Hurgotron | In ascii, I'm missing the persistent-net.rules file in /etc/udev/rules.d - it doesn't get autogenerated like I was used to, and manually generating it has no effect. Some helpful hint would be appreciated. | 16:00 |
---|---|---|
gnarface | Hurgotron: you trying to run without udev? | 16:02 |
gnarface | you need udev or eudev or vdev or some other such alternative for this to work | 16:03 |
Hurgotron | got eudev... | 16:04 |
gnarface | is it running? | 16:07 |
Hurgotron | udevd --daemon is running. | 16:10 |
gnarface | strange | 16:10 |
gnarface | strange | 16:12 |
gnarface | you're sure there's an ethernet device attached, right? | 16:12 |
Hurgotron | well I currently have eth0 - eth5. I just added a card with 4 interfaces and was not happy with the sequence, so I wanted to change it | 16:13 |
gnarface | did it used to work? | 16:17 |
gnarface | do you have the /lib/udev/write_net_rules binary? | 16:18 |
gnarface | i can't point to any instances where this stopped working for me that i know of | 16:19 |
gnarface | it suggests something else has gone wrong with your install... | 16:19 |
Hurgotron | the binary is there. And I didn't have it working on ascii yet, this is my first ascii box. The old Devuan has 70-persistent-net.rules | 16:19 |
gnarface | are you configuring the network with /etc/network/interfaces, or are you using a gui tool? | 16:20 |
Hurgotron | with /etc/network/interfaces | 16:21 |
gnarface | you using "auto" or "allow-hotplug" ? | 16:21 |
Hurgotron | auto | 16:21 |
gnarface | hmmm | 16:21 |
gnarface | what does dmesg say about the ethernet devices? | 16:22 |
gnarface | it's possible you're missing drivers or firmware | 16:22 |
Hurgotron | I'm able to use the interfaces. | 16:24 |
Hurgotron | ah. At least I see why the manually created file had no effect... | 16:25 |
Hurgotron | udevd[468]: Error changing net interface name eth3 to eth1: File exists [etc.] | 16:25 |
gnarface | oh, that's evidence | 16:25 |
Hurgotron | but - I'm used to that working... | 16:25 |
gnarface | could something you changed in the startup process be manually configuring the network before udevd can get to it? | 16:25 |
Hurgotron | not really. initrd probably, but I regenerated that one too | 16:26 |
gnarface | had you changed any boot order stuff? | 16:26 |
gnarface | moved a dhcp client up perhaps? | 16:27 |
gnarface | maybe too far up? | 16:27 |
Hurgotron | No, didn't mess with boot order. | 16:27 |
gnarface | what about your kernel command-line? i'm running out of ideas... | 16:28 |
gnarface | usually misbehavior like this is related to missing packages, missing hardware support, cavalier distro-mixing | 16:29 |
gnarface | something of that sort | 16:29 |
Hurgotron | Command line: BOOT_IMAGE=/boot/vmlinuz-4.9.113-vs2.3.9.7 root=/dev/mapper/vg0-debroot ro quiet | 16:29 |
Hurgotron | ok, I have a slightly nonstandard kernel. Trying. | 16:29 |
gnarface | well if you can use the devices in spite udev tripping up naming them, then i wouldn't expect it to be the kernel... but i don't know for sure you couldn't have done something to upset it | 16:30 |
gnarface | i think there IS a dhcp client embedded in the kernel now that could in theory at least be getting triggered too early somehow... | 16:31 |
gnarface | it's not a problem you should be having with a stock kernel | 16:31 |
gnarface | frankly it's not a problem you should be having with that vserver kernel patch either, unless you did something else to it | 16:32 |
Hurgotron | now running stock 4.9.0-7-amd64, I moved the net.rules file away. Nothign autogenerated. Trying to trigger generation manually. | 16:33 |
gnarface | weird | 16:33 |
gnarface | but the network is already up? | 16:33 |
Hurgotron | well, I do have dhcp running to connect to the machine... | 16:33 |
gnarface | it does seem at least possible it could be a culprit, if it's taking control of the ethernet devices before udev can even get to them | 16:34 |
gnarface | but if /etc/network/interfaces has static configs for all the network devices then that shouldn't happen | 16:35 |
Hurgotron | easy to check | 16:35 |
Hurgotron | good one! | 16:40 |
Hurgotron | without dhcp, I'm getting an autogenerated 70-persistent-net.rules | 16:40 |
gnarface | interesting | 16:40 |
Hurgotron | now modifying to my taste. | 16:41 |
gnarface | do you know if it's that kernel dhcp client or if it's a userspace one? | 16:41 |
gnarface | i think something is still wrong if it's preempting udevd | 16:41 |
gnarface | but i don't really know for sure | 16:41 |
gnarface | the important part is you can get it working | 16:42 |
Hurgotron | userspace. isc-dhcp, I had problems with that one before, requesting Ip twice etc. Maybe the same issue? | 16:42 |
gnarface | hmmm | 16:43 |
gnarface | maybe, but not an issue i've seen | 16:43 |
gnarface | that's actually usually caused by a misconfigured dhcp server | 16:43 |
gnarface | or... two of them | 16:43 |
Hurgotron | I consider the important part to fix the underlying issue, so that others don#t waste time on it... | 16:43 |
gnarface | warring dhcp servers on the lan | 16:43 |
gnarface | you might want to check your network and make sure the dhcp server you know about is the only one | 16:44 |
gnarface | i think there's cases where certain versions of windows might by default provide a fully operational wifi gateway with dhcp ... | 16:44 |
gnarface | it *shouldn't* be requesting an ip twice | 16:45 |
gnarface | it's far more likely it's requesting an ip *once* and getting two responses | 16:45 |
Hurgotron | please check the chanlogs for DHCPDECLINE. Definetely no server issue - it works with eberything else and before upgrading to ascii | 16:45 |
Hurgotron | since this is the second time I have issues with dhcp causing weird issues, there might be some common problem. | 16:47 |
gnarface | i don't have any access to channel logs that you do not also have | 16:47 |
Hurgotron | I just meant, if you want to know what i was talkign about :) | 16:48 |
Hurgotron | me and unmy had the problem that ascii was requesting an IP from the DHCP server, then giving DHCPDECLINE, requesting a new one. I somethines had both IPs set after that. | 16:49 |
gnarface | it could be a bug but it still sounds like a misconfiguration to me | 16:50 |
Hurgotron | and my issue is not solved. machine still can't rename the interfaces | 16:50 |
gnarface | like the first one it gets has an invalid network mask or something | 16:50 |
gnarface | but then maybe after the second one changes it, the first one becomes valid again | 16:50 |
Hurgotron | it went away after changing allow-hotplug to auto. | 16:51 |
gnarface | allow-hotplug causes other issues for me too | 16:53 |
gnarface | i can't say i've tested it extensively | 16:54 |
gnarface | personally i think it was a mistake to make it the default even for non-removable devices | 16:55 |
unmy | there have to be some conflict between network service - isc client and sysvinit | 17:00 |
unmy | udhcpc working just fine | 17:01 |
Hurgotron | ok, I'm now renaming my ethernet interfaces to lan0 .. lan5 in order to avoid the eth renaming collisions. *sigh* | 17:17 |
Hurgotron | not what I wanted, but... I can work with it. | 17:17 |
gnarface | Hurgotron, unmy: which DHCP server are you using? | 17:37 |
gnarface | you may be right but i still suspect there is another variable either way | 17:38 |
gnarface | or more people would be seeing this | 17:38 |
gnarface | that said, there are a few DHCP clients available in the repo | 17:39 |
gnarface | they're probably all worth comparing | 17:39 |
gnarface | but if i had to bet money on one behaving the most correctly according to the specifications it would be the ISC one | 17:39 |
gnarface | i've seen *all kinds of* whack misbehavior coming from various DHCP servers however, built into various plastic box routers and a couple windows boxes... | 17:40 |
gnarface | and more than once, a rogue DHCP server that avoided 2 double-checks for rogue DHCP servers | 17:41 |
unmy | gnarface, dnsmasq from openwrt updated to latest every month | 17:43 |
unmy | and no windows in my lan | 17:43 |
gnarface | hmmm | 17:43 |
unmy | just other "smart" devices and windows just from time to time to fix somebody laptop/pc | 17:43 |
Fjalar | Has anybody here ver installed LTSP? Running into trouble with ltsp-build-client on Ascii. | 20:28 |
Pr0metheus | any suggestion for a browser? Firefox has no sound after removing pulseaudio, Web hangs on videos,... | 20:56 |
debdog | Pr0metheus: just as a hint: there's apulse available, a "PulseAudio emulation for ALSA" | 21:01 |
gnarface | Pr0metheus: doesn't firefox-esr behave better? | 21:06 |
Pr0metheus | gnarface: the new viersion of firefox works very fast, but without pulseaudio it has no sound, i think I'll try debdog's suggestion and see if it can do anything | 21:07 |
gnarface | oh, i guess the new firefox-esr is broke again too | 21:08 |
gnarface | so nevermind | 21:09 |
debdog | hmm, hasn't there been an attempt for a browser comparison page somewhere? cannot find it atm | 21:11 |
Pr0metheus | debdog well most pages mention chrome firefox web and midori (the last I have never used) and some other forks iceweasel | 21:12 |
debdog | usually I use https://www.seamonkey-project.org/releases/#contrib (Contributed builds, Linux/x86_64 (some of FF's plugins work with it)) | 21:19 |
debdog | other's I've heard of: http://linux.palemoon.org/ http://www.basilisk-browser.org/ | 21:21 |
Pr0metheus | debdog thanks for the suggestions, I'll give palemoon a try | 21:28 |
gnarface | we need a proliferation of firefox forks | 21:30 |
gnarface | or just... not even firefox forks. just a bunch more browsers based on the gecko renderer | 21:31 |
debdog | I wonder where' the catch is with this browser https://vivaldi.com/ it sounds all too good? | 21:48 |
gnarface | probably like midori it's just missing a complete CSS or JavaScript implementation | 21:50 |
golinux | debdog: The UI is proprietary. | 21:50 |
debdog | ahh | 21:51 |
golinux | Everything else is open. | 21:51 |
golinux | iirc | 21:51 |
Pr0metheus | debdog I also read telemetry cannot be disabled | 21:51 |
golinux | It is very slick but troublesome that it takes so long to load | 21:52 |
zdzichu | this is basically chromium fork by people from opera browser | 21:52 |
golinux | I occasionally use it for testing but not as a daily driver. | 21:52 |
Pr0metheus | is anyone on ceres? is eudev updgrade working for you? | 21:57 |
zdzichu | heh, the BSD people knows it, too: Benno Rice: The Tragedy of systemd -- BSDCan 2018 | 21:58 |
gnarface | are you getting a package conflict about some polkit stuff when trying to switch to it, Pr0metheus? | 22:00 |
gnarface | or maybe it wants to remove udev and your running kernel along with it? | 22:00 |
Pr0metheus | no, it fails to install eudev | 22:01 |
Pr0metheus | there is an inconsistency with the libraries | 22:01 |
Hurgotron | zdzichu: That piece is rather positive about systemd. But misrepresenting the facts, IMO. | 22:01 |
gnarface | Pr0metheus: that's what i mean | 22:02 |
Hurgotron | I wanted to write a repyl, but lack the time currently. | 22:02 |
gnarface | Pr0metheus: i think that's the package conflict i'm referring to | 22:02 |
golinux | I think that the pkgs in ceres are older than the ones in ascii | 22:02 |
Pr0metheus | oh great:P | 22:02 |
Pr0metheus | I downgraded after the upgrade because eudev got broken | 22:02 |
golinux | That's going to change with beowulf. All development will go into unstabe . | 22:03 |
golinux | And work it's way to where it needs to go. | 22:03 |
golinux | -' | 22:03 |
gnarface | the eudev package in ceres seems to be newer | 22:03 |
zdzichu | Hurgotron: haven't seen it yet; you mean https://www.youtube.com/watch?v=6AeWu1fZ7bY ? | 22:03 |
gnarface | maybe some polkit stuff isn't tho0ugh... | 22:03 |
Pr0metheus | better to save manpower | 22:03 |
golinux | You can compare version #s here https://pkginfo.devuan.org/ | 22:03 |
golinux | pick the most recent one wherever that is | 22:04 |
gnarface | i'm still using regular udev on ceres, but there were problems switching to it in ascii that required weird workarounds in some cases (rm /etc/init.d/udev) | 22:04 |
Pr0metheus | the packaged version though seem to require previous versions of libeudev | 22:04 |
Pr0metheus | I believe eudev on ceres works for me (the previous version) | 22:05 |
golinux | Then downgrade | 22:07 |
Pr0metheus | yes | 22:07 |
golinux | Work will start after the August holidays in EU. Nothing is happening atm. | 22:07 |
Hurgotron | zdzichu: yes, that's the one. | 22:08 |
Pr0metheus | yeap I know I am in EU:P | 22:08 |
golinux | It's a bit frustrating for those of us who don't turn off. | 22:09 |
Pr0metheus | lol | 22:09 |
Pr0metheus | debdog: apulse: Firefox 58 tabs crashing when trying to play audio ... | 22:13 |
Pr0metheus | are you guys using synaptic or anything else instead? | 22:14 |
gnarface | wait you got apulse from the repo? i thought it wasn't in there... | 22:18 |
gnarface | it's working for me in firefox 60 | 22:18 |
gnarface | but it's a copy i built myself | 22:18 |
Pr0metheus | yeap it's there but I didn't download it, just read the apulse site | 22:20 |
gnarface | wait, how do you have firefox 58 in ceres right now anyway? | 22:21 |
gnarface | you should have firefox-esr 60.1.0esr-2 and firefox 61.0.1-1 | 22:22 |
Pr0metheus | I don't I assumed since apulse doesn't work with Firefox 58 because some function calls are not possible, it wouldn't for 60... | 22:22 |
Pr0metheus | ok let me try to install it, I hope it doesn't break alsa:P | 22:22 |
gnarface | my apulse build is apparently from january and it's working fine | 22:22 |
gnarface | so if you're having problems it might be with the newer apulse version | 22:23 |
Pr0metheus | hm, will try to reboot, no sound in firefox yet | 22:24 |
gnarface | you've got sound with other stuff, right? you checked speaker-test, right? | 22:24 |
Pr0metheus | yeap other things work | 22:24 |
gnarface | hmm. weird | 22:24 |
Pr0metheus | let me reboot | 22:24 |
Pr0metheus | nope apulse doesn't work for me | 22:31 |
gnarface | does it just not work with firefox? | 22:32 |
gnarface | or does it not work with anything? | 22:32 |
Pr0metheus | tried also the xfce plugin for pulse audio and it didn't work, was waiting to connect | 22:32 |
gnarface | you could try an older apulse build maybe? | 22:38 |
gnarface | i'm curious if the version i have would work for you | 22:39 |
Pr0metheus | did you build a deb? | 22:39 |
gnarface | yes | 22:40 |
Pr0metheus | you can send it over and I can try | 22:40 |
gnarface | i wouldn't, because i'm not sure it is anonymous | 22:43 |
gnarface | but it appears i didn't keep a copy anyway | 22:43 |
gnarface | the best i'd be able to do is send you a tarball of files | 22:43 |
gnarface | minus the package | 22:43 |
Pr0metheus | it's ok | 22:45 |
gnarface | how'd you build yours? | 22:46 |
golinux | Pr0metheus: Search the forum and dng for past discussion og apulse. | 22:46 |
golinux | og > of | 22:46 |
golinux | I know that people have used it successfully in the past. | 22:46 |
Pr0metheus | palemoon works with sound | 22:48 |
Pr0metheus | and seems to work fast | 22:48 |
Pr0metheus | one funny thing is that after moving from stretch to ceres, in an attempt to fix some issues I removed and reinstalled packages. That resulted in removing all Devuan icons I used for the XFCE menus:P | 22:53 |
gnarface | yikes | 22:54 |
Pr0metheus | bye bye firefox, pulseaudio, systemd, virtualbox, gnome-most but it seems that seahorse is irreplaceable | 23:17 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!