:43] loser_walex: Is it so that I edit the environment only for that command? [16:44] I'd like to install the package that includes "biblatex-software". How can I know which package I need to install? === ace_me1 is now known as ace_me [16:46] Antoine_: apt-cache search is the usual tool [16:48] sarnold: I thought it looked only in the name of the packages [16:49] apt search biblatex-software returns texlive-bibtex-extra. Does it mean texlive-bibtex-extra includes biblatex-software? [16:49] Antoine_: apt-cache search by default searches descriptions, but if you've disabled downloading those (maybe disabling translations? I forget, I did it once and immediately regretted it..) [16:49] Antoine_: so, once you've got some possible choices, you can try apt-cache show texlive-bibtex-extra or apt-file show texlive-bibtex-extra [16:50] Antoine_: apt-file show requires configuring a new service, but it's *really* useful, so it might be worth takinmg the time to set that up [16:50] Oooh, very nice. I didn't know apt show [16:51] So I confirm that biblatex-software is included in texlive-bibtex-extra :) [16:58] What is the difference between installing a library via apt and via pip3? [17:00] Do I get libraries that are better tested through apt? [17:04] Antoine_: I'm not super-good at this one.. they're different. installing the pip3 ones migth break software you've installed via apt. installing apt ones might not work for software you install via pip. I think the think is to always use virtual_env when using pip and *never* use pip to install things "outside" a virtualenv [17:04] Antoine_: but it'd be best to find advice from someone who actually knows pythonland [17:06] hi [17:07] sarnold: Ok I see :) [17:07] Thanks for explaining! [17:26] Antoine_, best to ask in #python [17:30] Antoine_, definitely +1 to what sarnold said: NEVER use pip to install things systemwide, you will break a bunch of stuff. Always use virtualenv or similar. [17:34] antoine: I've even used pip3 outside a venv for dependencies which I feel I will reuse for everything. Never ever use pip3 as root though and try to avoid installing the same packages via pip3 and apt [17:35] antoine: worst case scenario just find the package directory in your user directory and rename it python3 will create a fresh one and your renamed one can be used for selective diagnostics or as a pointer to reinstalling oackages you need via pip [17:40] Habbie: yay, thanks for the confirmation [17:40] :) === dingir__ is now known as dingir_ [18:22] Is there a way to limit memory/ram usage on a certain process/application? [18:25] Intelo: several, none of them great; ulimit -m in your shell can set kernel rlimits for child processes; you can also set cgroup resource limits, systemd is the easy way to get there https://www.freedesktop.org/software/systemd/man/systemd.resource-control.html#Memory [18:30] hm [18:37] are there known issues with the recent bionic grub2/shim update? [18:38] i have two identical haswell boxes, only difference is the serial numbers and that one has a prev revision mobo [18:39] after running a normal apt-get upgrade one of them seems to lock up after selecting any grub menu item [18:39] livecd, windows all boot fine [18:40] selecting the default menu item just leads to a black screen, booting in recovery mode gets stuck on "loading inital ramdisk" [18:41] how to --exclude multiple path with rsync ? [18:41] i've also taken the drive from the system box, booting either drive leads to a hang in one box, booting either drive works just fine in the other [18:41] system/sister box [18:43] shibboleth: this is ~nine days old and not much to go on https://bugs.launchpad.net/ubuntu/+source/grub2-unsigned/+bug/1927857 [18:43] Launchpad bug 1927857 in grub2-unsigned (Ubuntu) "Update to grub 2.04-1ubuntu44 breaks booting" [Undecided,New] [18:43] i'd be happy to do some further troubleshooting, problem is that i get no indication what's wrong [18:45] hi there, which package provides make-kpkg in 21.04? i connot find kernel-package anymore [18:45] Ky0sh1r0: it was removed; iirc the removal bug said that the kernel's makefile targets are better [18:45] I figured my problem [18:45] sarnold, this applies to bionic? also, the issue isn't the same, grub starts/i get the menu [18:46] shibboleth: dang :/ [18:46] shibboleth: with some recent grub changes, we're pushing the same package through multiple releases now [18:46] sarnold: thx [18:49] i pulled a drive from a retired box (xenial), boots fine [18:49] also, also: i did a memtest just to be sure, no issues [18:53] robertparkerx, rsync has an exclude flag, add paths to a "list-file" [18:55] sarnold, just for sh*ts and giggles: if i was to boot from a livecd and chroot into the installation, how would i use apt to pull the previous grub/shim and install it? [18:55] shibboleth this https://paste.gg/p/anonymous/0e95f7eae1ed40a18910695253b7dc9f still copy folders like strands/images [18:56] the exclude list paths must correspond to pwd [18:58] shibboleth: it might already be in /var/cache/apt/archives -- you can also pull updates with apt get install grub2=versionumber commands, if the packages are still on the archives. (they 'age out' when they aren't referred to by any of the package lists) [18:59] shibboleth: if the old ones are already gone from the mirrors, https://launchpad.net/ubuntu/+source/grub2-signed and https://launchpad.net/ubuntu/+source/grub2-unsigned are I think the packages to use to find the old binaries on launchpad [19:07] i'll try that, but... is there any way to get some debug output from grub to determine what goes awry? [19:08] if the prev version works it will be a tough sell to file a bugreport with.... "it just hangs" === ghoti_ is now known as ghoti [20:09] I can't seem to figure out how to change the size of my status bar text, like the size of the clock font. [20:09] Do I really have to edit the userchrome.css file for that? [20:17] shibboleth: I don't recall seeing any 'usual' grub debugging stuff; and yeah, "it just hangs, here's what it said" is really unsatisfying as a bug report, but we've got to start somewhere [20:52] Hello. Je galère à trouver un outil pour transformer des adresses mails en "codage" pour éviter les scans de bots [20:52] Si quelqu'un a ça sous la main, merci [20:55] floown: on parle anglais ici [20:55] floown: tu peux me demander en perso [20:55] ah :-) [21:00] sorry = [21:02] floown: pas de probleme [21:02] on dit monsieur: [21:02] "Pas de problemes" [21:02] pluriel [21:03] oui, d'accord [21:06] When I run "ssh -X " i am able to run x11 apps, but not via sudo. I only get an error regarding auternication. How can i get arround this? [21:06] error: https://paste.ubuntu.com/p/fMMnCnR7TB/ [21:11] Aavar: running X applications requires having the XAUTHORITY environment variable available and pointing to a file that the process can read, in addition to permissions to connect to the X11 socket [21:17] I'm building various in-tree kernel modules. Each one, when I run "modprobe module-name", results in "could not insert 'module-name': Exec format error". [21:18] remline: check dmesg, there may be more detailed errors there [21:20] remline: the error message itself means "The binary image supplied in module_image is not an ELF image, or is an ELF image that is invalid or for a different architecture." [21:20] (from man 2 finit_module) [21:22] sarnold: There are. "Skipping invalid relocation target, existing value is nonzero for type 1...". I have read the corresponding source code in kernel/module.c, but the objdump of these modules seems fine. [21:23] remline: does dmesg provide more details? [21:23] sarnold: dmesg lists that "skipping invalid relocation target" message. [21:24] sarnold: objdump reveals file format elf64-x86-64, same as the .ko that comes installed from Ubuntu. [21:24] sarnold: dmesg provides already too much, you have a security breach, ideal for hackers to use these infos. [21:25] unixbsd: feel free to lock it down if you wish [21:26] sarnold: Also, objdump -h shows what seem to be perfectly nice section headers for my .ko. [21:27] remline: weird. are you building the kernel modules with the same compiler, linker, etc, as the main kernel? [21:28] sarnold: I originally had the problem on machine 1, so I am now testing in a live environment on machine 2 with all stock packages and configurations. [21:28] sarnold: Perhaps I am missing something simple... cp /boot/config... .config; make modules_prepare; make M=... [21:30] uname is 5.11.0-16-generic, gcc is 10.3.0 [21:32] Hi, I've just installed 18.04 on a VPS with the clean installer (mini.iso) and the guided disk encryption. How can I make it automatically decrypt upon restart? [21:32] remline: do you have the necessary linux-headers-* packages installed too? iirc that's how struct randomization and so on stuff is passed around [21:32] hyperknot: just curious, what does that provide you over an unencrypted disk? [21:33] sarnold: I'm only doing it so that no traces are left on the shared host after me [21:33] hyperknot: aha, cool [21:34] sarnold: Yes, I do have the linux-headers packages installed. I had no troubles in groovy, so either I forgot how to do this (entirely possible!), or something changed in hirsute that I need to account for. [21:36] hyperknot: the /etc/crypttab file can have a command or a file placed in one of the fields to read the key from one of those, rather than require manual entry.. I'm not entirely sure how you'd manage the file, if it perhaps can live in the initramfs or something similar.. [21:36] remline: dang. I was hopeful for that one :/ [21:36] sarnold: my problem is that at this time only GRUB or something is loaded, isn't it? [21:37] hyperknot: I'm not entirely sure :/ I've only done one encrypted boot machine, and I type the password every boot :) heh [21:38] hyperknot: see /etc/cryptsetup-initramfs/conf-hook [21:39] and maybe the steps here https://help.ubuntu.com/community/Full_Disk_Encryption_Howto_2019#Post-Installation_Steps [21:40] this how-to is actually for encrypted /boot, so a different scenario than you'll have, but still similar enough to be useful, i think. [21:40] tomreyn: I went through the howto but it confuses me that there are so many parts which I didn't have to do with the 18.04 installer [21:40] I just followed the installer and everything was done for me [21:40] because you didn't want encrypted /boot [21:41] sure [21:44] hyperknot: this looks simpler https://askubuntu.com/questions/996155/how-do-i-automatically-decrypt-an-encrypted-filesystem-on-the-next-reboot#answers [21:45] sarnold: Might there be some new aspects to module signing in hirsute that could trigger this error? [21:45] does it usually take awhile to import sql from command line when it is large >10gb [21:46] It has been like 15-20 minutes [21:46] remline: if the signature is invalid I thin kthat'd return a ENOKEY 126 Required key not available [21:46] tomreyn: thanks! how can I see which partition is encrypted? [21:47] wrong channel [21:47] dev/mapper/vda5_crypt is in fdisk -l, is this the one I need to decrypt? [21:47] or it's /dev/vda5 [21:50] sarnold: Yeah, that would make sense. My module is unsigned, and hopefully that is still allowed, even though it will "taint" the kernel. [21:51] remline: unsigned --> depends on your secure boot settings; if secure boot is enabled, module signatures will be enforced [21:51] sarnold: I'm on BIOS, so I've assumed there is no secure boot. [21:51] remline: aha, probably a safe assumption [21:52] sarnold: The only other difference I have found between my built module and Ubuntu's build module is that Ubuntu's has .BTF and .gnu_debuglink sections. [21:53] remline: btf encodes type information for ebpf programs, it probably isn't mandatory [21:53] remline: same for gnu_debuglink, I don't actually know what that one is used for, but it sure has the sound of an optional thing :) [21:54] sarnold: My first impression is that its a way to strip debug symbols from the .ko, but still have debug symbols available elsewhere (a "separate debug info file") [21:55] remline: hmm. I wonder, could you use the dkms system for your module? I've never tried making a module rebuild via dkms, but that's a very common way to get built-on-the-computer modules to load [21:55] hyperknot: /dev/vda5 [21:56] thanks [21:56] hyperknot: /dev/mapper/vda5_crypt is the unlocked device [21:56] sarnold: I did investigate dkms, but found that the module I'm building (radeon) has too many drm dependencies to handle in dkms. [21:57] remline: oh bugger [22:02] sarnold, confirmed [22:02] sarnold: Thanks for brainstorming this with me. I think I'll install groovy in a live environment and see if my memory that building modules worked is correct. :) [22:02] shibboleth: the older grub worked, newer grub failed/ [22:03] remline: I'm happy to be a rubber ducky, sometimes I even guess correctly :) [22:03] i loaded a previous filesystem snapshot and did a grub-install /dev/foo, thereby restoring the previous version of grub/shim [22:03] hi. am I to take the fact that thinkbook 14 g2 itl is certified ubuntu hardware as a very good sign that I can buy https://www.dustinhome.no/product/5011207521/thinkbook-14-g2-itl-20vd and just install ubuntu lts via usb, and just have everything magically work? [22:03] (it is a version of the thinkbook 14 g2 with only iris xe, no nvidia.) [22:05] sarnold, 2.02-2ubuntu8.21 works, 2.02-2ubuntu8.23 just locks up while (i assume )loading the initramfs [22:05] bernalex: https://certification.ubuntu.com/hardware/202008-28154 .. it says it was certified using the oem kernel; that may or may not be 'easy' to do during install [22:06] now, as for why this only happens on one of two identical boxes: i have no earthly idea [22:06] heh, you anticipated my next question.. [22:06] shibboleth: any chance you can file a bug report on this? [22:06] sarnold: care to elaborate? [22:06] i don't have an lp account and i have no idea what to report [22:07] shibboleth: just a "2.02-2ubuntu8.23 fails on my hardware, the last lines are ...; 2.02-2ubuntu8.21 works fine" kind of note [22:07] bernalex: the specific variant being sold there https://psref.lenovo.com/Detail/ThinkBook/ThinkBook_14_G2_ITL?M=20VD003EMX [22:07] bernalex: there's an -oem flavor of kernel that's got a bunch of additional drivers [22:08] tomreyn: it's not for sale here unfortunately. [22:08] sarnold: should there be any problems installing ubuntu, and then installing that kernel? [22:09] bernalex: not sure what you mean there. i'm saying this is the very model which they sell at https://www.dustinhome.no/product/5011207521/thinkbook-14-g2-itl-20vd [22:10] bernalex: I'm not sure, I *hope* the installer makes it easy to select between them, but I haven't actually seen the installer in a few years [22:10] Rawr [22:10] tomreyn: ah, right. I misunderstood. then I don't know what you are trying to tell me. :) [22:10] sarnold: hah. I haven't installed ubuntu since it was called dapper drake... [22:11] bernalex: hehe [22:11] bernalex: that was just a free added service ;-), in case you want to double-check the specifications. [22:12] tomreyn: ah, I see. the specs are pretty detailed on the site I linked too. but there's always stupid stuff that should work that doesn't work, so I thought I'd get some hands on experience from this chan, if any available. [22:13] bernalex: right. the model you chose is a different one than the one that was certified. [22:13] tomreyn: yes, I know. [22:14] :) ok [22:14] this one doesn't have any nvidia dedicated gfx. [22:15] my old laptop died and it's kind of depressing how I have to pay the same amount today that I did in 2013, and even have to suffer downgrades in some areas. [22:16] actually, I'll have to pay lots more for this one than my old dead laptop, which I bought on sale. [22:17] I guess I'll have to get this one and pray that it works with ubuntu. -_o_- the only laptop I can find with similar specs and a similarly good gnu+linux story is a dell xps 13 9305, but that's about 1.4 of the price, and would take 30 days before shipping. [23:18] sarnold: Well, I'm perhaps proven wrong. I can't get my compiled module to insert on groovy, either. modprobe reports "exec format error", although there is no corresponding error in dmesg. [23:20] remline: argh. I wonder if it used to work on groovy and something changed, or if you're perhaps remembering an earlier release when things were easier [23:22] remline: What module? [23:23] jeremy31: For experimenting, I'm trying several "in-tree" modules. The actual module of interest is radeon. [23:24] remline: If you are trying modules from other kernels that will happen [23:26] The vermagic of the module needs to match the kernel or you get exec format error [23:27] remline: which Radeon/AMD GPU are you using? You're doing all this stuff because some problems with your GPU? [23:28] jeremy31: Indeed, that is a strange bit, too. There is some confusion between Ubuntu's 5.8.0-25 and upstream's 5.8.14. [23:29] tatertots: Fairly ancient: RV515. There is a timer bug that I like to patch manually. [23:30] remline: is that is external to the kernel, it is easier to download the source code of the kernel you have and do it [23:31] cf. bug 1791312 . I'm sure people would like to fix the bug, but there are too many hardware variations to play whack-a-mole on. [23:31] Error: Could not gather data from Launchpad for bug #1791312 (https://launchpad.net/bugs/1791312). The error has been logged [23:31] ubottu, what happened!? [23:31] I am only a bot, please don't think I'm intelligent :) [23:47] Hmm, probably relevant is that CONFIG_MODULE_FORCE_LOAD is not set for the ubuntu kernel. Maybe that means that modprobe -f is useless to avoid the vermagic error. [23:57] remline: ah i see, well good luck in your endeavors