=== mrbutthead9 is now known as mrbutthead
[01:41] <LuckyMan> does ubuntu comes with a default ssh key pair?
[01:42] <LuckyMan> I seem to have a key and I can't remember creating it
[01:43] <rbox> no
[01:43] <rbox> that would be ridiculous
[01:43] <LuckyMan> lol
[01:44] <LuckyMan> well I might have created it for some reason
[01:44] <rbox> probaly so you could ssh to somewhere with a key
[01:44] <LuckyMan> it might have been an incomplete configuration to access my raspbery pi
[01:45] <LuckyMan> I never put the public key on it
[01:46] <LuckyMan> I'm sure it wasn't for github, I just checked
[01:49] <LuckyMan> how do I delete it?
[01:50] <rbox> you mean aside from using rm?
[02:48] <demeplais> Hello all
[02:50] <demeplais> i need help
[02:53] <LuckyMan> demeplais, what do you need?
[02:53] <nelgin> He just said, help :)
[02:53] <LuckyMan> lol
[02:54]  * LuckyMan sends some help to demeplais 
[02:54] <LuckyMan> I hope that solved it!
=== mrbutthead5 is now known as mrbutthead
=== rue__ is now known as ruenoak
=== Non-BEAST is now known as Non-ICE
=== Juesto is now known as Juest
=== mrbutthead1 is now known as mrbutthead
=== not_phunyguy is now known as phunyguy
=== NewtonTrendy is now known as NewtonPumpkin
[08:38] <yziquel> apt-key is undergoing deprecation in 24.04. however, methods/gpgv.cc in apt package source code still tries to use apt-gey in GPGVMethod::VerifyGetSigners(). or so it seems. anyway, this breaks my sudo aptitude update with an "Unknown error executing apt-key" when I follow instructions for the docker repo. how do i fix that ?
[08:42] <mgedmin> what hapens if you use apt-get instead of aptitude?  I don't think aptitude is supported on ubuntu
=== diogeness_ is now known as diogeness
[08:42] <yziquel> mgedmin ah ? not supported ? guess it's a leftover of my old 2000s debian days, then...
[08:42] <yziquel> i'll try apt-get.
[08:42] <mgedmin> I mean, aptitude is in universe
[08:43] <mgedmin> and over time apt gained many of its features
[08:44] <yziquel> mgedmin sudo apt-get update seems to follow the same code path as i get the same 'Unknown error executing apt-key'. i like the aptitude TUI, though.
[08:52] <yziquel> mkay. this is going to be hellish to debug. oh, well...
=== kimiamania2 is now known as kimiamania
[09:43] <fweht> my vscode keeps crashing, is there any logfiles i can see more details?
[09:44] <abdelhak> facing almost the same issue
[09:44] <fweht> journalctl has nothing helpful
[09:45] <abdelhak> when open up vscode + postman + firefox, my laptop just freezes
[09:45] <fweht> ah, for me its when i try to open a certain directory
[09:45] <fweht> its weird because i just navigate into the directory, im not even open it with vscode, so im not sure if the error comes from vscode or from the file selector
[09:46] <fweht> but in a different application, like firefox, when i try to open the same directory, it works
[09:51] <abdelhak> @fweht there are few issues on the vscode officail repo in Github addressing a freezing and/or crashing problems in linux
[09:52] <fweht> thank you!
[09:52] <abdelhak> you are welcome
[09:53] <abdelhak> also this issue appeared a week or two ago
[09:54] <abdelhak> and for my freezing problem, i found out few solutions, one of them is to increase the swap space
[09:54] <abdelhak> but i didn't try it yet
[09:58] <fweht> for me its really the file selector interface.  when i open the folder from the terminal, it works, like `code-insiders /path/to/folder`
[10:02] <Rhonda> I have the issue that a dist-upgrade from 22.04 to 24.04 fails.  On most systems it worked flawless, but on some it doesn't.  Right now it failed "dpkg-divert: error: cannot divert directories" on libc6 install.  When I look into the preinst of the new package, it wants to divert /lib64, which is an actual directory on that system.
[10:02] <Rhonda> Does anyone know how to fix that issue?
[10:09] <abdelhak> sorry, this my first install of the distro, and i installed the latest version
[10:20] <taeaad> Does dconf-editor work for Xfce?
[10:39] <fweht> my vscode crashes with a lot of lines like this (where the number counts up), what does it mean? https://www.irccloud.com/pastebin/ZU1fzTJe/
[10:40] <mgedmin> Rhonda: why is /lib64 an actual directory on your 22.04?  it's not a standard configuration, do you know what created it?  does it belong to some package?
[10:40] <mgedmin> on my 22.04 /lib64 is a symlink to /usr/lib64
[10:41] <mgedmin> dpkg -S /lib64 says it belongs to the libc6 package
[10:41] <Rhonda> mgedmin: I don't know, no.  It's historic, and /lib also exists as directory and not symlink to /usr/lib
[10:42] <Rhonda> With the help of busybox I got that fixed for /lib64, but then on the next step it breaks with a segfault in the new postrm script.
[10:43] <mgedmin> ouch
[10:47] <mgedmin> btw when you say dist-upgrade, I hope you actually meant do-release-upgrade?
[10:49] <Rhonda> No, I don't, because do-release-upgrade doesn't support upgrading to 24.04 yet because there is no 24.04.1 release yet
[10:51] <mgedmin> um, are you from an alternate reality?
[10:52] <Rhonda> I don't know, am I? :)
[10:52] <mgedmin> curl https://changelogs.ubuntu.com/meta-release-lts | grep 24.04.1
[10:52] <mgedmin> 24.04.1 was released a month ago
[10:52] <Rhonda> Oh, when did that happen without me noticing. :)
[10:53] <mgedmin> (and prior to that you could upgrade to 24.04 by doing do-release-upgrade -d)
[10:53] <Rhonda> I'm so used to doing manually the sed -i -e s/jammy/noble/g that I haven't checked whether that's not needed anymore. :)
[10:53] <sixwheeledbeast> I wouldn't do that
[10:54] <sixwheeledbeast> 24.04.1 is released you should do-release-upgrade
[10:54] <Rhonda> sixwheeledbeast: Yes, I just found out about that.
[10:54] <sixwheeledbeast> it was released and then recalled but it's now available again.
[10:55] <lotuspsychj3> and make a backup before upgrade :p
[10:55] <Rhonda> Oh. apt install usrmerge  # now the symlinks are proper  \o/
[10:55] <sixwheeledbeast> even so you can force the 24.04 upgrade with the upgrade tool
[10:55] <mgedmin> huh, I thought usrmerge was part of the standard install
[10:55] <Rhonda> lotuspsychj3: In case it really breaks I'm going to reinstall, things are in puppet.  But reinstall is more cumbersome on that hoster so I'd like to avoid that.
[10:55] <mgedmin> at least my changelog has no record of me manually installing it
[10:57] <mgedmin> ah, hosting providers, let me tell you about a /usr/sbin/run-parts shell script that tries to add a random delay to stagger cron jobs before exec'ing the real /bin/run-parts
[10:57] <Rhonda> It's also priority required, so I'm also unsure why it wasn't on there.
[10:58] <Rhonda> mgedmin: yeah, we got all weird stuff at times. :)  But that's one of the few corner cases installs somehow
[10:59] <mgedmin> this badly-written script was using the IP address as a random seed, but wasn't expecting hostname -i to start returning IPv6 addresses, which happened after a do-release-upgrade to 22.04
[10:59] <mgedmin> so then every day cron started emailing me bash syntax errors until I figured out the reason
[11:01] <Rhonda> haha
[11:02] <Rhonda> Well, we had "issues" with libc6 upgrade "breaking" puppet runs for postgres.  It checks some things, but wasn't aware of the collate related changes that postgres 16 no warns about, producing additional output and thus didn't know what to do. xD
[11:03]  * Rhonda hugs apt install usrmerge tightly.  It seems to have fixed the issue.  Or maybe do-release-upgrade additional magic on top of that.  Either or, they both are golden.
[11:03] <mgedmin> yay
[11:04]  * mgedmin was desperately trying to remember the package name, but thought it was usrmove, couldn't find any trace of it in apt policy/apt search, and mistakenly decided that it was maybe no longer used/needed/no longer even existed
[11:06] <Rhonda> xD
[11:10] <fweht> i narrowed my problem somewhat down, vscode crashes when i open a folder in the file selector (like not fully open it in vscode, just navigating into the folder) and it happens only when there is a valid json file in the folder.  so my question, is this related to vscode or to the file selector (which would be part of the operating system)?  i assume vscode has no access to files during the selection process?
[11:11] <fweht> but why would the file selector even parse a json file before opening it?  thats so weird
[11:25] <skraitoz> hi all
[11:25] <skraitoz> :)
[11:25] <skraitoz> how is ubuntu
[11:31]  * skraitoz so no one here , find me at https://ascension.gg elune as skraitox then see ya all ... .
=== LuckyMan is now known as Guest8683
[11:56] <fweht> can i downgrade a snap package when i reinstalled it in the meantime, i.e. i have no cache of older versions?
[12:07] <BluesKaj> Hi all
=== noze` is now known as noze
=== HER is now known as hernan
=== hernan is now known as HER
[13:59] <brenndo> tox release on echo listen to physical devicecommands if not its remote ... sudo apt install tox    habs im apt angehängt
[14:01] <brenndo> so sieht ein virenprogramm aus das braucht nichmal interface und is unbeatable
[14:04] <brenndo> echo hide
[14:04] <brenndo> echo listen
[14:05] <brenndo> alles im paket tox angehängt
[14:07] <brenndo> haha zB als copyinfo
[14:11] <brenndo> `
[14:11] <lotuspsychj3> !ops | brenndo
[14:11] <ubottu> brenndo: Help! Channel emergency! (ONLY use this trigger in emergencies) - CarlFK, DJones, el, Flannel, genii, hggdh, ikonia, krytarik, mneptok, mwsb, nhandler, ogra, Pici, popey, sarnold, tomreyn, Unit193, wgrant
[14:12] <brenndo> expected expression
[14:13] <brenndo> ok ask i help
=== RandomZero5 is now known as RandomZero
[15:25] <yziquel> is there any way to get apt-get to fetch packages not from an HTTP url but from an unix socket ?
=== fling_ is now known as fling
[16:43] <tomreyn> yziquel: the sources.list(5) man page lists supported URI types. socket is not one of them.
[16:57] <hggdh> question: is IPv6 supported for the Ubuntu repo mirrors in Azure? Or where can I ask about it?
[16:58] <ravage> i guess for azure infra you would need to ask Microsoft
[17:00] <ravage> https://p.haxxors.com/e8kkkd4w.txt
[17:00] <ravage> but it looks v4 only
[17:07] <hggdh> ravage: no, not really. Canonical uses the Microsoft CDN; how the CDN is used is Canonical's responsibility
=== fling_ is now known as fling
[17:08] <ravage> well. the IP i get is owned by Mictosoft
[17:09] <ravage> also this is a pure community support room. so if it is managed by Canonical you would seek support by them
[17:10] <hggdh> the IP is owned by MSFT cuz it is under Azure. But yeah, I need to find who to ask this in Canonical
[17:11] <hggdh> (also keep in mind that ALL *.archive.ubuntu.com are managed by Canonical)
[17:11] <ravage> only because there exists an alias does not mean the infrastructure it leads to is managed by the same entity
[17:13] <CosmicDJ> hggdh: what? that is simply not true. Almost all archive mirrors are not operated by Canonical, see https://launchpad.net/ubuntu/+archivemirrors
[17:14] <ravage> at least the country mirror URLs mostly are not
[17:14] <ravage> de.archive.ubuntu.com just because thats one i know for sure it not
[17:16] <ravage> *is not
=== catties is now known as kitties
[17:33] <enigma9o7> I'm curious how ESM support works, I thought security issues were typically fixed, but I see 18.04 is listed as 'vulnerable' and not vulnerable/wip or needs eval, etc.  https://ubuntu.com/security/CVE-2024-47176
[17:33] -ubottu:#ubuntu- CUPS is a standards-based, open-source printing system, and `cups-browsed` contains network printing functionality including, but not limited to, auto-discovering print services and shared printers. `cups-browsed` binds to `INADDR_ANY:631`, causing it to trust any packet from any source, and can cause the `Get-Printer-Attributes` IPP request to an attacker controlled URL. D... <https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2024-47176>
[17:34] <enigma9o7> This is one of the CUPS related CVE's from yesterday I just decided to follow up on out of curiosity.
[17:34] <enigma9o7> "ESM provides 10 years of security updates for Ubuntu Main packages and 23,000+ Ubuntu Universe packages, including additional security updates from Canonical for critical and high priority CVEs. "
[17:37] <enigma9o7> I did try "sudo pro fix CVE-2024-47176" and it says no fix available.
[17:37] -ubottu:#ubuntu- CUPS is a standards-based, open-source printing system, and `cups-browsed` contains network printing functionality including, but not limited to, auto-discovering print services and shared printers. `cups-browsed` binds to `INADDR_ANY:631`, causing it to trust any packet from any source, and can cause the `Get-Printer-Attributes` IPP request to an attacker controlled URL. D... <https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2024-47176>
[17:37] <enigma9o7> although I don't think that's necessary when doing regular updates anyway.... but it's listed in the status info so I tried it.
[17:41] <topcat001> I never use (and disable) the auto detection features of cups.
[17:43] <topcat001> I tend to also disable avahi when possible. It is another vector.
[18:05] <tomreyn> enigma9o7: "Security updates for ESM releases will be released shortly." https://ubuntu.com/blog/cups-remote-code-execution-vulnerability-fix-available
[18:06] <tomreyn> those will likely need to be backported, which can involve a relevant development effort.
=== codyshepherd_ is now known as codyshepherd
=== Hobbyboy|BNC is now known as Hobbyboy
=== nightstrike_ is now known as nightstrike
=== isnessness is now known as ugjka
=== rpittau_ is now known as rpittau
=== effortDee_ is now known as effortDee
=== Menzador_ is now known as Menzador
=== sa_ is now known as sa
[18:15] <tomreyn> enigma9o7: also, the specific cve id you mentioned (2024-47176) was evaluated as medium priority which is not (guaranteed to be) covered by ESM according to the "What’s covered?" paragraph on https://ubuntu.com/security/esm
[18:16] <tomreyn> ... which may contradict the statement more to the top of the same page - "ESM enables continuous vulnerability management for critical, high and medium CVEs."
[18:22] <mgedmin> "In this instance, the coordinated disclosure date had to be moved up, balancing the need for adequate preparation and delivering updates as soon as possible. This affected our plan to simultaneously release security updates for all Ubuntu versions, including ESM."
[18:27] <pikaciu> Hello ? Is anybody here ?
[18:28] <tomreyn> pikaciu: yes, there are people here
[18:29] <pikaciu> Hello! I have a suggestion regarding the default software included in Ubuntu's ISO. Many users appreciate the usefulness of Neofetch for quickly displaying system information. Would it be possible to replace some less-used applications, like the GNOME games, with Neofetch in future releases? It would improve customization and provide users with more relevant tools. Thank you for considering this!
[18:32] <JanC> I'm not sure Neofetch is the best tool for that really?
=== kitties is now known as catties
[18:32] <tomreyn> pikaciu: (primarily volunteer driven) IRC is not a good place to post feature (or change of default) requests. you could maybe file a bug report against neofetch for this purpose. but i'd not get my hopes up too much that is would replace other software on this
[18:33] <tomreyn> i assume this ends the conversation?
[18:35] <JanC> if you want to discuss this more, maybe #ubuntu-discuss is a better place also
[18:36] <tomreyn> they left, unless you meant me, then i'm aware, thanks.
[18:37] <JanC> yeah, I didn't see until after  :)
[18:37] <chuckyoufarlie> I can ssh to a slurm host, and from there run -
[18:37] <chuckyoufarlie> srun -p gorman-gpu --x11 --gres=gpu:1 --pty /bin/bash
[18:37] <chuckyoufarlie> to get a shell on a gpu node. From that shell on the gpu node, I can run glxgears and other X apps.  But, when I try to run emacs, it fails with -
[18:37] <chuckyoufarlie> ; emacs
[18:37] <chuckyoufarlie> (emacs:3839301): Gtk-WARNING **: 13:51:56.544: cannot open display: localhost:90.0srun: error: _half_duplex: wrote -1 of 1748
[18:38] <chuckyoufarlie> The node is running Ubuntu 22.04.2. I've updated emacs.
[18:40] <chuckyoufarlie> Any help would be appreciated
[18:41] <JanC> sounds like you are trying to run GUI version of emacs?
[18:42] <tomreyn> and in parallelization?
[18:43] <tomreyn> package emacs-nox provides the non GUI variant
[19:28] <TygerTung> hello
[19:30] <chuckyoufarlie> @JanC @tomreyn I'm trying to help a user who wants to run emacs on a gpu-node to debug his code.
[19:31] <tomreyn> chuckyoufarlie: that's kind of you.
[19:31] <chuckyoufarlie> I've told him not to do that.
[19:31] <chuckyoufarlie> I told him to just run gdb in the shell. But he's more comfortable with emacs.
[19:32] <tomreyn> <tomreyn> package emacs-nox provides the non GUI variant
[19:32] <chuckyoufarlie> Also to note, I can ssh directly to the node (because I set it up. Not all users have direct ssh access to nodes) and run emacs via X.
[19:32] <chuckyoufarlie> I'll see if he will run emacs in the shell. He probably won't like it.
[19:35] <tomreyn> i have not worked with slurm / srun, so i can't immediately tell why glxgears and other X apps seem to start on an X display which is available, and the emacs GUI won't
[19:35] <chuckyoufarlie> I'm more concerned with why other X apps work, but not emacs.
[19:35] <tomreyn> maybe you just need to export the right DISPLAY
[19:36] <chuckyoufarlie> But it's already exported. I mean, I can run other X apps and 'echo ${DISPLAY}' returns - localhost:90.0
[19:36] <dvitek> I have encountered a substantial slowdown (~1000%) running python3 -c "import timeit; print(timeit.Timer('for _ in range(0,1000): pass').timeit())" on Cisco UCS servers when upgrading from ubuntu21 to ubuntu22.  Disabling hyperthreading in the BIOS seems to reduce the slowdown to "only" 40% or so instead of 1000%.  The same experiment on other
[19:36] <dvitek> hardware shows the upgrade makes no difference.  Anyone seen anything like this?
[19:37] <dvitek> I have confirmed on about a dozen cisco ucs servers (rack and blade) and a dozen supermicro (rack) servers.
[19:38] <dvitek> every version <= 21 is fast, every version >= 22 is slow
[19:39] <dvitek> even containers running old python binaries on 22 exhibit the 22 slowness, so I don't think it's in userland
[19:40] <chuckyoufarlie> @dvitek are you running python3 in both cases?
[19:40] <dvitek> yes.  I even have fixed python binaries compatible with various OSs so that I can get apples to apples comparisons
[19:41] <chuckyoufarlie> Have you tried running python3 with strace?
[19:41] <dvitek> and even say a ubuntu18 container on a ubuntu22 host is slow
[19:41] <dvitek> yeah, it makes something like 250 syscalls in 112 seconds on the slow systems
[19:41] <dvitek> I've used perf, the time is spread around with no smoking guns
[19:41] <dvitek> barely any context switching
[19:42] <chuckyoufarlie> Are the python3 binaries linked against the same libraries?
[19:42] <dvitek> when using the ubuntu18 container, yes, even glibc should be identical
[19:43] <dvitek> even ran the experiment with various livecds to confirm it could repro in a totally clean environment
[19:43] <chuckyoufarlie> And, to be clear, you're only seeing this on the Cisco UCS hosts?
[19:43] <dvitek> correct
[19:44] <dvitek> and "gets much better" once hyperthreading is disabled, even though HT did not have this effect on old versions
=== Sidewyz1 is now known as Sidewyz
[19:45] <tomreyn> try mitigations=off
[19:45] <dvitek> I had someone on my team try that already.  also tried changing various kernel scheduler parameters back towards what they used to be in ubuntu21 -- no dice
[19:46] <chuckyoufarlie> Have you tried taking a host that is already version ≤ 21, updating it, then doing an in-place upgrade to 22?
[19:46] <dvitek> All the upgrades were done as clean installs.  None of the <= 21 installs exist anymore, but I can boot off installation media to run experiments.
[19:47] <dvitek> also, sysbench's cpu test suffers -- 25% slowdown -- but nowhere near as badly as this python test
[19:54] <dvitek> I haven't isolated the difference since they both seem cpu bound, but I imagine the python test is doing a bunch of junk on the heap that sysbench would avoid
[22:42] <pedahzur> Is there a quick and easy way to "export" a set of PPAs to which you have subscribed to another machine? I have a bunch of files in my /etc/apt/sources.list.d/ Instead of finding all those third-party repos, and running apt-add-repo again, it would be nice to be able to "export" on my old system, and "import" on the new system.
[22:42] <ravage> Copy the files?
[22:43] <tomreyn> that'll miss the signing keys
[22:43] <pedahzur> ravage: Yeah, I can do that, but I'll miss the keys, as tomreyn observed.
[22:43] <tomreyn> but you can copy those, too
[22:44] <ravage> But maybe check if you need all those PPAs
[22:44] <ravage> If you use more than 3 something is maybe not right :)
[22:44] <tomreyn> if they're PPAs, though, then "apt-add-repository ppa:this/that" can actually be easier / faster
[22:45] <pedahzur> True. Maybe I should just start keeping a list. And, not all the repos have [signed-by=...] in their .list file, so finding the key might be a pain.
=== sam__ is now known as Guest3604