FatPhil | what's all this with /run/user/$UID tmpfss? Who creates those, and is it configurable? | 09:10 |
---|---|---|
FatPhil | Things that make you go hmmm... "/run/user/$uid is created by pam_systemd ...". | 09:18 |
FatPhil | "Should mention that it is called $XDG_RUNTIME_DIR, documented at standards.freedesktop.org/basedir-spec/basedir-spec-latest.html" | 09:22 |
FatPhil | 2 blanks here: | 09:22 |
FatPhil | root@thotspaz:~# find /etc -type f -print | xargs grep '/run/user/' | 09:22 |
FatPhil | root@thotspaz:~# find /etc -type f -print | xargs grep 'XDG_RUNTIME_DIR' | 09:22 |
FatPhil | and: root@thotspaz:~# dpkg -S pam_systemd | 09:23 |
FatPhil | dpkg-query: no path found matching pattern *pam_systemd* | 09:23 |
* drawkula would prefer eacxh user to have her own ~/tmp, ~/run, ~/var, ... and even ~/etc instead of those silly dotfiles | 09:47 | |
drawkula | soemeday, when I'm boared to death, I might try to prototype a system that way... | 09:47 |
drawkula | ist shure will not even be a sidenote in unix's history... | 09:48 |
drawkula | :-( | 09:48 |
drawkula | but the FS is self similar in several other places. I never understood why this had to stop in the users' dirs! | 09:48 |
drawkula | aaaaand a package manager for user packages would be nice to have too! | 09:49 |
FatPhil | having it host-specific does make sense, as I'm in ~ from 5 different hosts currently, and they musn't tread on each other's toes. | 09:49 |
helios21 | hi. well I think the FHS is pretty borked anyway | 09:50 |
drawkula | than have hostid subdirs... | 09:50 |
drawkula | .-) | 09:50 |
drawkula | I gave up sharing $HOME | 09:50 |
drawkula | too risky | 09:50 |
FatPhil | I think there's some smell no matter where you put such things. | 09:50 |
drawkula | someone said Kmail would trash all mail if you use a shaed home and dare to start 2 MUAs | 09:51 |
FatPhil | programs like firefox really don't cope well with the concept of shared home (and multiple instances of themself running) | 09:51 |
drawkula | but that was in KDE3 days | 09:51 |
drawkula | $THEY trashed unix years ago | 09:52 |
drawkula | the console only beards only needed time to notice it | 09:52 |
drawkula | :-) | 09:52 |
FatPhil | Some of us are still shouting at clouds... | 09:52 |
drawkula | http://www.nsa-cloud.com/ | 09:53 |
drawkula | my cloud is named sdf.org | 09:54 |
drawkula | :-P | 09:54 |
gnarface | i think you guys are making a big deal out of nothing. XDG_RUNTIME_DIR is being set somewhere to /run/user/whatever. if you don't set it it defaults to /tmp/$USER | 09:55 |
gnarface | or something like that | 09:55 |
gnarface | just figure out what is setting it to something you don't like, and change it | 09:55 |
gnarface | it's not being put anywhere in /run/ here, so whatever is doing it to you guys is something optional you requested yourself | 09:56 |
FatPhil | drawkula: I've gone one up - my cloud is called asdf.org | 09:57 |
gnarface | (probably a window manager with obscene ties to freedesktop.org would be my guess) | 09:57 |
FatPhil | gnarface: I can't see that setting anywhere - I grepped all over /etc for it | 09:57 |
gnarface | FatPhil: it might be in your WM configuration | 09:58 |
gnarface | it might actually be getting set from something that your WM runs at startup | 09:58 |
FatPhil | I do notice libpam_systemd is installed, I might try to purge that and see if it makes a difference | 09:58 |
gnarface | or it might have been added to your login scripts in ~/ even i don't know for sure | 09:58 |
gnarface | i don't have libpam_systemd here | 09:59 |
gnarface | but it is ceres so ymmv | 09:59 |
FatPhil | the whole concept of modifying scripts in a ~/ that is shared by half a dozen machines to be a frightening concept. | 10:01 |
FatPhil | modulo grammar | 10:01 |
FatPhil | hmm, actually that package isn't installed. got led down a blind alley. | 10:03 |
gnarface | sure but it builds character :) | 10:03 |
FatPhil | could it be consolekit that's making those tmpfss? | 10:03 |
gnarface | i really don't know but i wouldn't bet against it. i don't have that installed here either. (not using graphical login) | 10:03 |
gnarface | and i'm pretty sure i've never seen enlightenment (any version) or blackbox set XDG_RUNTIME_DIR to anything outside of /tmp. actually i recalled it usually being undefined and seeing a warning about it but i'm seeing here with a freshly updated ceres that something has defined it to /tmp/[some random string] | 10:05 |
gnarface | what WM are you using? did you check it for an environment variables configuration panel? | 10:06 |
FatPhil | I have no idea why consolekit was default selected - it brings in a whole load of crap. Seems to be pulled in by policykit, which is 'i A' on my system, apparently due to "elogind Recommends policykit-1" | 10:07 |
gnarface | that's ascii? | 10:07 |
gnarface | check the release notes | 10:07 |
gnarface | it's the graphical login manager that is pulling all that in | 10:07 |
gnarface | slim or lightdm or something like that | 10:08 |
gnarface | kdm | 10:08 |
gnarface | whatever your window manager defaults to | 10:08 |
gnarface | then it brings in some conflicting packages but you have to purge one of them | 10:09 |
helios21 | Now I found it in my training slides. Search engine startpage.com (Google) completely failed me on that: https://cr.yp.to/slashpackage/studies.html | 10:09 |
gnarface | https://files.devuan.org/devuan_ascii/Release_notes.txt | 10:09 |
gnarface | FatPhil: you should really read this whole thing ^ but at least read the part under "### Session management and policykit backends" | 10:10 |
helios21 | I like the filesystem layout proposed at https://cr.yp.to/unix.html | 10:10 |
FatPhil | gnarface: xdm there | 10:10 |
FatPhil | we don't do "Desktop Environment"s | 10:16 |
FatPhil | {console,policy}kit gone. I don't think I need elogind either by the looks of it. | 10:21 |
gnarface | hmm. xdm may do something different, i'm not sure | 10:22 |
FatPhil | Anything which says "Poettering touched me" I'm happier with it off the system. Alas the pulse cancer has metastasized too much. | 10:22 |
gnarface | meh, i'm still getting by without it. | 10:23 |
FatPhil | looks like xine and ffmpeg hard depend on lilbpulse | 10:23 |
gnarface | ah, but! | 10:23 |
gnarface | libpulse does not hard depend on pulseaudio, the daemon | 10:23 |
gnarface | so you can just uninstall or stop it | 10:23 |
gnarface | though some people actually like what it does | 10:24 |
gnarface | if you know enough about alsa with a couple extra shell tricks, you can get by without it for most things | 10:24 |
gnarface | the one thing steam "requires" it for is pretty egregious too | 10:24 |
FatPhil | OK, yeah, we don't have the daemon, by choice. | 10:24 |
gnarface | (since you could basically do the same thing by piping arecord to aplay through netcat) | 10:25 |
FatPhil | I know it solves some people's problems. You're lucky you prized OSS out of my cold dead hands and got me onto ALSA! | 10:25 |
FatPhil | holy jeebus, my webserver's under a completely mindless attack at the moment. the URLs being requested are seemingly random. might be hard to fail2ban this guy... | 11:58 |
FatPhil | one thing that is common is '.php' though, but I don't like configuring too many arbitrary rules. | 11:59 |
gnarface | it is free traffic | 12:11 |
gnarface | just redirect all 404 errors to a page with a bunch of banners :-p | 12:11 |
gnarface | if they're trying to break in, that means they're looking at responses | 12:12 |
gnarface | it also means you can just session cookie them and not give them a page unless they accept the cookie, then set a timer/counter on the server-side to also throttle the rate they can retry | 12:13 |
gnarface | or just give them a max hit count, or something like that | 12:13 |
gnarface | i mean, if someone hits three dozen 404 errors in the space of 10 seconds, they're probably doing something abusive | 12:14 |
xinomilo_ | googlebot usually :P | 12:34 |
FatPhil | most of the probes were POST, so I might just forbid that on the server that serves / | 12:45 |
FatPhil | it even tried a sql injection in the user-agent more than once! | 12:51 |
HumanG33k | hello one of you is good as setup postfix+dovecot+ldap ? | 14:13 |
HumanG33k | i need a jedi master | 14:14 |
enyc | fsmithred: hrrm you aren't here ;p | 15:37 |
Evilham | HumanG33k: you'd do best to ask a concrete question instead of asking to ask https://www.dontasktoask.com/ | 17:44 |
Evilham | lol, just found this by Rick: http://catb.org/~esr/faqs/smart-questions.html | 17:51 |
fsmithred | enyc, what's up? | 19:05 |
fsmithred | brb | 19:05 |
enyc | fsmithred: hrrm not sure if devuan specific, need to reboot and do more test,s, having toruble with vlc video playback on upgraded-to-beowolf desktop | 19:09 |
enyc | pulseaudio was mentioned in errors, as was messages about vaapi | 19:09 |
enyc | this is 'radeon' card.... | 19:09 |
enyc | pulseaudio seems to be sinstalled | 19:09 |
enyc | fsmithred: wantto reboot, load latest kernel, check boot logs, make clear observations, look up errors, first... but thought i might meniton as you sounded interested to see what goes wrong in practice | 19:10 |
James1138 | Here is where you can get the most up to date kernels - https://kernel.ubuntu.com/~kernel-ppa/mainline/ | 19:12 |
James1138 | They also work with Devuan. | 19:12 |
enyc | James1138: hrrm long time since i tried an ubuntu kernel image on debian/devuan | 19:12 |
enyc | mint also provides kernel 5.x ... | 19:13 |
James1138 | Devuan loves it. | 19:13 |
enyc | I remember leaving kernel recompiles on 486 machine overnight =) | 19:13 |
James1138 | I am running v4.9.184 but a new one just came out 7/10/2019. | 19:14 |
fsmithred | do liquorix kernels still exist? | 19:16 |
James1138 | https://liquorix.net/ | 19:16 |
fsmithred | those would probably work in devuan, too. I know lots of debian folked used to use liq kernel | 19:17 |
fsmithred | enyc, tell me a little about your upgrade. What desktop? | 19:19 |
James1138 | Most recent is v5.2/2019-07-08 00:07 | 19:19 |
fsmithred | I tried migrating stretch/kde to beowulf a couple days ago, and it was hell. | 19:19 |
fsmithred | because of metapackages | 19:19 |
James1138 | https://kernel.ubuntu.com/~kernel-ppa/mainline/v5.2-rc7/ | 19:20 |
enyc | fsmithred: had MATE and xfce available on ascii, i did a huge package clearup, dev packages, even removed task-mate-desktop and vairous mate and xfce packages ... lots of atp-get --purge autoremove and deborphan | xargs dpkg -P ....... | 19:20 |
fsmithred | I actually did get it to work after the fourth 'aptitude full-upgrade' | 19:20 |
enyc | fsmithred: ultivately dist-upgraded into beowulf | 19:20 |
enyc | fsmithred: and then, later, installed task-mate-desktop | 19:20 |
fsmithred | ok, sounds about right | 19:20 |
enyc | fsmithred: so i have beowulf with MATE only | 19:20 |
fsmithred | cool | 19:20 |
enyc | with all the mountains of packages i had' there was some arguing to get a dist-upgrade path but not true breakage | 19:21 |
fsmithred | I have a beowulf with xfce that works. Upgraded refracta/ascii (no metapackages) | 19:21 |
enyc | hence i started just removing swathes | 19:21 |
fsmithred | when we get the task packages and desktop-base done, it should be easier. | 19:22 |
enyc | task packages need doing because of different edvuan dependencies? | 19:22 |
fsmithred | yes | 19:23 |
fsmithred | exactly | 19:23 |
enyc | so what sort of things colud be wrong with my beowulf installed pkg set? | 19:23 |
fsmithred | some packages might be missing and you might not notice right away. | 19:23 |
fsmithred | then try something and it doesn't work, some function is not there | 19:24 |
fsmithred | if your user and shutdown/reboot from the desktop and can mount removable drives, you're doing well | 19:24 |
enyc | removable dirves ok | 19:25 |
enyc | about to test reboot from desktop | 19:25 |
fsmithred | the desktop you're using to talk to us here? | 19:25 |
fsmithred | guess not | 19:26 |
pav5088 | How to I get the current bug list for beowulf? "beowulf" as a tag doesn't seem to work in the bug tracking system. | 19:31 |
enyc | fsmithred: pulseaudio package installed, but not starting as part of user session ...... | 19:35 |
enyc | fsmithred: in MATE | 19:35 |
enyc | fsmithred: this suspicious to me | 19:35 |
enyc | I appear to have much the same pakcages with word "pulse" installed, i kept a backup of COLUMNS=256 dpkg -l flom ascii... | 19:36 |
enyc | I just no longer have libpulse-dev xfce4-pulseaudio-plugin ... and now of 12.2-4 insntead of 10.0-1+deb9u1 versions largely | 19:37 |
James1138 | https://packages.debian.org/sid/xfce/xfce4-pulseaudio-pluginHere you go Enyc - | 19:42 |
James1138 | https://packages.debian.org/sid/xfce/xfce4-pulseaudio-plugin | 19:42 |
James1138 | sorry | 19:42 |
James1138 | For Buster / Beowulf - https://packages.debian.org/buster/xfce4-pulseaudio-plugin | 19:43 |
fsmithred | will check my pulse pks in minute | 19:51 |
fsmithred | libpulse-mainloop-glib0:amd64 libpulse0:amd64 | 19:54 |
fsmithred | blueman Depends libpulse-mainloop-glib0 | 19:55 |
enyc | fsmithred: I have gstreamer1.0-pulseaudio libpulse-mainloop-glib0 libpulse0 libpulsedsp pulseaudio pulseaudio-utils | 20:23 |
enyc | but nonetheless, a pulseaudio *session* just isn't starting! | 20:23 |
fsmithred | I have very little experience with pulse. There should be a setting in the config file (/etc/pulse-something) | 20:26 |
fsmithred | and check the startup apps for the desktop. I don't recall where that is in mate. | 20:26 |
enyc | well you see if wonder if that now relies on systemd-logind or some other systemd session manager 'thing' | 20:27 |
fsmithred | oh | 20:27 |
fsmithred | you have consolekit? | 20:27 |
enyc | ii consolekit 0.4.6-6 | 20:27 |
enyc | yes, and login, logout, rebooting form menu, mounting removables, all works | 20:27 |
fsmithred | we have elogind, but it might be a pain to switch to that | 20:28 |
ba55 | is there a download for the beowulf iso? | 20:28 |
ba55 | i got an RC1 a while ago, dont see it now | 20:28 |
fsmithred | no beowulf isos yet | 20:28 |
ba55 | ahh - i wonder if the RC was pulled | 20:28 |
fsmithred | no, there's been no rc1 for beowulf | 20:28 |
ba55 | yeah there was lol | 20:29 |
fsmithred | there's been no beta for beowulf, and that will precede the rc | 20:29 |
fsmithred | no, there wasn't | 20:29 |
ba55 | i ran it | 20:29 |
fsmithred | if it said rc, then it must have been ascii | 20:29 |
fsmithred | there was a mini.iso | 20:29 |
fsmithred | for beowulf | 20:29 |
fsmithred | and I'm pretty sure it no longer works because of kernel changes | 20:29 |
ba55 | got ya oh well | 20:29 |
ba55 | i built a new pc | 20:29 |
ba55 | i cant set it up yet till this comes out lol | 20:30 |
ba55 | want to do 100% fresh | 20:30 |
fsmithred | if you really want beowulf, which isn't fully ready but can work just fine... | 20:30 |
ba55 | nah i can wait :) | 20:30 |
fsmithred | install a bare ascii and upgrade | 20:30 |
fsmithred | or use one of the derivatives that has a beowulf iso | 20:30 |
ba55 | i'm sure it will be out sooner than later | 20:30 |
ba55 | what do u mean by that second comment?? | 20:31 |
fsmithred | yeah, we should have installer isos soon | 20:31 |
fsmithred | some of the devuan derivative distros have beowulf isos | 20:31 |
fsmithred | refracta, miyolinux, I think gnuinos, and maybe star or crows (not sure) | 20:31 |
ba55 | ahh got ya - yeah ill hold off on those | 20:31 |
ba55 | i tried antix the other day | 20:31 |
fsmithred | lol, ok | 20:31 |
ba55 | man thats a bitch | 20:32 |
fsmithred | in what way? | 20:32 |
ba55 | for old slow pc's i get it | 20:32 |
ba55 | but its so basic | 20:32 |
ba55 | the window managers etc | 20:32 |
ba55 | i was just isntalling vm's seeing what other deb based distros without sysd looked like | 20:32 |
fsmithred | sometimes using just a window manager is a relief. | 20:33 |
fsmithred | very little to go wrong with it | 20:33 |
fsmithred | check back in a few weeks and we might have a beta | 20:34 |
fsmithred | or | 20:34 |
fsmithred | we might all be on vacation | 20:34 |
ba55 | lol | 20:35 |
ba55 | im counting on a few days for the iso ;) | 20:35 |
fsmithred | oh, you haven't been here very long | 20:36 |
ba55 | ive been around under diff nicks | 20:36 |
fsmithred | I can probably get desktop-base put together in a week or so, but all the task- packages need to be done | 20:37 |
ba55 | so how long after stretch came out, did ASCII release? | 20:37 |
fsmithred | a year | 20:37 |
ba55 | wow | 20:37 |
fsmithred | which was half as late as jessie | 20:37 |
fsmithred | and I'm counting on beowulf being half as late as ascii was | 20:37 |
ba55 | well then, maybe i should just go deb for now then switch when it's released :) | 20:38 |
fsmithred | so, I feel confident we'll have a release before the end of the year | 20:38 |
ba55 | awesome | 20:38 |
ba55 | its nice to have someone give me legit answers, i appreciate it | 20:38 |
fsmithred | I can tell you that the system is stable long before we get the installer isos debugged | 20:38 |
fsmithred | stable in the sense of "it works" | 20:38 |
ba55 | so stable.... say i installed ascii and upgraded... is there gna be any major downsides u know of? | 20:39 |
ba55 | i got a laptop i can test on | 20:39 |
fsmithred | depends on how you installed ascii | 20:39 |
ba55 | i would do it as u instruct me to :p | 20:39 |
fsmithred | if you selected a destkop environment from tasksel in the installer... | 20:39 |
fsmithred | you're in for a world of hurt | 20:39 |
ba55 | yeah so this machine im on , wont work | 20:39 |
ba55 | its MATE | 20:39 |
fsmithred | if you installed your desktop environment in individual pieces... | 20:40 |
fsmithred | upgrade won't be too bad | 20:40 |
enyc | fsmithred: hrrm pulseaudio -D can just start pulse manually no rpoblem seemingl | 20:40 |
ba55 | alright so say, from fresh install - how would i do it? | 20:40 |
fsmithred | if you install just a bare cli system, upgrade and then add desktop, it'll probably work pretty easily | 20:40 |
enyc | fsmithred: but i'm now locking-up on vlc still, screns black out, etce tc | 20:40 |
ba55 | got ya | 20:40 |
ba55 | well if u feel like it, (or get time) pm me a list of what i need to do for the upgrade | 20:40 |
fsmithred | metapackages cause issues because not all the deps are ready | 20:41 |
ba55 | id be happy to try it and report back | 20:41 |
fsmithred | check the guides for upgrade to ascii. It should be similar. | 20:41 |
ba55 | cool cool | 20:41 |
fsmithred | https://devuan.org/os/documentation/dev1fanboy/ | 20:42 |
ba55 | wow it seems pretty simple | 20:43 |
fsmithred | I've upgraded refracta to beowulf and it wasn't bad. It's xfce without metapackages. | 20:43 |
fsmithred | I tried stretch/kde to beowulf a few days ago and it was hell. | 20:43 |
ba55 | so say i install fresh with no destop | 20:44 |
ba55 | upgrade | 20:44 |
ba55 | THEN i can install mate | 20:44 |
ba55 | and it should be ok? | 20:44 |
enyc | ba55: maybe ;p | 20:44 |
fsmithred | yeah, pretty sure. I haven't tried mate in beowulf, but I think it's ready | 20:44 |
ba55 | heh well, when the rest of my parts show up today | 20:44 |
ba55 | i will be trying! | 20:44 |
fsmithred | Candidate: 1.20.4-2 | 20:45 |
fsmithred | mate-desktop ^^^ | 20:45 |
enyc | ba55: for me, "pulseaudio -D" as user isn't auto starting | 20:45 |
enyc | ba55: for me, i'm having video accell issues | 20:45 |
enyc | but these may not be devuan specific issues | 20:45 |
ba55 | hmm i dont use pulse audio anymore | 20:45 |
fsmithred | enyc, do you normally use non-free vid drivers? | 20:46 |
enyc | fsmithred: no | 20:46 |
ba55 | i switched to JACK | 20:46 |
enyc | fsmithred: radeon | 20:46 |
ba55 | and audio has been superb | 20:46 |
* enyc removed the vaapi packages etc. | 20:46 | |
fsmithred | firmware-amd-graphics is in non-free | 20:48 |
enyc | oooooooooooh hadn't considered that bit | 20:48 |
fsmithred | yeah, a few people have run into problems without that | 20:49 |
ba55 | bbiab guys, gotta get back to work :/ | 20:49 |
enyc | well, i'm not using user/kernel non free driver | 20:49 |
enyc | unless you go to some ... new risc architecture etc you are rather stuck with thins ilke intel microcode proprietary let alonge video ! | 20:50 |
enyc | anyway | 20:50 |
enyc | without mesa/vdpau driver whatnots, vlc working now | 20:50 |
fsmithred | wow | 20:50 |
enyc | fsmithred: whyso wow? | 20:51 |
fsmithred | usually you need to add that stuff to fix vid problems | 20:51 |
enyc | fsmithred: i'm going to leave it like this for now and see if other traditional issues i have had, go away | 20:51 |
enyc | fsmithred: even on ascii, if i did alot of videos, usage, chormium, etc iwould eventually get osme sort of drm errors in kernel nad general system crash | 20:52 |
enyc | well, xorg/video crash and genealyl reboot needed | 20:52 |
enyc | be nice to see this new arrangement (even without vdpau accell) might actually be more reliable now! | 20:52 |
fsmithred | I get that when I open too many browser tabs | 20:52 |
enyc | the other thing i will re-try sooner or alter is just phasing out chromium! | 20:53 |
enyc | a friend of mine has a sepratae cgroup/whatever and constrains chromium's ram because otherwise it is like a gas that expands to fill all available.... | 20:53 |
enyc | on my mx-linux on another disk, i had firefox working preety well ... | 20:53 |
fsmithred | choosing a browser these days is like choosing a president - whichever one you least want, pick the other | 20:54 |
slvr | links has been good to me | 20:54 |
enyc | slvr: yes i used that too ;p | 20:54 |
enyc | slvr: lots of packages.debian.org/[something] and maunalyl saving debs ;p | 20:54 |
fsmithred | bbiab | 20:56 |
guru | Is beowulf stable enough in a production environment? | 21:46 |
enyc | guru: depends on your use case i suspect | 21:58 |
enyc | guru: a lot seems to work fine, som fun with dependencies and desktops | 21:58 |
guru | webserver | 21:59 |
enyc | 19:38 < fsmithred> so, I feel confident we'll have a release before the end of the year | 21:59 |
enyc | fsmithred: btw, no, i have separate dual-pentium-III thing with old xen and screen sessions and allsorts ;p | 21:59 |
fsmithred | kayak? | 21:59 |
enyc | guru: very likely to be fine i suspect! | 22:01 |
enyc | guru: try it | 22:01 |
fsmithred | pretty sure we don't mess with apache at all | 22:01 |
fsmithred | or nginx | 22:02 |
enyc | of course, debian buster , and consequently devuan beowulf is tried much less in production already , so you could choose to wait for next point release, or devuan formal release ... | 22:02 |
guru | Yeah seems like I should stick to ascii to be safe | 22:02 |
enyc | fsmithred: i guess the fun could come with ... removed or outdated/broken sysvinit start scripts | 22:02 |
fsmithred | yeah, I don't know if they're removing them | 22:03 |
enyc | r_state [radeon]] *ERROR* rv770_restrict_performance_levels_before_switch failed | 22:03 |
enyc | radeon 0000:01:00.0: ring 0 stalled for more than 10208msec | 22:03 |
enyc | radeon 0000:01:00.0: GPU lockup (current fence id 0x0000000000049a5f last fence id 0x0000000000049a85 on ring 0) | 22:03 |
enyc | i may well do twell to try a neewer/upstream kernel | 22:04 |
enyc | but first i try radeon.dpm=0 nomodeset | 22:04 |
fsmithred | FWIW, I just installed apache2 onto the machine behind me (beowulf/xfce) and then connected to it with a browser on the computer in front of me. Didn't even have to get out of my chair. "It works" | 22:06 |
fsmithred | ! | 22:07 |
fsmithred | try installing in a VM to see if everything you need installs easily. | 22:07 |
fsmithred | btw, computer in front of me is still running jessie, because it works. | 22:10 |
FatPhil | i think everything I can see is running jessie (2 Pis, 1 Pi3, an Asus lappy, and an MSI cubie mini thingy) | 22:17 |
Centurion_Dan | hi fsmithred! | 22:21 |
fsmithred | hi Centurion_Dan | 22:24 |
fsmithred | you back home? | 22:24 |
Centurion_Dan | nope, just back at the inlaws where I have decent connectivity and power... the campervan doesn't have an inverter and I don't have a dc-dc converter for my laptop. | 22:25 |
Centurion_Dan | It was nice having a couple of days without the children though ;-) | 22:26 |
slvr | get a 14 volt laptop | 22:26 |
guru | fsmithred: so far so good thank you :) | 22:27 |
fsmithred | he's got a couple of semaphore flags tucked behind the driver's seat | 22:27 |
enyc | oh this is getting silly | 22:27 |
enyc | i now have some non display related issue entierly | 22:27 |
fsmithred | what issues? | 22:27 |
enyc | even without changig kernel command line i now cant boot 4.19.0-5-amd64 even afer reinsatlalng package | 22:27 |
enyc | it get so ar booting and gets sotck on a mountan of:- | 22:27 |
enyc | WARNING: Device /dev/sd?? not enetialised in udev hataase even ater waiting 100000000 microseconds. or so | 22:28 |
fsmithred | are you in chroot now? | 22:28 |
fsmithred | or you get that on booting? | 22:28 |
enyc | i get that on booting | 22:28 |
enyc | i can boot a 4.9.0-9-amd64 | 22:28 |
fsmithred | I was getting that error in chroot while running update-grub | 22:28 |
enyc | just not the 4.19.0-5-amd64 | 22:28 |
enyc | wich has now become broken ;-( | 22:29 |
enyc | but only very recently | 22:29 |
enyc | did a dpkg-reconfigure grub-pc and reinstalled grub | 22:29 |
fsmithred | ok, not a uefi problem | 22:29 |
enyc | did a --reinstall of the linux-image-4.19.0-5-amd64 package wahteer it was | 22:29 |
enyc | yes , not uefi machine | 22:29 |
enyc | oh | 22:29 |
enyc | it EVENTUALLY boots 4.19.0-5-amd64 | 22:29 |
enyc | just very very slow at first | 22:30 |
enyc | i left it for a bit... | 22:30 |
fsmithred | yeah, I saw the same with update-grub - those messages took a long time to finish | 22:30 |
enyc | fsmithred: recent udev upgrade? whats changed very lately? | 22:30 |
fsmithred | a couple months ago, I ran into that and I think I ended up mounting /run/udev to the chrooted system | 22:30 |
fsmithred | and that made the warnings go away | 22:31 |
fsmithred | I thought it was related to the grub-efi-amd64-signed package at the time, but I guess not. | 22:31 |
* enyc slaps enyc | 22:32 | |
enyc | [ 402.700697] radeon: `o' invalid for parameter `dpm' | 22:32 |
enyc | BOOT_IMAGE=/vmlinuz-4.19.0-5-amd64 root=/dev/mapper/clone_vg-root ro radeon.dpm=o nomodeset | 22:32 |
enyc | 0 not o ;p | 22:32 |
* enyc slaps enyc | 22:32 | |
guru | I'm kind of scared to upgrade from ascii to beowulf on my laptop. Should I take the plunge? ;) | 22:32 |
guru | Someone flip a coin | 22:33 |
fsmithred | guru, wait | 22:33 |
guru | Smart man hah | 22:33 |
fsmithred | I assume you need to use your laptop, right? | 22:33 |
guru | Yup hah | 22:33 |
fsmithred | and it's working now? | 22:33 |
enyc | guru: i took complete disk image backup on this desktop system disk ;p | 22:33 |
fsmithred | Don't fuck with it. | 22:33 |
enyc | +1 to fsmithred | 22:33 |
* fsmithred still has a wheezy server in production | 22:34 | |
guru | fsmithred: Solid advice thanks :) | 22:34 |
fsmithred | for your amusement, if you're hot to play with beowulf, put this on a usb: https://get.refracta.org/files/experimental/refracta10-nodbus_amd64-2019-04-16.iso | 22:35 |
guru | :) I'll do that thanks | 22:36 |
fsmithred | I think that one does not have a network manager of any kind. | 22:38 |
fsmithred | dhclient eth0 | 22:39 |
guru | Isn't wheezy EOL? | 22:39 |
enyc | guru: ELTS for wheezy exists | 22:39 |
enyc | guru: but doesn't cover all packages | 22:40 |
guru | Debian 8 will also receive Long Term Support for five years after its initial release with support ending on June 30, 2020. The supported architectures include amd64, i386, armel and armhf. | 22:40 |
guru | ah I see | 22:40 |
guru | fsmithred is playing the long game :P | 22:40 |
* enyc finally boots complexity mochinu with ... radeon.dpm=0 nomodeset ;p | 22:41 | |
fsmithred | yeah, wheezy died some time ago. I thought about updating to jessie - the main app that is needed would continue to work, but after that, there's no deb package | 22:43 |
fsmithred | and it's lvm on raid, so I decided not to mess with it | 22:44 |
enyc | [ 356.084078] [drm] VGACON disable radeon kernel modesetting. | 22:44 |
enyc | [ 356.084138] [drm:radeon_init [radeon]] *ERROR* No UMS support in radeon module! | 22:44 |
fsmithred | enyc, stackexchange says remove nomodeset | 22:49 |
enyc | fsmithred: indeed! | 22:52 |
enyc | seems like modern radeon requires kms | 22:52 |
fsmithred | you have new vid card in old machine? Or was the pIII something else? | 22:54 |
enyc | piii/machine is that with irssi etc | 22:55 |
enyc | nothing todo with desktop with error | 22:55 |
fsmithred | and another search hit says to add radeon.modeset=0 | 22:59 |
fsmithred | which I'm pretty sure does the same thing as nomodeset | 23:00 |
enyc | hrrm yesand not useful with new kernels | 23:01 |
enyc | fsmithred: going to try newer kernels ;p | 23:02 |
fsmithred | enyc, does that install have anything in /dev when it's not running? | 23:05 |
fsmithred | all I have in the install that gave me that error message about devs not in udev database has only /dev/null | 23:07 |
fsmithred | standard debian install has some static links in /dev. I don't know if that would make a difference. | 23:07 |
enyc | fsmithred: hrrm good question | 23:27 |
enyc | fsmithred: booting this 5.2 kernel finally ;p | 23:27 |
enyc | ssame nonsens with not initialiezd in ude database etc | 23:27 |
enyc | was very helpful to boot into older 4.9 kernel, then sort out linux kernel installs | 23:28 |
enyc | as grub-probe etc semeed to have massive runtime hangups similarly | 23:28 |
fsmithred | I'm thinking eudev is involved. Might be helpful to know what conditions trigger that warning. | 23:47 |
guru | enyc: I made a backup image like you did and upgraded. Glad I did the performance is significantly better. | 23:59 |
enyc | guru: im still arguing with kernels and radeon | 23:59 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!