debdog | Guest4445: from my POV the simplest solution is a DM which supports autologin. this does start X automagically as well | 00:06 |
---|---|---|
debdog | I just can't recall atm which DM I've used for that | 00:07 |
debdog | aight, giong back to BF2 | 00:08 |
Guest4445 | If i put startx to bashrc loop happens ie x is started several times | 00:21 |
rustytaco | slim can do it | 00:25 |
rustytaco | even if it is a really bad idea | 00:25 |
rustytaco | why would you want every process running as root? | 00:25 |
Guest4445 | Dont needmultiuser | 00:26 |
Guest4445 | Slim doesnotautostart in single usermode | 00:27 |
debdog | what is single user mode? | 00:28 |
rustytaco | Lets do this a simpler way.. What bad idea are you trying to accomplish that needs every process to run with the ability to damage the OS, rather than being contained to a user? | 00:29 |
rustytaco | im not scrolling up 1400 pages | 00:29 |
Guest4445 | You can find info on single usermode in internet. Generally you edit inittab to get it | 00:30 |
rustytaco | runlevel 1 | 00:30 |
dan9er[m] | A-B problem | 00:30 |
rustytaco | What ya trying to accomplish tho? There's probably a better way | 00:31 |
rustytaco | thats the neat thing with all this mess, theres at least 300 ways to skin a cat, and we have to tools for em all! | 00:31 |
dan9er[m] | Guest4445: Like taco said, what are you actually trying to do ("A") that you _think_ you need single user mode for ("B")? | 00:32 |
Guest4445 | I just want only root user,no multiuser | 00:32 |
dan9er[m] | Why tho | 00:33 |
dan9er[m] | The only situation this makes sense to me is rescue disk/mode | 00:33 |
debdog | not even M$ does that | 00:33 |
dan9er[m] | I'm sure that if you tell us why, and it is sensible, we'll understand | 00:34 |
debdog | still, slim should be able to do autologin as $any_user, even root | 00:35 |
Guest4445 | It's not rescuemode,it's one user mode | 00:35 |
Guest4445 | No other users are allowed | 00:36 |
dan9er[m] | You keep dodging the question | 00:36 |
dan9er[m] | What do you need single user mode for | 00:36 |
debdog | just stop that discussion, please | 00:36 |
debdog | he wants it, we try to help accomplish it | 00:37 |
Guest4445 | Slim cannot start from single usermode, you first have to givepassword | 00:37 |
debdog | Guest4445: I am pretty sure it should™ be able to https://duckduckgo.com/?t=palemoon&q=slim+autologin&ia=web | 00:38 |
debdog | if it does not work it prolly is abug | 00:38 |
Guest4445 | It says give root pass or press cttrl dformajntenance | 00:40 |
Guest4445 | So slim.is just not started automatically | 00:40 |
debdog | Guest4445: you have to configure it accordingly, of course | 00:41 |
debdog | it doe not do that by default | 00:41 |
debdog | you have to enable autologin in the config file and tell it which user | 00:42 |
rustytaco | Yes you need to be in runlevel 3 for a X session to run as a user. | 00:44 |
rustytaco | If we knew what you're trying to run, we could probably be more helpful. | 00:44 |
rustytaco | If you want to run X and everything under it as root, you'll have to bypass some bits put in place to make the normal multiuser desktop environment work, fwiw. But why single user mode? Why X programs as root? | 00:45 |
rustytaco | You need to be in runlevel 3 for slim to start -- or manually start it in your single user (runlevel 1) init scripts -- just use update-rc.d to set it to run in runlevel 1 and it might work. | 00:46 |
rustytaco | still i'd like to understand the reasoning as there's probably a lot better ways to do what you seek to do | 00:46 |
Guest4445 | As i told i need rt capabilities for latency free work, so root is need andleast running servicez and desources | 00:50 |
debdog | no, that info is at least a 20 years old | 00:51 |
Guest4445 | Rhel manual says that scheduler can only manage rootprocesses | 00:54 |
debdog | I'd suggest to just try it the "usual" way. then, if the latncy is too big try to find the bottleneck and improve it | 00:58 |
Guest4445 | Already got someimprovements just trying if root will givebenefits | 01:01 |
debdog | maybe, on really old hardware | 01:02 |
debdog | I mean _really_ old | 01:02 |
fsmithred | Guest4445, look at live-config scripts for autologin | 01:03 |
fsmithred | https://git.devuan.org/devuan/live-config/src/branch/master/components/0140-xinit | 01:04 |
fsmithred | for auto-startx | 01:04 |
fsmithred | 0160-sysvinit for console autologin | 01:04 |
fsmithred | the xinit script makes a file in /etc/profile.d that contains startx (and some other niceties) | 01:05 |
fsmithred | the sysvinit script edits /etc/inittab getty lines | 01:05 |
fsmithred | boot a live iso and look at inittab if you don't like sed heiroglyphics | 01:06 |
fsmithred | rustytaco, default runlevel in debian distros is 2 and 2-5 are all multi-user graphical | 01:07 |
Guest4445 | Dont know what to do. Whils in single user mode it says give pass or press ctrl d for maintenance. If i puf startx inbashrc x is started several times inloop | 01:07 |
rustytaco | probably because single user mode isnt intended for what you're doing | 01:10 |
rustytaco | its for repairing a broken install | 01:10 |
fsmithred | why do you need to use single? | 01:12 |
Guest4445 | Donf know but indeed a manual from 2000 told try single user mode to improve rt | 01:13 |
Guest4445 | I don'tneedmultiuser | 01:13 |
debdog | bashrc is called by ever shell started. things that should only be started at login do go into .profile (or ~/.bash_profile) | 01:14 |
debdog | *every | 01:14 |
rustytaco | thats 20 years ago | 01:16 |
rustytaco | The kernel has changed immensely since then | 01:16 |
rustytaco | Schedulers all around have improved immensely and become pluggable even. | 01:16 |
rustytaco | thats like 2.2.x era kernels | 01:16 |
rustytaco | Ignore the "multi user" thing | 01:17 |
rustytaco | What are you trying to do? Audio stuff? | 01:17 |
fsmithred | not all services are started in S | 01:17 |
debdog | 20 years ago I was able to build a studio PC with a latency lower than 10 ms (max) without "single user mode" | 01:18 |
rustytaco | you can give a single process realtime priority without any strife | 01:18 |
* rustytaco points at the audio processing rack running on his radio serer | 01:18 | |
Guest4445 | Rt is about experimenting. I recompiledkernel 15 times with cutting all notneeded it helped, single user may help too | 01:18 |
rustytaco | you can give real time priority without all this single user nonsense tho.. | 01:19 |
debdog | all right then. back to slim (or any some other DM). just tell it to do autologin and which user | 01:19 |
Guest4445 | It waz possible like 3msthat time as per article,i now can get 1ms or 2.1ms for largerprojects | 01:20 |
rustytaco | Ja do that, set things up to give rt priority to the needful processes and let the kernel do what it was made to do | 01:20 |
* rustytaco watches his passive radar running without any fuss of single user or recompiling kernels. | 01:21 | |
Guest4445 | Not that simple. U must shut down all unneeded | 01:23 |
Guest4445 | Including multiuser | 01:23 |
Guest4445 | It indeed helps | 01:23 |
debdog | out of curiousity, "where" do you try to reduce latency, Guest4445? somthing like sound or reading sensors or user interaction, like keyboard an mouse? | 01:31 |
Guest4445 | For sound purposes. U can read rhel about rt how to get ultimats results | 01:35 |
dan9er[m] | I flashed Daedalus desktop iso on my USB, and it can't mount /cdrom | 04:21 |
dan9er[m] | It looks like once I boot in, all the USB ports die | 04:25 |
dan9er[m] | I'm gonna try burning an actual DVD | 04:28 |
grayrock | i cp'd the rc1 x64 daedalus desktop iso to a usbstick earlier today. it booted ok. the partitioning messages seem to be fixed. choosing mate on tasksel and continuing gives an error, so i aborted there. the "D" in the text graphic of "DEVUAN" is mangled for me (not important). note that this is a report, not a complaint. :) | 04:34 |
dan9er[m] | Black screen on boot | 05:00 |
dan9er[m] | Wtf | 05:00 |
gnarface | is that the installer disk? try the live iso | 05:01 |
gnarface | regular install may be missing some non-free firmware and drivers | 05:01 |
grayrock | oh, btw, i'm installing offline, completely disconnected from any net, so that might be relevant. | 05:02 |
gnarface | grayrock: what was the error you got? | 05:02 |
gnarface | dan9er[m]: (you should be able to install them, but a the live iso might make it easier to figure out which ones) | 05:03 |
grayrock | gnarface: i foolishly didn't make a note of it. i'll try it again and do that. | 05:04 |
grayrock | gnarface: https://dpaste.com/2Y2CLUU8Q | 05:53 |
grayrock | out of curiosity, i went back a step and chose lxqt, which installed normally. | 05:54 |
grayrock | amd xfce would always install normally. so it's definitely mate-specific. | 05:57 |
grayrock | and* | 05:58 |
rustytaco | hmm | 06:03 |
rustytaco | my new compy will be hear soon | 06:03 |
rustytaco | 64kb of ram! | 06:03 |
gnarface | grayrock: uh, sorry i'm too paranoid to click on anything but paste.debian.net, but it matters very little in this case; the person you want to be showing the error to anyway is fsmithred | 06:54 |
gnarface | the most i could do is speculate on a cause, but i'm sure it's not user error | 06:55 |
gnarface | (thanks for testing these images by the way, they need it) | 06:55 |
grayrock | gnarface: ok, no prob and reluctance understood. | 07:01 |
grayrock | https://paste.debian.net/1284648/ | 07:05 |
gnarface | grayrock: oh, that's weirdly unspecific. i was thinking it'd be another missing library package or something. only thing i can suggest is that you alt+fN over to an unused terminal and enable it so that you can see the actual console errors as they appear | 07:09 |
gnarface | whatever is failing should have thrown an error of some sort, and if it's not showing in the UI it should show up in the first unused virtual terminal (i think it's F4 but i'm not sure) | 07:10 |
grayrock | yeh, i should have thought of that. i will do that when i get a chance. | 07:11 |
rrq | grayrock: possibly a side note, but "console productivity" conflicts with some desktop environments, so don't do moth. | 07:11 |
onefang | Default virtual terminals are 1 to 6, with X on 7. | 07:12 |
gnarface | onefang: i thought that was the old days, i thought X moved to 1 but also i thought ncurses mode was different | 07:13 |
gnarface | grayrock: it might matter whether you're starting in X11 mode or ncurses mode | 07:13 |
rrq | mmm the installer has vt1-4, with 1 being the dialog flow and 4 being a tail of the log file, leaving you 2 and 3 | 07:13 |
grayrock | oh, that's relevant, because i usually choose console productivity. i'll try it without. | 07:13 |
onefang | Oh, you meant the installer. | 07:13 |
gnarface | onefang: yea this is happening in the daedalus installer | 07:14 |
grayrock | heh, i'll just push buttons until i see output. | 07:14 |
onefang | OK, I was talking about the installed OS, at least up to chimmy. I haven't tried deadulus yet. | 07:15 |
rrq | more specifically the kernel offers 64 virtual terminals nowadays, but sysvinit only uns login on 6 of them, and X traditionally starts on 7. | 07:18 |
rrq | you can easily run "openvt -c 23 bash" to get a bash on vt23 if you like | 07:18 |
rrq | (but who does?) | 07:19 |
rrq | r | 07:19 |
* onefang searches my keyboard for the F23 key. Is it near the ANY key? | 07:19 | |
rrq | you may need to run "chvt 23" to shif the monitor | 07:20 |
rrq | t | 07:20 |
onefang | I'm gonna need a bigger desk. | 07:20 |
onefang | Actually I had moved dmesg to 1, and only use vt2, and even then only rarely. | 07:24 |
onefang | Everything else is done in X with lots of tmux windows and other things. | 07:24 |
grayrock | looks like it's just a missing package on the iso. debian-mate-default-settings is needed as a dependency for mate-session-manager. probablt gets fetched when installing with net enable, but i install offline. | 08:50 |
grayrock | multiple-typos* | 08:50 |
grayrock | and it's vt4 that shows /var/log/syslog. vt1 is the installer, vt2 and vt3 are ash shells. | 08:54 |
* grayrock realizes someone already said that | 08:58 | |
rrq | supposedly the desktop iso includes anything dependent and recommended from task-mate-desktop; is debian-mate-default-settings something extra? | 09:22 |
grayrock | syslog gave that as the reason for the error installing mate in offline mode. | 09:24 |
rrq | ah; it's a non-default choice option | 09:24 |
rrq | the default choice is mint-mate-default-settings | 09:25 |
grayrock | oh, well it mentioned all three: mint, ubuntu, and debian. it couldn't find any of them. | 09:25 |
rrq | yes; not sure why... looks like a bug in the "dependency resolver"... | 09:29 |
grayrock | "the following packages have unmet dependencies: mate-session-manager depends: mint-mate-default-settings or ubuntu-default-settings or debian-default-settings" | 09:33 |
rrq | oh.. hmmm mint-mate-default-settings is the default choice, but it's missing from debina | 09:33 |
grayrock | mate-default-settings* | 09:33 |
rrq | hmm; the the default option is more complex than the first of options... | 09:35 |
rrq | debian repository only has one *-mate-default-settings, namely debian-mate-default-settings, and that one is not syntactivally the default one that mate-session-manager depends on | 09:43 |
grayrock | i searched md5sum.txt on the iso for "mate-default-settings" and came up empty. | 09:44 |
rrq | yes the iso logic picked the default one to resolve the dependency, and then drop that one since it doesn;t exist in debian | 09:44 |
grayrock | i was installing offline, btw, so it would have only looked on the usbstick. | 09:45 |
grayrock | and it looks like none of them are on the usbstick. | 09:46 |
rrq | well the debian repository, toatally, only has one of them. the other ones might exist within other package collections, and the maintainer of mate-session-manager seems to know of them all | 09:48 |
grayrock | so i assume that adding any one of them to the iso would fix the problem. | 09:50 |
rrq | mmm it would hide this problem, yes, but it'd be better to expand the logic for "the default choice" beyond that mere syntactical consideration | 09:52 |
rrq | it should pich "the first choice that has an implementation within the repository" (perhaps) | 09:54 |
rrq | pich=pick | 09:54 |
rrq | rather than just "the first choice mentioned" | 09:54 |
grayrock | even if the programming logic is perfect, it will still need to find debian-mate-default-settings.deb (pr one of the others) on the installation media (when installing without internet) | 09:55 |
rrq | yes; I'm talking about the logic that decides what to put on the iso | 09:56 |
rrq | https://git.devuan.org/devuan/installer-iso/src/branch/master/pool/review.lsp | 09:57 |
grayrock | oh. i see. | 09:58 |
rrq | line 192 picks "the first of options as the default choice".. and that's obviously too weak logic | 09:59 |
rrq | (or thereabouts.. too weak logic) | 10:02 |
djph | aww, no daedalus torrent yet. boo. :( | 10:09 |
* rrq just saw a bug on line 245.. (this bug?) | 10:14 | |
rrq | mmm at least P rather than x makes the iso inlcude debian-mate-default-settings | 10:18 |
grayrock | i'm gonna hit the sack but i'm glad you're homing in on it. goodnight. | 10:20 |
grayrock | installng a second instance of devuan 5 causes the common swap partition to be unrecognized in the first instance because of the UUID change. i don't remember that happening before. | 16:03 |
rrq | grayrock: 5.0.0-rc2 are available | 16:17 |
grayrock | cool | 16:21 |
grayrock | btw, i notice daedalus comes with nftables instead of iptables. i know zip about nftables. it would really be nice if it was configured with all incoming ports closed by default. | 16:34 |
grayrock | unless servers were installed by the installer anyway. | 16:36 |
grayrock | i like that ssh-server is now an option in tasksel. before it was installed by default, which always seemed like a security risk to me. | 16:39 |
rrq | you can install and use iptables | 16:42 |
grayrock | i know, but i figure i might as well just start using nftables since that's the furure. | 16:42 |
grayrock | future* | 16:43 |
rrq | well, I'm still waiting for it to handle ipset | 16:43 |
rrq | until then the iptables synta is more powerful | 16:43 |
grayrock | in that case, it would be great to have iptables and ufw, and all their dependancies, available on the installer iso. i like to get a firewall up before exposing a new install to the internet. | 16:45 |
* grayrock is paranoid | 16:45 | |
onefang | Justifiably paranoid. | 16:46 |
rrq | I think iptables is in the desktop pool | 16:47 |
rrq | and ufw | 16:47 |
rrq | i.e. on the desktop iso | 16:48 |
grayrock | great, that's what i use. | 16:48 |
rrq | ipset is there as well | 16:48 |
grayrock | laters. gonna go play with rc2... | 17:08 |
grayrock | rc2 report: works great, mate install is flawless. minor issue: openssh-server and openssh-sftp-server are installed even if you don't select ssh server on tasksel menu. otherwise nice release candidate. lotta hard work you guys did. thanks. | 20:10 |
grayrock | ufw can be installed just by copying four .debs from the installation media pool and installing them with dpkg: libip4tc2_1.8.9-2_amd64.deb, libip6tc2_1.8.9-2_amd64.deb, iptables_1.8.9-2_amd64.deb, and ufw_0.36.2-1_all.deb | 20:20 |
* grayrock don' need no steenking dvd | 20:20 | |
Xenguy | Thanks for the feedback grayrock | 20:35 |
buZz | soooo, i distupgraded to dae<tab> today, but now 'slim' doesnt allow my useraccount to login, while logging in as root goes fine | 21:11 |
buZz | 'failure to execute login command' :/ | 21:12 |
buZz | and ofcourse, slim.log is completely empty :) | 21:12 |
buZz | 'login' from a rootshell works fine though, also logging in on terminal works fine | 21:13 |
buZz | maybe i'll just try a different login manager, lets see | 21:14 |
buZz | ... are there even others? :D | 21:17 |
buZz | enabling autologin in slim doesnt prevent it :( | 21:18 |
buZz | got it! | 21:24 |
buZz | it was a full disk, lol | 21:24 |
* buZz writes 'add diskfull msg to slim login failure' to maybeTODO.txt | 21:30 | |
Xenguy | : -) | 21:31 |
dan9er[m] | I just installed Devuan Chimera on my desktop, and my wifi card doesn't work. (I can access internet via Android USB tether.) How do I check what firmware I need and which package provides it? It's a TP-Link Archer T8E | 21:37 |
buZz | dan9er[m]: if you already have the driver installed (probably) you likely get a message about trying to load firmware but failing in 'dmesg' | 21:38 |
buZz | with a filename :P then just apt search filename (without filename extension i guess) | 21:39 |
buZz | i -believe- thats a Broadcom Corporation BCM4630 | 21:41 |
dan9er[m] | Yeah just looked it up, sorry | 21:42 |
buZz | is it firmware-brcm80211 ? | 21:42 |
dan9er[m] | I don't see bcm4630 in that package's description | 21:43 |
buZz | i do see > Binary firmware for Broadcom/Cypress 802.11 wireless cards | 21:44 |
buZz | which it is, i guess? | 21:44 |
dan9er[m] | Well i've installed that and restarted, still no network | 22:11 |
buZz | aw, well , maybe try to lookup other people with bcm4630 | 22:14 |
buZz | i mean, their online postings | 22:14 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!