[14:57] <xsisec_> Hi Guys what is the difference from install cuda from https://developer.nvidia.com/cuda-downloads?target_os=Linux&target_arch=x86_64&Distribution=Ubuntu&target_version=22.04&target_type=deb_local or  sudo apt install nvidia-cuda-toolkit nvidia-cuda-toolkit-gcc?
[17:02] <xubuntu2d> Did anyone noticed that slack system tray no longer works in xubuntu 23.04?
[17:02] <xubuntu2d> Is there a place i can start debugging this?
[18:31] <sharky> hi folks, hoping someone can help. I upgraded from 22.10 to 23.04 and clicking "login" prompted a loop which blanks the screen briefly then pops the login screen back. sing recovery mode I've removed all autologin stuff from lightdm and commented out nopasswdlogin from /etc/group, which at least gives me a full password prompt, but when I enter my
[18:31] <sharky> correct password, it says it's incorrect. I've changed/confirmed it's correct using recovery mode, but it still doesn't work. Baffled, stumped, frustrated. Full thread here [https://askubuntu.com/questions/1467818/upgrade-to-xubuntu-23-04-cant-login-popup-loop-tty-login-info-doesnt-work]; any ideas really gratefully received. Thanks all.
[18:31] <sharky> *Using recovery mode.
[18:34] <sharky> I've not yet tried the "drastic measures" approach answer here [https://askubuntu.com/a/84697/68396] but I'm not 100% convinced it would do much anyway. Maybe it would...
[18:34] <diogenes_Vx15> sharky, try to create a new user and login as a new user.
[18:34] <sharky> ooh ok, good idea
[18:34] <sharky> presumably in recovery mode also?
[18:35] <diogenes_Vx15> try normal mode.
[18:35] <sharky> can you do that in the normal login prompt?
[18:36] <sharky> options at login are my username or "Other...". If I enter a new user "user2" and a password, it says "password incorrect" still
[18:36] <diogenes_Vx15> just hit Ctrl+Alt+f1/2/3... and there you do sudo adduser or sudo useradd
[18:37] <diogenes_Vx15> then you login there as a new user and run: startxfce4
[18:38] <sharky> at tty1/2/3 the first prompt is "Computername login: _"
[18:39] <sharky> per the GUI issue, if I enter my existing user & pw, it says incorrect
[18:39] <sharky> also doesn't work for user=root
[18:40] <sharky> but feasibly I can do this in recovery:
[18:40] <sharky> useradd username -m -s /bin/bash
[18:40] <sharky> passwd username
[18:40] <sharky> adduser username sudo
[18:40] <sharky> ??
[18:40] <sharky> (thanks for the ideas btw)
[18:41] <diogenes_Vx15> the man page says thay it's recommended to use adduser.
[18:41] <diogenes_Vx15> that^
[18:44] <sharky> https://askubuntu.com/questions/161074/i-accidentally-deleted-administrator-account-and-the-other-accounts-are-asking-r#161075 per this I'm getting "the user 'user2' does not exist", so I'll try useradd now
[18:45] <sharky> ok done, rebooting
[18:45] <sharky> password incorrect.
[18:46] <sharky> something's seriously screwy with the password manager it seems... looks like it works normally in recovery mode, "mount -o rw,remount /" seems to work fine, everything confirms it's working, user2 was added to home and then to sudo.
[18:47] <sharky> is there any difference between tty1, 2, 3? I've never really known what the point of multiple tty's is. Anyway, original user & user2 passwords both don't work in tty nor login GUI
[18:48] <diogenes_Vx15> sharky, no difference but i think it woild be a good idea for you to backup your data.
[18:49] <diogenes_Vx15> eventually you might want to reinstall.
[18:49] <diogenes_Vx15> thay being said, i'm out, good luck.
[18:57] <sharky65> to remove my root password in recovery mode (set errantly), do I just follow this [https://askubuntu.com/a/84697/68396] i.e. nano -B /etc/shadow and delete the encrypted password between the first & second colons?
[18:58] <sharky65> I'm about to do this for my main and recently-created (likely not needed) second user, but just wondering if there will be any bad outcomes for doing it to the root account?
[19:03] <sharky65> Just removed passwords in shadow for main accounts, GUI login says "Failed to authenticate", password box greyed out, so that is having some effect.
[19:03] <sharky65> went back into recovery, set passwords for both users, save, exit, resume boot, login GUI, password incorrect. GAH.
[19:06] <sharky65> Anyone know if I can use an xubuntu 23.04 live USB to repair an install, is that a thing?
[19:17] <sharky65> possible progress, writing as much for myself as anything else: ages ago I tried to have my yubikey usable as a 1 touch login which never worked but evidently left deadly cruft in the system. Guide here [https://support.yubico.com/hc/en-us/articles/360016649099-Ubuntu-Linux-Login-Guide-U2F]. Have removed "auth       required   pam_u2f.so"
[19:17] <sharky65> from /etc/pam.d/sudo and it no longer says password incorrect, I've plugged in the yubikey, enter the password, tap the yubikey, and it does the screen blank loop. So. Progress, somewhat.
[19:30] <sharky65> ok, amazing, I can now login to tty but not to GUI still, it does the blank loop. So conceptually this is/was yubikey interfering with password PLUS possibly a display issue.
[19:37] <sharky65> following this [https://askubuntu.com/questions/1465176/login-screen-does-not-appear-after-upgrade-to-23-04] I already have nvidia and xserver-xorg-video-nouveau AND video-nvidia-470 installed so ostensibly no need to remove the nvidia drivers just yet, maybe just need to select the noubveau drivers while in tty...
[19:54] <sharky65> which is proving hard to do. lspci -vnnn says kernel driver in use: nvidia. I feel like I shouldn't need to purse nvidia drivers, just select the nouveau one...
[20:00] <lessiyana> I want to lead a lesbian in KDE #kde-lgbtq
[20:06] <lessiyana> I want to lead a lesbian in KDE #kde-lgbtq
[21:23] <sharky> I presume I didn't miss anything while I was offline again but please re-paste it if I did
[23:00] <Unit193> sharky: I'd be interested in what happened, fwiw.  Also, what pam modules do you have installed?
[23:01] <sharky> pam was indeed the login issue, related to yubikey
[23:01] <sharky> https://support.yubico.com/hc/en-us/articles/360016649099-Ubuntu-Linux-Login-Guide-U2F I commented out everything to do with pam_u2f
[23:02] <Unit193> Gotcha, seemed like the direction you were going.
[23:02] <sharky> but this just fixed my password issue, I'm now back at the login loop, can access & login to tty, purged nvidia driver, nouveau driver installed but not in use, seemingly no way to enable it.
[23:03] <sharky> https://askubuntu.com/questions/1467818/upgrade-to-xubuntu-23-04-cant-login-popup-loop-tty-login-info-doesnt-work full issue here, but asking a new one since it is/was 2 separate things.
[23:04] <Unit193> Nothing in /var/log/lightdm/ or .xsession-errors?  Did you try the ol' rm ~/.Xauthority ~/.ICEauthority trick?
[23:04] <sharky> new question here https://askubuntu.com/questions/1468208/xubuntu-22-10-to-23-04-gui-upgrade-login-loop-nvidia-drivers-purged-no-driver
[23:05] <sharky> I will now, one sec
[23:05] <Unit193> Ah, looks like the loop is before lightdm, meh.
[23:06] <Unit193> And just for kicks:  apt list ?config-files  list anything?
[23:08] <sharky> var log lightdm: session pid=1107: exited with return value 127. seat seat0: session stopped. stopping display server, no sessions require it.
[23:08] <sharky> (possibly cos Im
[23:08] <sharky> I'm in tty though?)
[23:09] <sharky> tough to tell where this loop occurs relative to the gui/tty experience.
[23:10] <Unit193> Right, but if it shows lightdm then it's not the drivers at least, even if you can't login to your user.
[23:10] <sharky> "m ~/.Xauthority ~/.ICEauthority trick?" - any further intel on this, it's new to me
[23:10] <Unit193> !xhangs
[23:10] <sharky> well I can login in tty...
[23:10] <Unit193> Dang, no link... :/
[23:10] <sharky> ok 1 sec
[23:10] <Unit193> Nah if lightdm isn't even showing up, then it's not that.
[23:10] <Unit193> Any feedback on the apt list command?
[23:11] <sharky> I should do the rm trick at root level?
[23:11] <sharky> haven't gotten around to apt list yet, sorry - no ability to copy & paste commands across machines ;)
[23:12] <sharky> apt list: anything I'm looking for in particular? scrolling now
[23:12] <Unit193> Hopefully not much output, but look for nvidia stuff.  Also, you could use termbin if you have internet.
[23:12] <Unit193> apt list ?config-files  are removed packages with files remaining.
[23:13] <sharky> libnvidia-compute-460/lunar 470.182.03-0ubuntu1 amd 64
[23:13] <sharky> libnvidia-compute-465/lunar 470.182.03-0ubuntu1 amd 64
[23:14] <sharky> lots more nvidia at the end also,
[23:14] <sharky> nvidia-compute-utils
[23:14] <sharky> nvidia-dkms
[23:14] <sharky> nvidia-kernel-common
[23:14] <sharky> nvidia-prime
[23:14] <sharky> nvidia-settings
[23:15] <Unit193> So yeah, there's leftover files.  My thinking here is that it's possible that the config that says to use nvidia is still present, but logs should say.
[23:16] <Unit193> If you're *sure* you can get rid of those,  apt purge ?config-files also works.
[23:16] <sharky> any concern that'll also purge all the other config-files?
[23:17] <Unit193> It'll absolutely purge all packages that have been removed but have files leftover, yes.
[23:17] <Unit193> (For packages still installed, it won't touch them.)
[23:17] <sharky> at this stage nvidia's not doing anything for me, so i'm not averse to chucking out all it's cruft. Notwithstanding it might have been nice to keep the settings in case I can reinstall it later.
[23:18] <Unit193> Last nvidia driver I used was 304. :/
[23:19] <sharky> purging now. eeks.
[23:20] <sharky> now wondering whether I should have tried to boot into an older linux kernel...
[23:20] <sharky> will reboot now & if it fails: "rm .{X,ICE}authority" do I just run this at root in tty?
[23:21] <Unit193> That'd be as your user, they're in your own home dir.
[23:21] <sharky> gotcha
[23:21] <Unit193> But that only helps if lightdm pops up, but your user throws you back to the login screen.
[23:24] <sharky> ah. backed up & deleted now, config file purge didn't change anything, will reboot and check to see if authorities did
[23:24] <sharky> then try oldest linux kernel
[23:25] <sharky> it just feels like there should be a way to request the gpu uses a specific installed module
[23:26] <sharky> no improvement, rebooting for older kernel
[23:27] <sharky> no change there either
[23:27] <sharky> alas
[23:34] <sharky> var log lightdm: "session pid=1107: exited with return value 127. seat seat0: session stopped. stopping display server, no sessions require it. Process exiting with return value 0. XServer 0: X server stopped. Releasing VT 7. Removing X server authority /var/run/lightdm/root/:0. Seat seat0: display server stopped."
[23:34] <Unit193> Nothing specifically is jumping out at me.  You could cat /var/log/lightdm/x-0.log | nc termbin.io 9999  and I could take a look, or you could ask in #ubuntu and hope the more exposure has better luck.
[23:34] <sharky> This feels like the doom loop. Session ending, giving up x server, the getting dumped back to login greeter.
[23:35] <Unit193> Sounds like there should be a log before that, of what went wrong.  Either in that, or x-0-greeter.log :P
[23:35] <sharky> permission denied, port 9999 name or service not known. possibly have to open port on my router?
[23:37] <Unit193> Urgh, .com not .io.  I always do that and not sure why, sorry.
[23:38] <Unit193> That'll pastebin your log and the link should be short enough to hand copy.
[23:38] <sharky> haha no worries, already googled it and found it was .com, same error, will look at my eero app. after downloading the eero app...
[23:39] <Unit193> Permission denied pasting, or trying to read the file as non-root?
[23:39] <sharky> ugh, silly sausage, didn't add sudo
[23:39] <sharky> https://termbin.com/hnir
[23:39] <sharky> what a wonderful service
[23:41] <sharky> "vesa: Refusing to run on UEFI" aligns with someone someone else said on a previous thread with no answer
[23:42] <sharky> https://www.linuxquestions.org/questions/slackware-14/vesa-refusing-to-run-on-uefi-again-4175687045/ this suggests nomodeset could be the cause, potentially I can edit my grub. now to remind myself where grub lives...
[23:43] <Unit193> You'll want to poke at /etc/default/grub and then update-grub (as root)
[23:44] <Unit193> Also cat /proc/cmdline  to see what you booted with.
[23:44] <sharky> yup, there now. nomodeset present. IDK what that relates to, presumably changed it based on a bug years ago and left it.
[23:45] <Unit193> ...Been there, done that.  "Do I still need this?  *Why* do I even have it?!"
[23:46] <sharky> booted with 5.11 kernel just to try it but removing nomodeset should apply to all kernels I select I believe. Let's find the F out!
[23:47] <sharky> trying 6.2 again now. splash screen is NOTABLY higher res.... so is login screen. Palpable excitement.
[23:47] <sharky> password entered. blank screen. login popup again. arsebiscuits.
[23:48] <sharky> tty is much higher res as well. weird that nomodeset has improved the res but it still won't login/work
[23:48] <sharky> tty login fine still
[23:50] <sharky> ok, lspci -vnnn|less shows gpu kernel driver in use: nouveau. Hooray! But somehow I guess this wasn't the issue then...?
[23:50] <Unit193> Check ~/.xsession-errors?
[23:53] <sharky> https://termbin.com/fgg3
[23:54] <sharky> "xfce4-session: symbol lookup error: xfce4-session: undefined symbol: xfce_screensaver_new" seems innocuous but is the last thing...
[23:54] <Unit193> xfce4-session: symbol lookup error: xfce4-session: undefined symbol: xfce_screensaver_new  Uhh...is something not updated?
[23:54] <Unit193> No, that's xfce4-session crashing.
[23:55] <Unit193> That sounds like a new xfce4-session without an updated libxfce4util, that symbol is new.
[23:55] <sharky> before the update, it gives the user a warning about lock screens. I wonder?
[23:55] <Unit193> LANG=C apt list | grep ed,loc
[23:55] <sharky> potenially I could update libxfce4util?
[23:57] <sharky> https://termbin.com/nxpcx
[23:58] <Unit193> ...Not what I was expecting, OK..  Uh, exo/now 2021.11.16  that's not a version string that ever had..
[23:59] <Unit193> apt-cache policy libxfce4ui-2-0 xfce4-session  ?