[00:06] <guiverc> imi, `sudo apt full-upgrade`   (upgrade has restrictions on what it can upgrade; thus `dist-upgrade` and `full-upgrade`) - same as with debian
[00:06] <imi> thanks
[00:07] <imi> what are the restictions? why are the restrictions there? is it safe to use apt full-upgrade without those restrictions?
[00:09] <guiverc> restrictions can vary on release.. best to refer to `man` which will show manual ref page for your actual release (and your preferred command, apt, apt-get etc)
[00:10] <guiverc> apt particularly has been improving so more recent releases can do more...  `apt-get` is pretty static so release matters far less
[01:20] <metbsd> server iso cannot boot up the frame buffer
[01:20] <metbsd> is there a text only installet iso?
[01:20] <metbsd> someone gave me before i forgot
[01:31] <cheater> sarnold: hey, are you around by any chance?
[01:42] <metbsd> server iso cannot boot old hardware
[01:49] <Maik> metbsd: what is old hardware? And if it's 32 bit, that's not supported anymore.
[01:49] <guiverc> netboot iso
[01:52] <Maik> those do not exist anymore afaik
[01:56] <parabyte> this is powered by 4 pin molex power connector and is x86 based
[01:56] <parabyte> anyone know what form factor this board is?
[01:56] <parabyte> https://paste.pics/434e6495ad51a8814ef04e687263f535
[01:59] <Maik> parabyte: probably wrong channel to ask since it's not a ubuntu related support question
[01:59] <parabyte> Maik, I guessed that, i was just hoping there was someone on here who knew exactly what the board was intended for etc
[01:59] <Maik> simple
[02:00] <Maik> !discuss
[02:00] <Maik> ;)
[02:00] <Maik> night peeps, i'm off
[02:29] <zealsham> if i have a list of subdomains and i want grep for particular string like "NXDOMAINS"  when i run the dig on it and if the grep matches that particular subdomain should be written to a file. how do i do that in bash
[02:34] <sarnold> zealsham: for d in foo.com bar.com baz.com google.com ; do dig "$d" | grep NXDOMAIN && echo "$d" > "$d" ; done
[02:35] <zealsham> thank you @sarnold
[02:44] <chrometiger> Freshly installed ubuntu 20.10 on my pi.  no audio out of the headphone jack.   is there any fix to this. also I see that I dont have raspi-config  and apt-get doesn't notice it.  anyone know how I can fix these 2 problems?
[02:46] <sarnold> raspi-config may be specific to raspbian
[02:48] <Maik> sarnold: it is
[02:49] <Maik> on the headphone jack, try switching the output in Sound Settings
[02:50] <metbsd> its 64 bit hardware but its bios only
[02:50] <metbsd> server iso halts
[02:50] <metbsd> dont know why
[02:51] <chrometiger> Maik: thats the thing, in settings it doesn't give my that option
[02:53] <chrometiger> I have Multichannel Output - built-in Audio listed twice ....  Headphones Built-in Audio .....  and Analog Output built-in Audio   none of which work
[03:00] <sarnold> chrometiger: dang :( I didn't try any of the audio on my 3b+ stuff, so I'm not sure ohw well it's supposed to work.. we've got a bunch of their firmware blobs though, I had the impression it ought to work
[03:57] <cheater> sarnold: do you know of a good way we can get ubuntu to fix that bug in the info package in 18.04?
[03:57] <sarnold> cheater: the fastest way to get it fixed is to figure out what changed between 18.04 and 20.04 to fix it, then backport that fix to the 18.04 package
[03:58] <cheater> i'm unable to backport it.
[03:58] <sarnold> cheater: this describes the process of getting fixes into releases https://wiki.ubuntu.com/StableReleaseUpdates
[03:58] <cheater> the best i can do is to tell the maintainers somehow.
[03:59] <sarnold> ah; you can do that part by 'ubuntu-bug info'
[03:59] <cheater> oh. thanks!
[04:33] <jckfrst1> anyone running ubuntu touch?
[04:34] <Bashing-om> !touch | jckfrst1
[04:36] <jckfrst1> !ass
[05:01] <jep> Hi everybody ... greetings from Denmark, Europe ... I'm running Virtual box on OSX Catalina host and installing Ubuntu server 20 as guest .. I've installed the following packages on Ubuntu server like this: sudo apt-get install dkms build-essential linux-headers-generic linux-headers-$(uname -r) ... after that I've mounted the guest editions CD and successfully run: sudo
[05:01] <jep> /mnt/cdrom/VBoxLinuxAdditions.run  ... it is installing guest additions ... no errors  ... rebooting and nothing happens ... what am I doing wrong ???
[05:06] <cheater> what's a good way to make apt-cache search only print package names, and not the description?
[05:10] <cheater> never mind - found a better way to do what i want to do
[05:18] <sarnold> cheater: if you're still trying to get all the -doc packages installed.. I had an idea earlier, but it made some stupid mistakes along the way...
[05:18] <sarnold> cheater: I had to run off to deal with something.. and now i'm trying to clean up that mess
[05:18] <cheater> oh i found a way to do it now
[05:18] <sarnold> cheater: so this isn't as useful as I wanted, but maybe it's useful dpkg -l | awk '$1 = "ii" {print $2"-doc";}'
[05:18] <sarnold> ah, good
[05:19] <cheater> one sec
[05:20] <cheater> sarnold:  https://pastebin.com/k6E41jXL
[05:21] <sarnold> cheater: aha! nice
[05:21] <cheater> i keep my full desktop config inside one ansible playbook
[05:21] <cheater> it's over 1k lines now...
[05:21] <sarnold> I *never* use comm, I don't use it often enough to get the hang of it, and because I don't get the hang of it, I never go to it first, heh
[05:22] <cheater> i just googled "bash find lines that exist in both files"
[05:22] <cheater> :)
[05:23] <sarnold> see I'd wind up doing something silly like cat a b | sort | uniq -c | awk '$1 ~ /2/ { print $0;}' except I bet the awk needs fiddling and who knows what else is broken with it...
[05:24] <cheater> yeah that wouldn't work :)
[05:32] <jep> Hi
[05:33] <cheater> sarnold: i reported the bug using ubuntu-bug btw, thanks a lot.
[05:33] <cheater> sarnold: and there's a workaround, you can just download the info file that's missing from gnu.org.
[07:13] <TomTom> good morning. maybe today somebody has a good idea. the current realtek eth driver has a problem during phy probing and requires that the module gets removed and added again and the link being reset. where to put these commands so that every boot starts up networking only after these commands?
[07:18] <bibble> TomTom: which kernel ? 5.10 may not have any problems.
[07:19] <TomTom> bibble: i am currently using 5.10.15 and it helped for two or three reboots, but from this morning the PHY part is missing again.
[07:19] <TomTom> `WantedBy=network.target` should be executed before networking? right?
[07:19] <catbeard> https://docs.fedoraproject.org/en-US/quick-docs/understanding-and-administering-systemd/#common-service-parameters could modify the Before/After in networking.service
[07:21] <devslash> im using ubuntu 20.10. i have 2 monitors. secondary on the left and primary on the right. every time I boot up Ubuntu, it shows the login prompt for Ubuntu on the secondary monitor. does anyone know how to fix that other than unplug the 2nd mointor which id rathrter not have to do
[07:22] <bibble> TomTom: multiple times Bluetooth has vanished mysteriously, but shutdown, switch off PSU, switch off mains, wait a few minutes. It reappears for some reason. Very strange.
[07:23] <bibble> Wonder if motherboard caching magic can something mess things, if they worked previously.
[07:29] <TomTom> in my case its also a quite old board. probably also a bios thing. but i've found multiple entries about the PHY probing problem for RTL chips also using newer boards... i try the way of reloading the module and setting the link status before the networking target. i need the reboot to be reliable and having a reachable network afterwards. standing in the basement and turning it on/off until it works is not an option B-)
[07:55] <coredumb51> Good morning
[07:56] <coredumb51> How could I debug shorcuts not activating??
[07:56] <coredumb51> I've set several shortcuts on alt+shift+[1-0] but 8 and 9 are not working for some reason
[07:57] <coredumb51> gives alt+shift+( and ) on my keymap
[07:57] <coredumb51> Any idea??
[08:45] <ash123> hello, I am having frequent screen freeze issues on my laptop with ubuntu 20.04. I have to restart the laptop everytime it freezes. How can I know the root cause?
[10:14] <eeos> Hi! An external USB disk formatted ext4  gets mounted /media/user/"label" but mounted by root, so the user cannot actually use it ....
[10:14] <eeos> Is that a design choice?
[10:48] <metbsd> i didn't know that ubuntu iso is so sensitive about usb writer
[10:49] <metbsd> looks like etcher is the only one that works
[10:49] <liowenex> Use DD
[10:49] <metbsd> rufus don't work anymore
[10:53] <JivanPal> Rufus doesn't just write the image to the drive, it extracts the files and places them in a partition on the stick, and does different things for different distros based on problems people have had trying to get USB sticks to work with those distros' installers
[10:53] <JivanPal> No need for that these days
[10:54] <JivanPal> Etcher just dd's the image to the drive and verifies the checksums match afterwards
[10:54] <metbsd> dd for linux and etcher for windows.
[10:56] <JivanPal> Would still recommend Etcher on Linux because it checks the checksum of the drive content matches the source file automatically afterwards
[10:56] <metbsd> yah
[10:56] <metbsd> i'll stop using rufus now
[10:57] <metbsd> not reliable
[11:11] <coredumb51> How could I debug shorcuts not activating??
[11:11] <coredumb51> I've set several shortcuts on alt+shift+[1-0] but 8 and 9 are not working for some reason
[11:11] <coredumb51> gives alt+shift+( and ) on my keymap
[11:11] <coredumb51> Any idea??
[11:11] <ikonia> manutally run the shortcut command
[11:11] <ikonia> check the log ?
[11:12] <coredumb51> command works fine
[11:12] <coredumb51> same command is used for all the other keymaps
[11:12] <coredumb51> just ( ) that don't activate
[11:12] <eeos> hi! An external usb HD with ext4 partitions is mounted under /home/user/media .... but has root.root ownership, and the user cannot make any modifications. Is this a design choice?
[11:13] <coredumb51> it "beeps" feels like the shortcut tries to do something else ...
[11:13] <coredumb51> same issue with ctrl+alt+(
[11:13] <coredumb51> ctrl+alt+) works fine though
[11:14] <eeos> coredumb51: are there some core key bindings connected to the bindings you are trying to overwrite?
[11:15] <coredumb51> eeos if there is I've missed them...
[11:15] <coredumb51> also it's supposed to tell you when you overwrite one right?
[11:16] <eeos> coredumb51: I suppose so ....
[11:17] <coredumb51> can't fathom what's wrong :/
[11:17] <eeos> coredumb51: by any chance, could this be a problem of keyboard setting?
[11:17] <eeos> coredumb51: you know, wrong keyboard layout identified ....
[11:17] <coredumb51> it works all fine
[11:17] <coredumb51> it detects and set the shortcut fine as well
[11:20] <coredumb51> so
[11:21] <coredumb51> shift+alt+( and ctrl+alt+( don't work
[11:21] <coredumb51> shift+alt+) don't
[11:21] <coredumb51> ctrl+alt+) does
[11:22] <coredumb51> everything else custom is fine
[11:25] <ikonia> ~~~~~
[11:27] <eeos> wait a sec coredumb51 .... it sould be ctrl + shift+alt+( ....
[11:28] <eeos> depending on your keyboard there is no such a thing as ( without shift
[11:29] <eeos> coredumb51: have you tried ctrl + alt + 9 and ctrl + alt + 0? do they work?
[11:29] <eeos> coredumb51: (depending on your keyboard layout, of course)
[11:30] <eeos> My question: when an external usb HD with ext4 partitions is connected, it is mounted under /home/user/media .... but has root.root ownership, and the user cannot make any modifications. Is this a design choice?
[11:30] <coredumb51> eeos yes 9 and 0 work
[11:32] <coredumb51> eeos by bad it's indeed ctrl+alt+8 that doesn't work
[11:34] <westor> maybe you need to remove -r and allow non-registered users come in ?
[11:45] <eeos> westor: are you answering my question?
[11:45] <westor> no i'm talking to #ubuntu moderators
[11:46] <westor> a friend had a problem he cannot connect until he registered his name on freenode
[11:49] <jeremy31> westor: have them use webchat.freenode.com
[11:50] <westor> yes he trying via https://webchat.freenode.net/
[11:50] <westor> error: Cannot join channel (+r) - you need to be identified with services - see https://freenode.net/kb/answer/registration
[11:50] <jeremy31> westor: https://freenode.net/kb/answer/registration
[12:28] <raecarruth> hello
[12:28] <raecarruth> can i get wine to recognize a usb device?
[12:28] <raecarruth> or rather, how do i go about getting wine to recognize a usb device?
[12:35] <kill-animals> hello
[12:39] <kill-animals> how can I list all directories containing two or more files?
[12:41] <cbreak> find . -type d -links 4
[12:41] <cbreak> increase the number until you're bored
[12:42] <cbreak> (or write an exclude pattern to not list directories with 1 or 0 files)
[12:43] <cbreak> maybe something like  find . -type d -not \( -links 2 -or -links 3 \)
[12:54] <Towser> I forgot my crypt password, how can I get it to boot again
[12:58] <Towser> I can't boot up because it asks for a crypt password
[13:21] <BluesKaj> Howdy folks
[14:03] <cbreak> towser: did you consider deleting everything and reinstalling?
[14:24] <bart-> good afternoon
[14:24] <bart-> i have a simple question is it possible to use the touchscreen in the terminal too?
[14:46] <toothe> playing around with an ubuntu docker container for the firs ttime....I don't have ip
[14:46] <toothe> is there a way to install it?
[14:46] <toothe> nevermind...got it...
[14:48] <kyentei> I did an Ubuntu 20.04 server install, then added the KDE Neon repositories on top and installed neon-desktop. However, my brightness controls don't seem to be working. I've tested both the 5.4 and 5.8 HWE kernel, the mainline 5.10.15 kernel, and I've tried all the acpi_backlight kernel options. Backlight works in Arch, and I think on Ubuntu live as well. I'm using the amdgpu graphics drivers. Does anyone have a recommendation where I could look next?
[14:48] <kyentei> I should add; manually echoing values to /sys/class/backlight does not seem to be working either
[14:49] <ikonia> why didn't you just install kubuntu
[14:49] <kyentei> And install neon-desktop on top of that, you mean? I figured starting from a clean minimal base would be better.
[14:50] <ikonia> why would it be easier
[14:51] <ikonia> you've installed a server targeted OS - installed parts of a desktop setup then expecting desktop functionality
[14:51] <ikonia> why not just install kubuntu, the desktop setup that's fully setup and ready to go
[14:51] <ikonia> and then work any problems from there
[14:51] <kyentei> I don't want Kubuntu, I want KDE Neon.
[14:52] <ikonia> isn't kubuntu shipping with KDE and neon as default ?
[14:52] <kyentei> No. KDE Neon is the latest KDE Plasma on top of Ubuntu LTS. Kubuntu is not.
[14:52] <kyentei> And the KDE Neon installer (calamares) doesn't allow for custom partitioning. Or at least, not without crashing.
[14:53] <ikonia> I'm looking at the repos now and the versions seem pretty close on packages
[14:53] <ikonia> I guess if you want support with the neon packages, contact the neon guys
[14:54] <kyentei> But it's not neon though, is it. I am just missing some part of the Ubuntu default packages to allow proper brightness controls and I'm asking if someone else has suggestions that I haven't thought of yet. it has nothing to do with KDE Neon specifically
[14:54] <ikonia> you don't know that though
[14:54] <ikonia> as you don't know what the neon repos pull in
[14:55] <ikonia> and you're using bleeding edge stuff
[14:56] <kyentei> Sure I do, I've looked at what I installed. It's not bleeding edge for anything other than KDE Plasma. Besides that, it's all Ubuntu 20.04
[15:03] <ikonia> and it's not supported
[15:03] <ikonia> I'd looke at what the kubuntu-desktop package pulls in
[15:03] <ikonia> and see what you're missing
[15:05] <kyentei> I'll have a look at that indeed, thanks
[15:22] <swear000> hi
[17:27] <Monotoko> Aight, I'm confused: https://pastebin.com/uVJU3SGY
[17:27] <Monotoko> file, owned by me
[17:27] <Monotoko> *folder
[17:27] <Monotoko> doesn't give me permissions to access it
[17:28] <Monotoko> root can access, I've done chown -R monotoko:monotoko local
[17:28] <Monotoko> and yet my account still can't get to it
[17:40] <arooni> xbacklight -set 50 ;; doesnt seem to do anything no ubuntu 20.04
[17:40] <arooni> newly installed;
[18:28] <ash123> hi all! I installed Ubuntu 20.04 on my pc. Sometimes my laptop's touchpad doesn't work. I have to restart everytime this happens. Can you please help?
[18:29] <arooni> xbacklight on ubuntu 20.04 doesnt seem to do anything
[18:30] <arooni> intel integrated gpu on my lenovo t420
[18:31] <oerheks> ash123, how about the urls i gave you yesterday?
[18:31] <oerheks> !logs
[18:32] <ash123> I didn't find any such settings in my bios
[18:33] <oerheks> maybe there is a bios update available, did you check?
[18:34] <ash123> for freezing issue, i set idle=nomwait so far from: https://askubuntu.com/questions/1047374/hp-envy-x360-ryzen-5-status-of-known-issues
[18:37] <ash123> and regarding bios, I dont have windows 10 anymore so not sure how to update.
[18:38] <oerheks> depends on the vendor of the motherboard
[18:39] <oerheks> dmidecode -t 2
[18:39] <oerheks> or dmidecode -t baseboard
[18:42] <ash123> https://pastebin.com/CPUFLFPY
[18:46] <oerheks> hp, one can put an update on usb/fat32 and run it from the biosmenu, AFAIK
[18:48] <ash123> one thing i noticed when touch pad doesn't work is that Ubuntu doesn't recognize the hardware i think. I couldn't see touchpad in settings.
[20:19] <SirOsisOfThelive> Hi, I installed ubuntu server 20.  i guess i used lvm to make 1 drive out of 4 600 Gig drives.   i thought i told it to use all drive space.   lsblk -f shows only 1.46G FSAVAIL   how do  i access the rest of the drive space?
[20:33] <Sayona> Hi, what repo do you use to install php71,72,73,74 ?
[20:43] <tomreyn> Sayona: none that's supported here. but Ondřej Surý provides a PPA for this purpose https://deb.sury.org/ https://launchpad.net/~ondrej/+archive/ubuntu/php
[20:43] <Sayona> thanks tomreyn  :)
[20:44] <Sayona> is safe to use ondrej nu?
[20:44] <tomreyn> SirOsisOfThelive: you can create LVs, logical volumes, then file systems on those.
[20:45] <tomreyn> Sayona: maybe. "not supported here" ;-)
[20:45] <Sayona> :P
[21:10] <LiveWire> I was tyring to run some video benchmarks, installed the 3070 finally lastnight.. For some reason driver issues from 1660S to it, would of thought it was pretty auto. 3dmark will run, but takes like 10minutes now to get system info..
[21:10] <LiveWire> damn it, wrong window yet again
[21:13] <LiveWire> but while im here, can someone tell me why when I open Chrome I have to enter a damn password. Ive cleared my basic pw from keyring
[21:19] <WaV> https://askubuntu.com/questions/31786/chrome-asks-for-password-to-unlock-keyring-on-startup
[21:20] <LiveWire> WaV: ty, i actually googled quite a bit before I asked
[22:25] <sybariten> hey! (l)ubuntu 18 here, on a rather fresh machine. Live OS though. And having mysql install problems (from apt). Here's the end of the install process, with an error: https://paste.ubuntu.com/p/KGXP4dh39N/
[22:26] <oerheks> interesting, installing mysql in a live iso..
[22:27] <oerheks> ?
[22:28] <sybariten> oerheks: yeah, well
[22:28] <superschnell> sybariten: you'll have to find out why /usr/share/mysql/mysql-systemd-start pre  failed
[22:28] <Habbie> sybariten, line 19 has two hints
[22:29] <sybariten> superschnell: so would thatr be something that existed _before_ i started installing mysql?
[22:29] <Maik> what use is it to install it in a live session anyway....
[22:29] <Habbie> sybariten, no, it wouldn't
[22:29] <superschnell> sybariten: I have no idea what you mean
[22:29] <oerheks> Maik,  one needs a hell of a mem for that
[22:29] <superschnell> sybariten: it starts, it fails
[22:30] <sybariten> Habbie: the service is, uh, loaded and active:failed .... dont really knwo what that means
[22:30] <oerheks> seriously, did you take a daily iso or just the download?
[22:30] <Habbie> sybariten, often one of those commands will tell you also why
[22:30] <oerheks> http://cdimage.ubuntu.com/bionic/daily-live/current/HEADER.html
[22:30] <oerheks> else go figure out what you are doing :-D
[22:30] <sybariten> oerheks: me? It's ... hmmm ... i cant remember. I was just sdome 18.04 lubuntu iso
[22:31] <sybariten> or how do you mean?
[22:31] <superschnell> sybariten: don't be distracted by the package manager messages, it just ends by reporting a prerequisite failed, but that's just from the package manager's perspective. It wants to install X, then Y, but can't install Y because X failed and is required for Y. Look at why X failed, not Y.
[22:32] <superschnell> mysql-server-5.7 has a "post installation script", and that means that after unpacking some files to your root filesystem, it attempts to do stuff. One of those things is starting a program. It failed. It showed you what program it tried to start. You can now try to find the reason in logs mentioned or try to start it yourself (possible with a parameter that says it should run in the foreground, not
[22:33] <superschnell> the background, see its man page) and see what it says.
[22:33] <superschnell> Until, we know little more than you do.
[22:33] <superschnell> Until then*
[22:34] <sybariten> superschnell: aha aha....  the journalctl said this, but i dont know if theres anything there that raises an eyebrow  https://termbin.com/pmrp
[22:34] <superschnell> Modern LiveCDs no longer have the same problems precipitated from a readonly medium that they used to, however, it isn't out of the question that the unusual mixed read-only/writable filesystem merger yields some unforeseen issues. But we won't know till you check.
[22:35] <Habbie> sybariten, there's no mysql in there
[22:35] <superschnell> ok
[22:35] <sybariten> Habbie: no i know...
[22:35] <sybariten> Maik: if it works its practical when you want to do things fast
[22:35] <superschnell> sybariten: try /usr/share/mysql/mysql-systemd-start pre
[22:35] <superschnell> Just guessing
[22:35] <TJ-> sybariten: the script that failed should output some info, see its source-code https://git.launchpad.net/ubuntu/+source/mysql-5.7/tree/debian/additions/mysql-systemd-start?h=ubuntu/bionic-updates
[22:36] <superschnell> sybariten: as root, obviously
[22:36] <sybariten> Hm! Weird. It has a permission error. Even though i sudo. Coming up..
[22:37] <TJ-> sybariten: what does this show? "journalctl -u mysql.service"
[22:37] <superschnell> sybariten: interesting.
[22:37] <superschnell> What error?
[22:37] <TJ-> That script is badly written; it ought to return different exit codes for each error
[22:37] <sybariten> https://paste.ubuntu.com/p/gfkKmMK3j7/
[22:38] <superschnell> sybariten: I suggest sudo su
[22:38] <superschnell> then trying again
[22:38] <sybariten> aha ok
[22:38] <sybariten> maybe i should had done that for the whole install...
[22:38] <sybariten> TJ-: hmmm "--- no entries ---" i think, if thats what you're after?
[22:38] <superschnell> Perhaps, I have no idea if this will work. It's curious in any cse.
[22:39] <superschnell> case*
[22:39] <TJ-> sybariten: ha, yes, but no, not helpful!
[22:39] <TJ-> superschnell: from what I can see in the source-code, the command that fails is "mysqld --verbose --help --innodb-read-only"
[22:41] <sybariten> superschnell: same error actually! Interesting.... cause i can definitely ls the dir and cat files in it.  https://paste.ubuntu.com/p/XQGN8tm9BB/
[22:41] <sybariten> How can things be out of permission when one is root?
[22:42] <superschnell> sybariten: a read-only mount?
[22:42] <TJ-> sybariten: how about "ls -ld ls -lrt /etc/mysql/conf.d"
[22:43] <Habbie> TJ-, double ls
[22:43] <superschnell> sybariten: I must say I don't really get it
[22:43] <superschnell> sybariten: you could use strace to find out. Perhaps the process tries to switch context and becomes a different user first
[22:43] <superschnell> Then it is no longer root, and fails
[22:44] <superschnell> do stat /etc/mysql/conf.d as root
[22:45] <superschnell> It should yield some useful information. Perhaps the daemon attempts to run as user 'mysql' and then is refused acces.
[22:45] <superschnell> access*
[22:45] <superschnell> Keyboard is irritating today. Or rather, my typing.
[22:45] <TJ-> Habbie: oops
[22:45] <sybariten> superschnell: hmmm i dunno, i think although it comes from a very special filesystem ( as in being read from a USB memory ) i think from the loaded OS's point of view it should behave like a normal OS ... more or less.... if you understand what i mean
[22:46] <superschnell> sybariten: I know, I used to be a livecd distro co-dev
[22:46] <sybariten> What i mean is, in my experience theres typically not a lot of things that are read-only (as opposed to normally) in a loaded live os
[22:46] <sybariten> superschnell: oh, cool! Yeah then you know
[22:46] <superschnell> But before I blame it on fs peculiarities, I'd like to narrow down the circumstances of the error step by step
[22:47] <TJ-> sybariten: the problem looks to be that the systemd service is running as the mysqld user and doesn't have access to that directory
[22:47] <superschnell> I haven't seen errors caused by that on livecd/liveusb systems for a while
[22:47] <TJ-> sybariten: "ls -ld   /etc/mysql/conf.d"
[22:47] <superschnell> So what TJ- said basically
[22:48] <TJ-> maybe the traverse mode is missing
[22:48] <sybariten> should i try a pure ubuntu 18 instead, on a usb memory, instead of lubuntu?
[22:49] <superschnell> sybariten: no, please just find out what the error is
[22:49] <sybariten> okidoki. Whats your tip for my next step?
[22:49] <superschnell> There is no need to jump to conclusions yet, we're just guessing at this point
[22:49] <sybariten> yeah i understand
[22:49] <superschnell> stat /etc/mysql/conf.d
[22:49] <Habbie> ls -ld /etc /etc/mysql /etc/mysql.conf.d
[22:49] <superschnell> One of those
[22:49] <TJ-> Could also be an apparmor issue
[22:50] <Habbie> dmesg
[22:51] <superschnell> And you can still eventually run strace after you've listed its uid/gid ownership and permissions
[22:52] <sybariten> so stat tells me some memory related stuff regarding a file?   https://paste.ubuntu.com/p/JdFmv2gm6T/
[22:52] <Habbie> oops
[22:52] <Habbie> ls -ld /etc /etc/mysql /etc/mysql/conf.d
[22:52] <Habbie> but, your output has all we need on that
[22:52] <sybariten> ah ofcourse.... yes one sec
[22:52] <sybariten> ok
[22:52] <Habbie> so permissions on dirs are not the problem
[22:52] <Habbie> the stat call covers the bit i typoed
[22:54] <superschnell> Well, at this point I'm inclined towards strace, but one might consider it overkill
[22:54] <TJ-> Looks to be related to Bug #1435368
[22:55] <sybariten> Habbie: okay yeah no its rw by root so thats good i guess?
[22:56] <TJ-> sybariten: try this: "sudo apparmor_parser -r -T -W /etc/apparmor.d/usr.sbin.mysqld"
[22:56] <Habbie> sybariten, mysqld may not run as root; but group and other have r and x so reading should not be a problem
[22:56] <superschnell> TJ-: this is 5.7 though
[22:56] <superschnell> might not matter
[22:56] <TJ-> sybariten: then try "sudo systemctl restart mysql"
[22:56] <TJ-> superschnell: I found that bug because it is referred to in the package's postinst script
[22:57] <superschnell> ok
[22:57] <superschnell> sudo systemctl stop apparmor
[22:58] <superschnell> ;-)
[22:58] <sybariten> superschnell: TJ- well apparmor is certainly something i dont have previous (explicit) experience of. The apparmor thing ran silently (i was root and didnt sudo), but the restart gave an error. Job for mysql.service failed because the control process exited with error code. See "systemctl status mysql.service" and "journalctl -xe" for details.
[22:59] <Lvl4Sword> What's the method to get alt codes to work on Ubuntu? I've got AltGR working, but it only accepts 2 characters
[23:00] <TJ-> sybariten: OK, so it isn't an obvious missing apparmor profile or doesn't appear to be (you forced its install and the mysql service still failed)
[23:00] <superschnell> sybariten: I see no issues disabling it, at least temporarily for purposes of installing mysql, certainly not on a live system
[23:00] <superschnell> TJ-: oh, did I misunderstand?
[23:01] <TJ-> superschnell: not that I can tell :)
[23:01] <superschnell> Well, should he temporarily disable it then, or?
[23:01] <sybariten> TJ-: ah, so that was what we were trying. Apparmor is something i can disable, and try an install after that?
[23:01] <superschnell> That might work?
[23:02] <TJ-> Go for it :)
[23:02] <superschnell> sybariten: it's just a live system, I don't see the harm. I don't see you falling victim to a 0-day exploit right now ;-)
[23:03] <sybariten> hehe ok. Yeah in theory the machine shoulnd even be exposed to the internet so ....  and i'm also only doing test stuff (on a machine that i reboot now and then) . So, systemctl stop apparmor ?
[23:03] <superschnell> If you're already root, yeah, why not
[23:03] <superschnell> Otherwise, put sudo in front
[23:04] <sybariten> But what is the thing i want to try after that - an install with apt-get as i did first, or this  /usr/share/mysql/mysql-systemd-start pre  ?  I still don't understand how much is yet installed
[23:04] <superschnell> apt-get -f install
[23:04] <superschnell> I'd try that first, we'll see what happens
[23:06] <sybariten> hmmm dependency problems....  https://paste.ubuntu.com/p/xbCps7m4CW/    does an apt-get -f install without any package name always go back to something i was doing previously ?
[23:07] <Habbie> sybariten, yes, -f will try to fix what is broken
[23:09] <superschnell> But that means switching off apparmor didn't work or the error is caused by something else
[23:12] <sybariten> yeah...
[23:13] <Sayona> Hi question: What do you preffer in ubuntu: mariadb or mysql?
[23:13] <TJ-> sybariten: I'll try it in a virtual machine
[23:13] <sybariten> TJ-: cool, i also have that on a win machine actually .... vmware
[23:13] <sybariten> with the same ISO i mean
[23:15] <TJ-> the only thing I can imagine is this is some kind of problem with the cow/tmpfs of the live image
[23:15] <oerheks> Sayona, tons of discussions and comparison like https://kinsta.com/blog/mariadb-vs-mysql/ -- https://www.guru99.com/mariadb-vs-mysql.html
[23:15] <oerheks> use the one you like best?
[23:15] <superschnell> TJ-: yeah
[23:16] <superschnell> TJ-: but it's unlikely on a standard live system, it's just weird
[23:16] <TJ-> all we need is an strace log
[23:16] <oerheks> Sayona, and crossposting polls, please don't
[23:16] <superschnell> TJ-: wouldn't it just show us error 13 again, or would it show us more?
[23:16] <superschnell> I guess we shouldn't assume
[23:16] <Sayona> thnks
[23:16] <Sayona> :P
[23:17] <superschnell> sybariten: did you choose "install" rather than "try" or something?
[23:17] <superschnell> I'm not aware that would really make a difference in this day and age, but o
[23:18] <superschnell> ok*
[23:18] <sybariten> oh, and for the record....  if theres a suspicion there would be an actual HW problem with the USB memory i'm running off....  i had the same problem last night, with another computer with another USB memory with the same lubuntu on. I did not dig as deep into that error yesterday ofcourse but it started much in the same way.
[23:19] <TJ-> it does seem to be apparmor, here's another instance https://github.com/kubernetes/minikube/issues/7906#issuecomment-619484044
[23:19] <sybariten> superschnell: at what stage? bootup ?  no i think i let it go past 30seconds this time so it should have done try. I mean i arrived at tdesktop just as usual, with zero interaction
[23:23] <TJ-> here's the upstream bug https://bugs.mysql.com/bug.php?id=82281
[23:24] <TJ-> sybariten: try this: "aa-disable /usr/sbin/mysqld"  then "systemctl start mysql"
[23:25] <sybariten> shoiuld i have aa-disable ?
[23:25] <sybariten> cause i don't , as an executable
[23:25] <TJ-> sybariten: ah, not by default
[23:26] <sybariten> apparmor-utils or something?
[23:28] <TJ-> sybariten: lets just try removing it! "apparmor_parser -R -T -W -v /etc/apparmor.d/usr.sbin.mysqld"
[23:28] <sybariten> what does that do?
[23:29] <TJ-> removes the apparmor profile for mysql
[23:29] <TJ-> this is to prove apparmor is the cause, so you an get the package configured, and then apt will be able to install packages again
[23:30] <sybariten> ok so lemme try the parser command first
[23:31] <sybariten> removal succeede but it said warning failed to create cache
[23:32] <superschnell> TJ-: any idea why stopping apparmor didn't work? I might learn from that. Not too familiar with apparmor
[23:32] <TJ-> that's fine, ignore that
[23:32] <TJ-> superschnell: did it actually get stopped though?
[23:33] <superschnell> TJ-: well his output was silent, but I suppose we don't really know
[23:33] <TJ-> Looking at the apparmor profile it has
[23:33] <TJ-> # Allow config access
[23:33] <TJ->   /etc/mysql/** r,
[23:33] <superschnell> Yeah, I believe the bug report you mentioned immediately
[23:33] <sybariten> TJ-: shall i try a new apt-get -f install , as root?
[23:34] <TJ-> but from the way I read the 'pre' action it wants to *write* to /etc/mysql/mysql.conf.d/<somefile> - if so that'd explain it
[23:34] <superschnell> Just that I like to drill these sorts of incidents into my mind, because I'm oldschool and I when I am root, I expect full access without some security toolkit interfering, it's just a tic or an anachronistic preference I suppsoe
[23:34] <TJ-> sybariten: just try "systemctl start mysql" which'll re-execute that 'pre' action
[23:35] <TJ-> superschnell: but it isn't YOU being stopped, it is the 'mysql' user :)
[23:35] <superschnell> TJ-: then it should be root as well, not user mysql, but that wasn't mentioned in the report, so presumable it runs as root
[23:35] <TJ-> superschnell: or more accurately, the mysqld process
[23:35] <superschnell> presumably* ffs
[23:35] <sybariten> TJ-: mmmyeap, no complaints now.
[23:36] <TJ-> sybariten: OK, so that confirms that the apparmor profile is the issue...  it doesn't affect 18.04 mysql-server install in an LXD container, and I'd assume both would pull in the same version from the archives which makes this weirder
[23:36] <TJ-> going to try the virtual machine now
[23:36] <sybariten> ok
[23:40] <TJ-> Reproduced :)
[23:41] <sybariten> Ah you get the same error? Well thats reassuring  :)
[23:43] <superschnell> TJ-: Does apparmor work the same way in your LXD container?
[23:43] <superschnell> sybariten: turns out you had no way of preventing this and it's a damn bug
[23:44] <superschnell> Well, bug, it's a failure to synchronise expecations between two software package development teams
[23:44] <superschnell> No less annoying though
[23:44] <TJ-> # Allow config access
[23:44] <TJ->   /etc/mysql/** r,
[23:44] <TJ-> oops, wrong one!
[23:46] <TJ-> OK, the problem is the overlayfs, apparmor is denying an open() on upper/etc/mysql/conf.d
[23:50] <superschnell> So that's a new bug?
[23:51] <TJ-> no, it is Bug #1804334  - I'm going to edit the bug title to point out this is a live / overlayfs issue
[23:51] <sybariten> superschnell: aha? A bug on which side, so to speak, do you think?
[23:52] <superschnell> sybariten: the problem is that apparmore doesn't allow access to a postinst script
[23:53] <superschnell> apparmor*
[23:53] <sybariten> Aha okay
[23:53] <sybariten> but.... this is then different when comparing a live OS vs an installed (l)ubuntu?
[23:54] <superschnell> It appears to, because the prefix path isn't in the apparmor config, looks like
[23:55] <superschnell> TJ-: correct me if I'm wrong though
[23:56] <superschnell> Isn't it a bit of a PITA to coordinate every release one makes with the apparmor team, or am I mistaken about that requirement
[23:57] <TJ-> sybariten: correct, because the Live env is a combination of a read-only file system image with an overlay tmpfs (RAM) to make it work
[23:57] <superschnell> If you circumscribe application and script permissions this narrowly, your circumscriptions effectively become a component of upstream development, do they not?
[23:57] <TJ-> no - it is up to the package maintainers to write and ship the apparmor profiles
[23:57] <superschnell> Right
[23:57] <TJ-> this is very much NOT an expected scenario!
[23:58] <TJ-> I've updated that bug report with a preface about why it occurs
[23:58] <superschnell> Hmmm, so you can still always install mysql with any sort of apparmor profile manually
[23:58] <superschnell> without*
[23:58] <superschnell> TJ-: nice work man
[23:59] <TJ-> we don't generally expect people to be installing a database in a RAM file system :)
[23:59] <sybariten> TJ-: aha. i see. And i guess none of this would be different if i tried with a ubuntu iso instead of lubuntu
[23:59] <TJ-> sybariten: correct