libera/#devuan/ Monday, 2023-03-20

brocashelmsfox: elogind is systemd's logind forked for systems not using it (e.g. runit, sysvinit, openrc, s6)00:02
brocashelmsfox: there are alternatives that you could also try like seatd or consolekit200:02
sfoxwhy would i want that?00:02
XenguyI'm not sure if it is relevant in this case, but when I did an apt-get dist-upgrade from ASCII to Beowulf...00:03
brocashelmelogind is a hard dependency of many packages, and without something else in its place, your system can get nuked (especially if you use gtk/qt programs)00:04
XenguyThere were some funky dependencies that were not resolved...00:04
brocashelmas of chimaera, elogind is pushed more than ever00:04
Xenguyfsmithred helped me get that sorted out manually (I would not have been able to resolve it on my own) and the gist of it was (I think) I needed to replace consolekit with elogind00:05
brocashelmremoving elogind automatically makes apt prompt you with an install of consolekit2 (at a minimum, you lose programs like network-manager, gvfs, and udisks2 if you don't need those)00:05
Xenguy^^ Hope I got that right00:05
brocashelmso if you don't even have consolekit2 installed, then you are prompted with more core packages being removed due to that hard dependency on either of those two (elogind and consolekit2)00:05
brocashelmso "disabling" elogind is not an option IIRC00:06
XenguySo I get the impression that elogind is the preferred more recent software of the two, but again I am no expert at all00:06
sfoxso if it's a package dependency,00:06
brocashelmelogind is the systemd of non-systemd distros now00:06
sfoxcan i stop it from locking and rebooting my computer all the time with rc-update disable elogind?00:06
sfoxelogind is super buggy00:07
brocashelmit's bloated00:08
brocashelmlike i said, hard dependency. removing it is possibly the worst thing you can do (aside from dbus) if you want a functional desktop00:08
sfoxI can't use my laptop without having to open the lid, login, wait for the it to suspend, press power button, hope nothing broke, login again, hope nothing else screwy happens00:08
brocashelmmaybe check with the acpi packages there?00:09
brocashelmdepending on your DE, xfce's power manager is pretty straight-forward about what to do when you press a button or close the lid00:09
sfoxi'm not using a DE00:12
sfoxi was using ACPId in beowulf00:12
brocashelmi see there's repowerd available00:15
brocashelmit's a daemon for power management00:15
Xenguysfox: Again this may or may not be relevant to your use case, but to suspend my laptop, I use the 'pm-suspend' command; it is found in the 'pm-utils' package...01:00
XenguyThere are ways to map this to the 'function' keys of some laptops, but running the command itself often works just fine, depending on your hardware01:01
gnarfacesfox: not sure what's best in your case. do you have a /etc/elogind/sleep.conf?01:11
gnarfacei would disable it as a test, just to help diagnose the actual issue01:13
gnarfacebut i still think there's some line in a config that you can use to disable just the one feature that's tripping up01:14
sfoxlet me check01:27
sfoxactually rc-update delete elogind seems to of fixed the issue01:28
gnarfacewell that just disables the whole thing01:29
gnarfacebut i can't be sure wasn't doing something you still want it to do01:29
sfoxgnarface, I don't have a /etc/elogind/sleep.conf01:29
gnarfacehmm, odd01:29
gnarfaceand there's no [sleep] section in the main config?01:29
sfoxthere is, but everything's commented out01:30
sfox/etc/elogind/logind.conf [Sleep] section01:30
gnarfacecheck the docs on the commented options, maybe one of them just disables the extra sleep state request that's firing01:30
sfoxwell why do i even want elogind?01:31
sfoxI'm not sure i was using it in beowulf01:31
gnarfacewell maybe you don't, but it's necessary for some graphical logins01:31
sfoxis it needed for slimlick?01:31
gnarfaceno idea01:31
gnarfacetry "apt-cache rdepends elogind"01:32
gnarfaceor wait, maybe it's just depends not rdepends, i always get them confused01:33
gnarface"aptitude why elogind" should also work01:33
gnarfaceanyway, you can just disable it if that's working for you01:33
gnarfacei usually don't use it here but i don't use a graphical login either01:33
gnarfacethere's consolekit as an alternative but it doesn't support all the same graphical logins01:34
gnarfaceif you're just using "startx" to get into Xorg you probably don't need it01:34
sfox please if you want me to look01:34
gnarfaceit would be rdepends i guess sorry01:37
sfoxi don't have ukui or 'podman' or mate installed01:38
gnarfacethe going theory here is that you have some redundant hook firing in both elogind and acpid which is causing the issue. you should in theory be able to disable the one in elogind without disabling the whole thing, but off the top of my head i don't know how01:38
sfoxi don't seem to have an dependencies using elogind01:38
gnarfacemaybe it was pulled in as a recommend01:39
sfoxoh i see01:39
sfoxi can't remove it eithe01:39
gnarfacego ahead and disable it and reboot to see if anything breaks01:39
sfoxotherwise a whole bunch of packages which aren't even using elogind get uninstalled01:39
sfoxwhy would slim need elogind?01:39
gnarfacei do think it's expected for slim to need elogind, but i'm not clear on why were' not seeing it in this list, maybe it's removed one or two steps up the dependency tree01:40
sfoxok back01:41
sfoxyeah everything seems to work fine01:42
sfoxit's just the out-of-order locking where the system suspends before slock finished executing01:42
gnarfacei think it's possible you may just loose some functionality like the ability to power down by UI as a non-root user01:43
gnarfacebut maybe anything not requiring root access will still work01:43
sfoxwell i have a real power button i could use and i'm the only user of this computer so i can just set a stickybit on those binaries01:44
sfoxspeaking of01:44
gnarfacethat's basically what things like elogind are for, they're for allowing root functionality to the non-root user for some specific power and session management tasks01:44
sfoxwhy do we need a whole daemon and xml based permissions to turn off a computer instead of a setguid stickybit on power managmeent executables?01:44
gnarfacei don't know man, i still just use su, the whole thing seems like a giant security hole-punch to me01:45
sfoxwell thanks for the help01:45
gnarfaceon the systems i'm not sharing with anyone, i just log in from the text prompt and run startx01:46
sfoxon the systems i do share, i just add users to wheel who manage the system01:46
gnarfacethe whole thing to me seems like a childish attempt to chase behavioral parity with commercial operating systems01:47
gnarfaceyou can uninstall it and all its dependencies if you don't use any graphical logins01:48
gnarfacebut then if you're also not running Xorg as suid root... obviously you lose access to your graphical power buttons. do we still need those when we have physical power buttons and the "shutdown" command? well it depends on who you're trying to impress, i guess.01:49
gnarfaceyou got two types of people out there basically; the type of person who adapts their mind to fit the task as necessary, and the type of person who will rage quit their job if the computer requires them to type a password to shut it down01:55
gnarfacei do still think there's probably a way to make elogind behave without throwing the whole thing out the airlock but i really don't know the specifics, and there's a possibility some hardware or driver level issue is the real culprit here01:59
gnarfacefsmithered might know more from working with this setup01:59
sfoxI'm open to exploring it but the description of what it is to me of something taken from systemd makes me biased against it02:09
sfoxassume it's some kind of poorly made and thought out EEE takeover from redhat02:09
gnarfacethat's basically the truth of it, but it's important to remember the enemy is systemd-logind in this case. elogind is what's actually trying to save people from systemd.02:12
gnarfaceor one of the things, anyway.02:12
gnarfacedesigned as a drop in replacement with out the explicit systemd tie-in02:13
gnarfacesfox: it would actually be interesting to find out if systemd-logind is misbehaving the same way on your hardware02:20
gnarfacethough if debian removed acpid that might be hard to test02:21
gnarfacebut it also seems like if the problem was that simple more people would be seeing it02:21
brocashelmseatd is a more minimal replacement for it, but elogind is the harder dependency02:30
brocashelmelogind is constantly fighting upstream decisions while seatd goes a simpler, minimalistic direction02:31
brocashelmkind of like switching from sudo to doas02:31
brocashelmif nothing else, try another distro that does not have systemd or elogind by default02:32
brocashelmantix, hyperbola, pclinuxos, etc.02:33
xrogaanI'm on Daedalus. On powerdown, the networking stuff tells me it can't configure networking interface. Can't find the exact error message though. ANybody with that issue?21:00
gnarfacexrogaan: seems strange that it's trying to configure it on power down... you using network-manager?22:16
debdogI am almost certain I have seen this on my laptop when the eth0 cable is pulled during shutdown. tried to reproduce it right now didn't work.22:30
debdogmaybe it has t be pulled at a certain point. I use wicd there22:30
debdogon Cimaera that is. not sure about Beowulf22:39
fsmithred'allow hotplug eth0' vs. 'auto eth0' issue?22:45
xrogaangnarface: it's not configuring it, but cleaning removing something22:45
xrogaanShutdown is too fast, I can't really read everything properly22:46
systemdleteThis is more about hardware.   My FX8350 on the PC I use the most is running around 79.2C.   The output from sensors says the critical level is 80C and the hysteresis is 77C.   There is a lot of dust clogging the heat sync on the CPU, and I think the fan blades may be dusty also.  I will clean it soon, but is this a serious problem for a22:46
systemdleteFX8350, being that it is said to be a hot-running CPU?22:46
xrogaansmell test would tell me it's something about mounting network file system, and a clean unmount at shutdown. But that's a stretch.22:46
gnarfacethere should be a way to get a log of it22:47
xrogaansystemdlete: yes22:47
systemdleteyes... it is serious or it is normal for FX8350 to run this hot?22:48
debdogIn my case nfs was not the issue.22:48
gnarfacesystemdlete: it's about 50 degrees higher than mine at idle22:48
xrogaantemp at idle or during use?22:48
gnarfacesystemdlete: i'd say it's a serious problem. clean that fucking fan22:48
systemdleteok, gotcha.   Will do sooner than later.   I need to get some compressed air.22:49
systemdletebest way to clean the fins of the sync?22:49
xrogaansystemdlete: it's very bad.22:49
xrogaanMax. Operating Temperature (Tjmax)22:49
gnarfacei have a really nice fan on mine, so it's reasonable to expect that you may only be 30 degrees higher than what you should be at idle, but i'd shut it down and take the fan off the heatsync then take compressed air and a vacuum cleaner to it22:49
fsmithredI'd expect 61C to be minimum operation temp on FX chip22:50
xrogaanit's stated as max though22:50
gnarfacefsmithred: nah i have the same cpu here with a noctua fan and it idles below 3522:50
fsmithredoh, ok22:50
gnarfacefsmithred: maybe with stock cpu cooler and that shitty clay junk as thermal compound you'd get 60 but i take my shit more seriously than that22:51
systemdleteI will also clean the intake and outtake fans on the case while I'm at this.   And I'll see if I can clean up the PS fan(s) if possible.22:51
gnarfacefsmithred: nice aftermarket cooler and arctic silver thermal compound beats everything stock22:51
fsmithredI'm used to looking at laptop temps. I can warm them up a lot making isos.22:51
systemdleteI have a wraith that came with the CPU.  But this is like the first version of the wraith I think.   There are better ones now.22:52
gnarfacesystemdlete: make sure all those fans even work still22:52
systemdletegnarface:  They turn and they are pretty quiet.22:52
xrogaantjmax means that reaching that temperature the CPU's thermal control system kicks in to reduce power and limit temperature.22:52
systemdleteand they do vary in speed sometimes.22:52
gnarfacesystemdlete: 80 idle sounds like temperatures from a laptop with a dead fan22:52
systemdletedesktop fan22:52
systemdletedesktop sysstem22:52
systemdletethis is a desktop.22:53
fsmithredclean dust first22:53
gnarfacedust in the main heatsync can be a serious problem for heat transfer22:53
systemdletethen what?22:53
fsmithredthen re-check temp.22:53
systemdletefsmithred, "clean dust first"--then what22:53
gnarfacestart with that, just do it good. then check the temperature again22:54
fsmithredIf you remove any dust, you will see a difference in temperature.22:54
systemdleteyeah, that was what I had in mind too22:54
systemdleteok guys, gotta shut this beast down.22:54
systemdletego get some compressed air.22:54
fsmithredget some cats and you can watch the machine shut down when it overheats22:54
systemdleteI'm also going to clean and replace the thermal paste (I only use a tiny drop, I know...)22:54
systemdlete(I think I'm allergic to cat hair and dog dander)22:55
systemdlete(no pets here at this time)22:55
fsmithredyeah, amd stock heatsinks use a sticky pad, not silver paste22:55
systemdlete(and I don't have any hair to speak of... LOL)22:55
systemdletefsmithred, so is it good to replace it with grease (thermal paste)?22:56
systemdleteis it better?22:56
systemdleteor leave it alone22:56
fsmithredI've done both22:57
fsmithredif cleaning the dust doesn't get it cooler, then maybe you should replace it22:57
systemdleteok, great.  thanks everyone.  You guys are the best, always.22:57

Generated by 2.17.0 by Marius Gedminas - find it at!