libera/#devuan/ Saturday, 2020-11-14

clortanybody have a fix for python (pip2, pip3) erroring when installing something, i always get12:17
clortsubprocess.CalledProcessError: Command '('lsb_release', '-a')' returned non-zero exit status 112:17
clortsolutions i found are not working for me with devuan ceres12:18
clortlsb_release is provided by package lsb-release12:24
clorthow is python so broken.  i can get around it by faking /usr/bin/lsb_release to return something sensible12:30
clortcan i install python-fontconfig without using pip?12:32
cgb_orig<- channel n00b seeking assistance12:36
onefangSeek and ye shall find.  Perhaps.12:38
clortwho changed all the mountpoints?  i used to have root mounted on /12:38
clortnow i have /media/user/UUID or something12:39
cgb_origattempting to resolve xhost unable to display error12:39
cgb_origwas a lot easier on Solaris in my day12:39
clortoh nvm sorry12:39
cgb_orighas anyone experienced/resolved "unable to open display" before and willing to share their expertise?12:48
onefangSome more detail on what it is you are doing might help us help you.12:49
clortwell i got a nice little python script working now that will tell me every installed font that has a particular character or codepoint \p/13:02
clortjust pip is all broken13:02
clortcgb_orig: you'll want to have the DISPLAY environment variable set to point to your active display, if launching an X app from a shell13:03
luser977cgb_orig: echo $DISPLAY comes up empty?13:03
fsmithredecho 'ALWAYS_SET_PATH yes' >> /etc/default/su14:23
clortanyone here using zram with a 4GB RAM machine?14:34
cgb_orig$DISPLAY=":0.0", "127.0.0.1:0", "localhost:0" - all respond "unable to open".  Have already xhost +'d15:13
clortcgb_orig: are you invoking the program from a local x-terminal?15:17
cgb_orig@clort: correct, sudo'd15:20
aitor_cgb_orig: are you trying to run X as root?15:21
cgb_orig@altor: correct15:22
aitor_do you have a ~root/.Xauthority?15:24
ShorTiethat don't work real well15:24
aitor_ShorTie: what do you mean?15:25
cgb_orig@altor - yes. @ShorTie - I'm aware of security implications - I use it for testing15:25
cgb_orig@aitor, apologies, small font15:26
ShorTienot saying it can't be done, but most stuff install for user i do believe15:28
clortwhat program are you invoking before you get 'unable to open' cgb_orig15:29
clortwhen you report an error, it helps to share what you did to generate the error15:29
clortbecause we cannot see inside your brain15:29
cgb_origResolved! xhost :local:root.  Will disable with xhost - once finished15:30
clortok15:30
ShorTieunable to open means to me your going to the wrong screen, like :1 instead of :015:31
cgb_orig@clort - indeed, thanks.  I was invoking a number of programs all generating "unable to display" - I will be more specific in further queries.  Thanks for your time and see you again.15:33
clortcheers15:43
aitor_did anybody give a try to Ubus?15:44
aitor_https://openwrt.org/docs/techref/ubus15:44
aitor_I'll change my question: did anybody give a try to OpenWrt in general?15:51
clorti'm pretty sure there is an openwrt chat15:55
aitor_i'm looking a t the forum, thanks15:57
aitor_clort: yes... https://webchat.freenode.net/16:00
clortmaybe #openwrt or ##openwrt16:00
aitor_forget the link above16:03
aitor_there are three channels: #openwrt, #openwrt-devel and #openwrt-adm16:03
aitor_i have .deb packages for libubox and libubus including some working examples16:08
aitor_bbl16:18
unixbsdConcerning a harddisk, with ntfs, is it possible to format ntfs with mkfs.ntfs with badsector (excl. during format)?20:12
unixbsdis the  main package site offline, and down currently?21:00
unixbsddeb.devuan.org seems unstable currently, for 10 min ago.21:01
systemdletedevuan ascii (host not vm) system occasionally causes screen "flashes" (quick off-on? I think).  This has been happening occasionally, but today, it actually caused a freeze-up of sorts.  The system would update the clock, but the overall system seemed to be frozen.  I couldn't input or use the mouse.  (I forgot to try to get a console, sorry.)21:15
systemdleteThe kernel log shows a ton of messages like:  "radeon 0000:01:05.0: GPU lockup (current fence id 0x00000000037f8025 last fence id 0x00000000037f80bc on ring 0)"21:16
masonsystemdlete: Overheating?21:16
systemdleteI ddg'd this a bit and discovered that other people have run into this.21:16
masonCould be a bug, but I'd tend not to expect a bug to show up over time.21:16
systemdletehttps://gitlab.freedesktop.org/drm/amd/-/issues/85621:17
masonsystemdlete: https://bugzilla.kernel.org/show_bug.cgi?id=62721 has an interesting idea at the bottom21:18
systemdleteI know I need to upgrade to beowulf, but I've been busy doing other things first.  It is on my list.21:18
systemdletethanks mason.21:19
masonDon't know if it'll help, but it's a relevant knob.21:19
systemdleteI might try it if I experience it again.   Did you look at the link I posted?  I think that is a more complete analysis.21:21
systemdletebtw, this is not a laptop.  I think the user in your link was running a laptop, so your overheating theory makes sense (laptops do all sorts of upsetting things)21:23
systemdleteI admit that this ASUS board has, from time to time, been problematic.  Just yesterday, the onboard ethernet port went out, and I had to reseat/jiggle the connector until it started working again.21:24
systemdleteMost of the time, this system runs just fine.  I suppose the onboard radeon chip might be an issue though.  Have you heard of this?21:24
systemdleteGuess I should offer up my kernel version also:  4.9.0-14-amd6421:27
systemdleteI was running a number of VMs also (about 6 or 7 I think)21:27
systemdleteThere was plenty of RAM left, though.  Only about 16gb of the 32gb were in use, per htop.21:27
masonsystemdlete: I looked at your link. I didn't see discussion but I'll look again.21:39
masonOh, following it to the original bug, there's more data.21:39
systemdlete(they migrated to a new site)21:46
systemdletehere is another message from the kernel log just before I recycled the box:  " [drm:radeon_ib_ring_tests [radeon]] *ERROR* radeon: failed testing IB on GFX ring (-110)."21:46
systemdletemore info:  Kernel log looks like this is rs780 chip, fwiw21:53
WafficusHi there, I installed Jami on Devuan via apt, but running "jami" doesn't seem to launch it22:08
WafficusI found the package in Debian, but it doesn't appear to be named anyhting differently22:10
Wafficushttps://packages.debian.org/sid/jami22:10
DeknosWafficus, try running "gnome-ring"?22:18
Deknoscould be that the binary is still called that name.22:19
Wafficusyes you're right22:20
Wafficusit is gnome-ring for some reason22:20
WafficusI think its cause they changed the name to Jami recently22:20
Wafficusthanks a ton!22:20
systemdleteuh... I wonder, maybe this is important?  I have no /lib/firmware which means there is no amd microcode updates.  Is this correct (devuan ascii) for my amd system?22:37
systemdletemason: If a microcode update fixes the gpu issue, then maybe this is why I am having this problem?22:37
gnarfacedepends on how old your system is22:37
systemdletehow old... well, this is a fx8300 processor I think22:38
systemdletefx8350 maybe?22:38
gnarfacenot old enough.  you need the microcode updates for security, at least22:38
systemdletebut I guess you want to know the chipset22:38
gnarfacecan't tell if it'll be relevant to your other issue22:38
gnarfaceno, FX tells me all i need to know - the FX series are the oldest AMD processors vulnerable to the spectrum/meltdown vulnerabilities.22:39
gnarfaceso you will at least want the microcode upates for that, regardless of it helps with any other GPU issues (i doubt it actually)22:39
systemdleteso I need to install it then?22:39
systemdletehelp me here22:40
gnarfaceyea.  phenom and earlier don't do speculative execution, but from what i read, the FX series has preliminarily implemented some of it22:40
gnarfaceso yes, install it22:40
systemdleteah, I see.22:40
systemdletewhat did you mean by "not old enough" then?22:40
gnarface"not old enough to predate speculative execution vulnerabilities"22:41
systemdleteat any rate, it won't HARM anything to install it and use it, right?22:41
gnarfaceif you had a pentium II it wouldn't be vulnerable either :-P  follow?22:41
gnarfacei've never seen it harm anything.  on a mobile processor i suspected it made the idle a bit hotter, but i can't be sure that wasn't just impending fan failure (the fan failed completely soon after that)22:42
systemdlete"either?"  (sorry, trouble parsing by this old man here)22:42
gnarfacea pentium II wouldn't be vulnerable22:42
gnarfacejust ignore the either.  it's a superfluous word in that sentence.  i'm sorry i use too many words.22:42
systemdleteok22:42
systemdleteI get confused by these things, sorry.22:43
gnarface(actaully any single-core chip would not be vulnerable because the vulnerability fundamentally relies on multiple cores)22:43
systemdletespectre?22:43
gnarfaceyea.  but not ALL multiple-core systems are vulnerable, just the ones that do "speculative execution" or "branch prediction"22:43
gnarfacewhich was really popular after the core2duos but then quickly became very unpopular recently so new ARM chips are avoiding it22:44
systemdleteistm that boot log mentions spectre22:44
systemdletebut I don't have microcode installed22:44
gnarfacewell the chips ship with some microcode22:44
systemdleteis the spectre bit in the kernel itself then?22:44
gnarfacethe updated microcode just replaces it on the fly22:45
systemdleteI mean, compiled in?22:45
systemdleteistr it was, but maybe my memory fails me (as it often does these days)22:45
gnarfacethere is kernel components too i think, yes, that also had to be patched.  update your kernel.22:45
systemdleteI'm running the latest I think, for ascii anyway22:45
systemdlete4.9.0-14-amd6422:46
gnarfacei'd assume the beowulf kernel has the patch, but i can't tell you for sure that 4.9 one does22:46
gnarfaceactually i would assume it does but i'm just not 100% sure22:46
gnarfacei'd expect it to be mentioned in the changelog though22:46
systemdleteI know I need to update to beowulf.  That is becoming increasingly obvious.22:46
gnarfacewell the ascii-backports kernel may also have it22:46
coniferhi, I used to run firetools in the past but when I switched to beowulf I could run them for a day or two, and now they complain about a possible permissions problem22:47
gnarfacei still have some ascii systems around here but all headless22:47
gnarfaceconifer: any specifics about the permissions of what?22:47
coniferand I don't really have an idea how I broke it22:47
systemdleteyou mentioned additional heat when running with the microcode package installed.  This is a desktop, and it has a fan, heatsync, etc etc.22:48
gnarfacesystemdlete: it was a couple degrees warmer when idling.  it was not a big deal.22:48
systemdleteIn fact, I think I have about 6 fans total (including the PS)22:48
coniferI get a gui message "cannot run firejail sandbox, you might not have the correct permissions to run this program"22:49
systemdletemy PS is deluxe, apparently, it has 2 fans (one inside and one outside)22:49
coniferfirejail itself runs fine22:49
coniferi.e. I run apps in firejail without problems22:49
gnarfaceconifer: it was working fine before, or did you migrate from Debian?  i don't know anything about firejail really, but permissions issues are common when migrating from Debian22:50
systemdletenext time the flashing and freezing starts again, I'll reboot.  I'll install the microcode package now though so it will be ready at next boot.22:50
gnarfacesystemdlete: the primary complaint people are having is that it hampers performance for some workloads22:51
coniferit's a fresh beowulf installation. I don't have good experience in upgrading neither debian nor devuan22:51
fsmithredconifer, is firejail-profiles installed?22:52
fsmithredin buster/beowulf it's a separate package22:52
coniferyes, it is, the backports version22:52
gnarfacesystemdlete: (10-20% for Intel chips... enough that even Kernel developers are risking the security problems by avoiding the patch, but for AMD FX chips, probably not as much as that)22:53
systemdleteinteresting, thank you for that info gnarface.  I note that firmware-amd-graphics is installed already.  So I think you were right that the GPU issues are probably not addressed by the cpu microcode.22:53
coniferI tinkered with net.core.bpf_jit_harden kernel.kptr_restrict and kernel.yama.ptrace_scope but it doesn't seem to matter in this case22:53
systemdleteall good info gnarface.  thanks again and again.22:54
gnarfacesystemdlete: no problem22:54
gnarfaceconifer: you have the backports versions of both packages, right?22:54
gnarfaceconifer: (usually backports stuff is intended to be used as a set)22:55
coniferfiretools don't have a backports version22:55
fsmithredfirejail and firejail-profiles22:55
coniferbut firejail & profiles are both from the backports22:56
fsmithredok, good22:56
coniferI can run firetools as root, but it doesn't run the apps as intended (at all)22:56
fsmithredI made an alternative to firetools because I didn't like it. Couldn't edit which apps showed up in it. Is it still like that?22:57
coniferwhen I run it as root I get "QStandardPaths: wrong ownership on runtime directory /run/user/1000, 1000 instead of 0"22:59
conifernot sure if this helps, but this is what I see in terminal22:59
fsmithredoh, are you using 'su' or 'su -' to get root?22:59
conifersu without dash22:59
fsmithredthe behavior of su changed23:00
fsmithredif you want the old way23:00
fsmithredecho 'ALWAYS_SET_PATH yes' >> /etc/default/su23:00
gnarfaceconifer: uid 1000 belongs to the first regular user you created.  uid 0 is root.  that error seems fairly straightforward.  it's saying it wants the directory owned by root23:00
coniferand firetools doesn't start any apps, term shows "The new log directory is /proc/27928/root/var/log"23:00
fsmithredyou can't start apps with firejail as user?23:00
gnarfaceconifer: oh.  yea, probably you used su instead of "su -" to make that directory, so you didn't inherit root's environment and it made it as your regular user23:01
coniferwhen I run it with "su -" I get "qt.qpa.screen: QXcbConnection: Could not connect to display"23:01
gnarfaceconifer: root's environment doesn't have DISPLAY defined, so you'll have to set it manually there23:02
conifer"Could not connect to any X display" and it doesn't run at all23:02
gnarfaceconifer: DISPLAY=":0.0" [command]23:02
coniferOK, but I shouldn't need to be root to run firetools at all, right?23:02
fsmithredright23:02
fsmithredyou should not want to run any desktop apps as root23:03
coniferexactly23:03
gnarfaceconifer: i honestly don't know enough about firetools.  if you think not, i think you're right because that sounds sane.  but you might need to still set up the directories as root initially23:03
gnarfaceconifer: if you correct the permissions on that directory to be whatever it actually wants, i assume it will start working again.23:04
gnarfaceconifer: (in the future when you use "su" always use "su -" now instead)23:04
fsmithredthat way you get root's path23:05
coniferwhich dir do you mean? /run/user/1000 ?23:05
gnarfaceconifer: yea, because that's the one from the error you reported, but there may be others23:05
gnarfaceconifer: (remember, i dont' know anything about firejail)23:05
gnarfaceconifer: (or firetools, other than what i've gleaned by context here)23:06
gnarfaceconifer: i'm somewhat suspicious that /var/user/1000 is even the right name, honestly.  if it wasn't supposed to be using 1000 for the UID, i can't imagine it was supposed to be using 1000 for the name either.  what does that directory actually show for current ownership and permissions?23:07
coniferalso, my display seems not to be 0.0 ;_;23:07
gnarfaceit should be unless you have 2123:07
gnarface2*23:07
gnarfaceare you running multiple Xorg instances?23:07
gnarfaceyou might also get permissions errors from Xorg if you don't run "xhost +" or "xhost +localhost" or something like that first23:08
gnarfacesu -c should bypass that though23:09
fsmithredthere is /var/run/user/100023:09
* gnarface does not have either 23:09
fsmithredmaybe it's an elogind thing?23:09
gnarfacemaybe23:10
conifer_run/user is owned by root, 755, /run/user/1000 is owned by user, 70023:10
gnarfacebut this should have worked: DISPLAY=":0.0" su -c [command]23:10
gnarfaceconifer: 700 would mean only that user can access it23:11
gnarfaceconifer: is it the right user though?23:11
coniferit's a laptop, there is only this one user and root23:12
coniferI don't think I have multiple xorgs...23:12
fsmithredI'm not sure I understand what you're trying to do23:12
coniferrun firetools as a regular user, it used to work in the past23:13
fsmithredok. Do you use that just to start apps in firejail, or do you use some of the other functions it has?23:13
conifer...or am I seriously not remembering things?...23:14
fsmithredbeen a few years since I tried firetools, but I always ran it as user.23:14
fsmithredit put a dock on the desktop with a few app buttons23:14
coniferI used it see if the apps and jails closed correctly23:15
fsmithredok, my script doesn't do that part. It just starts the apps.23:15
coniferand shut the jails with misbehaving apps23:15
fsmithredhow would one do that manually?23:16
fsmithredI'm running 2 apps in firejail and I have 4 firejail processes running.23:16
coniferwithout firetools I kill the app or its jail (whichever works) either from console or gui task manager23:17
fsmithredI just closed ff and the two firejail processes for it went away on their own23:18
fsmithredbut nothing was misbehaving23:18
coniferyeah, they usually do. but sometimes the app seem to close, but its firejail remains open with some memory still occupied23:19
fsmithredYou could try this and see if it does enough of what you want: https://sourceforge.net/projects/refracta/files/Extras/firemenu-1.2.deb/download23:19
fsmithredI actually have a newer script that's better, but it's not packaged.23:20
fsmithredand it needs yad to run (like zenity but more)23:20
fsmithredhttps://sourceforge.net/projects/refracta/files/Extras/firemenu-1.2.deb23:21
fsmithredwget the second one23:21
fsmithredHere's the updated script: http://distro.ibiblio.org/refracta/files/extra_packages/firemenu23:23
fsmithredI need to go make dinner. bbl.23:23
conifersure, np, enjoy your meal :)23:24
coniferthanks for your help so far23:24
systemdleteDoes anyone know approximately when Chimaera is due to arrive?  I know it is dependent on how many hands are working on it, etc, etc, and various unknowns.  But is it a month away, 6 months, a year... just broadly speaking is what I am asking.23:30
systemdlete(And I know you hate this question:  Are we there yet?  Are we there yet?)23:30
gnarfacesystemdlete: not until after Debian's goes stable.  so... you have a while probably.23:30
systemdleteIt sounds like devuan is now "all caught up" with the official release levels of Debian then?23:31
systemdleteI just don't want to do an upgrade only to find out it is out of date a few weeks later.23:32
gnarfacesystemdlete: well for the moment it's caught up, but that doesn't mean Chimera will be released hot on the heels of Bullseye.  it'll probably at least take a few months like the others did.23:32
systemdletesure, sure.  I figure as much.23:32
gnarfacesystemdlete: no, it's not even something you need to worry about until after you hear Debian's next release gets marked stable23:33
systemdleteThere's all that careful surgery that needs to be done, to cut away the rot and infestation...23:33
Jjp137if you want something remotely concrete, the hard freeze for Debian Bullseye would be on 2021-03-12 (source: https://release.debian.org/bullseye/freeze_policy.html)23:33
gnarfacesystemdlete: there's no conceivable situation where Chimera will go stable before Debian's corresponding release does.23:33
Jjp137but yeah even then you'll be waiting for quite a few more months23:33
systemdleteOK. thanks all.23:33
gnarfaceyou're actually pretty much in the sweet spot for the right time to upgrade to beowulf right now.  make a backup first to be sure obviously, but really right now, beowulf is fresh but not raw.23:34
gnarfaceascii is starting to get stale23:35
systemdleteyeah23:35
systemdleteboy is it ever...23:35
gnarfacegtx 1060 still needs beowulf-backports kernel to work23:37
gnarfaceso depending on your hardware you might still need backports23:37
systemdletegnarface:  No need for backup, as I'll be using separate partition(s).  I have never directly clobbered an existing system for the last 15 years.  For one thing, it is nice to keep the older version around for reference (cross-mounting).  If anything goes terribly wrong, the old partition is still there so I can do things like get online and run down the source of the problem(s), etc.23:37
gnarfacesystemdlete: hehe, sounds like famous last words to me, but i trust you know what you're doing.23:38
systemdleteGoodness gnarface.  I don't think I have any hardware here from the current era.23:38
gnarfaceyou'll probably be fine then23:39
systemdleteI do backups of the host and most VMs (I don't bother backing up most test VMs)23:39
gnarfacebeowulf has been very stable looking here23:39
systemdleteSo even if I need it, backups are available to me.23:39
systemdleteI am running beowulf on a couple of VMs already, and yes, it is fine.  Any issues are probably stemming from the host still being ascii (stale, as you say)23:39
systemdleteI mean, I could just update the host kernel to the backports kernel... it is the same (4.19.0) as beowulf kernel I think.23:40
gnarfacemight help, but i honestly couldn't be sure23:41
systemdleteBut it seems like it would be a good idea to get off ascii already.  I don't want to be 2 releases downrev when Chimaera finally arrives.23:41
gnarfaceyea i think it's the same version or nearly though23:41
systemdletepackage level is different.  But the same kernel release number.23:41

Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!