libera/#devuan/ Tuesday, 2021-12-28

gnarfacei'm actually highly skeptical that upstream default change fixed anything for anyone00:00
gnarfacein every test case i have here, it either did nothing or crippled the network at boot00:00
Kingsyabsolteuly prefect. that worked a treat.00:00
gnarfacethat was it?  sweet.  always nice when it's something simple at least00:01
Kingsyhehe! awesome00:01
systemdletegnarface:  I have a pcmcia wireless G card myself.00:02
gnarfaceKingsy: the long network stop time might be something fixable too, maybe related to a mismatch in your system hostname and dhcp server's configuration, or something like that00:02
* systemdlete not sure why he blurted that out, but now he feels so much better that he said it00:03
Kingsyah thats possible. less concerned at the moment.00:03
gnarfacesystemdlete: pcmcia g?  how'd you find one of those?? i didn't think they even existed00:03
systemdleteMy older brother gave it to me about 25 years ago00:03
systemdleteHe was getting a newer one and saw I could use it00:03
gnarfaceKingsy: you'd have to make sure /etc/hostname, /etc/hosts, and the dhcp & dns server's opinion of your hostname match00:03
Kingsyhow does the "sleep" functionality work out of the box on devuan? if I close my laptop it sleeps but when I open it again it doesnt wake. or at least it wakes but I don't get i3 back. I need to hard reboot00:03
Kingsygnarface: thanks for this!!00:04
systemdleteI used it for a while in a laptop (if you can call it that).  It weighed almost as much as our dog.00:04
systemdleteThe dog was much easier on my lap than that machine.00:04
gnarfaceKingsy: sleep is an ongoing problem no different from in debian.  when it fails to wake up properly it's usually to do with some driver bug (NVidia proprietary drivers are very common culprits here, but so are many wifi devices) - hibernate is known to work better if your swap partition is >= to your RAM amount00:05
Kingsygnarface: how can I stop it sleeping then?00:05
Kingsyhappy to just turn off the functionality for now. or switch it to hubernate.00:05
rrqKingsy: if you have elogind its logind process handles th elid switch by default00:05
systemdletegnarface:  More than happy to mail it to you.00:05
systemdlete(I even have the box it came in)00:06
gnarfaceKingsy: i think the window manager calls acpid scripts, it should provide a interface to disable sleep i think00:06
KingsyI don't know what elogind is, so I am guessing I don't have it.. hah! :D00:06
rrqKingsy: if so, edit /etc/elogind/logind.conf and configure all HandleLid* to "ignore"00:06
gnarfaceKingsy: elogind is a devuan created drop-in replacement for systemd-logind00:06
gnarfaceKingsy: (no, you don't need it unless you're using certain graphical login prompts)00:07
Kingsywell /etc/elogind/logind.conf <- does exist for me00:07
Kingsybut everything inthe file is commented out00:08
rrqthose comments also documents the default setup00:08
Kingsyah ok!00:09
rrqyou'll need to restart the logind service after changed conf00:11
Kingsyone more question, haha cna you tell I am fresh on devuan from a different distro? I am using runit, why do I even have anything in /etc/init.d/* ? I was expecting /etc/sv/ and /var/service/00:11
Kingsyrrq: thanks somuch for the info!00:11
Kingsyrrq: for example you said you'll need to restart the logind service.. its not in /etc/sv/  so how is runit even aware of it?00:12
rrqI'm not famililar with runit, but I'd assume it 'assimilates' the sysvinit conf's for convenience00:13
Kingsyheh then I am confused why /etc/sv/ even exists.00:14
Kingsyhm I mean no, looking at a file in /etc/init.d/ those are not runit scripts.00:14
gnarfaceKingsy: debian did something weird upstream of us with several of the alternative init systems before systemd was rolled out, which included, amongst other things, bolting on compatibility for the sysvinit scripts to all of them.  it's usually not a hugely complex task to undo.00:16
Kingsygnarface: I am not sure what you mean. sysvinit is included regardless of what init system you choose?00:18
Kingsyhehe sorry if I am being dumb here.00:19
gnarfaceKingsy: ah, no you're not being super dumb.  your analysis is hyperbolic but not inaccurate in facts.00:19
gnarfaceKingsy: just don't freak out; they did have sane reasoning behind it, this wasn't the same guard that did it as that rolled out systemd.00:20
gnarfaceKingsy: and, like i said, you can get a bare runit installed, you just need to talk to someone here who has done it00:20
rrqyes, the case is that packages that distribute a service include start/stop configurations for that service; those files belong to the service package and not the init system (package)00:21
gnarfaceKingsy: (embarrassingly enough, i actually always liked sysvinit so this isn't something i've explored.  i just know you're not the first to ask nor would you be the first to do it - for runit or openrc)00:21
Kingsyah ok! sounds like I'll need some more info on it I suppose. so just curious really. So when a package is installed, lets say... nginx. does that come with the appropriate runit files which will allow runit to control the daemon? and if so where? with void linux for example. new runit services were in /etc/service. in order to allow runit to manage them you simply just did a ln -s /etc/service/nginx00:23
gnarfaceKingsy: uh... so here's the thing.  it's supposed to but for runit, that work might largely have not been done yet for many packages.00:24
gnarfaceKingsy: on the plus side, i hear the task is much simpler for runit than for sysvinit00:24
gnarfaceKingsy: (tbh that work was never even done for all of the sysvinit packages, and i'm not sure nginx even got one)00:24
Kingsywhat do you mean? all packages much have at least an sysvinit script right? are we just saying that if a package doesnt have a runit service. the sysvinit will be included?00:25
gnarfaceKingsy: what i'm saying is, less than 100% of the packages that need a sysvinit script even have them.  but it's closer to 100% than with runit.00:26
gnarfaceKingsy: i know how it sounds, but trust me this is not a reason to panic either.00:26
Kingsyif a package doesnt have a sysvinit script.. how does it even function and do its job?00:27
gnarfaceKingsy: yea you just have to start it manually.00:27
Kingsyah got you.00:28
gnarfaceKingsy: but making a basic sysvinit script isn't a difficult task, and neither is hard-wiring a process to start without an init script by just throwing the bare command into /etc/rc.local00:28
Kingsyyeah or writing a runit service to do it too. not the end of the world.00:28
Kingsygnarface: is there any documentation on runit within devuan?00:28
gnarfaceKingsy: and remember, only userspace daemons, aka "services" actually need a init script.  not literally every packages00:28
Kingsyyeah true00:29
gnarfaceKingsy: dunno about documentation.  in general for documentation, first you check the man pages, then you check /usr/share/doc/[package name], then you check for a package named "[package name]-doc", then you check the homepage link in the package description00:30
gnarface(and that's not just for devuan; those instructions apply to basically any debian derivative, and probably mostly applies to most other distros too)00:30
Kingsythis is great. thanks for the info! I am quite enjoying my devuan experience so far. and its nice there is a good bunch of folks on here. makes it more pleasant00:31
Kingsygnarface: chers00:31
gnarfacecheers :)00:31
systemdletegnarface:  Any way to get 1920x1080 with nouveau?  I've looked at several posts online but I can't figure this out.  I tried adding video=1920x1080@60 to the grub linux command line, but no help there00:33
systemdleteThis is not terribly important, but it would be nice if not too much effort.00:34
gnarfacesystemdlete: i'm not sure it's gonna; make sure the card could even do that physically.  i forget how much ram was on that one.  but also check the xorg log to make sure it's detecting the right amount of ram and if not, specify the right amount of ram.  (i used to have that problem with some cards)00:34
systemdleteah, ok00:34
Kingsysystemdlete: yeah its probably ahardware limitation, what does xrandr say for the output?00:35
gnarfacethe xorg log should actually list all the valid detected resolution&refresh range combos00:35
Kingsyoh good to know!00:35
gnarfaceoh, i don't know if this is still a problem, systemdlete, but i used to have noticed, especially with CRTs, some limitations in selecting some resolutions unless i also specified the manufacturer-supplied manual's definition of the valid horizontal and vertical refresh rates for the display, and not trusted EDID to it00:38
systemdletethat's xrandr output00:39
gnarfacewhen i used to do this, i was also passing nomodeset to grub though, to avoid unrelated driver bugs00:40
Kingsyyeah I meanthat says to me you are limited by hardware to 1680x1050500:40
Kingsygnarface: might know more but I cant imagine it being a driver or config issue.00:41
gnarfaceyea it might even be a hardware issue00:41
gnarfacei think it won't go below 640x480 either00:42
gnarfacenvidia hardware is weird00:42
KingsyI am starting to hate nvidia tbh. I was going to get a gaming laptop with a nvidiagtx in it. but I am starting to think thats a bad idea. no idea how radeon compares performance wise to the proprietry nvidia drivers mind.00:43
gnarfacethe newer batch of AMD cards is sick00:43
systemdlete(I hope that's the right extract from the xlog)00:44
gnarfacei just switched after many decades, and 3rd party game compatibility is only almost up to par with nvidia but i don't regret it.  the linux drivers are just so much higher quality and better integrated with the distro's native tools00:44
systemdleteso I am SOL on this, huh?00:45
gnarfaceyea, it's probably never gonna do 1920x108000:46
systemdleteI mean, again.  This is only a test box.  So I don't care too much. I'll probabbly be ssh'ing in from another system most times anyway, or hitting openwrt via its web interface.00:46
systemdleteI appreciate the help though.  Educational, if nothing else.00:46
systemdleteThis is one old board!  15+ years old, per the firmware info.00:47
gnarfaceyou could try manually putting in a custom modeline to the xorg.conf00:47
systemdleteThat can do damage to my monitor?00:47
gnarfaceit'll either work or not.  you can't blow up the screen on a LCD like you could with a CRT i don't think...00:47
systemdleteoh, ok.00:47
systemdletemaybe worth trying then00:47
gnarfaceeven most the later generation CRTs were pretty solidly protected against oopsies like that and would just hardware disable themselves on the fly if fed something out of range00:48
systemdleteIf you don't see me here for a few days, you'll know why...00:48
systemdletethanks again Kingsy and gnarface00:49
gnarfacebut, crafting a proper modeline isn't something i can really help with.  the one time i had to use one and did so successfully it was because the display was unique enough to have a forum of users supporting it and someone else posted the right modeline00:49
HurgotronI think it's been 20 years ago since I last crafted a mode line for X11 :)  I seem to remember there was a tool for that, the fine tuning at least00:59
Hurgotronfsmithred: I think that was it, thanks! Couldn't find it anymore01:02
fsmithredI see it at /usr/bin/xvidtune but I think what you really need is a modeline generator. You can probably find one online.01:03
systemdletewell, since I am sharing the same monitor between 3 PC's... I'd think that the timings for the 2 that are getting full resolution would be suitable as timings for the one that isn't getting it01:12
systemdleteor is that flawed thinking?01:12
fsmithredI think if all three vid cards can do the same settings, then it should be fine.01:15
HurgotronIf you are using the same OS on all computers, yes, that should work. If it doesn't the hardware likely doesn't support the designated mode line(s) for the monitor. But you can probably get it to work with a lower frame rate, or somesuch.01:15
duds-Is GNOME over Wayland well supported? I find it weird that GNOME itself isn't an option in the installer considering Debian does, GNOME Flashback is an odd replacement (would be great as another option, though)03:07
golinuxYou can set up GNOME post-install03:13
* golinux can't imagine why anyone would want to but different strokes03:14
Xenguy"Debian's Not GNOME"03:20
golinuxGetting closer everyday (and offtopic)03:23
systemdleteI am running apcupsd (maybe that's a problem by itself, idk?) for 2 UPS's.  I have each one configured in a separate file in /etc/apcupsd.  I modified the init script to launch both instances.  After a while, both seem to have a crash or the like.03:44
rrqHurgotron: try "cvt - calculate VESA CVT mode lines"03:46
systemdleteNow, I've violated a rule about specifying usb devices in the config files.  Mine are both USB, so how can the configs know how to distinguish these if they are both USB?  That's why I specified the device in the DEVICE statement.03:46
systemdleteI've looked all over the web for a solution; it seems a lot of people have this same issue when they run multiple UPS's and apcupsd.03:46
systemdleteWould I be better off with NUT instead of apcupsd?03:47
systemdleteor some other solution?03:47
systemdleteby "after a while," I mean several hours or even several days.03:49
systemdleteI should probably hit something like #apcupsd huh?03:51
fluffywolfI've only used apcupsd on one ups at a time, on rs232, and it worked fine...03:57
gnarfacei think apcupsd might need some help03:58
gnarfacei recall having some sort of issue with it aside from it being a pain in the ass to figure out how to get usb even working03:59
gnarfacethe memory is very vague but crashing does ring a bell04:00
gnarfacei think i also ran into an issue with it missing support for newer models04:01
gnarfacebut then i ran into an issue with cheaper models not having the right features and gave up on it completely04:01
systemdletethanks. At least I know I am not the only person losing their mind over it04:04
systemdleteNow that I've re-read the "doc" for the umpteenth time, I realize that it is not some cardinal sin to spec the DEVICE field.  It just says to leave it blank is the best solution for MOST installations.  THAT part of the doc is silent on multiple UPSs.04:05
systemdleteI can't find a apcupsd channel.  The last release was 5 years ago.04:06
fluffywolfI stopped using it not because there was anything wrong with it, but I got a ups that talked SNMP, and wrote a little script to do what I needed through that.04:09
gnarfacesystemdlete: my analysis was that it was never updated after the era of ubiquitous usb connectivity for apc upses04:49
gnarfacesystemdlete: yea, that entered into how hard it was to figure out what to do, but in the end it's a really simple config change that should be the current default in practicality04:50
systemdletewell, I'm going nuts now.04:51
systemdleteI mean nut(1)04:51
systemdleteI'm looking into how to configure it now04:51
systemdletemaybe I'll have better luck with it. It seems to be in maintenance.04:51
systemdleteunlike apcupsd04:51
systemdletegnarface:  You make it sound like it is a really simple fix?04:55
systemdleteI mean, I'm looking at nut now, but if you have a better solution using apcupsd, I'll reconsider.04:55
systemdletefluffywolf:  apcupsd worked fine here also.   For one UPS.  It was when I tried to follow their instructions for multiple UPSs that the world got complicated04:57
systemdleteIt sort of worked, but every so often it crashed.04:57
systemdleteI'm not knocking the program, btw.  All I'm saying is that it is not stable enough, or at least for my config.04:57
systemdletenut supports multiple UPSs in a variety of configurations.04:58
systemdlete(from what I am reading about it)04:58
onefangI'm sure you'll nut it out.  B-)04:59
fluffywolfI learned the other day that "nut", as a verb, did not mean any of the things I thought it did.05:01
systemdletewell, I just discovered a possible reason apcupsd wasn't working quite right -- xfce power daemon was running05:10
systemdleteit might have been querying the units at the same time my own scripts were querying them and caused some kind of collision?05:10
systemdleteI've disabled power manager for now.05:11
systemdletedisabled it in startup05:12
systemdletenot sure how to kill it05:12
systemdletegotta restart I guess... bbl05:23
XenguyMay the fourth be with you05:24
onefangWrong time of year.  It's almost Hippy GNU Ear!05:27
XenguyMy bad, all for the GNU though05:36
rktaI can see null bytes in my /var/log/boot, is this normal?12:52
debdogI don't even have that file12:56
debdogthe kernel messages during boot go into syslog12:57
HurgotronThe /var/log/boot has the init startup stuff12:59
plasmoduckis there an Australian package mirror I can use for devuan?13:06
onefangplasmoduck: Nope.13:10
Mirko[CHIP]Hi there13:14
plasmoduckI'm not getting the best speeds13:18
plasmoduckwhat about if I set one of the Australian Debian packages13:19
onefangPackage mirrors typically redirect to for the bulk of packages shared with Debian.13:21
rrqplasmoduck: you could try nz.devuan.org13:50
rrqno nz.deb.devuan.org13:51
plasmoduckrrq: in my sources.list I have do I change it to ?14:19
plasmoduckrrq: I see exists15:15
fsmithredplasmoduck, I think nz is the only country code that actually does what it advertises.15:21
bb|hcbplasmoduck: try nz because its in new zealand, while au is in europe (a default, because no mirror is there)15:22
plasmoduckau is austrralia15:27
bb|hcbyes, but because there is not australian mirror it points to in europe15:38
plasmoduckah ok16:07
metalaI have reinstalled my Devuan/chimaera on my laptop but I am hitting an issue w/ Docker + docker-compose. My previous install was beowulf -> chimaera. Dmesg shows "cgroup: Unknown subsys name 'elogind'". When I kill elogind and umount /sys/fs/cgroup/elogind then docker-compose works.17:57
metalaThe weird thing is that in my previous install (I still have it on an external drive), I also have elogind and /sys/fs/cgroups/elogind mounted, but I don't have the issue.17:58
metalaSame docker version, same docker-compose version.17:59
metalaAnother thing, I have installed podman, but it seems it depends on systemd, it writes to stderr "ERRO[0000] unable to write system event: "write unixgram @0002c->/run/systemd/journal/socket: sendmsg: no such file or directory""18:00

Generated by 2.17.0 by Marius Gedminas - find it at!