Guest81544 | what is the difference between vixie cron and cronie? | 00:29 |
---|---|---|
gnarface | other than vixie cron is in the repo, i don't know | 00:32 |
Guest81544 | Cronie (sys-process/cronie) is a fork of now outdated vixie-cron done by Fedora, which is still maintained. Because of it being a fork it has the same feature set the original vixie-cron provides. Additionally cronie comes with an anacron implementation which is enabled, by default, through the anacron USE flag. Be aware of the configuration differences as noted in bug #551352 when migrating from another cron system. Expected jobs | 00:40 |
Guest81544 | may not run at all. | 00:40 |
ullet | github added more microsoft javashit garbage | 00:43 |
gnarface | Guest81544: anacron is also in the repos. "now outdated" sounds like Redhat FUD | 00:44 |
gnarface | some stuff doesn't need to be changed | 00:44 |
Guest81544 | so | 00:46 |
Guest81544 | looks like cronie-cron is really vixie-cron post-redhat-fud | 00:46 |
Guest81544 | so my migrations should go smooth | 00:46 |
gnarface | well, probably, but i'd still read about whatever bug #551352 is to be sure | 00:47 |
mason | Red Hat wrote a bunch of the software you're running. Don't be sore losers. | 00:56 |
mason | In fact, that wording... That's Gentoo FUD from https://wiki.gentoo.org/wiki/Cron | 00:58 |
mason | Anyway, the Debian FUD is probably the most relevant: https://wiki.debian.org/cron | 01:00 |
MinceR | and so the hed rat PR campaign starts here | 01:02 |
mason | No need. Just run the software. :) | 01:03 |
MinceR | and pay the fees for the support contracts | 01:03 |
MinceR | oh, and you must also love it | 01:03 |
mason | No, no, just run the software you're running now. Much of it was written by my co-workers. | 01:04 |
mason | All free. | 01:04 |
mason | bbiab | 01:04 |
DocScrutinizer05 | redhat CEO(?) drafted the *huge* plot of all this, some 10+ years ago | 01:04 |
DocScrutinizer05 | basically the software lock-in | 01:05 |
MinceR | it will have to be replaced eventually, since either hed rat won't maintain it or will "maintain" it in a hostile manner | 01:05 |
ullet | i want evidence | 01:05 |
MinceR | especially if the Linux-Destroying Foundation manages to kill off the kernel and the developers don't fork it | 01:05 |
DocScrutinizer05 | >><mason> All free.<< | 01:07 |
DocScrutinizer05 | !freee | 01:07 |
infobot | well, freee is http://www.youtube.com/watch?feature=player_detailpage&v=ZTdUmlGxVo0&t=3296 | 01:07 |
DocScrutinizer05 | either you're the customer or you are the product. Nothing ever is "for free" | 01:09 |
DocScrutinizer05 | definitely nothing RH does and offers is | 01:10 |
MinceR | there is such a thing as "free software", but it's antithetical to redhat/ibm | 01:10 |
gnarface | alright, i apologize for the off-the-cuff "RH FUD" comment. i didn't mean to send this channel spiraling off-topic | 01:11 |
MinceR | no need, we've seen lots of RH FUD against all competitors to cancerd already | 01:12 |
gnarface | yea but this conversation solidly belongs in #debianfork now | 01:12 |
DocScrutinizer05 | id did since >><mason> Red Hat wrote a bunch of the software...<< already | 01:15 |
gnarface | the relevant part is just that vixie cron is still the default in ceres currently, so it's not going away soon | 01:15 |
DocScrutinizer05 | very good | 01:16 |
Guest81544 | redhat re-implement simple UNIX tooling into some huge abomination that only they can understand, forcing support contracts | 01:22 |
Guest81544 | so it's not really support contracts. It's the IT equivalent of a mafia stickup. We run the joints around here, pay your dues | 01:24 |
mason | Wow, I missed all the hate. There are some amazing brains at work in here. I'm sorry I missed the show. | 03:13 |
MinceR | if you missed it, how can you know what it was like? | 03:14 |
mason | MinceR: You're going to change the world, young man. I can see it in your eyes. | 03:14 |
MinceR | unlikely | 03:16 |
mason | Nah, you can get far hating everything. Proven tactic. | 03:17 |
onefang | I'm happy to be missing it. | 03:17 |
MinceR | ah yes, ibm is everythign | 03:18 |
MinceR | s/gn/ng/ | 03:18 |
* slvr reloads the tv box with devuan | 03:39 | |
mason | slvr: We're doing that here. Works well. No desktop environment, no PulseAudio, various browsers and vlc for various media. | 03:40 |
slvr | I just want nfs to work again | 03:40 |
mason | What NFS issue do you have? We're sharing media with NFS here. | 03:40 |
mason | Beowulf on both ends, but we've also had CentOS, Debian, and Ubuntu in various roles, serving and consuming. | 03:41 |
slvr | no problems in devuan. mint changed something with how it mounts and so I'm ditching mint. | 03:41 |
slvr | if you put it in fstab..... no booterama | 03:41 |
ullet | q | 03:42 |
mason | Hm, not familiar with Mint. Is it a systemd hang-up or something? (Did you specify _netdev in fstab?) | 03:42 |
slvr | ip:path mountpoint nfs defaults 0 0 | 03:42 |
slvr | same way I've done it for 20 years | 03:42 |
mason | slvr: You can try adding _netdev to the mount options field to see if that helps. | 03:44 |
mason | FWIW, what you've got is what I use, except the TV computer mounts read-only. | 03:44 |
tuxd3v | slvr, IP:Exported/FS /Mount/Point nfs nfsvers=3,proto=tcp,hard,intr,rw,local_lock=all,noexec,nosuid,rsize=16384,wsize=16384 0 0 | 03:44 |
slvr | the share is marked ro on the server | 03:44 |
mason | slvr: Marking it ro on the client will give you a "read only" error instead of a "failed write" error. | 03:45 |
slvr | it's just there to make an accidental keypress harmless | 03:45 |
slvr | my workstation has it mounted rw | 03:46 |
mason | Heh, same pattern here again. Next you'll tell me you copy your DVDs with vobcopy. But anyway, try _netdev as that might help. | 03:46 |
slvr | why bother? Chasing a moving target is not what I want to do in my living room. | 03:47 |
slvr | I have a server rack of netbooty varmints for that. | 03:47 |
mason | Random note, I'd recommend starting with Beowulf. I had ASCII on my TV box and the version of VLC in it was problematic. | 03:48 |
mason | Beowulf was stable enough at the time (moreso now) that I figured I'd just run that rather than backporting one package. | 03:48 |
slvr | I ran beowulf on my servers for quite a while and ascii on my desktop. Now that my desktop has moved to beowulf I'm distro hopping the family computers over too. | 03:49 |
mason | cool | 03:49 |
slvr | the wheezy->jessie->ascii->beowulf process is unbelievably painless in lxc containers. | 03:50 |
slvr | sed your sources, update/upgrade, repeat. reboot at the end. | 03:51 |
meep_____ | slvr: oh really? | 03:55 |
meep_____ | I use LXC containers with Devuan | 03:55 |
meep_____ | slvr: are you running openrc in your lxcs? | 03:55 |
slvr | uh... I should know that.... haha | 03:56 |
slvr | probably not | 03:56 |
slvr | nope | 03:57 |
meep_____ | slvr: that would be interesting because openrc is container/bare-metal aware | 03:58 |
meep_____ | At least reporting wise | 03:59 |
meep_____ | And the way debian packaged it it is a bit screwed up | 03:59 |
ullet | no vim3 yet on https://pkgmaster.devuan.org/devuan/dists/beowulf/main/installer-arm64/current/images/device-tree/amlogic/ | 05:42 |
meep_____ | What is vim3? | 05:49 |
meep_____ | The text editor? | 05:49 |
tuxd3v | irs a ARM board: https://www.khadas.com/vim3 | 05:51 |
ullet | it's fast and can run off any dc voltage i have lying around | 05:53 |
meep_____ | Wtf is a deep learning module | 05:54 |
ullet | i'd love to run devuan on it, but i'm somwhat mentally retarded | 05:55 |
meep_____ | That can be a problem | 05:57 |
meep_____ | Are you looking to augment yourself to compensate for your mental deficiencies? | 05:58 |
ullet | it just means i can't do everything i'd like to, but who can? | 05:58 |
gnarface | eventually you'll learn to recognize the feeling of being stupid as a sign you're actually getting smarter | 06:15 |
gnarface | people who actually are stupid don't notice | 06:16 |
ullet | so it's the khadas guys who have to get their .dtb into the mainline linux-image? | 06:29 |
tuxd3v | ullet does you already own the board? | 06:29 |
ullet | yeah! | 06:29 |
ullet | it's fun | 06:29 |
ullet | was a tough call between it and the odroid n2 | 06:30 |
ullet | but i went with the vim3 for the wide voltage range for input, and i enjoy building my own cooling solutions | 06:31 |
ullet | khadas should have done the cpu facing down though | 06:31 |
Xenguy | Fuck vim3 for even choosing that name, that's ridiculous, it has nothing to do with Bram right? | 08:15 |
meep_____ | I have the new emacs4 | 08:19 |
meep_____ | And my brother has gnu6 | 08:19 |
ullet | ++ | 08:26 |
ullet | so is the .dtb provided by the manufacturer or wut | 08:26 |
scramblez | Hi All, I installed devuan using a qcow2 image and then added xfce4, but I'm unable to start it: "/usr/bin/startxfce4: 118: exec: xinit: not found" | 11:52 |
scramblez | Tried various things, all failing in the same way. | 11:52 |
ullet | is xorg installed scramblez ? | 11:52 |
scramblez | Yes, I installed it | 11:53 |
ullet | you should have /usr/bin/xinit executeable then | 11:53 |
ullet | is it there? | 11:53 |
scramblez | Hmm ... no :( | 11:54 |
scramblez | I followed the wiki page to install xorg | 11:54 |
scramblez | Except I didn't install the video-dummy and input-void parts, instead I installed radeon and libinput | 11:55 |
ullet | you should also have package xinit installed, which provides /usr/bin/xinit | 11:55 |
scramblez | Wouldn't that be part of x11-server-utils? | 11:56 |
ullet | dpkg -l |grep xinit | 11:56 |
ullet | ii xinit 1.4.0-1 arm64 X server initialisation tool | 11:56 |
ullet | will show you if you have it installed. i don't know offhand what the dependency structure looks like | 11:57 |
scramblez | I just installed it | 11:57 |
scramblez | Hmm ... now different error: parse_vt_settings: Cannot open /dev/tty0 (Permission denied) | 11:58 |
ullet | are you starting xfce from the framebuffer console? | 11:59 |
scramblez | Yes | 12:00 |
scramblez | It should still start with "startxfce4" | 12:00 |
ullet | does plain 'startx' work? | 12:01 |
scramblez | Hmm ... startx also failed: xf86OpenConsole: Cannot open virtual console 1 (Permission denied) | 12:01 |
ullet | you're running as user? | 12:01 |
scramblez | Yes, su'ed into user from root | 12:02 |
ullet | in some versions i can only start x as root | 12:02 |
scramblez | I see | 12:02 |
scramblez | It shouldn't be so though. I thought the xserver-xorg-legacy wrapper was meant to deal with this and allow plain users to start it. Should the plain user be in a wheel group or something? | 12:03 |
scramblez | Ah! Hold on ... | 12:04 |
scramblez | We have lift off! :-) | 12:05 |
ullet | what did you change to fix, scramblez | 12:05 |
iv4nshm4k0v | scramblez: Plain user should have access to the relevant devices; say, /dev/fb0, /dev/dri/, /dev/input/. So far as I can tell this is usually accomplished by chacl'ing the permissions on them. | 12:05 |
scramblez | Thanks guys, I had to make a minor change in xserver-xorg-legacy for "needs_root" - it was auto, I set it to yes. | 12:06 |
scramblez | It started now. | 12:06 |
iv4nshm4k0v | scramblez: Then it runs via a set-uid root wrapper (so that it can access the device nodes I've just mentioned.) | 12:07 |
iv4nshm4k0v | IME, plain getty + login + "common" PAM modules give no additional permissions, so X has to be run as root. | 12:07 |
iv4nshm4k0v | OR, the user can be added to the relevant groups(5) instead. | 12:07 |
scramblez | Yes, that's what confused me, I thought the xserver-xorg-legacy wrapper would deal with that. But it seems the wiki did not mention I had to install xinit too. I thought it would have been dragged in as a necessary dependency. | 12:08 |
scramblez | All good now. Thanks again. | 12:08 |
scramblez | This is more of an xfce question: is there an applet for network connections? | 12:09 |
iv4nshm4k0v | If bin/startxfce4 is part of some bigger package, I expect its dependency on xinit be Recommends: at most; and my preference is to disable automatic installation of so "recommended" packages. | 12:10 |
ullet | /8 | 12:10 |
scramblez | But then how would startxfce4 run if it needs xinit? | 12:10 |
ullet | i'm going to have to patch irssi to catch those | 12:11 |
iv4nshm4k0v | scramblez: $ dpkg -S startxfce4 ? | 12:11 |
scramblez | It lists the man page and binary. | 12:12 |
scramblez | part of xfce4-session | 12:13 |
iv4nshm4k0v | (Or, rather, bin/startxfce4.) My point is that it's /permitted/ for packages to break somewhat if Recommends: are not installed. | 12:13 |
scramblez | Right, I see. | 12:13 |
scramblez | What's your opinion of a Display Manager for xfce4? SDDM, LightDM, something else? | 12:14 |
iv4nshm4k0v | (In this case, xfce4-session works just ok without xinit, /provided/ it's started via some desktop manager; and not from user's command line.) | 12:15 |
scramblez | iv4nshm4k0v: yes, I see why xinit is not a prerequisite. | 12:15 |
iv4nshm4k0v | I'm not an xfce4 user, but even after some 20 years of using XDM I still see no reason to switch away from it. | 12:17 |
iv4nshm4k0v | (Then again, I've toyed with KDE and Gnome back c. 2000, and found no use for them, either.) | 12:17 |
ullet | openbox for me right now. | 12:18 |
scramblez | OK, I seem to not be able to log out of the GUI ... | 12:18 |
ullet | i also like using fluxbox textfiles to startup and position my applications | 12:18 |
scramblez | Ahh yes, FB an old favorite | 12:19 |
Uki | hi ! some guidance please, i have the base/bare system installedto have the gui (openbox) what should be installed next , what is the optimal sequence? | 12:19 |
ullet | having performance problems with xfce on arm, past couple of years | 12:19 |
ullet | so it's not for me | 12:19 |
scramblez | ullet: how did you log out from it? | 12:19 |
scramblez | The logout menu options do not work | 12:19 |
ullet | of xfce? | 12:19 |
scramblez | Yes | 12:19 |
scramblez | :-/ | 12:20 |
ullet | i don't know | 12:20 |
scramblez | Hmm ... shutdown works, but log out won't. Odd. :-/ | 12:20 |
scramblez | Oh! Hold on, it took now. | 12:21 |
scramblez | V odd | 12:21 |
scramblez | OK, it seems to work now, well enough for a qemu image anyway. | 12:22 |
scramblez | I noticed the qemu image would not boot with EFI boot, only legacy. | 12:23 |
scramblez | Would I need to installed it from an ISO within qemu to get it to work with EFI? | 12:23 |
buZz | EFI works within qemu? :O | 12:25 |
buZz | oh it does, amazing | 12:25 |
scramblez | You need to point it to " -bios /usr/share/edk2-ovmf/OVMF_CODE.fd" | 12:26 |
scramblez | then it will boot into an EFI boot menu instead of the legacy bios | 12:26 |
scramblez | However, the image you're trying to boot needs to be EFI-able | 12:26 |
scramblez | The qcow2 image I downloaded seems to be legacy bios only. | 12:27 |
scramblez | I'll try again to see what I get | 12:27 |
Uki | Hello ! Some guidance please, i have the base/bare system installed, To have the gui working (openbox) what should be installed next , what is the optimal sequence? xorg 1st? | 12:28 |
buZz | Uki: its likely that that openbox installation already pulled in xorg | 12:29 |
buZz | especially if you say its running already ;) | 12:29 |
Uki | i have nothing else but the base syste | 12:30 |
* ullet gives the Openbox salute | 12:30 | |
buZz | Uki: dpkg -l | grep xorg <- probably that shows its installed already ;) | 12:31 |
buZz | (thats not the only way to see, nor a efficient one) | 12:31 |
Uki | it isnt iinstalled | 12:32 |
buZz | then how is your gui being displayed? :P | 12:32 |
buZz | ooo 'To have the gui working' | 12:33 |
buZz | Uki: you can just apt install openbox directly | 12:33 |
Uki | will it pull all else needed? or should xorg come 1st? | 12:34 |
buZz | it will pull everything in that it depends on | 12:34 |
Uki | in the same install ill need nvdia driver can be done afterwards? | 12:35 |
buZz | yeah you could later do that, or directly | 12:35 |
buZz | apt install nvidia-driver openbox | 12:36 |
buZz | ullet: fluxbox here :) | 12:36 |
Uki | E: unable to locate | 12:39 |
buZz | Uki: apt update first? :D | 12:40 |
Uki | failure resolving ... | 12:43 |
buZz | fix internet first? :D | 12:43 |
scramblez | Uki: did you go through this: "https://friendsofdevuan.org/doku.php/community:minimal_xorg" and then this: "https://friendsofdevuan.org/doku.php/community:openbox-guide" | 12:44 |
* scramblez goes to fetch more coffee ... | 12:45 | |
Uki | connection is working its pinging hosts | 12:47 |
Uki | but not reaching mirror | 12:48 |
scramblez | Uki: is dhcp working? Do you have a gateway configured? netstat -rn | 12:50 |
scramblez | need to run | 12:51 |
scramblez | bye and thanks for your help guys | 12:51 |
Uki | checking that | 12:52 |
xinomilo | fix resolv.conf | 12:52 |
xinomilo | add some dns there | 12:53 |
Uki | worked on install , why isnt working now? | 12:53 |
Uki | networks had gateway as dns but not working, same in resolv,conf, only worked after adding dns to resolv.conf, on install it only proceded with auto option , manual wouldng go, just for the record. back to the gui install | 13:42 |
ullet | 20 | 13:43 |
Uki | xorg installed , now the package s, suggested above "nvidia-driver" depends on other packages, where to start? | 14:03 |
buZz | Uki: 1) apt update | 14:12 |
buZz | :) | 14:12 |
Uki | got that way back, ,,, about nvidia driver, figure it out , what pulls it all up is: xerver-xorg-video-nvidia | 14:16 |
Uki | at the end of instakking it messege come up, about conflict with nouveau, it says reboot , is really just it? | 14:18 |
Uki | installing* | 14:18 |
buZz | yes, that should be all | 14:19 |
Uki | morfe | 14:24 |
Uki | oops :b ok wiith all that instaleed, + openbox and slim , i get no logging screen | 14:26 |
Uki | just text loggin | 14:26 |
Uki | tried X -configure failed to initialze module NVIDIA , number of created screen does not match number of detected devices, i have no clue about it, how to fix? | 14:36 |
buZz | what nvidia card is it? are you sure its still supported by nvidia's driver? | 14:36 |
Uki | GF108 [GeForce GT 630] | 14:38 |
buZz | thats pretty old | 14:40 |
buZz | ~8 years | 14:40 |
Uki | about that | 14:40 |
buZz | i pretty much doubt the nvidia-driver in devuan still supports that | 14:41 |
xinomilo | nouveau might be better | 14:41 |
Uki | wont run what i want | 14:42 |
Uki | log says no screen found , what is missing? | 14:48 |
onefang | GF108 [GeForce GT 630] (rev a1) is exactly what I run. Devuan ASCII desktop. | 14:48 |
buZz | onefang: with nvidia-driver ? | 14:49 |
buZz | maybe nvidia-legacy-390xx-driver still supports that old card | 14:49 |
buZz | or nvidia-legacy-340xx-driver | 14:49 |
buZz | or nvidia-legacy-304xx-driver | 14:49 |
buZz | rtfm ;) | 14:49 |
onefang | Yep, with nvidia driver. | 14:50 |
buZz | which one? | 14:50 |
buZz | clearly not 418.74 :D | 14:50 |
Uki | i got 390.116 | 14:50 |
buZz | ah yeah, ascii | 14:51 |
onefang | I think the top level package is xserver-xorg-video-nvidia version 390.116-1 then it just pulls in everything else. | 14:52 |
onefang | Been awhile since I installed it. lol | 14:53 |
Uki | (EE) no screens found ..... something is missing | 14:53 |
Uki | it did pull lots of stuff | 14:54 |
onefang | Install recommends as well, maybe even suggests. | 14:54 |
buZz | Uki: nvidia often removes old GPUs from drivers | 14:55 |
buZz | but well, onefang says it works | 14:55 |
onefang | From my ancient install script - apt install xinit x11-xserver-utils xserver-xorg libgl1-nvidia-glx nvidia-driver | 14:57 |
onefang | I debootstrap my Devuans. | 14:57 |
onefang | I use OpenSim 3d virtual worlds a lot, I develop for them. Also works for Blender. | 14:59 |
onefang | So yeah, nouveau wont cut it. | 14:59 |
buZz | hmhm | 15:00 |
buZz | tbh nouveau did get a ton better over last ~10 years | 15:00 |
ullet | do they do opengl-es backend onefang ? | 15:00 |
Uki | i goy apititude, xerver-xorg-legacy is in purple there IdA | 15:04 |
Uki | got* | 15:04 |
Uki | i follow the link suggested above dummy did not install, but i did add it | 15:05 |
onefang | (II) NVIDIA GLX Module 390.116 Sun Jan 27 06:24:32 PST 2019 | 15:06 |
onefang | (II) NVIDIA(0): NVIDIA GPU GeForce GT 630 (GF108) at PCI:2:0:0 (GPU-0) | 15:06 |
buZz | yeah good idea onefang | 15:06 |
buZz | Uki: pastebin your /var/log/Xorg.0.log to paste.debian.net or somethign :) | 15:06 |
Uki | got on boot error for nvidia something in the 1st line | 15:09 |
Uki | install command i think | 15:09 |
Uki | 2 lines with EE | 15:11 |
Uki | no devices detected | 15:11 |
Uki | no screens found | 15:11 |
xinomilo | check this wiki for help : https://wiki.debian.org/NvidiaGraphicsDrivers | 15:12 |
Uki | i have no clue how to paste bin:( | 15:12 |
onefang | I need food, back later. Good luck. | 15:13 |
xinomilo | open a new pad eg. https://pad.dyne.org and paste | 15:13 |
xinomilo | avoid pastebin.com crap | 15:14 |
Uki | the driver i have support my card :) | 15:14 |
Uki | http://us.download.nvidia.com/XFree86/Linux-x86_64/390.116/README/supportedchips.html | 15:14 |
Uki | GT 630 | 15:14 |
Uki | i got only console text on that machine | 15:16 |
user282069 | can we select regional mirrors for sources.list? | 15:40 |
user282069 | with deb.devuan.org | 15:41 |
onefang | That can only be done manually, by referring to https://pkgmaster.devuan.org/mirror_list.txt and finding one in a nearby county / same country. | 15:42 |
user282069 | yep thank you | 15:42 |
user282069 | found it just then onefang o/ | 15:42 |
openbsdtai123 | Do you think that Sagemath is main repository of Devuan for Arm64/rpi3b? I will likely need to install this big soft on pi. | 16:40 |
user282069 | thank you everyone for contributing to this wonderful project :] cheers o/ | 16:58 |
fsmithred | openbsdtai123, it looks like it's there: https://packages.debian.org/buster/sagemath | 17:15 |
systemdlete | OK, what stupid thing have I done/not done here: http://paste.debian.net/1145535 ? | 18:05 |
systemdlete | For some reason, it always fails on eth1, with or without the post-up directive | 18:05 |
systemdlete | when I run service networking start, it does not add the additional route I specified. I can do it, however, directly from the command line without error. | 18:06 |
systemdlete | I googled this, tried some of the suggestions, but nothing seems to work for me. | 18:06 |
systemdlete | btw, I installed the ifupdown-extras package | 18:06 |
slvr | Can you chain commands in that file? | 18:16 |
systemdlete | ? | 18:16 |
slvr | might need quotes, or to move it out into a script to use || | 18:16 |
systemdlete | oh | 18:16 |
slvr | might need paths | 18:16 |
systemdlete | no, according to the man pages, this is acceptable, and even encouraged | 18:17 |
slvr | sorry, just guessing | 18:17 |
systemdlete | examples on man pages don't use paths | 18:17 |
systemdlete | thanks | 18:17 |
gnarface | systemdlete: i'm not completely sure, but i think you need to use iproute2 if you want multiple gateways. ifconfig assumes there's only one uplink | 19:15 |
gnarface | systemdlete: i note these are both local ip ranges though, so you may simply be in error specifying both... | 19:16 |
systemdlete | Even if I comment out the post-up line, I still get an error. | 19:17 |
gnarface | systemdlete: comment out the last TWO lines and see if there's no error | 19:17 |
gnarface | systemdlete: (actually maybe commenting out lines 14 and 21 instead might work, based on this logic) | 19:18 |
systemdlete | No difference | 19:18 |
gnarface | hmmm. | 19:18 |
gnarface | dunno then, sorry | 19:18 |
gnarface | but i think it is to do with the gateways | 19:18 |
gnarface | what is the actual error? | 19:19 |
systemdlete | tried that also, no difference | 19:19 |
systemdlete | RTNETLINK answers: File exists ifup: failed to bring up eth1 | 19:20 |
systemdlete | (that's on 2 lines, actually) | 19:20 |
systemdlete | At boot, both interfaces do come up, and the routing table is correctly populated. But I have to add the additional route manually. | 19:22 |
systemdlete | And those gateway declarations on lines 14 and 21 are definitely necessary because without them, the router table does not get filled in correctly | 19:23 |
iv4nshm4k0v | (Late to the party.) systemdlete: Could you please clarify what you intend to achieve with this interfaces(5) configuration? | 19:25 |
tuxd3v | systemdlete, policy routing? https://tldp.org/HOWTO/Adv-Routing-HOWTO/lartc.rpdb.simple.html | 19:25 |
systemdlete | for those who came late: The same errors occur even without the post-up command. So we can set that aside for the moment. The question really is, why won't eth1 come up after boot? | 19:26 |
systemdlete | (It comes up upon boot, but it won't restart anytime after boot) | 19:26 |
systemdlete | as far as what is to be achieved (for the moment, that is): Two physical interfaces, each with a different gateway. | 19:27 |
iv4nshm4k0v | systemdlete: The problem, AIUI, is that your system refuses to send IPv4 packets towards "outside" networks via /two/ default gateways. | 19:27 |
iv4nshm4k0v | I'm somewhat rusty on this specific configuration; I seem to recall that indeed policy routing was necessary, but I may be wrong. | 19:28 |
iv4nshm4k0v | Basically, and unless I be mistaken, an IPv4 routing table cannot have more than a single "0.0.0.0/0" entry. | 19:30 |
iv4nshm4k0v | Having routes to different networks (such as 192.168.50.0/24 and others) is of course possible. | 19:30 |
systemdlete | Here is the current configuration: http://paste.debian.net/1145562/ | 19:31 |
systemdlete | Note that the post-up is commented out. | 19:31 |
systemdlete | Here is the router table, immediately upon reboot: http://paste.debian.net/1145564 | 19:32 |
systemdlete | So specifying the additional gateway for the 2nd interface doesn't seem to do damage, but I admit I see the error upon boot that eth1 is not up | 19:33 |
iv4nshm4k0v | systemdlete: Note that you still have two "gateway" entires. I've just checked, # ip route add default via X.X.X.X results in "RTNETLINK answers: File exists". | 19:33 |
tuxd3v | 192.168.57.0 0.0.0.0 255.255.255.0 U 0 0 0 eth1 | 19:33 |
systemdlete | ^^ | 19:33 |
tuxd3v | do you see the problem :) | 19:33 |
tuxd3v | you are sending data with the same default gateway | 19:34 |
tuxd3v | yes by 2 diferent ifaces | 19:34 |
systemdlete | there is only one default gateway I see | 19:34 |
tuxd3v | but with same default gw | 19:34 |
iv4nshm4k0v | When ifup(8) fails to add a default IPv4 route (such as when there already is one), it deems the interface "down" (though that's technically "not wholly up") as the result. | 19:34 |
systemdlete | but... | 19:35 |
systemdlete | it works! | 19:35 |
tuxd3v | but you are saying that eth1 doesn't work | 19:35 |
systemdlete | No I didn't | 19:35 |
systemdlete | I said I get an error from the command, that's all | 19:35 |
systemdlete | In addition, I can not use the post-up command, but again, one thing at a time | 19:36 |
iv4nshm4k0v | systemdlete: That's the point; comment out the "gateway" line on eth1 and it will /still/ work and will /not/ complain. | 19:36 |
systemdlete | nope | 19:36 |
systemdlete | it still complains | 19:36 |
iv4nshm4k0v | systemdlete: Without second "gateway"? | 19:36 |
systemdlete | without second gateway, no route is added for eth1 | 19:37 |
iv4nshm4k0v | I suppose post-up won't be executed if "gateway" fails. | 19:37 |
systemdlete | forget post-up for now | 19:37 |
tuxd3v | hehee | 19:37 |
systemdlete | but yes, that is probably true | 19:38 |
iv4nshm4k0v | systemdlete: That's seriously strange; you have three entires in your FIB now, | 19:39 |
systemdlete | I agree it is strange. | 19:39 |
iv4nshm4k0v | one for eth0 "gateway", one for eth0 "address", and one for eth1 "address". | 19:39 |
systemdlete | Especially since I see other examples of very similar configs on various posts out there | 19:39 |
systemdlete | the first is the default gw | 19:40 |
iv4nshm4k0v | systemdlete: If you remove eth1 "gateway", a route will /still/ be created for eth1 "address". | 19:40 |
tuxd3v | maybe you can configure via interfaces( in redhat is possible.. ), in debian/devuan maybe better to stick with policy routing | 19:41 |
systemdlete | ok, so I removed that gateway statement; the route table still looks the same (upon reboot) | 19:42 |
iv4nshm4k0v | systemdlete: And no error, right? | 19:42 |
iv4nshm4k0v | The problem with having two gateways would be that when you, say, ping 1.2.3.4, what will prevent the system from sending a packet from 192.168.**56**.4 via 192.168.**57**.1, or from 192.168.**57**.4 via 192.168.**56**.1? | 19:43 |
systemdlete | no error, but still get the same routing table. OK, but then the routing table wasn't wrong after all, right? | 19:43 |
iv4nshm4k0v | (The answer is, obviously: policy routing.) | 19:43 |
tuxd3v | because you need 2 routing tables | 19:43 |
tuxd3v | :) | 19:43 |
tuxd3v | policy routing ;) | 19:44 |
iv4nshm4k0v | systemdlete: Depends on what you're trying to achieve? | 19:44 |
systemdlete | system only has one routing table I ever knew of. There can be more than one? | 19:44 |
tuxd3v | policy routing, yes | 19:44 |
tuxd3v | that's the objective :) | 19:44 |
systemdlete | I want anything destined for the 56 network to go out on 56, anything for 57 to go out on 57, and anything else to go out on default (which is also 56) | 19:44 |
gnarface | that's what iproute2 is for | 19:44 |
tuxd3v | see this: https://tldp.org/HOWTO/Adv-Routing-HOWTO/lartc.rpdb.simple.html | 19:44 |
iv4nshm4k0v | systemdlete: Then you should only use one gateway, which is 56.1 in this case. | 19:45 |
systemdlete | which is what I have now | 19:45 |
systemdlete | one default gw, yes | 19:46 |
systemdlete | OK, thanks to you all. I guess that 2nd gateway statement was throwing it all out of whack. Once that was removed, that much works. I will try to get this additional route added, per the man pages | 19:51 |
systemdlete | (the one for subnet 50) | 19:51 |
tuxd3v | create a second routing table and assign iface eth1 to that routing table | 19:52 |
tuxd3v | in that way you could manage 2 diferent gateways, in a universal way | 19:52 |
systemdlete | I might actually have it all working already, folks... | 19:53 |
iv4nshm4k0v | tuxd3v: Assuming that 192.168.57.1 /is/ a gateway. | 19:53 |
systemdlete | using that post-up command | 19:53 |
tuxd3v | I believe you can do it with postup commands also. | 19:53 |
tuxd3v | I don't do it for a while, but I used to do so | 19:54 |
iv4nshm4k0v | systemdlete: If 192.168.57.1 is a gateway to a private 192.168.50/24 network, likely so. | 19:54 |
systemdlete | All good now. It was just that gateway statement that was making it all go wonky | 19:54 |
tuxd3v | yes because you have in the same routing table, 2 gateways.. | 19:54 |
tuxd3v | you can only have one per routing table | 19:55 |
iv4nshm4k0v | Curiously, it's possible to have several IPv6 default (::/0) gateways, and have it fail in way more interesting ways. | 19:55 |
iv4nshm4k0v | (Without source-based routing, that is.) | 19:55 |
systemdlete | so, for your viewing pleasure: http://paste.debian.net/1145573 | 19:56 |
systemdlete | And thanks again for your assistance. I did make that one mistake. But it turns out I won't need policy routing or iproute2 or a magic carpet | 19:57 |
systemdlete | (thought the magic carpet might be interesting...) | 19:57 |
systemdlete | s/thought/though/ | 19:57 |
systemdlete | so much for snarkishness | 19:57 |
tuxd3v | systemdlete, policy routing is indeed a nice tool to have in your arsenal, don't downplay it :) | 20:02 |
systemdlete | I'll look at it. | 20:04 |
user282069 | devuan-keyring is already the newest version (2017.10.03). << going ascii to beowulf per the guide; where third step says i should get a new keyring pkg | 20:27 |
user282069 | seems like the pkg is fine and the dev1fanboy guide is off | 20:35 |
psarria | guys, in devuan, can i delete the file /etc/debian_release ? is it in use for any purposes ? | 20:44 |
gnarface | psarria: what does this tell you? dpkg -S /etc/debian_release | 20:53 |
psarria | absolutely nothing | 20:54 |
gnarface | psarria: this is on ascii? | 20:55 |
psarria | yes, ascii | 20:55 |
gnarface | psarria: was this an upgraded system? i think it might just be leftover cruft. | 20:56 |
psarria | that's right, it's an upgraded system | 20:56 |
psarria | so i'm going to delete it | 20:56 |
gnarface | psarria: i think now the things that used to read that read /etc/debian_version and /etc/os-release instead | 20:57 |
gnarface | my google searches suggest that disappeared well before ascii | 20:57 |
psarria | i didn't know | 20:57 |
gnarface | news to me too, but i don't seem to have that file anywhere | 20:58 |
gnarface | it used to be part of the base-files package | 20:58 |
gnarface | this might be the historical info https://news.ycombinator.com/item?id=23051972 | 20:59 |
gnarface | maybe don't *completely* delete it though, just move it elsewhere temporarily until you're sure nothing you have left installed needs it | 21:00 |
psarria | gnarface, ok, thanks a lot, i'm going to rename it and eventually will be deleted after a time without problems | 21:02 |
tuxd3v | ullet, ViM3 seems not yet on Mainline kernel.. | 22:08 |
tuxd3v | Also uboot doesn't mension it.. | 22:08 |
ullet | they say i can try 5.6 with 'patches from' https://github.com/khadas/fenix | 22:09 |
ullet | but their script needs x86 to build it | 22:09 |
tuxd3v | yeah, not mainline :D | 22:10 |
tuxd3v | I was adding A311D to edev1, and I noticed that :) | 22:10 |
ullet | ah | 22:11 |
ullet | their support guy says their dtb is meson-g12b-a311d-khadas-vim3.dtb | 22:11 |
tuxd3v | but stuff about arm --> #devuan-arm | 22:11 |
ullet | can you open the channel to unregged chatters tuxd3v | 22:12 |
tuxd3v | I can't | 22:12 |
tuxd3v | register your-self. its very yeady :) | 22:13 |
tuxd3v | yeady -> easy | 22:13 |
ullet | got a free email that's not blacklisted? | 22:13 |
tuxd3v | in uboot is present: khadas-vim3_defconfig | 22:19 |
tuxd3v | but not in the kernel, only Vim3L | 22:19 |
tuxd3v | arm trusted firmware doesn't have a g12b, only g12a(v2.3), maybe its in mainline.. | 22:22 |
tuxd3v | Anyone with a Rock64 board here? :) | 22:25 |
ullet | hrm :/ not me | 22:26 |
ullet | modified Khadas's build script to pull devuan packages instead of debian tuxd3v . but idk. it's grabbing the i386 packages | 23:17 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!