--- Log opened Fri May 25 00:00:31 2018 00:02 < AuroraAvenue> How do I install this appimage on Solus-OS ? https://drawpile.net/download/#Linux 00:08 < oerheks> Just mark the Appimage in file properties as executable and run it. 00:11 < revel> How do they work, anyway? Some binfmt magic? Some sort of self-extracting shell script thingie? 00:11 < AuroraAvenue> oerheks, that doesn't do anything - as the program associated with AppImage , does nothing. 00:12 < revel> Oh, you can't just `./some_appimage` ? 00:12 < AuroraAvenue> hangon ... 00:13 < AuroraAvenue> Can't ~I just click 'Run' ? 00:14 < AuroraAvenue> so basically nothing is happening :( 00:16 < AuroraAvenue> Okay So I DON'T KNOW WHAT I AM DOING ... 00:16 < AuroraAvenue> and I need Halp ?!! 00:16 < AuroraAvenue> jgdxjugdzjgd@twelve-ten-thurs-rmount-gardens ~/Downloads $ ./Drawpile-2.0.10.AppImage 00:16 < AuroraAvenue> ./Drawpile-2.0.10.AppImage: error while loading shared libraries: libkeyutils.so.1: cannot open shared object file: No such file or directory 00:16 < AuroraAvenue> jgdxjugdzjgd@twelve-ten-thurs-rmount-gardens ~/Downloads $ 00:17 < ||JD||> AuroraAvenue: calm your tits sister 00:17 < rascul> looks like you need to install libkeyutils 00:17 < AuroraAvenue> I know that.. 00:17 < AuroraAvenue> HOW ? 00:17 < revel> Using your package manager. 00:17 < ayecee> MOAR CAPS 00:17 < rascul> use your package management tools to search for and install it 00:18 < AuroraAvenue> Wats a pacage man. tool ? 00:18 < ayecee> which distribution are you using 00:18 < oerheks> chmod a+x Drawpile-2.0.10.AppImage && ./Drawpile-2.0.10.AppImage 00:18 < AuroraAvenue> Solus Os. 00:18 < ayecee> wtf is solus 00:18 < revel> Some Linux distro. Has its own package manager, too. 00:18 < ayecee> great 00:19 < AuroraAvenue> It's Irish ! 00:19 < revel> So... Can't just say "apt install libkeyutils" or whatever. 00:19 < ayecee> AuroraAvenue: here you go, https://solus-project.com/articles/package-management/basics/en/ 00:19 < ayecee> everything you ever wanted to know about solus package management, and some things you didn't want to know too 00:19 < AuroraAvenue> https://paste.ubuntu.com/p/2bk7DynyjG/ 00:20 < AuroraAvenue> sudo eopkg search libkeyutils 00:20 < AuroraAvenue> oh 00:20 < AuroraAvenue> right 00:21 < revel> Probably don't need sudo just to search for packages. 00:21 < AuroraAvenue> AAAAAArgh ! 00:22 < ayecee> dude, you're stressing me out 00:22 < AuroraAvenue> libkeyutils is not in their repo's :( 00:22 < revel> Just search for "keyutils" instead, mebbe. 00:22 < revel> Or possibly use a more "mainstream" distro if it's really not there. 00:24 < AuroraAvenue> sumthing sumthing process 11044: https://paste.ubuntu.com/p/3bcCrgpMkp/ 00:25 < AuroraAvenue> ALL I want to do is LIVE-AR ! FFS 00:25 < AuroraAvenue> **live-art 00:25 < ayecee> maybe use the distribution that the maintainers of the software use 00:25 < rascul> you would have to rebuild dbus with the correct options 00:26 < ayecee> that's just for the backtrace 00:26 < triceratux> AuroraAvenue: whats the fascination with solus ? what makes you think it suits your use case ? 00:26 < AuroraAvenue> So I have to re-install my operating system, now - Do I ? 00:26 < revel> Or just adding /var/lib/dbus/machine-id could work. 00:26 < AuroraAvenue> AAAAAArfgh ! 00:26 < rascul> oh, yeah, i missed the backtrace part 00:26 < ayecee> AuroraAvenue: well, no, you could struggle through this instead. 00:26 < ayecee> how much are you attached to solus? 00:26 < AuroraAvenue> okay dookie 00:27 < AuroraAvenue> It irish - I'm Irish - call me a Paddy! 00:27 < ayecee> is that "i'll stick with it through thick and thin", or "i'll switch if it gives me problems" ? 00:28 < AuroraAvenue> that's " I'd die for it", pal. 00:28 < ayecee> okay. just remember that. you're choosing the hard way. no more complaining. 00:29 < ayecee> start by seeing the manpage for dbus-uuidgen, like suggested. 00:29 < ayecee> you're in uncharted territory here. 00:29 < AuroraAvenue> but at least I shall have a bulb of Solus - which is more than most in the globe :D 00:29 < rascul> dbus-uuidgen > /var/lib/dbus/machine-id 00:29 < rascul> looks like you can do that to maybe make it work 00:29 < ayecee> or at least get to the next hurdle :P 00:29 < domhnall> not sure how active, but there is a #solus and #solus-dev channel. (not sure if that was already discussed either) 00:30 < AuroraAvenue> no their not active. 00:30 < AuroraAvenue> Anyway - I am creating a headache - so I'll just get my hat ... 00:30 < ayecee> that's the spirit 00:31 < rascul> oh 00:31 < notmike> jim: you know the real problem with leftism is feelings of inferiority and oversocialization 00:31 < rascul> i was hoping he would try that dbus-uuidgen before leaving 00:31 <@jim> AuroraAvenue, do you want a linux dist that makes it easier to install utilities and libs that might be needed for something you're doing or trying to do? 00:31 < CompanionCube> notmike: ..what 00:31 < rascul> the real problem with leftism or rightism is that they're real things 00:31 <@jim> the problem with being left is there's always someone to my right 00:32 < ayecee> jim: only if it's irish, apparently. 00:32 < notmike> CompanionCube: the industrial revolution and it's consequences have been disastrous for the human race 00:32 * domhnall wondering if them using #solus and asking here prompted the op mode...(doubtful, but i dont know) 00:32 < CompanionCube> ...sigh, anprim. 00:33 < triceratux> AuroraAvenue: have you asked yourself whats wrong with original vanilla debian ? 00:33 < triceratux> oops 00:33 < ayecee> domhnall: no, more like forgot to remove it from earlier. 00:33 < domhnall> ayecee: knew there was something reasonable about it... 00:39 < ayecee> it's a takeover! 00:39 < ayecee> aw. 00:39 < Blueking> are there ways to check what hdd a folder are using ? 00:39 < ayecee> Blueking: df dir 00:40 < MrElendig> Blueking: findmnt 00:40 < MrElendig> with apropiate flags 00:40 < MrElendig> also, they are directories not folders 00:40 < ayecee> they can be both 00:41 < Blueking> what this tell you ? /tmp tmpfs tmpfs rw,nosuid,nodev,noexec,noatime,size=67108864k 00:41 < ayecee> that it's not a hdd 00:41 < Blueking> is it in ram ? 00:41 < ayecee> but a tmpfs, in ram, yeah 00:42 < Blueking> tested software transcoding of 4k movie on plex server and this lags :P 00:42 < ayecee> the bottleneck is probably not disk io then 00:43 < Blueking> ok :) 00:43 < Blueking> bed time :) 00:53 < TRS-80> Evening, fine people! :) 00:54 < ayecee> greets to the rest of you too 00:54 < nekoseam> evening 00:54 < nekoseam> i'm far from fine though 00:54 < infinisil> nekoseam: Hello far from fine 00:55 < ayecee> nekoseam is very coarse this evening 00:56 < TRS-80> I have been trying to get my head around Debian package management, specifically, how to install the newer version of some software without breaking Debian (which I seem to have done at this point) 00:56 < MrElendig> depends on the software in question 00:56 < MrElendig> eg installing py 3.6 alongside 3.5 is really easy 00:57 < MrElendig> replacing glibc, not so much........ 00:57 < TRS-80> hmm yeah see your point 00:57 < MrElendig> (goes for any distro and packaging system) 00:57 < MrElendig> nix sort of being an outlier 00:57 < TRS-80> I think I have managed to get myself into some dependency hell by not handling sources correctly 00:58 < TRS-80> so best way just to be very careful about dependencies? check manually? 00:59 < TRS-80> say for example with something I am compiling from source for latest version. In this case GNUCash 00:59 < TRS-80> but also I want to learn in general so as not to cock it up again in the future 01:00 < TRS-80> and I been reading a lot on it already but still can't get my head around it 01:02 < Bashing-om> TRS-80: Have you seen the debian manual : http://www.debian.org/doc/manuals/debian-reference/ch02.en.html#_more_readings_for_the_package_management 01:02 < Bashing-om> http://www.ubuntugeek.com/reference-guideubuntu-package-management-using-dpkg.html ? 01:02 < TRS-80> Bashing-om: yes I have but maybe not those sections, I'll check those out, thanks 01:02 < mices> what channel can i discuss video editing 01:04 < TRS-80> I also learned about CheckInstall, which sounds awesome 01:10 < MrElendig> it isn't 01:10 < MrElendig> since it doesn't handle deps at all 01:10 < MrElendig> (and a lot of other issues with it= 01:11 < TRS-80> Yeah I definitely broke Debian (https://wiki.debian.org/DontBreakDebian) as I seem to be on some FrankenDebian combination of testing/unstable :/ 01:12 < TRS-80> MrElendig: well by now I do understand difference between dpkg/.deb and apt-get for instance (latter will also go out to listed repositories and resolve dependencies, where former is more like a local install) 01:12 < triceratux> wipe it all away & boot the siduction 18.3.0 liveiso 01:12 < ayecee> replace all the old boring problems with new and exciting problems 01:13 < TRS-80> triceratux: I need to reinstall Debian stable but just don't have time for that disruption right now 01:13 < TRS-80> so I live on with little niggling bugs in the meantime 01:15 < domhnall> TRS-80: bugs or conflicting deps? 01:15 < TRS-80> mainly like I said I also don't want to do the same thing again in future. But I think that "break Debian" page has got me sorted out in that regard. I will be much more careful messing around with sources in future. 01:15 < Bashing-om> TRS-80: If the package manager is not happy ---- no one is happy :P .. what shows ' sudo apt -f install ; sudo dpkg -C ' ? 01:15 < TRS-80> domhnall: conflicting deps, now I'm afraid to update, etc. 01:19 < TRS-80> Bashing-om: 'sudo dpkg -C' returns nothing; can't find reference for apt -f flag, what does that do? 01:20 < domhnall> TRS-80: oh okay, kind of left the looming question about how the dependency hell happened and what the conflicting pkgs are. (src vs binary, v1 vs v2) 01:22 < TRS-80> domhnall: OK maybe I'm not in dependency hell so much as I am on unstable and that's not what I wanted and I know I can't really go back to stable without reinstalling 01:23 < domhnall> yeah 01:23 < rascul> well you might be able to, but not easily ;) 01:23 < TRS-80> rascul: how, by picking through everything very carefully manually? 01:24 < TRS-80> I think that's probably a little beyond my GNU/Linux abilities presently, plus sounds like a giant pain in the rear (but what do I know). 01:25 < rascul> replacing a running distro with another is mostly doable but messy and a time consuming manual process 01:25 < TRS-80> figured :) 01:25 < rascul> the other option is just change your sources to stable and wait for the next release ;) 01:25 < domhnall> rascul: educational one too :) 01:25 < rascul> next major release i mean 01:25 < TRS-80> domhnall: yeah, one day I will install Arch or Gentoo or whatever for learning purposes, but today is not that day on my daily driver desktop that I need lol 01:25 < rascul> which, being debian, will probably be sometime in the next 20 years or so 01:27 < domhnall> TRS-80: not uncommon thinking I'd say. Though with Arch pacman doesnt install from src as apt does...but portage does and is cross-platform. 01:28 < rascul> i wouldn't normally recommend replacing a running distro though, better to reinstall 01:29 < TRS-80> domhnall: I mean, I know there are some differences and one day I will dig into those and I have heard you will learn a lot, but for now I just can't get into it with school and life schedule/demands right now 01:29 < TRS-80> rascul: oh don't worry, I wasn't planning on it, knew you were sort of half joking 01:29 < domhnall> TRS-80: I hear ya, just making small talk about the flagships of Linux in general... the package managers. 01:30 < TRS-80> another project I would like to get up relatively soon is new server build which will be capable of multiple VMs... then I can play/learn aplenty without breaking anything 01:32 < TRS-80> domhnall: yeah maybe I should take it as such and enjoy the convo and sort of file it away for later but I'm kind of stressing a bit about school and some other things and so added complexity right now is just sort of pre-emptively blocked out like a painful childhood lol 01:32 < domhnall> heh 01:34 < TRS-80> coming to this channel to shoot the breeze is how I am choosing to spend what little R&R time I have just this moment, after all 01:36 < TRS-80> one thing I would like to throw out there is that GNUCash is apparently like the only (or maybe, first?) F/LOSS accounting software that will connect to your bank and download transactions automatically just like QuickBooks does via OFX / Direct Connect and I am very excited about that as it appears getting finances in order is going to end up being even easier than I had anticipated! Yay! :D 01:37 < popnfloss> shut up 01:37 < TRS-80> popnfloss: ? are you saying this is news to you also? 01:37 < rascul> not the only, there's a few but i can't recall names 01:38 < rascul> i tried briefly to get that setup though with gnucash and i failed, but i didn't try very hard 01:38 < meyou> kinda surprised banks use an open api for that 01:38 < meyou> figured intuit would've made it proprietary 01:38 < TRS-80> rascul: yeah hence my throwing in there "first" at recalling I believe what GNUCash claim was 01:38 < TRS-80> meyou: you and me both man 01:39 < rascul> if/when i can get it setup with gnucash or something else open source then i'll start using it ;) 01:39 < rascul> until then, it's not worth the effort for me to manually import the crap 01:40 < TRS-80> some good info here: https://wiki.gnucash.org/wiki/Setting_up_OFXDirectConnect_in_GnuCash_2 and here: https://wiki.gnucash.org/wiki/OFX_Direct_Connect_Bank_Settings including links to some other online databases where I was able to pull my bank's info up right away 01:40 < TRS-80> http://www.ofxhome.com/ that's it, pretty neat 01:41 < TRS-80> database of OFX info on hundreds of banks 01:41 < misternumberone> hi, I need to install Debian 9.4 using the DebianLive CD. However when I type sudo apt-get install debian-installer-launcher, "FATAL -> failed to fork" right after "processing triggers for menu" and then when sudo debian-installer-launcher, "no suitable d-i initrd image found, aborting" 01:41 < rascul> TRS-80 i'll look into it again sooner or later, maybe next time i'll get it working 01:41 < TRS-80> rascul: maybe time to give it another try some time, from research I have done it looks very doable 01:41 < TRS-80> yeah man 01:43 < TRS-80> there is also a Bayesian filter that you can train to automatically apply the expenses to the correct account after a while, as well as good duplicate recognition apparently, so it looks pretty automated, I'm super excited about getting it going 01:44 < TRS-80> will still end up doing some manual splits of different categories on one receipt like at Walmart for example but those few will be an order of magnitude less work than importing manually much less manually entering receipts 01:45 < TRS-80> misternumberone: why not just let it do it's thing? I don't understand? I recall the process to be pretty painless / automated? 01:46 < misternumberone> TRS-80: debian-installer-launcher is not installed by default on the DebianLive CD, I need to install from a full live image, but it does not contain the installer 01:47 * first-order just used the netinstall media. 01:47 < rascul> misternumberone why must it be from a live cd? 01:47 < TRS-80> ^ that's what I did as well and it was a breeze (netinstall) 01:48 < TRS-80> just plug in an eethenrnet cable though as wifi drivers likely proprietary (assuming laptop) 01:49 < R088Y> @search scientific advertising 01:49 < first-order> Or download the wifi driver and stick it on a flash drive or something, like I did. 01:49 < rascul> TRS-80 you can get the ones with nonfree stuff that contain those drivers 01:49 < misternumberone> rascul: TRS-80: first-order: My fundamental problem is that my installation will require installing the GRUB 2 bootloader to a floppy disk, and the debian installer CD iso does not contain the kernel module floppy and therefore cannot detect my floppy drive. 01:49 < rascul> TRS-80 for example https://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/9.4.0+nonfree/amd64/iso-cd/firmware-9.4.0-amd64-netinst.iso 01:50 < rascul> misternumberone oh my, a floppy disk! 01:50 < misternumberone> I would like at this point to use network boot as well, but I would need an installation image for loading that contains the ability to use my floppy drive 01:50 < rascul> that is certainly an interesting requirement nowadays 01:50 < TRS-80> >treating one of the greatest features of Debian like a bug! If I wanted to take that approach, I would install Ubuntu, Mint, etc! XD 01:51 < rascul> misternumberone you might find more help from ##debian 01:51 * first-order also doesn't have hardware old enough to need boot disks in use atm. 01:52 < TRS-80> misternumberone: you really don't have an USB drive laying around? at least to get started and get internet to then try and find your driver? 01:52 < first-order> Although optimally if I were running hardware old enough to not support USB boot, I would personally just run Puppy on it. 01:52 < first-order> Or even TinyCore or DSL. 01:53 < first-order> Pretty much Slot 1 to hardware as old as a 486 would be a perfect usage case for DSL. 01:54 < first-order> IIRC DSL was even touted as being able to run on a 486 with A6MB RAM. 01:54 < first-order> *16MB 01:55 < misternumberone> TRS-80: I can boot into either a DebianLive iso on CD, a Debian Installer iso on CD, or network boot. I have internet access in all cases by ethernet. first-order: I need the standard packaged software library of debian, as many dependencies are unavailable for the fringe platforms. 01:57 < first-order> Although you could also get away with running Wheezy or even Jessie on Slot 1 or Socket 7 hardware. 01:57 < TRS-80> misternumberone: this foe may be beyond me, as I am but a low level wizard presently 01:58 * first-order wouldn't mind testing that idea on a Thinkpad T20...... 01:58 < misternumberone> At present I believe I need either a way to download and install the floppy kernel module in the installer CD image, or a way to run the installer in a DebianLive image 01:59 < TRS-80> first option is the way I think I would try to go, although I don't really know how to do that on GNU/Linux (yet) 01:59 * TRS-80 gets popcorn 01:59 < misternumberone> first-order: would it be possible to use the Wheezy LTS installer to install debian 7 with floppy support in the installer, then upgrade to stretch for support for current packages without backports? 02:00 < first-order> Haven't tried that, don't have any hardware in use old enough to justify Wheezy at this point. 02:01 < first-order> Got Stretch running fine on Haswell hardware, but it's also modern compared to what I think you're trying to run. 02:02 < first-order> Ironically, I had trouble getting Ubuntu 16.04 to detect my wireless adapter, while Debian Stretch detects it without issue. 02:02 < misternumberone> The concern I have is that that would require upgrading first to jessie and then to stretch, and transitioning the installation from legacy service to systemd at the same time 02:02 < date_night> Debian is best distro! 02:03 < first-order> There is no 'best distro,' it's entirely subjective to usage case. 02:03 < misternumberone> it would seem better to start in a stretch systemd installation to begin with, but I need an installer that supports my floppy 02:03 < jim> dis-tro... distro duck! 02:03 < date_night> wait people still use floppy? 02:04 < Dagmar> No. 02:04 < first-order> That being said, there isn't much to gain from running Ubuntu to justify it over Debian. 02:04 < Dagmar> Polish. 02:04 < first-order> Or Manjaro over Arch. 02:04 < misternumberone> date_night: Dagmar: I need to install GRUB to floppy disk in order to boot my system 02:04 < Dagmar> Fewer people that will know you waste a lot of time recompiling things? 02:04 < syb0rg> misternumberone, try installing fdutils from the live system 02:05 < first-order> Actually, Manjaro gets the shit bitched outta them for taking forever to run security updates downstream. 02:05 < date_night> misternumberone: what computer would that be? that requires floppy to boot 02:05 < jim> apparantly some do... we had someone yesterday asking about drivers 02:05 < syb0rg> it might install the floppy module as a dependency 02:05 < misternumberone> syb0rg: the floppy disk works fine in the live system. the problem is the live system does not contain the installer 02:05 < syb0rg> really? that's lame 02:06 < syb0rg> I assumed debian was like ubuntu with the installer available in the live version 02:06 < syb0rg> that's a really convenient feature to be missing 02:06 < jim> misternumberone, live which? 02:06 < misternumberone> syb0rg: I thought that would make sense as well and the debian website implies it, but i can't find how to run the installer from it, it seems missing 02:07 < date_night> Hmm i thought debian live iso contains the installer, guess not then... 02:07 < misternumberone> jim: DebianLive 9.4 CD 02:07 < jim> I thought so too 02:07 < misternumberone> from https://www.debian.org/CD/live/, it seems not to contain the installer 02:08 < syb0rg> " A "live install" image contains a Debian system that can boot without modifying any files on the hard drive and also allows installation of Debian from the contents of the image. " 02:09 < jim> best thing to do is ask them 02:09 < misternumberone> I tried to install sudo apt-get install debian-installer-launcher, and encountered errors so i came. The page states "If you intend to install Debian from the downloaded live image, be sure to have a look at the detailed information about the installation process." however the installation manual does not mention live CD image. 02:09 < first-order> Netinstall enables you to install everything from the web, or go Arch-style, installing a base console system, and then installing a desktop from there. 02:10 < syb0rg> what errors did the live installer give you misternumberone? 02:10 < syb0rg> was that what you originally mentioned yesterday where it said the module floppy didn't exist? 02:10 < misternumberone> when I type sudo apt-get install debian-installer-launcher, "FATAL -> failed to fork" right after "processing triggers for menu" and then when sudo debian-installer-launcher, "no suitable d-i initrd image found, aborting" 02:10 < syb0rg> try running sudo apt-get update first? 02:11 < misternumberone> yesterday, I was trying to use the regular installer image, not live, and it did not contain the floppy kernel module for floppy drive support 02:11 < misternumberone> syb0rg: yes 02:11 < syb0rg> ok, well at least it can use the floppy drive from the live version, that is a good sign. Now you just need to get that pesky installer... installed. 02:11 < syb0rg> odd choice by debian not to have it included by default 02:12 < shalok> I'm considering striping my disks for increased performance. It seems this is done with either RAID0 or LVM2 striped LVs. Which should I use? 02:12 < banisterfiend> hi, i am messing with iptables. I hvae two custom chains -- how do i get the rules in chain A to always override the rules in chain B ? 02:12 < domhnall> banisterfiend: which one was entered first? 02:12 < banisterfiend> domhnall it's not determined which chain is defined first, they can be turned on and off at will 02:13 < syb0rg> misternumberone, how much RAM does this machine have? 02:13 < banisterfiend> but i always want chain B (which may or may not be added before chain A) to have higher precedence than chain A 02:13 < syb0rg> it looks like it might be running out of RAM, in which case you just need to create a swap file/partition to expand your memory before installing 02:15 < misternumberone> syb0rg: I noticed that some say that error occurs when out of memory. The system has 384 MB SDRAM and does not support more. however the live image does not have a swap file. Will it work to make a swap partition on the hard drive I will later install to and try to use that? 02:15 < syb0rg> yes :-) 02:16 < syb0rg> that is exactly what you should do next 02:16 < jim> misternumberone, how did you find out it didn't have the floppy drivers? 02:17 < syb0rg> he pasted in the error message yesterday jim 02:17 < dannylee> .>> 02:17 < syb0rg> it said so when he tried to install grub2 to /dev/fd0 02:18 < misternumberone> jim: I attempted to install GRUB to /dev/fd0 and discovered that /dev/fd0 did not exist 02:18 < misternumberone> despite the floppy drive working in windows 02:19 < jim> ok 02:23 < jim> misternumberone, were you talking about this yesterday? 02:24 < misternumberone> jim: I asked here about the regular installer yesterday 02:25 < jim> ok... was there something special going on, like this being in a vm or floppy drives are somehow special? 02:26 < jim> what media is it you're trying to install debian onto? 02:27 < misternumberone> it is a normal floppy drive, the motherboard is ASUS P28-VT 02:27 < misternumberone> i want to install debian to the hard drive 02:28 < misternumberone> the reason I need a boot disk is because I will be using full disk encryption and that requires a separate boot disk in cases where the motherboard firmware BIOS is on a read-only memory device such as is the case with the ASUS P28-VT 02:30 < syb0rg> misternumberone, in that case you probably aren't gonna want that swap partition on the hard drive - do you have a usb stick for that? 02:30 < Dan39> misternumberone: non-luks then, right? :) 02:33 < dannylee> i type root and my password...but it wont let me into root..i`m using fedora 27 lxde 02:33 < saderror256> hi 02:33 < Dan39> dannylee: anything more to the story? 02:33 < saderror256> so i just used dd to backup my os 02:33 < saderror256> with 02:33 < misternumberone> syb0rg: Dan39: LVM inside LUKS, and the idea would be create unencrypted swap partition now for the live CD, then overwrite all that with the encryption when installing 02:33 < saderror256> dd if=/dev/sda of=backup.iso 02:33 < Dan39> dannylee: google for how to recover 02:34 < saderror256> and when i run it in virtualbox it doesnt work. 02:34 < saderror256> why? 02:34 < syb0rg> ah, fair, misternumberone. 02:34 < dannylee> i just want tooo get into root...too delete history 02:34 < lavamind> would it be considered safe to disable ext4 write barriers when using a journal device that has power-loss protection? eg. Intel DC S3700 NVMe device 02:35 < Dan39> misternumberone: im a fan of plain non-luks, so that people can't tell it is encrypted 02:35 < syb0rg> dannylee, I have never used Fedora, but are you sure root has a password? Try: sudo su (fedora uses sudo, right?) 02:35 < Dan39> yea it uses sudo -_- 02:35 < Dan39> if it's installed heh 02:35 < Dan39> probably is 02:36 < dannylee> yes ive done that 5 time...yes 02:36 < syb0rg> saderror256, what is the error you are getting? 02:36 < Dan39> dannylee: thats not the root password then 02:36 < syb0rg> what specifically doesn't work? 02:36 < Dan39> dannylee: if you do `sudo su` you have to type your user's password 02:36 < dannylee> ok what is my password 02:36 < syb0rg> ... 02:36 < Dan39> -_- 02:37 < syb0rg> are you trollin' there bud? 02:37 < dannylee> no 02:37 < jim> misternumberone, does your situation allow you to be flexible about which part of the disk is encrypted? 02:37 < syb0rg> dannylee, why do you think we can tell you your password? 02:37 < saderror256> syb0rg, no errors. it looks as if it works but if i boot in through virtualbox it says no bootable disks found 02:37 < dannylee> i just change my root password yesterday 02:37 < misternumberone> jim: I prefer LUKS over entire disk for encryption so that the data in areas like /boot or swap partition does not reveal information 02:38 < syb0rg> saderror256, so you are making an image of the virtual device, then attaching that image to a new machine and trying to boot to it? 02:38 < dannylee> i even tryed my old password 02:38 < saderror256> i guess so yes 02:38 < jim> misternumberone, ok, so why not do that? 02:39 < syb0rg> saderror256, can you mount the image you made and see if everything looks right? 02:39 < saderror256> sure 02:39 < dannylee> lxde is a bit of a weeeeny 02:39 < misternumberone> jim: I am doing that I said above? 02:40 < dannylee> ok thanks...my kids might have mess me up... 02:40 < syb0rg> dannylee, so your real problem is that you forgot the root password? 02:40 < jim> well in one paragraph you said you were doing full-disk encryption 02:40 < saderror256> syb0rg, wont mount, im going to flash to usb and try that then 02:40 < dannylee> no 02:40 < dannylee> no 02:40 < saderror256> no 02:40 < syb0rg> lol. 02:40 < dannylee> i know my password 02:41 < saderror256> dannylee, is it "passwordcakeswordwithkittensinabucket15567338370-==+[]{}" 02:41 < dannylee> my daughter is satan.. 02:41 < misternumberone> well I am, LUKS placed over entire disk is full disk encryption, right? 02:42 < Dan39> misternumberone: well there is luks headers telling everyone the data is encrypted 02:42 < syb0rg> saderror256, I guess another question would be - why do you need to do this? Can't you just make a clone of the machine? 02:42 < jim> misternumberone, what if you did what you were going to do on the floppy, on a usb stick instead? 02:43 < syb0rg> Dan39, is it possible to store the header on a separate device (and still be able to use the encrypted device)? And if so, would the storage medium just look like it contained noise? I was curious about that earlier 02:44 < misternumberone> Dan39: yes this is true, and I thought of asking you earlier from something you said, how is it possible that you could have a full disk encryption, and it not be possible to tell that the device is encrypted/ 02:44 < Dan39> syb0rg: you dont need a header 02:44 < syb0rg> oh, really? 02:44 < rcm888> where is xrdp channel? 02:44 < syb0rg> So that is just a convenience feature? 02:44 < Dan39> https://wiki.archlinux.org/index.php/Dm-crypt/Encrypting_an_entire_system#Plain_dm-crypt 02:44 < saderror256> syb0rg, well im actually making an operating system for fun, lol, but i thought saying i was doing a backup with dd wouldve been easier for people to understand 02:44 < misternumberone> jim: it is not possible for the computer to boot directly from a USB device 02:44 < Dan39> syb0rg: afaik yea 02:44 < jim> rcm888, there is a bot, alis, that can assist you in looking for channels on the Freenode irc net. To start, /msg alis help 02:45 < Dan39> "Contrary to LUKS, dm-crypt plain mode does not require a header on the encrypted device: this scenario exploits this feature to set up a system on an unpartitioned, encrypted disk that will be indistinguishable from a disk filled with random data, which could allow deniable encryption." 02:45 < syb0rg> lol okay saderror256. My question still applies though. If you are creating the OS in a VM and you want to back it up or make a copy, can't you just do a full clone of the machine? 02:45 < syb0rg> Cool hobby, btw :-) 02:45 < syb0rg> sounds ambitious 02:46 < saderror256> syb0rg, but i want to format it into a .iso format, so it can be bootable from a computer 02:46 < Dan39> syb0rg: hmm there is also like you said luks "detached header" 02:46 < syb0rg> that is really cool Dan39, never knew that 02:46 < Dan39> https://wiki.archlinux.org/index.php/Dm-crypt/Specialties#Encrypted_system_using_a_detached_LUKS_header 02:46 < rcm888> jim: \/msg alis list xrdp ? 02:46 < saderror256> syb0rg, is it more hobbyist if im only 13 and im doing this :P i have lots of free time, school just went out 02:47 < jim> rcm888, for channel names, yes... I forget how to search channel topics 02:47 < misternumberone> Dan39: and the patterns of the data in the disk could not be shown to be similar to the patterns of an encrypted disk that does have a header? I'm all for trying to use dm-crypt without LUKS and instruct GRUB to recognize it somehow, i'm just curious about the benefits 02:47 < saderror256> syb0rg, okay it was flashed, im going to go ahead and test it on another computer 02:48 < syb0rg> Okey dokey saderror256 02:48 < saderror256> selected boot device failed :( 02:48 < Dan39> misternumberone: sounds right 02:49 < syb0rg> saderror256, try looking at the data on the drive and see if it looks correct 02:49 < rcm888> jim: I didnt find anything related to xrdp 02:49 < saderror256> for once a post with 101 upvotes ons stack overflow didnt work for me 02:49 < syb0rg> and check the partition flags 02:49 < Styil> saderror256, what are you trying to do/ 02:49 < Dan39> misternumberone: luks definitely has a detectable signature on the disk, thats like the point. plain dm-crypt wont. 02:49 < saderror256> bad superblock, unable to access 02:50 < syb0rg> I think he means Dan39, without the header is the data truly indistinguishable from noise. At least that is what I got out of the question 02:50 < Dan39> oh 02:50 < jim> rcm888, ok, try /msg alis help list to see how to search the topic 02:50 < Dan39> well thats the point of encryption id say 02:50 < Dan39> but im sure its debatable 02:50 < saderror256> Styil, converting a system into an iso file so it can be burned onto a USB 02:51 < TRS-80> rascul: BTW looks like mid 2019 for next Debian release (Buster) if I am reading this correctly: https://lists.debian.org/debian-devel-announce/2018/04/msg00006.html 02:51 < jim> saderror256, what if you copied files to it instead, and made an image out of that? 02:51 < Loshki> I can think of at least one way to distinguish encrypted data from noise 02:51 < syb0rg> saderror256, also, what bootloader are you using and are you using UUIDs 02:51 < Styil> Hmm, never done that before, but wouldnt that just involve dding the partition into a .iso file? 02:52 < Dan39> like id say any HDD should have some kind of pattern of some sort on it, or be all zero's if brand new. if the entire thing looks random, then it's safe to say someone either purposely overwrote the entire drive with randomness, or it is encrypted 02:52 < rcm888> jim: /msg alis LIST * -topic xorg 02:52 < rcm888> jim: /msg alis LIST * -topic xrdp 02:52 < saderror256> syb0rg, grub Legacy, i am unsure about the UUID's but i think genfstab does it for me? 02:52 < Dan39> so which is better? i don't know, now you have me doubting myself :P 02:53 < jim> rcm888, ok :) did that produce results? 02:53 < Dan39> id say all randomness is better than showing as LUKs 02:53 < syb0rg> I'd agree if you are looking for deniability Dan39 02:53 < rcm888> jim: 0 found for xrdp, some found for xorg 02:53 < syb0rg> saderror256, no clue, I don't know much about confuring bootloaders 02:53 < syb0rg> *configuring 02:54 < jim> xrdp, isn'/t that the remote desktop thing? 02:54 < Dan39> then i can play it off as im a dumb citizen running windows who just doesnt want my friend im going to give the HDD to to steal my identity, not that im running encrypted linux system to revolt against the government that has banned encryption 02:54 < misternumberone> Dan39: what intrigues me is the possibility that one could refuse to clarify whether the drive was encrypted or simply overwritten with random data many times, in both cases leaving for traces only the ghosts of files the drive stored before it was encrypted. 02:54 < saderror256> well syb0rg we are on par, me neither... 02:54 < rcm888> jim: yes 02:54 < saderror256> i tried again using "dd if=/ of=./image.iso" and it says i cant use files 02:54 < misternumberone> and the drive not reveal which is the case 02:55 < syb0rg> I doubt that is the error message you are getting saderror256 =P. Better to be specific about exactly what you tried and exactly what the error message was 02:55 < jim> rcm888, I see no reason why you wouldn'/t get a pretty good response here, as long as you added informative details to your question 02:55 < saderror256> error reading '/': Is a directory 02:56 < saderror256> sorry im lazy to alt tab, virtual machines are fullscreen atm so i have to continue escaping 02:56 < syb0rg> ah I didn't look at the if 02:56 < syb0rg> dd wants a block device input 02:56 < syb0rg> like /dev/sda 02:56 < saderror256> wait... 02:56 < saderror256> last time i tried sad1 02:56 < syb0rg> yeah 02:56 < saderror256> *sda1 ha 02:56 < saderror256> should i do /dev/sda this time? 02:56 < Dan39> saderror256: what are you trying to do now? -_- 02:56 < syb0rg> probably, Dan39 02:57 < syb0rg> oops 02:57 < syb0rg> saderror256, 02:57 < saderror256> the same thing im trying to do 02:57 < Dan39> write hdd image to usb? 02:57 < rcm888> jim: Installed xrdp xorgxrdp from sources onto ubuntu server 16.04 lts + lxde. blank screen when connecting. 02:57 < saderror256> Dan39 02:57 < syb0rg> yes, try dd if=/dev/sda of=./image.iso 02:57 < saderror256> Dan39, to iso 02:57 < Dan39> thats not really an iso id say 02:57 < Dan39> make it a .img :P 02:57 < Styil> saderror256, make sure you run it with sudo if you havent 02:57 < Styil> meh 02:57 < Styil> same thing 02:57 < Dan39> but extension dont really matter, anyways... 02:58 < saderror256> Styil, its root 02:58 < syb0rg> probably accurate Dan39, but he is just gonna burn it to a usb 02:58 < Dan39> syb0rg: what distro is on the HDD? 02:58 < syb0rg> Hm? 02:58 < saderror256> i have a feeling this wont work though, /dev/sda.... wait, its done... yep, i doubt it will work 02:58 < Dan39> sorry 02:58 < Dan39> saderror256: 02:58 < syb0rg> =P 02:58 < saderror256> Dan39, on the virtualbox is Arch 02:58 < syb0rg> fair, I just did that to you 02:59 < Dan39> saderror256: oh im not sure that's a great idea 02:59 < Dan39> saderror256: i wouldnt dd an install from inside a virtualbox 02:59 < Styil> ^ 02:59 < saderror256> oh wait.... "dd: writing to './image.iso': No space left on device 02:59 < Dan39> tarball the os 02:59 < saderror256> Dan39, what what and what? how will that do me any good? 03:00 < syb0rg> saderror256, either enlarge your virtual hard drive or make a shared folder and output the image to that. Or, listen to the advice these guys are giving you 03:00 < Dan39> saderror256: haha 03:00 < syb0rg> assuming you have space on your physical device 03:00 < saderror256> i think what is doing is just dding the disk drive, which will just overflow... 03:00 < Dan39> saderror256: does your computer have a cd/dvd drive? or have another USB? 03:01 < saderror256> Dan39, well, yes... 03:01 < Dan39> just do a proper install maybe? write install iso to disk/usb, boot from it, install to other USB 03:01 < saderror256> oh wait 03:01 < saderror256> i need to mount first woops 03:01 < saderror256> ls 03:01 < syb0rg> saderror256, you are correct actually, enlarging the virtual drive will not work 03:01 < syb0rg> go with the shared folder idea 03:01 < saderror256> syb0rg, okay 03:01 < syb0rg> and make sure the virtual drive is a manageable size 03:01 < Dan39> saderror256: why dont you do a proper install...? 03:02 < saderror256> Dan39, i used my own install script...? 03:02 < Dan39> ok 03:02 < Dan39> like i said, tarball it 03:02 < Dan39> :P 03:02 < Dan39> saderror256: oh btw, you can usually add additional storage drives to VM 03:02 < saderror256> Dan39, but again, what good will that do, because i mean, it just tarballs it, i dont think distro makers like ubuntu tarball their oses... ill tarball it, but i dont get your point 03:03 < saderror256> Dan39 vm is mounted to ~/src on my host system so, its expanded 03:03 < Dan39> saderror256: well it's not proper, but if all you want to do is move the install to a USB stick, it would work 03:03 < Dan39> or not even tarball 03:03 < Dan39> cp -a 03:03 < Dan39> partition/mkfs the usb 03:03 < Dan39> cp -a 03:03 < saderror256> ok im getting confused 03:03 < saderror256> slow down 03:04 < Dan39> theres nothing special about a linux install. it is files on a filesystem. you can copy the files into a new filesystem on your USB, then you just have to install/update grub and possible fstab 03:04 < saderror256> wait, what about grub-mkrescue? 03:05 < Dan39> im not familiar with it, but if you have a VM install, im not sure id want to copy the grub from it, id want to do a fresh grub install while chrooted into the usb 03:05 < saderror256> wait, what im thinking is i can copy the whole os (i think you mentioned that earlier, lol) 03:05 < Dan39> oh hmmm 03:05 < Dan39> grub-mkrescue might help 03:06 < saderror256> just mentioned that 03:06 < saderror256> grub-mkrescue 03:06 < Dan39> right, im researching it after reading what you just said 03:06 < saderror256> oh 03:06 < saderror256> i think i have to do it on the host system not the vm though 03:07 < Dan39> usually i just grub-install for USB installs, but sometimes normal grub install on a USB stick can be annoying if it doesnt show up as the 1st drive 03:07 < saderror256> what does -a do? 03:07 < Dan39> maybe grub-mkrescue fixes that issue 03:07 < saderror256> brb 03:08 < de-facto> i am trying to find a simple oneliner to bloack _all_ facebook traffic from my box. Any ideas which would be the best approach? 03:08 < de-facto> i am trying to do something like this "for NET in $(whois -h whois.radb.net '!gAS32934' | sed -n 2p) ; do sudo iptables -A OUTPUT -s $NET -j DROP ; done" 03:08 < de-facto> yet it doesnt work properly yet 03:08 < solidfox> ubuntu sure updates quite a lot 03:09 < solidfox> every day I have at least 2 or 3. and sometimes over 100 03:09 < Dan39> saderror256: well for cp, -a per the manual is -a, --archive same as -dR --preserve=all 03:09 < syb0rg> solidfox, better then not updating enough. =P You could choose to have security updates auto-install, and do other updates when you feel like it 03:10 < solidfox> syb0rg, I don't like auto install because sometimes I need to install a program, but you can't if you're updating. 03:10 < jim> misternumberone, what I can suggest is if you were to go to the #debian channel, be concise and informative, as well as generally cool. (don't worry, you have been here... the thing is, they get hammered ever day pretty much, you name it, they get hammered by it... if you were to give them full details on the situation, they might be able to suggest something... I'm out of ideas 03:10 < Dan39> saderror256: for most arch installs that will work perfectly. if by some crazy chance you are running selinux (like redhat/fedora does) then you would want to also add -Z, though it's not required since you could just touch /.autorelabel to fix that on the 1st boot 03:10 < solidfox> syb0rg, but I do keep up 03:11 < drakonan> What is the best single board computer for router firewall and qos ? 03:11 < solidfox> drakonan, I also wonder the same thing 03:11 < solidfox> I'd like to make my own firewall 03:11 < saderror256> hi back 03:12 < saderror256> okay so some files failed but they were in /sys/ thats okay right? 03:12 < Dan39> yea thats fine 03:12 < misternumberone> jim: well as part of my current attempt, I'm trying to temporarily use and then stop using a swap partition in the live image. I created one, but top shows 0 swap still. How can I instruct the live image to use the swap, while then being able to stop using it later so I can overwrite the partition? 03:12 < saderror256> thanks tty, cant view my old logs :/ 03:12 < Dan39> hahaha 03:13 < Dan39> i feel your pain 03:13 < drakonan> I want to run pfsense on an arm platform but isk if the source is avail or not 03:13 < drakonan> Idk* 03:13 < syb0rg> swapon /dev/sd[letter][number] misternumberone 03:13 < syb0rg> then when you are done with it, swapoff /dev/sd[letter][number] 03:14 < Dan39> misternumberone: you could just create a swap file inside the encrypted swap filesystem 03:14 < syb0rg> the encrypted fs doesn't exist yet Dan39, he is gonna make it when he installs 03:14 < Dan39> right, sorry 03:14 < Dan39> i realized he mentioned that already after i hit enter, sorry 03:15 < Dan39> saderror256: after copying the filesystem over to usb, if on arch now you can arch-chroot /path/to/usb, then run grub-install /dev/sdX 03:16 < jim> saderror256, so what you're trying to do is image your hd? 03:16 < saderror256> yes 03:17 < Dan39> well he said he is trying to copy an arch installed in a VM to a USB stick, and make it bootable, but do correct me if i'm wrong saderror256 03:17 < Dan39> him trying to make an image of it was just his attempt at how to do that afaik :P 03:18 < saderror256> Dan39 correct 03:19 < Dan39> saderror256: ok so lets say sdb is your USB stick, you copied all the files from VM to filesystem /dev/sdb1 ? 03:20 < saderror256> yeah 03:20 < saderror256> i have to go here, so goodbye, sorry lol 03:20 < jim> misternumberone, one thing they said, is that if the modules are available, you should be able to modprobe floppy (I don't know how early (or even if) the modules come into play in the installer 03:20 < Dan39> so now just genfstabd and grub-install and it should work 03:20 < syb0rg> good luck with your project, saderror256 03:21 < saderror256> syb0rg, thanks! 03:21 < Dan39> damn i was hoping to see him get it working 03:21 < Dan39> it was like 1 more step 03:21 < Dan39> chrrot, grub-install and done pretty much 03:22 < Dan39> of course id like to get him using UEFI, but maybe next week :P 03:22 < syb0rg> Dan39, for a portable image I prefer BIOS 03:22 < Dan39> syb0rg: oh, i prefer both 03:22 < Dan39> ;) 03:22 < syb0rg> hmmmm 03:23 < Dan39> i tested it recently, works great 03:23 < syb0rg> like install two bootloaders? 03:23 < Dan39> will boot both UEFI and legacy to same OS 03:23 < Dan39> kind of 03:23 < syb0rg> I mean I know UEFI has some legacy support 03:23 < Dan39> grub-efi, grub-pc 03:24 < Dan39> it gets a bit iffy deciding if dos vs gpt partition 03:24 < syb0rg> ok, so do you use an msdos partition table and set it up to boot with grub-pc, but have grub-efi and an efi partition for machines that can use it? 03:24 < Dan39> i did dos and my PC booted to UEFI just fine still 03:24 < Dan39> syb0rg: correct 03:24 < syb0rg> Interesting approach, is that a standard thing for live images? 03:24 < Dan39> no idea 03:24 < Dan39> i did this for a full install, not live image 03:25 < Dan39> i tested it with syslinux first actually, someone else had an article that way 03:25 < Dan39> but grub is easier for me usually 03:25 < syb0rg> yeah that is intriguing, I wonder if it would confuse some machines though 03:26 < Dan39> i dont think so 03:26 < syb0rg> I know some machines don't like EFI on MSDOS, but then you would just fall back to grub-pc 03:26 < Dan39> cept some UEFIs that dont like booting EFI from msdos partition 03:26 < Dan39> right, beat me to it :P 03:26 < syb0rg> hehe 03:26 < Dan39> syslinux was annoying actually 03:27 < Dan39> compared to grub 03:27 < syb0rg> never messed with any bootloader other than grub 03:27 < syb0rg> and even then I mess with it minimally 03:27 < syb0rg> just install and pray it works 03:27 < Dan39> heh 03:27 < Dan39> it is a relief to finally get good with grub 03:27 < syb0rg> I bet 03:28 < Dan39> for years when i only used linux at home i dreaded touching anything to do with the bootloader 03:28 < syb0rg> Yeah I dread it a bit, it is a very fragile black box to me 03:29 < syb0rg> but I can do the whole chroot and reinstall deal, so it isn't too terrifying 03:29 < syb0rg> I can always reinstall it if I break it 03:29 < Dan39> heh 03:30 < Dan39> back then i also thought there must have been some more magic that the installers were doing. turns out there isn't.... 03:30 < Dan39> only a little magic that grub does, and then a bunch of files in a filesystem 03:31 < syb0rg> Yeah I'm sure I could learn it, but I have avoided it so far 03:31 < syb0rg> probably should get around to it some day 03:31 < syb0rg> maybe in the 2030s.... 03:31 < xor> I input my password on a login input from a terminal, does anyone know where those logs are stored, so I can clean them? 03:31 < Dan39> wonder how much legacy BIOS will still be around by then 03:32 < Dan39> xor: look where all the others logs are maybe? -__ 03:32 < Dan39> that would be /var/log 03:32 < Dan39> if systemd, rm those journals too 03:33 < xor> Dan39, I did that already, any other places? 03:33 < Dan39> ummm 03:33 < Dan39> hmm 03:33 < Dan39> in /etc might be a file 03:33 < Dan39> the utab or something, i dont remember 03:34 < Dan39> maybe not 03:34 < Dan39> im thinking of wtmp 03:35 < Dan39> which looks like is in /var/log 03:35 < Dan39> i say just rm those bad boys, i dont know off hand how to delete entries from wtmp :| 03:35 < Dan39> but maybe google does haha 03:36 < xor> I did a grep for part of my password, which now I also have to remove from bash history, and found nothing on /etc and /var/log, at least not in text format. 03:36 < de-facto> so great that journald is binary 03:36 < Dan39> yea i said delete the journals in /var/log 03:37 < xor> de-facto, Not so great, but at those files are gone. Or so I think, I don't know if there are caches somewhere else. The command itself doesn't pull up any entries now. 03:39 < Dan39> xor: think again, binary doesnt mean there isn't text in there that is grep'able though, run `strings` on your journal files and check it out :P 03:39 < syb0rg> xor, I could have mentioned this sooner but shred -u is a good secure delete tool if ya didn't know 03:39 < xor> Dan39, I removed all the files on var/log/journal. 03:39 < syb0rg> although rm'ing the log files is probably good enough unless you're paranoid 03:40 < xor> syb0rg, ahhh I see. 03:40 < xor> I might just change the password then. ^_^ This one is very non ergonomic, if that makes sense. 03:40 < Dan39> syb0rg: good idea, should be shredding 03:41 < syb0rg> it makes sense xor, I know all about non-ergonomic passwords 03:41 < Dan39> i meant to say de-facto think again, not xor <_< 03:41 < Dan39> not xor, har har. 03:41 < syb0rg> get it together Dan39 =P 03:42 < xor> Dan39, ^_^ 03:42 < xor> Well, thanks for the tips folks. Have a good night. 03:42 < Dan39> xor: user login attemps, at least on my system, are actually plain text in the journal files :P 03:43 < rypervenche> He left :P 03:43 < de-facto> huh? i just think journald uses a horrible format and interface, thats all 03:43 < Dan39> de-facto: i see, ok. i agree 03:43 < syb0rg> what is the rationale for binary logs, again? 03:44 < Dan39> when i have to wait minutes to view my logs from last month, i dont like it, instead of just less'ing the file timestamps from last month 03:44 < dannylee> now i can get into root with my terminal..but i still can~t get in with a graphical login...ok rm .bash_history./.this did work ok 03:44 < Dan39> syb0rg: who knows, i've read the argument a couple times. i just accept it and have both :P 03:45 < Dan39> dannylee: you shouldn't be logging into graphical with root 03:45 < dannylee> ok 03:46 < syb0rg> it seems like everyone hates binary logs honestly, pretty annoying choice 03:46 < dannylee> yes i cant login with a graphical desktop 03:46 < Dan39> dannylee: as a non-root user? 03:46 < syb0rg> dannylee, good, don't (if you mean as root) 03:47 < dannylee> but i have root...i~m ok 03:47 < Dan39> now i know dannylee is troll 03:48 < dannylee> the termial is all i care about 03:49 < Dan39> maybe even a robot 03:49 < dannylee> the terminal is all i care about 03:49 < dannylee> sorry i just cant spell 03:49 < dannylee> tonight 03:49 < Dan39> it's trying too hard, definitely a robot 03:57 < misternumberone> jim: when I use modprobe floppy (in the installer) I get file not found error; syb0rg: When I run sudo swapon /dev/sda2, "swapon: /dev/sda2: read swap header failed" I created the linux-swap as 2nd partition in /dev/sda, after deleting 1 of 2 old partitions 03:57 < syb0rg> misternumberone, how did you make the swap partition? 03:58 < syb0rg> you might need to run mkswap /dev/sda2 03:58 < misternumberone> syb0rg: sudo parted -> mkpart primary linux-swap 2GB 10GB 03:58 < syb0rg> yeah you didn't make the filesystem I think misternumberone 03:59 < syb0rg> try mkswap, then swapon again 03:59 < syb0rg> well "filesystem" - swap space more accurately I guess 04:01 < misternumberone> thanks it works now I can try to install debian-installer-launcher 04:01 < syb0rg> cool tell us if it works 04:03 < ikus060> Hello, I want to access the sas controller chipset via i2c of gpio. But so far, my tentativ is failing. The chipset in question is MG9072. 04:06 < syb0rg> ikus060, the only part of that I don't follow is "my tentativ is failing" 04:06 < syb0rg> what does that mean? 04:06 < syb0rg> (I strongly doubt I can help you with this problem, but I am curious) 04:07 < cmj> the least 04:07 < kurahaupo_> ikus060: could you provide the English word for "tentativ" 04:07 < cmj> 'tentative', presumably 04:07 < syb0rg> that still doesn't parse 04:07 < ikus060> yes *tentative 04:08 < cmj> the least to be expected 04:08 < kurahaupo_> ikus060: tentative is an adjective 04:08 < syb0rg> maybe you meant something more like attempt?? 04:08 < kurahaupo_> Perhaps you mean "attempt"? 04:08 < syb0rg> hey =P 04:09 < syb0rg> in any case, what did you try and what error did you get? 04:09 < kurahaupo_> syb0rg: is beaten by someone on a swipe touch screen 04:09 < cmj> glad that's ironed out 04:10 < syb0rg> I wasn't beaten kurahaupo_, my comment was first 04:10 < ikus060> To rephrase, I have a supermicro backplane with chipset MG9072. This chipset is responsible to manage the Fan. I'm trying to access the chipset directly either via i2c or sgpio. But my *attempt" failed 04:10 < syb0rg> by a fraction of a second, I'll have you know 04:10 < supernov1h> Is it possible to create a folder with symlinks to files I want to make available to a specific user to essentially create public ssh access to a specific set of files in a large directory, which I don't want them to see 04:10 * kurahaupo_ christens syb0rg "captain slow" 04:10 < supernov1h> as in a jailed user with symlink visiblity to a chosen list of files 04:11 < syb0rg> lol, okay then kurahaupo_. 04:11 < cmj> do you have gpio and said support in the kernel, ikus060? 04:11 < Dan39> supernov1h: i dont think so, but tias 04:11 < ikus060> I think sgpio cannot be used to control the fan. I guess. So I'm looking into i2c. But I can't figure the bus to contact the chipset. 04:11 < syb0rg> and what error message did you get ikus060, how specifically are you trying? 04:11 < cmj> sometimes you might need to go through and see what hardware you have and load manually. 04:12 < ikus060> Well, i2cdetect is reporting this: https://pastebin.com/JVRmjjWE 04:13 < misternumberone> debian-installer-launcher installed without the error this time, but when I run it sudo debian-installer-launcher I still get "no suitable d-i initrd image found" 04:13 < kurahaupo_> ikus060: sorry I'm not meaning to nit-pick your English, it's just that I'd rather not make wrong assumptions about what you mean and waste time answering the wrong question 04:13 < cmj> i2c modules, also, gpio are not always built in most distros' kernels 04:13 < ikus060> I was expecting a response from one of the 0x20-0x2f range. 04:14 < ikus060> cmj: I'm running debian. I have i2c and gpio 04:14 < cmj> me too 04:14 < Dan39> misternumberone: maybe just do a manual install 04:15 < syb0rg> wow, you have no luck misternumberone 04:15 < Kharma> I'm installing an apt package which depends on an app I havw inatalled through python - python had the newest version so I installed with pip.. the app I'm currently installing with apt, is now installing the older apt version of the python app.. how to satisfy a dependency of apt with a python module, possible? 04:16 < Dan39> misternumberone: https://wiki.debian.org/Debootstrap https://www.debian.org/releases/stable/amd64/apds03.html.en 04:16 < cmj> that sounds destructive 04:17 < misternumberone> Dan39: syb0rg: I would really like to manually install debian the way a distro like arch is traditionally manually installed, but there is no guide in the documentation to installing without the installer 04:17 < cmj> ikus060: what chipset is MG9072? 04:17 < misternumberone> if I tried to use a guide for arch and just adapt it would that have a chance of working? 04:18 < Dan39> misternumberone: i just linked you 04:18 < Dan39> misternumberone: 2 links with info :p 04:18 < syb0rg> debootstrap looks pretty neat 04:18 < cmj> i see what it is now; out of my league 04:18 < misternumberone> thanks 04:22 < ikus060> cmj: It's the chipset on the backplane for SAS: https://ami.com/ami_downloads/AMI_Backplane_Controller_MG9072_Data_Sheet.pdf 04:23 < Psi-Jack> Man.. ProxySQL is pretty sweet 04:23 < phogg> Kharma: No. You would have to make your own .deb package out of the newer version, give it the same name and a compatible versioning scheme. 04:24 < phogg> Psi-Jack: and how does that work, then? 04:24 < phogg> asking for a friend 04:25 < Psi-Jack> phogg: Well, you can configure ProxySQL to effectively run in front of MySQL, and it can do fairly seamless failover, or load balancing between servers. 04:27 < phogg> Psi-Jack: Reading the feature list now. Looks nice, but is mysql-centric -> useless to me. 04:27 < Psi-Jack> In my case, I have two master-master Percona 5.7 servers running, with the auto-increment shift setup so they can each run nearly conflict-free, and then I have ProxySQL actually being the connection point to them and it load balances and connection pools the clients. 04:27 < Psi-Jack> Yes, it is MySQL-centric. 04:27 < cmj> speaking of, i like haproxy 04:27 < Psi-Jack> cmj: I like fabio better. 04:28 < cmj> url? (so i don't get hot dude with long hair) 04:28 < Psi-Jack> hahahaha 04:28 < Psi-Jack> google fabiolb 04:28 < cmj> ok thans 04:29 < Psi-Jack> fabio works with consul and vault to have live dynamic changes of routing rules and certificates. 04:29 < cmj> ah nice 04:29 < Psi-Jack> And.... It's actually faster than haproxy. 04:34 < Psi-Jack> Even has a noroute html file you can provide as a default if someone tries to hit a host header that doesn't have an actual route for it. Which I use to say "You're obviously unauthorized, so stoppit!" 04:34 < Psi-Jack> heh 04:40 * aBound mmmm banana cream pie 04:58 < supernov1h> do broadcast messages show up on the standard stdout if you're ssh'd into a server and see them 04:58 < supernov1h> or are they differentiatable from other streams 04:58 < supernov1h> sorry, pts* 05:03 < Dan39> supernov1h: they show up 05:03 < Dan39> most of the time 05:03 < Dan39> depends on distro :P 05:12 < supernov1h> interesting, because I have software that parses stdout on a server and I think a broadcast message would be a bit strange to handle 05:12 < supernov1h> another question though: can I symlink files to a user which doesn't have permissions to see the file? 05:14 < cmj> to a user? 05:14 < cmj> yes many can see the file exists without read permissions 05:15 < syb0rg> supernov1h, I don't use symlinks that often but don't they preserve permissions? 05:15 < supernov1h> cmj no I Want to provide a user access to a file that it doesn't have permission to view because it's owned by another user 05:16 < supernov1h> or is that what the any user can read permission is for... 05:16 < supernov1h> the third column... 05:16 < cmj> there is fs.protected_symlinks too 05:16 < syb0rg> supernov1h, whenever people talk about doing tricky stuff with permissions, access control lists come up 05:16 < cmj> what acces are you providing if they can't view it 05:16 < syb0rg> maybe you should look into them? 05:17 < syb0rg> I can't help you with those because I have never used them, but they may help you. 05:18 < supernov1h> cmj: the user should be able to download the file over ssh 05:18 < cmj> that's read permissions 05:18 < supernov1h> exactly, so the third column read should make it okay 05:18 < supernov1h> given the file tree up to that point is also accesible 05:24 < misternumberone> while installing the packages I would need to install using debootstrap (lvm, linux-util, cryptsetup, debootstrap, and more i haven't noted yet) and their dependencies, the live image ran out of disk space. I'm not sure how exactly the live image temporary storage works, but I have the swap partition now. Is there some way I can expand it? 05:26 < syb0rg> oh man. You *really* have no luck. 05:27 < misternumberone> well all of this is supposed to be easy to overcome I just have to keep trying 05:27 < syb0rg> yeah I mean you'll get there eventually. I'm not sure how live temporary storage works either, though. 05:43 < supernov1h> is there a way to tell what told the system to restart 05:43 < supernov1h> when I log in, my sshd_config line PrintMotd seems to be ignored, (its no) and it tells me the system needs to be restarted 05:46 < MrGrz> hi 05:54 < realist_> Hello, I’m part of a company that built a distro for data center networking, I’d like suggestions on places i might be able to let people know about our distro, example : websites that help you choose a distro, Linux news websites, popular forums…etc. 05:56 < aBound> https://distrowatch.com <<<< is generally a website that shows a few new distros on the sidebar to the right also lists more than a dozen. 05:57 < realist_> abounbd, i submitted it there, thanks 05:59 < syb0rg> realist_, reddit.com/r/linux 06:00 < syb0rg> news.ycombinator.com 06:00 < kbob> tuxmachines.org 06:00 < realist_> syb0rg: good one, thanks 06:00 < supernov1h> Is there a way to set a home directory in a /etc/pam.d/sshd like a CHRoot? 06:01 < [R]> supernov1h: like a chroot? 06:01 < supernov1h> yeah my chrootdirectory is ignored because pam is being used 06:02 < supernov1h> I want one particular user not to be using pam but you can't se that in sshd_config 06:05 < jim> expand it out all the way... what is this all about, and what -exactly- do you want to accomplish? 06:06 < supernov1h> I Want to control their home directory 06:06 < supernov1h> and their environment to think its the root 06:07 < supernov1h> yes I realise that it requires configuring PATH, the ssh service to be built a specific way etc 06:07 < jim> so you want their home dir to be chrooted? 06:07 < supernov1h> yes I guess if that wasn't obvious? 06:07 < supernov1h> I basically said that 06:08 < jim> if you do that, then there is -nothing- else, no executables, no libs, no services, no system-wise config files 06:08 < CrazyTux> hello, when I have installed multiple DEs on a distro, say Xubuntu on Ubuntu Mate, how can I set Xubuntu app not to appear when I log into Mate? 06:08 < supernov1h> I know that already jim 06:08 < jim> so you don't want them to execute any programs? 06:08 < supernov1h> are you capable of answering my question or not 06:09 < supernov1h> because this has turned the other way 06:09 < syb0rg> supernov1h, you strike me as the kind of person who doesn't want help 06:09 <@jim> probably, I am. 06:09 < CrazyTux> for example xfce terminal appears when I log into Mate. 06:10 <@jim> that's hostility, and it's the -one- thing I don't tolerate 06:10 < aBound> CrazyTux: Wouldn't it be more ideal just to remove Xfce from Ubuntu MATE? 06:10 < supernov1h> Right well I'm quite aware of what will happen to the user when they log in, but currently since pam is hijacking the authentication for ssh, a conventional chroot openssh style, is not working 06:10 < supernov1h> I just want to, without disabling pam, have this user chrooted 06:10 < CrazyTux> aBound, I want both of those DEs. 06:11 < jim> why don't you just delete the user? obviously you don't trust him 06:11 < [R]> jim: haha 06:11 < supernov1h> the user is a chinese manufacturer who need to log in and see a directory of binaries 06:11 < aBound> CrazyTux: Ideally, you want both DEs but you want to hide Xfce from the login screen? I assume that's what you're after. 06:11 < jim> tell him "sorry, but this arrangement didn't work out because I don't trust what you do and what you might do" 06:11 < supernov1h> so no they are not trusted 06:11 < CrazyTux> aBound, I was just curious whether there is any settings that can make it possible. 06:11 < CrazyTux> aBound, yes. 06:12 < jim> then just terminate the relationship 06:12 < CrazyTux> aBound, not at the login screen. 06:12 < supernov1h> unfortunately the cost of manufacturing locally is about 2^32 times grater 06:12 < jim> why are you supposed to do hours worth of work, essentially, securing your machine -for- him 06:12 < sn00bie> the remount doenst work because the device is busy, how can i solve that? 06:13 < supernov1h> you accuse me of being hostile while you remain condescending in almost everything you said toward me by the way, you do realise that 06:13 < CrazyTux> aBound, for example when I log into Mate, is it possible that apps like xfce terminal can be hidden? 06:13 < Triffid_Hunter> sn00bie: lsof -n | grep /mount/point wil tell you what's using it, unless you have swap or loop or some similar kernel thing accessing it 06:14 < jim> supernov1h, I'm just saying, if you don't trust him, why do you let him use your machine? he can get his own machine, and his own net, that does not involve yours 06:14 < aBound> CrazyTux: Sounds like you're trying to use the MATE DE yet hide all of the Xfce programs. 06:14 < CrazyTux> aBound, yes. 06:15 < jim> shell access is an act of trust, there's no way around that 06:15 < supernov1h> jim: we have a build server, this is just a way to access a folder on it 06:16 < jim> and if you chroot him the way you say, he can't even run bash 06:16 < supernov1h> he doesn't need to, he runs a windows program which does it in the background 06:16 < supernov1h> it's just a security measure 06:17 < jim> also, if you chroot him, his user can't read the pam configs 06:17 < aBound> CrazyTux: I'm not familiar or know if that's possible, as I only use Unity at the moment. I can't comment if you can hide it seeing as it's installed. 06:17 < CoCo_Kid594> sounds 2 me like he doesnt want him seeing his porn.. LOL.. :) 06:17 < CrazyTux> aBound, ok 06:17 < jim> I'm not sure he has to, but one things for sure, he can't 06:18 < jim> CoCo_Kid594, please avoid the shortenings, like: please spell out u as you, it helps people (particularly new english speakers) to understand, at least, most of what's going on 06:18 < CoCo_Kid594> it made me laugh.. i remember trying to CHROOT a FTP.. I just ban umm. 06:19 < aBound> CrazyTux: What about removing certain Xfce programs you don't want. 06:19 < CoCo_Kid594> thanks will do... 06:19 < jim> no, he's calmed down now 06:19 < CrazyTux> aBound, I will need them when I log into Xubuntu. Leave this topic. I asked this out of curiosity. 06:20 < domhnall> CrazyTux: similar to you, I have Xfce but also KDE. Only xfce I see when using kde is the xfce terminal. afaik. 06:20 < supernov1h> jim: no the shell doesn't need to execute a thing except ls and whatever lld ... /usr/bin/ls requires 06:20 < aBound> :P 06:21 < jim> if you chroot him, he won't have access to any part of the / partition... not /, not /usr, not /usr/bin, and not /usr/bin/ls... and not (as you say) the libs it needs 06:21 < swine_> are there any sendmail gurus around? 06:21 < swine_> i've got the weirdest problem 06:22 < cmj> is this 1996? 06:22 < domhnall> hah, what a bug? 06:22 < swine_> @cmj more like 2006 06:22 < jim> swine_, sendmail is ooollllddd stuff :) 06:22 < swine_> @jim i know. i'm stuck on a centos 5.3 box 06:22 < jim> but that's not bad if it works 06:22 < domhnall> swine_: on linux or a *bsd? 06:22 < [R]> sendmail is still a thing? 06:22 < domhnall> ah... 06:22 < swine_> linux 06:23 < swine_> the damn thing is seemingly pulling a relay out of its ass 06:23 < swine_> it insists on using mx0a-001c2c01.pphosted.com as the mail relay 06:23 * aBound sends [R] sendmail :P 06:23 < swine_> no idea where on earth this is coming from. i don't have mx0a-001c2c01.pphosted.com specified anywhere 06:24 < jim> supernov1h, I'd like to be able to help with this, and, I don't see any way to do it, because you've blocked out everything he might need by this chrooting 06:24 < supernov1h> okay what if he has access to echo 06:24 < supernov1h> and not bash echo 06:25 < supernov1h> actually no, he needs bash but I've already got that in the root with the linked libraries 06:25 < jim> swine_, it's possible your sendmail is acting like an open relay (pot o' gold for spammers) 06:26 < aBound> You're eating my Lucky Charms. 06:26 < jim> your ISP might be within their rights to take your connection down 06:26 < domhnall> swine_: and that really sucks being stuck on centos5. 06:26 * aBound leprechaun 06:26 < domhnall> heh...'stuck' 06:26 < jim> at a bare minimum of what legal remedies are available :) 06:26 * domhnall runs back t reading 06:27 < aBound> Noooo more reading. :P 06:27 < sn00bie> why can i overwrite an prg/file that is running as systemd process? if i start it with ./ i cant overwrite the prg / file. is the prg complet in ram with systemd? 06:28 < aBound> CentOS 5 is still supported? 06:28 < domhnall> no 06:28 < [R]> sn00bie: "prg/file"? 06:28 < jim> supernov1h, the only thing I can see would be to create a new dir structure and bind mount stuff in it, including your user's home dir 06:29 < aBound> I was about to say, given 7 is out. 06:29 < domhnall> that's a double bad...centos 5 AND sendmail. 06:30 * aBound cries 06:30 * domhnall thinks it'd be a nice setup for a lab target machine though... 06:31 < domhnall> swine_: but ignore my banter, im just typing because I can. 06:31 < jim> sendmail basically lets you do anything,,, so you would have to learn how to let the things you want to have happen, and deny whatever else 06:31 < aBound> Sounds like a firewall. 06:34 < jim> aBound, pretty much, yeah 06:34 < aBound> jim: Is it scriptable too? :P 06:35 < aBound> Blah, I can look it up. Teehee. 06:35 < jim> not sure... maybe yes 06:36 < aBound> Figures, as long as it's got commands more or less you should be able to write a bash file. 06:36 < aBound> :P 06:42 < nekoseam> How many of you use the fish shell? 06:43 < domhnall> newb questions, does ConsoleKit have a strong dependency on dbus? 06:44 < m0nk3y_b0y> what is the best linux distro / desktop environment to run on a personal laptop that is lightweight? 06:44 < [R]> domhnall: versus a weak dependency? 06:44 < domhnall> okay, m0nk3y_b0y beats my newb question. 06:44 < [R]> m0nk3y_b0y: whicchever one you like 06:44 < m0nk3y_b0y> lightweight linux distro / desktop environment 06:44 < domhnall> [R]: I dont know... 06:45 < epicmetal> m0nk3y_b0y: the heaviest is GNOME and anything based on it 06:45 < nekoseam> m0nk3y_b0y: Try BunsenLabs 06:45 < [R]> domhnall: what? 06:45 < Triffid_Hunter> m0nk3y_b0y: the one you enjoy the most.. personally I like gentoo with kde :P 06:45 < m0nk3y_b0y> epicmetal: what is a very light functionala desktop environment 06:45 < aBound> m0nk3y_b0y: That depends if you want a user friendly distro use: Lubuntu or Xubuntu. 06:45 < nekoseam> It's Debian + Openbox. Openbox is lighter than any DE 06:45 < Triffid_Hunter> m0nk3y_b0y: xfce, lxde have various balances of light and functional 06:45 < epicmetal> m0nk3y_b0y: I recently switched from GNOME to MATE and I like it. GNOME is the only one that does disk indexing properly though (or at all) 06:46 < nekoseam> I generally steer people away from Ubuntu since it spies on the user these days 06:46 < Triffid_Hunter> nekoseam: that's because it's a wm, not a de.. may as well use twm :P 06:46 < m0nk3y_b0y> back in the day I used Debian + Windowmaker 06:46 < nekoseam> Triffid_Hunter: That's like saying may as well use ed instead of neovim 06:46 < nekoseam> Openbox has GUI configuration which makes it convinient and user friendly 06:47 < aBound> Might as well use nano over ed. :P 06:47 < nekoseam> The only other window manager that has GUI configuration is Fluxbox 06:47 < epicmetal> m0nk3y_b0y: MATE specifically has slow computers in mind as part of their mission statement, which is nice 06:47 < m0nk3y_b0y> aBound: what makes those more friendly than Debian? 06:47 < nekoseam> aBound: Nothing. Except maybe default sudoers and certain media codecs installed 06:47 < aBound> m0nk3y_b0y: Install it and pretty much don't have to worry about it. :) 06:47 < nekoseam> and proprietary software of course... 06:48 < nekoseam> Debian was my first distro though and I never had any trouble learning it 06:48 < epicmetal> m0nk3y_b0y: as for distro, I always come back to Arch 06:48 < jmadero> hi all - I've got a bash script that is supposed to add a date to the beginning of a music title and instead it's replacing the title with the date: https://pastebin.com/MfccW6t3 06:49 < aBound> nekoseam: I can't say how much the open source drivers are if you use Nvidia or AMD. But if you're using Intel then I can see the reason for Debian. :P 06:50 < aBound> m0nk3y_b0y: It's just Ubuntu with lightweight desktop environments. 06:52 < m0nk3y_b0y> aBound: but why Ubuntut over Debian 06:52 < nekoseam> No reason 06:53 < Evidlo> should DISPLAY be set for prorgrams invoked by dmenu? 06:53 < nekoseam> Don't. Canonical can't be trusted these days 06:54 < domhnall> ah nevermind...missing posix_fallocate() support. 06:54 < aBound> m0nk3y_b0y: Ubuntu generally uses newer software over Debian unless there's backports. Configuring Debian would take a bit more time as nekoseam pointed out Ubuntu will include extra software like restricted extras, codecs and the likes. 06:54 < nekoseam> Robert Miller has passed away 06:54 < nekoseam> https://www.linuxjournal.com/content/rip-robin-roblimo-miller 06:56 < aBound> There isn't much difference, except Debian does have stable, testing and the other edition. That I can't seem to remember. :P 06:57 < NextHendrix> unstable 06:57 < domhnall> nekoseam: name sounds familiar, He was dev or something? 06:57 < aBound> Yesss unstable that's it. 06:58 < m0nk3y_b0y> aBound: is Lubuntu and Xubuntu the sameas Ubuntu? whats the difference 06:59 < aBound> m0nk3y_b0y: The desktop environments are different. Whereas Xubuntu uses Xfce by default Lubuntu will use lxde. 06:59 < m0nk3y_b0y> aBound: but they are Ubuntu right? 07:00 < aBound> m0nk3y_b0y: They're indeed Ubuntu, just using different DEs. As Kubuntu uses KDE. 07:01 < m0nk3y_b0y> aBound: ok thanks 07:02 < aBound> However, Kubuntu may or may not support "snaps" by default. But generally from my standpoint snaps aren't stable. 07:05 < aBound> m0nk3y_b0y: If your intention is to use Lubuntu or Xubuntu stick to LTS releases primarily point releases. :) 07:13 < jim> or use plain vanilla debian, then you have unrestricted access to any of them 07:14 < notmike> Ew 07:14 < xShyPotato> hello anyone online, I would like some clarifications on the blockchain technology 07:14 < jim> hi 07:14 < aBound> Hi. 07:15 < jim> oh, I don'/t know anything about blockchain 07:15 < xShyPotato> ahah okay, thanks anywya :) 07:15 < aBound> It's the cute_korean_girl oh snaps. :P 07:15 < jim> I know some folks who went to a seminar 07:15 < jim> xShyPotato, don't let me stop you from asking though 07:16 < aBound> Teehee. 07:16 < xShyPotato> is blockchain a p2p technology? so different people around the world host a blockchain server? 07:17 < Triffid_Hunter> xShyPotato: any particular one? git uses blockchain for example although I believe it predates the buzzword 07:17 < domhnall> jim: there's talks of implementing blockchain in US voting polls. Seems like a promising tool to get to know. 07:18 < aBound> https://www.google.com/url?q=https://en.m.wikipedia.org/wiki/Blockchain&sa=U&ved=0ahUKEwjv3OeejqDbAhWlKX0KHaqqBcsQFgggMAE&usg=AOvVaw1odt60O_Ijwg5LlaTq3N-b 07:18 < aBound> Oops. 07:18 < xShyPotato> none in particular, im studying blockchain in Big Data and I just don't really understand some of its technologies 07:18 < aBound> https://en.m.wikipedia.org/wiki/Blockchain?wprov=sfla1 07:18 < xShyPotato> thus im asking here hah 07:19 < Triffid_Hunter> xShyPotato: well blockchain simply refers to each chunk of data depending on a previous one using a hashing function, there's no servers or network protocol or anything involved 07:19 < aBound> No servers say wha! 07:19 < Triffid_Hunter> xShyPotato: many specific applications certainly add network protocols and such to it 07:21 < xShyPotato> Okay thanks a lo t:) 07:21 < aBound> Information overload. 07:22 * aBound brain slush 07:22 < domhnall> xShyPotato: in short though, it is an infrastructure, a bit complex/abstract to describe. 07:22 < xShyPotato> Got it! Thanks everyone 07:22 < nekoseam> Ubuntu (and derrivatives) are Debian with more media codecs, a different release model and proprietary software repos enabled by default 07:22 < nekoseam> It uses a bit more memory than Debian 07:23 < domhnall> xShyPotato: but you're right though...yes it is p2p system. 07:23 < aBound> I'm off to Wonderland. :P 07:24 < oiaohm> nekoseam: I would not say more media codeces. 07:25 < oiaohm> nekoseam: I have not as yet found a codec that is ubuntu that is not in debian 07:25 < oiaohm> nekoseam: ok you might have to enable debian non free but it there. 07:27 < nekoseam> oiaohm: not sure which codec is required for twitch to play but debian doesn't include them 07:30 < oiaohm> nekoseam: https://wiki.debian.org/MultimediaCodecs I guess you went down this page. 07:31 < oiaohm> nekoseam: from memory twitch was libavcodec-extra package install. 07:32 < oiaohm> nekoseam: ubuntu installs more codecs by default. 07:35 < nekoseam> oiaohm: that's what I mean 07:39 < swine_> so i managed to get postfix on this thing 07:39 < swine_> but for some reason it's sending ESTMP AUTH PLAIN requests even though the server wants AUTH LOGIN 07:40 < Triffid_Hunter> swine_: check the smart relay settings, should be sometihng in there for auth.. although usually postfix is used for delivering directly to the destination rather than going through a relay 07:41 < swine_> what's the best option for reliably relaying mail? 07:41 < swine_> ssmtp barfs too 07:47 < jim> swine_, reliably and open? 07:53 < Triffid_Hunter> swine_: ssmtp or msmtp work great for me, just have to set your login details properly 07:53 < Triffid_Hunter> I have postfix on my VPS handling mail for a few domains 07:56 < swine_> i'm able to login via telnet 07:56 < swine_> but these relays are not behaving in a way that sendgrid likes 07:56 < swine_> specifically sendgrid seems to want AUTH LOGIN 07:57 < nai> hi, is there a way to make the kernel print more error messages when it can't run /init? i'm trying to run a *very* minimal linux system in QEMU using a custom initramfs, and i just get a kernel panic "Requested init /init failed (error -2)" 07:58 < nai> i'd like to know if it's because it can't find /init at all, or it can't find a needed library, or something else 07:58 < Triffid_Hunter> nai: what's in your /init? is it an ash script or dynamically linked binary with missing libraries or what? 07:59 < Triffid_Hunter> nai: I've made initramfs by hand a few times, used busybox ash to handle my /init, kernel seemed quite happy with that. my busybox was dynamically linked to uclibc because I wanted to add a few extra things to the initramfs and didn't see the point in having everything statically linked against the same libc 07:59 < nai> Triffid_Hunter: i've made multiple tests, one with init=/bin/bash (after having copied the needed libs in /lib), one with a #!/bin/bash script, and now i'm trying it with a custom init program that i compiled using my custom libc (the very minimal kind, only handles system calls) 08:00 < Triffid_Hunter> nai: and did init=/bin/bash work? 08:00 < nai> Triffid_Hunter: no, none of my tests worked. most failed with error -2, one test failed with error -13 but i could not isolate the cause 08:00 < nai> i can't find documentation on these errors 08:00 < nai> or on how to make the kernel actually tell me why it failed 08:01 < Triffid_Hunter> nai: well I'd add a static busybox and use that for debugging.. boot with init=/bin/ash (add suitable symlinks) then poke at bash et al 08:01 < Triffid_Hunter> nai: or chroot into your image and play with it like that 08:01 < nai> Triffid_Hunter: that's a good idea 08:01 < Triffid_Hunter> nai: perhaps your initramfs is missing /dev/console and /dev/null ? those two nodes usually need to be static in the image, busybox mdev can make the rest dynamically during boot 08:03 < sn00bie> why can i overwrite an prg (executable file) that is running as systemd process? if i start it with ./ i cant overwrite the executabke file. is the prg complet in ram with systemd? 08:03 < nai> indeed, it's missing them since i haven't done anything at all to make them exist. i thought this was the job of init, but if init can't execute, how are we supposed to create these nodes? 08:03 < notmike> Og Bobby Johnson 08:03 < sn00bie> i hope that clearfy my question a little bit 08:04 < Triffid_Hunter> sn00bie: yeah you can overwrite files anytime you like in linux. the old one stays on the disk as long as any inode points to it or any process has an open filehandle to it, and is deleted when the last filehandle is closed and last link is removed 08:04 < nai> Triffid_Hunter: and now i'm feeling dumb but even chroot doesn't work. i'm doing "sudo chroot . /init" from within my initramfs dir, and i get "chroot: failed to run command ‘/init’: No such file or directory" 08:05 < nai> init is there and executable 08:05 < sn00bie> Triffid_Hunter okay, 2 examples. 1) if systemd start the programm i cant overwrite the file at any time. 2) if i start it from bash with ./ it isnt possible to overwrite the file 08:05 < kurahaupo_> nai: is it statically linked? 08:06 < kurahaupo_> sn00bie: there are three possible scenarios that may apply 08:06 < nai> kurahaupo_: no 08:06 < nai> but the only needed lib is in lib/ 08:07 < kurahaupo_> sn00bie: the simplest is that if you remove the file and create a new file with the same name, it can't affect any running process 08:07 < Triffid_Hunter> nai: sounds like a missing shared lib 08:08 < kurahaupo_> sn00bie: the second is that the loader makes an in-memory copy of the file when it starts; again, altering the file subsequently can't affect the running process 08:08 < Triffid_Hunter> sn00bie: no idea about systemd, don't use it 08:08 < Jonno_FTW> is there a channel for linux networking questions or is ehre is here fine? 08:09 < kurahaupo_> sn00bie: the third case is where the file is mmapped into the memory of the running process, in which case subsequent changes to the file will affect the memory of the process. Or such changes may be forbidden. 08:09 < nai> alright, i'll try to investigate that. no idea about my initial question though? is there no way to make the kernel more verbose about it? 08:10 < nai> i've tried adding "loglevel=7 debug" to the kernel's command line 08:10 < kurahaupo_> nai: run ldd init to see what shlibs it needs 08:10 < sn00bie> Triffid_Hunter everything is working, the only thing is i am surprise about that. x 08:11 < nai> kurahaupo_: i've done that, and as i said the only needed lib is in lib/ 08:11 < nai> i'll try to make a static executable 08:11 < kurahaupo_> sn00bie: which one applies depends on factors including whether the owner of the file is the same as the owner of the process 08:12 < kurahaupo_> nai: if it's not static, it will need ld-linux.so, even if ldd doesn't say so 08:14 < kurahaupo_> nai: you might like to try « strace chroot . /init » 08:15 < sn00bie> my program is starting at boot time with systemd, the program can update by himself he get the file from an server, exit the program, and systemd starts the program again and now the new prg is running. => i get suprised that the programm can overwrite by himself, and asked me why? if i start the program from bash it doenst work. 08:15 < nai> kurahaupo_: it explicitly shows that execve("/init") fails with "No such file or directory" 08:16 < nai> the "chroot" and "chdir" system calls succeed 08:18 < nai> i feel like my initramfs folder is simply missing some basic component that makes linux unable to do anything in it 08:39 < Dagmar> Being a folder could be a problem 08:40 < kurahaupo_> sn00bie: it's almost certainly the first case I mentioned 08:42 < nai> Dagmar: i archive it properly before giving it to QEMU, but the chroot should work 08:44 < kurahaupo_> nai: an ELF binary needs a loader much the same way that a script needs an interpreter, and at the same point in the execve. In the case of ELF, it's ld-linux.so (or similar - might vary with different distros) 08:44 < first-order> What exactly in the hell can one do with ancient hardware running DSL or TinyCore, vs. modern hardware running Debian, Fedora, or Arch, anyways? 08:45 < nai> kurahaupo_: i also copied my system's ld-linux.so to the lib folder 08:45 < first-order> Pretty much, hardware that might be better off running DOS, such as my currently-nonfunctional Solo 2100. 08:45 < nai> kurahaupo_: but in that case the "no such file or directory" should happen for ld-linux.so, not for init, right? 08:47 < kurahaupo_> nai: the program invoking your init wouldn't see the indirect call to ld-linux.so, it would just see ENOENT 08:47 < nai> i see 08:47 < kurahaupo_> I'm now wondering whether the fs is noexec 08:48 < kurahaupo_> Or SELinux is messing with you 08:48 < nai> just a subfolder in my normal fs 08:49 < nai> maybe just copying my ld-linux.so is not enough 08:53 < nai> YES! got it 08:53 < nai> it had to be in /lib64 08:53 < qrvpzvb> Say that I have two disks in an mdadm mirror. I then buy two more disks. How could I add those? From what I've read, if I had used level=10 at the start, you just can't add any new raid-devices, just replace. Otherwise, I'd have to create a standard level=1 array with the new disks and then a level=0 on top of those two mirrors; that'd mean that I'd have to create a new fs on top of that though. 08:53 < qrvpzvb> Am I correct? Are there other ways? 08:54 < nai> it actually had to match exactly the filename in the INTERP segment of my executable. well, guess my problem is solved for now, thanks everyone for the help, i'll come back later maybe 09:06 < Dagmar> qrvpzvb: It depends entirely on what you're trying to accomplish. 09:07 < qrvpzvb> Dagmar: I just said, I want to add two more disks and get more storage. 09:08 < Dagmar> You are perfectly allowed to make a second pair of mirrored disk and span the overlaid volume using LVM 09:08 < Dagmar> Things are generally much simpler if you just use mdadm for things which are actually RAID 09:08 < Kharma> Anyone in here self-host domainmod(.org) or something similar? 09:08 < Dagmar> i.e., mirrored disks, parity bits, etc 09:09 < qrvpzvb> Dagmar: so you're saying that there's no way to do that without LVM? 09:09 < Dagmar> qrvpzvb: Considering that you're talking about changing a mirrored pair to something else? 09:09 < Dagmar> Yes, you should be using LVM. 09:10 < Dagmar> ...because it's easier. 09:10 < jim> Dagmar, what is "span" in that context? 09:10 < Dagmar> jim: The same thing "span" means everywhere else. A filesystem that extends beyond a single disk slice. 09:10 < qrvpzvb> Ok; thanks! I guess LVM is unavoidable after all 09:11 < Dagmar> qrvpzvb: If you haven't already blessed part of the original pair of disks as a PV you have *no choice* but to backup your data and reformat 09:13 < Dagmar> If anything, LVM is a heck of a lot easier to use than mdadm. 09:14 < Dagmar> ...but its job is not to present hardware abstractions. Its job is to present storage abstractions. 09:14 < qrvpzvb> no it's ok I don't have any data; I'm just trying to understand the technologies before commiting 09:14 < jim> it's also been around for a couple decades 09:15 < Dagmar> Okay well get this straight right now then... LVM is for storage abstractions. mdadm _can_ do those, but it's main utility is in its ability to provide an abstraction layer for disks, as hardware 09:15 < qrvpzvb> LVM can't do raid though...? 09:15 < Dagmar> LVM can do RAID 09:15 < jim> you can do a real, hardware raid, then format the raid with an lvm pv 09:16 < Dagmar> ...but as far as I know it's mainly leveraging mdadm for that 09:16 < Dagmar> LVM is actually far easier to deal with than I suspect you realize 09:16 < Dagmar> ...but it's not without reason that most people generally do what I'm describing here. Using mdadm for hardware-level abstractions, and LVM for storage abstractions. 09:16 < jim> (or you can do the same thing with an mdadm raid, it's just less reliable) 09:17 < Dagmar> Also, despite what it might look like, LVM doesn't introduce any major overhead costs for its abstractions 09:18 < qrvpzvb> I don't worry too much about that 09:19 < qrvpzvb> I just want to minimize the different concepts and utilities I have to use 09:19 < Dagmar> Well then let's get you acquainted with the three layers of LVM so you can have an idea of what you're in for 09:20 < Dagmar> Lemme see how much juice my laptop has in it real quick 09:20 < qrvpzvb> sorry about that 09:20 < Dagmar> No need to be sorry 09:21 < Dagmar> You're asking perfectly reasonable questions 09:21 < Dagmar> You've got a reasonable use-case. 09:21 < Dagmar> You've just missed something (LVM) that could really help you 09:21 < Dagmar> ...and I want a cigarette and this will be easier to explain if I can use the laptop outside. :) 09:22 < qrvpzvb> heh 09:22 < Dagmar> Hmm... laptop claims it's fully charged despite it sitting for three weeks unplugged. We'll see 09:23 < notmike> It lies 09:25 < Dagmar> Well, it's a high-quality battery so it might be accurate. I just suspect the 100% it's claiming is the new 100% which is probably 96% from a month ago 09:25 < Dagmar> *shrug* 09:25 < Dagmar> qrvpzvb: Anyway, you do understand classic MBR/GPT partitioning, right? 09:27 < qrvpzvb> sure 09:27 < Dagmar> Mainly that primarily they're a guideline by which the kernel simply translates absolute sector accesses of block storage through a set of offsets 09:27 < Dagmar> Okay, good. 09:27 < Dagmar> LVM makes that pretty much obsolete. 09:28 < Dagmar> The lowest of the three layers in LVM is the Physical Volume abstraction. 09:28 < qrvpzvb> disks 09:28 < qrvpzvb> or other block devices 09:28 < qrvpzvb> I would assume 09:28 < Dagmar> This doesn't always mean an entire _physical_ volume, it's just where they come into play 09:28 < Dagmar> yeah 09:29 < Dagmar> LVM doesn't define things like partition tables do. Notably it doesn't have a table of slices with a handy flag that makes claims about what's in that slice 09:30 < Dagmar> What it does is writes its metadata to a sector or two about that slice, so the first step with physical volumes can happen on an entire disk, or just on a partition (because it'll happily read the block storage /dev/sdc1 just as it would /dev/sdc) 09:31 < qrvpzvb> which helps so that other OSes don't mess up your data thinking your drives are empty 09:31 < Dagmar> MANY people start by going ahead and declaring an actual partition (even if it's the only partition on a disk) and blessing that as a pv 09:31 < Dagmar> Yeah 09:31 < Dagmar> It's not necessary, but people do it 09:32 < Dagmar> People especially do it in say, a enterprise SAN environment because SAN managers are slightly prone to looking at slices that have been allocated and deciding they're unused and can be reclaimed/purged if they don't see a partition table 09:32 < Dagmar> *sourface* 09:32 < qrvpzvb> heh 09:32 < Dagmar> But anyway, the only necessary step to blessing something as a physical volume (a "pv") is the command `pvcreate /dev/whatevername` 09:33 < Dagmar> LVM will write the metadata declaring that storage to be a pv in there and it's done. The next time you boot, when LVM scans around looking for PVs it'll see that 09:34 < Dagmar> Note that a PV is almost equivalent to a partition in this respect. It can't span across multiple devices, but in practice this is not a problem because of the next layer 09:34 * Kharma ... takes note... >_> 09:35 < qrvpzvb> oh, so I don't even need an lvm.conf or anything; cool 09:35 < Dagmar> Volume Groups (aka VGs). You create a volume group using the command vgcreate, and the only things you really *need* to specify are the device names of the physical volumes that will be members of that group, like `vgcreate /dev/sda2 /dev/sdb /dev/sdc` 09:35 < Dagmar> Nah, you *used* to but that's not been necessary for awhile 09:36 < Dagmar> At one point some time ago I actually backed that file up and attempted to restore it to a new system with the same disks in a slightly different configuration and things went pear-shaped with a quickness. heh 09:36 < Dagmar> Anyway, once PVs are collected into a Volume Group LVM now has a large swath of blocks to play with that it organizes into "extents" which are roughly equivalent to blocks of storage 09:37 < Dagmar> Oh! Volume groups *do* get a name, just like LVs. 09:37 < Dagmar> I almost forgot that bit 09:38 < Dagmar> Generally people are very unimaginative about their volume group names, doing things like "vg01" and "vg02" 09:38 < Dagmar> You can use more descriptive names, but it'll result in longer device names 09:39 < Dagmar> Those names will figure into the device name you'll ultimately use to access the logical volumes, as /dev/mapper/vg00-* and so on 09:39 < Dagmar> I should amend that psudocode invocation of vgcreate to be "vgcreate vg00 /dev/sda1 /dev/sdb /dev/sdc" 09:40 < swine_> hey Dagmar 09:40 < Dagmar> Somewhat obviously volume groups can span multiple pvs 09:40 < Dagmar> sup 09:41 < Dagmar> Anyway, once you have created a volume group, you're at the point where allocating slices gets *really* easy 09:41 < swine_> pmed you. i'm surprised you're still around 09:41 < Dagmar> Like, way easier 09:41 < Dagmar> swine_: I am eternal. 09:41 < swine_> i was just reading the upanishads 09:41 < qrvpzvb> so, why would you need more than one vg? 09:41 < swine_> seems appropriate 09:42 < Dagmar> Generally it's purpose-driven 09:42 < Dagmar> For example, here I have a pretty massive setup that I am slowly transitioning to a fully enterprise-level thing with storage handled by it's own dedicated box 09:42 < Penguin> Dagmar: "its" 09:43 < Dagmar> ...but in the meantime, the VM host I have has one VG for the disk slices the VM host itself uses, and a second VG for all the disks that are basically just "massive storage space" 09:43 < Dagmar> Later I'll be able to transplant that entire second VG off the machine and onto the dedicated SAN box 09:44 < Dagmar> Then all those various mounts will just get replaced with NFS mounts or possibly iSCSI. I haven't decided hyet 09:44 < Dagmar> Once you've got your Volume Group established, creating a logical volume is also pretty easy... 09:44 < qrvpzvb> nice 09:45 < Dagmar> `lvcreate -L 4G -n isostorage vg00`` 09:46 < Dagmar> -L is --size, and it can be specified a couple of different cool ways 09:46 < Dagmar> That one's obviously a 4 gigabyte slice, but you can actually say -L 50%FREE 09:46 < Dagmar> ...or 100%FREE comes up a lot for folks 09:47 < Dagmar> That lvcreate command results in the creation of /dev/mapper/vg00-isostorage` 09:47 < Dagmar> ...which you can then run mke2fs on to format, mount it, and just go 09:47 < qrvpzvb> nice 09:47 < Dagmar> bleh mkfs. Old habits die hard 09:48 < Dagmar> There is a LOT in LVM that was designed around the concept of orthagonality and "reasonable assumptions" 09:48 < Dagmar> For example, the various scanning commands have longer and shorter forms, which result in more verbose or less verbose output 09:48 < Dagmar> You can get a list of physical volumes using `pvs` or `pvscan` 09:49 < Dagmar> You can get a list of volume groups using `vgs` or `vgscan` 09:49 < Dagmar> ...and you can get a list of logical volumes using `lvs` or `lvscan` 09:49 < Dagmar> Note that one of those reasonable assumptions pertains to the size of a logical volume 09:50 < Dagmar> If you don't specify a size, lvcreate will reasonable assume you just want to use all the space remaining (for example) 09:50 < learningc> Can I use hex numbers in for loop? 09:50 < Dagmar> s/reasonable/reasonably/ 09:50 < qrvpzvb> make sense 09:50 < qrvpzvb> learningc: yes 09:51 < qrvpzvb> learningc: or rather, what are you talking about? 09:51 < learningc> for number in {0x74200040..0x7420007c..0x4} does not work 09:51 < Dagmar> learningc: Not in a way that will treat them as numbers. If you pass a list of hex numbers to iterate over it'll be treated just like any other list of strings 09:51 < Dagmar> for number in "0x00 0x30 0xFF"; do blah done will run three iterations 09:51 < learningc> I want to pass these hex numbers to an application 09:52 < Ulrar> Hi, dumb question, what makes pam accept a login for a given user or not ? I'm trying to "convert" a system user into a regular user, so I gave it a password, a valid shell and changed it's uid and gid to be > 1000, but slim still says unknown user 09:52 < qrvpzvb> I don't think that bash has any support for hex numbers 09:52 < qrvpzvb> various utilities (like printf) do 09:52 < Dagmar> Bash is really only looking for whitespace-separated tokens there 09:52 < learningc> How can I get these numbers to iterate without listing them all? 09:53 < Dagmar> qrvpzvb: Oh, and before this can come up... lvm can do online resizing (only grow operations!) online safely 09:53 < Dagmar> learningc: Do they follow a specific sequence? 09:53 < Dagmar> "sequence" being the key word here 09:53 < qrvpzvb> oh... not shrinking; unfortunate 09:54 < Dagmar> You've got a `seq` command that can take a start number and an ending number and an optional increment value 09:54 < learningc> Dagmar, these numbers are register addresses, so I want to dump a range of them in 4 byte 09:54 < qrvpzvb> he did say from 0x742... to whatever in steps of 0x4 09:54 < Dagmar> qrvpzvb: Well, most filesystems don't support online shrink operations 09:54 < Dagmar> learningc: You can use seq to generate the list as a bunch of decimal numbers and then convert them to hex with printf inside the loop 09:55 < qrvpzvb> so you could use seq and then printf for the hexing 09:55 < qrvpzvb> yup 09:55 < Dagmar> qrvpzvb: The online resizing thing comes up often here, and it's safe so long as the filesystem isn't already damaged 09:55 < Dagmar> Mainly LVM politely wrappers other tools to do the actual resizing of the filesystem 09:56 < Dagmar> You can use lvextend to increase the size of an LV (once again it'll assume 100%FREE but you can tell it how large to go) and if you pass it -r it'll go ahead and run the command to resize the filesystem for you 09:57 < nai> for (( i = 0; i < n; i++ )) printf -v x %x "$i"; echo "$x"; done 09:57 < Dagmar> I used to be paranoid about it, but after about 30-40 needless downtime windows where I made backups and then crossed my fingers and did the resize online, then unmounted it and ran fsck to check, I got over the willies 09:58 < nai> seq is nonstandard and can be unpredictable 09:58 < Dagmar> If the filesystem is in some way knackered, the step where it calls for the resize will abort and complain at you 09:58 < qrvpzvb> thanks for the primer Dagmar; guess it's time to hit the manpages! 09:58 < Dagmar> nai: Well, it's "standard enough" on Linux machines 09:58 < qrvpzvb> c-style for loops are bash only as well 09:58 < qrvpzvb> (I think) 09:58 < Dagmar> It comes from GNU coreutils and I've not noticed any Linux distros using some other version 09:59 < notmike> Omg do you bros ever stop talking about Linux? 09:59 < Dagmar> I will say I CHECK that carefully before trying it on some weird thing like AIX 09:59 < Dagmar> motmike: You wandered into the wrong neighborhood, bruv 09:59 < nai> oops, forgot a "do" in there 09:59 < notmike> <3 10:00 < Dagmar> qrvpzvb: one last bit about LVM. The RAID parts are still "kinda" new, but that definition of "new" has a lot longer view than most people take... i.e., it's less than a decade old 10:01 < Dagmar> It's _probably_ past the point where if there were going to be any fs-corrupting bugs we'd have found them by now 10:02 < Dagmar> ...but since the people using RAID tend to be doing setups that are expected to be around for a *very* long time, people are still mainly using mdadm directly for RAID stuff and LVM for the rest 10:02 < Dagmar> ...because they originally set things up that way 10:03 < qrvpzvb> thanks for the heads up! 10:05 < Dagmar> The main utility I think is that perhaps the mdadm documentation is a bit more mature and plentiful 10:05 < Dagmar> ...and documentation is _necessary_ to avoid mangling terabytes of data at once 10:05 < Dagmar> ;) 10:06 < Dagmar> Ther'es also that people doing RAID specifically have reason to be paranoid about possible software bugs or RAM failures introducing bit-rot and or corruption 10:07 < Dagmar> ...and at that level PERC controllers are kind of easy to come by 10:07 < Dagmar> Vendors selling hardware with PERC controllers also make pretty promises about them, but the important thing is that they can be counted on to not magically change their metadata format if you have to buy a new one 10:08 < Dagmar> This used to NOT be the case (15-20 years ago) 10:08 < Dagmar> People would get Dwarf Fortress levels of "fun" by buying a new RAID controller only to find out that their old RAID controller kept its metadata about the array layout in a format the new controller has no idea about 10:09 < Dagmar> Technical lock-in like that is one of the reasons mdadm came about 10:10 < Dagmar> Granted, mdadm does have some additional risks if you're not using ECC RAM, and while those risks are *small* the resulting corruption still means it generally makes sense to use the PERC controller 10:11 < Dagmar> ...if for no other reason that if the controller develops a RAM problem and starts corrupting the array, you can just blame the vendor 10:11 < Dagmar> IF you decide to do virtual machines, not using LVM is basically inexcusable 10:12 < Dagmar> KVM and virt-manager are quite happy with LVM. You can pretty much hand virt-manager a volume group and create virtual disks right in the interface, which are allocated as individual logical volumes 10:13 < Dagmar> If push comes to shove, you can use dd to copy the underlying logical volume, image it straight to a _real_ disk and go put that in a real machine 10:13 < oiaohm> Dagmar: its possible to design ram controllers that prevent defective ram issues. Its required for satellites with the high radiation flipping bits. So a RAM problem truly should be blame vendor for poor design. 10:13 < Dagmar> oiaohm: Generally ECC takes care of it 10:13 < oiaohm> Dagmar: there is a level above ECC 10:13 < Dagmar> One bit gets flipped, ECC notices, corrects it, and reports it 10:13 < Dagmar> yeah but we're not building satellites here 10:14 < Dagmar> I've seen personally the kind of engineering that goes into those 10:14 < Dagmar> I am impressed 10:14 < oiaohm> Dagmar: there are rare raids that use dual ram 10:14 < oiaohm> Dagmar: ie mirror raid in the ram. 10:14 < Dagmar> The ISP I used to work for had a POP in the same building as the guys who were building the Mars rover 10:14 < Dagmar> Those were some very cool nerds to talk to 10:15 < oiaohm> Talking to those guys you learn what true durablity in computing looks like. 10:15 < Dagmar> oiaohm: I've also had to replace disks on some machines that went so far as to use triply-redundant controllers 10:15 < Dagmar> Oh gods yes 10:15 < learningc> ok, it seems that I can use hex with a regular for loop in c style: for ((number=0xa;number < 0xf;number++)) 10:16 < Dagmar> learningc: I'd be surprised if that works but if it works, great 10:16 < sauvin> In bash!? 10:16 < Dagmar> I haven't looked into how far one can push the bash code that makes C-style loops work 10:17 < Dagmar> I tend to stick with what I *know* works and why when I'm writing bash scripts so I can reduce the amount of unit testing I need to do 10:18 < learningc> but the only problem now is that the number is being converted to decimal, but I need it in hex 10:19 < Dagmar> Convert it back to hex using printf 10:19 < Dagmar> (I just checked and it does indeed allow hex) 10:19 < Dagmar> (Handy to know) 10:21 < nekoseam> hewwo frens OwO 10:22 < Dagmar> oiaohm: I remember very distinctly chuckling at one of the guys for saying a bit of foil (not saying he was wrong) protects the CPU from cosmic rays 10:22 < Dagmar> This is definitely not a thing we have to worry about with home setups 10:23 < learningc> How do I convert it to hex? 10:24 < nai> learningc: i showed you earlier 10:24 < nai> printf %x "$n" 10:25 < learningc> I need to do something like this: for ((number=0xa;number < 0xf;number++)){ readreg $number} where number is in hex representation, 0xa 10:25 < Dagmar> "0x%02X" may be more to your liking 10:26 < learningc> Ok thanks 10:26 < Dagmar> Having low-memory addresses being shortened makes debugging ugly 10:27 < Mikato> hi 10:29 < geirha> for number in 0x{a..f}; do 10:32 < Dagmar> Now you see why we suggested using printf in the first place 10:32 < Mikato> i need help with PTP Precision Time Protocol, i got two linux machine, one will be Grandmaster (as source time) and the second will be a boundary clock (slave), i read that i supose use ptp4l for grand master as a master and for slave as a slave, but on linux i got two config files in /etc/sysconfig/ptp4l.conf and in /etc/ptp4l.conf, so where i supose write my new config and what i supose write to config files when i want to run software synchronisati 10:32 < Mikato> on betwen two linux? 10:33 < hojuruku> Why are some opensource "code of conduct" documents being used to enforce tolerance of pedophiles. How? It says you must respect all sexual preferences: https://github.com/ContributorCovenant/contributor_covenant/issues/538 And what is this? 10:33 < hojuruku> http://www.b4uact.org/toronto-star-is-pedophilia-a-sexual-orientation/ 10:34 < hojuruku> The code of conduct document for freedesktop is being enforced. You must love sex with children and those who do it off away freedesktop.org infrastructure or you will be pro-actively banned for not being at peace with the boylover.org / freedestkop.org community. 10:34 < hojuruku> Google Chrome's code of conduct's wording doesn't normalize sex with children. I think freedesktop has a problem here. 10:35 < sauvin> Pedophilia is off-topic here. 10:35 < hojuruku> https://chromium.googlesource.com/chromium/src/+/master/CODE_OF_CONDUCT.md 10:36 < hojuruku> sauvin: no opensource code of conduct documents are being used to say you need to love pedophilia or GTFO. I don't want to GTFO because I am an anti-pedophilia-normalization activist and a opensource developer. 10:36 < jelly> you'll probably want to address specific projects 10:36 < hojuruku> I'm trying to get the source amended to say sexual preference excluding pedophilia. 10:36 < sauvin> Sexual preference is irrelevant here. 10:37 < hojuruku> it isn't - you can get banned for a private message for disrespecting pedophiles, after being trolled by a pedo lover daniels. that's what the github link is about. 10:37 < learningc> Dagmar, so I can do this? readreg printf 0x%02X $numbers where readreg is binary executable 10:37 < shalok> How does `mdadm --assemble /dev/md0 --name foo` know which drives belong in the array? There are no /etc/mdadm* files on my host... 10:37 < sud0x3> hojuruku: paedophilia, is a psychiatric disorder, wake the fuck up and stop posting shite 10:38 < hojuruku> shalok: through signatures.. 0.90 and 1.0 versions of mdadm puts the signatures at different places on the disk 10:38 < jelly> hojuruku: your own language however is incendiary, and distracts from your goal; at this point I 80% think you're just trolling 10:38 < Armand> jelly: 100% 10:38 * jelly conservative 10:38 < shalok> hojuruku: But I didn't even specify any disks... Is it just searching every disk in /dev/*? 10:39 < sud0x3> hojuruku: How often have you been impacted by this? and what yhave you done to resolve other than posting on IRC which in itself is a minority 10:40 < hojuruku> sud0x3: I wish you were right but my dad was accused of the crime of disrespecting pedophiles and so was others. Hear my dad tallk about how he was prosecuted for disrespecting Mark Newton and Peter Truong sate media praised gay marriage activists here: (mckee burns on austlii.edu.au case law) http://www.truthnews.com.au/web/radio/story/cultural_marxism_101 You can see the sexual positions the state media praised gay marriage activists used on 10:40 < hojuruku> the russian infant in the edward de sears indictiment linked to on here: https://quadrant.org.au/opinion/qed/2013/07/the-story-the-abc-doesn-t-want-you-to-see/ So the government is saying you have to love sex with children. It's a crime to link to tinyurl.com/abcplugspedos most popular video on my youtube channel. Put "abcplugspedos" into www.caselaw.nsw.gov.au to see what i mean because it disrespects pedophiles mark newton and peter troung. 10:40 < learningc> geirha, How can I jump by 4 bytes? for number in 0x{aa..ff..4}; do ? 10:41 < geirha> learningc: that one's non-trivial, so use c-style for-loop and printf for that 10:46 < learningc> geirha, It does not "seem" to work. for number in 0x{8..20};do should give me 8 9 a b c d e f 10 etc 10:48 < iflema> c style 10:49 < well_laid_lawn> ↑ 10:49 < kurahaupo> learningc: {n..m} expands to decimal numbers in all contexts; otherwise how should it decide which base? 10:50 < geirha> learningc: brace expansion was usable for your initial, simplified example, not for these less trivial ones 10:51 < Hermes> Hey! I was wondering if anyone know of a tool kind of like: https://www.youtube.com/watch?v=hqDGWfZhitc this, which shows reads and writes to a given file in realtime. 10:51 < kurahaupo> learningc: if you actually want substantive arithmetic -- including non-decimals -- then use the arithmetic « for ((start;test;step)) do … done » 10:51 < geirha> It's possible, but fairly unreadable, e.g. for number in 0x{0,1}{{0..9},{a..f}} 0x20; do 10:52 < nai> for (( i = 8; i < 20; i++ )) do printf -v x '0x%02X' "$i"; echo "$x"; done 10:52 < kurahaupo> nai: s/20/32/ 10:53 < nai> indeed -- that's what they want 10:53 < jim> learningc, you doing this in bash? try doing something like: seq from to where from and to are the starting and ending points -in decimal-, and when you get the right start and end points, put it in a for loop, and convert each number to hex in the loop 10:55 < kurahaupo> jim: you're not seriously suggesting using the "seq" command? 10:57 < AWSCUBER> kurahaupo 10:57 < nai> as i said earlier, seq is nonstandard and can be unpredictable 10:57 < learningc> humm seq sound too complicated 10:57 < AWSCUBER> i could run my znc in my shell 10:57 < AWSCUBER> $ znc --makeconf 10:57 < AWSCUBER> -sh: 29: znc: not found 10:57 < AWSCUBER> $ ./znc --makeconf 10:57 < AWSCUBER> -sh: 30: ./znc: not found 10:57 < AWSCUBER> $ 10:58 < well_laid_lawn> the enter key isn't used for punctuation 10:58 < geirha> AWSCUBER: that means there are no znc command in PATH, and no znc command in the current directory 10:58 < AWSCUBER> so geirha 10:59 < AWSCUBER> :/ 10:59 < Marikus> AWSCUBER how did you install it? 11:00 < AWSCUBER> i wanna instal znc in my shell 11:00 < AWSCUBER> i alrandy dowload it 11:00 < AWSCUBER> but i canot setupit 11:00 < AWSCUBER> $ ls 11:00 < AWSCUBER> eggdrop znc-1.6.5 11:00 < AWSCUBER> $ 11:00 < well_laid_lawn> the enter key isn't used for punctuation 11:00 < AWSCUBER> the eggdrop run weel 11:00 < eahm> why 165? 170 is out 11:00 < AWSCUBER> but znc not yet 11:01 < geirha> why not install it via the package manager? 11:01 < kurahaupo> learningc: if you have shell scripting questions, like "for to do a loop using arithmetic", #bash generally has the most reliable answers. 11:02 < kurahaupo> AWSCUBER: you don't need to keep PMing me, I figured it was you from the identical question 11:02 < AWSCUBER> i a sorry 11:02 < AWSCUBER> how can setup thz znc in my shell 11:03 < supernovah> Can you mount --bind as read only 11:03 < nai> AWSCUBER: what distro are you running? 11:03 < nicholasBPM> I have a remote mysql account that i only can connect to from localhost, i have ssh access to that server.. would it be possible to make some kind of mapping so i can access the remote mysql account direct from my local computer 11:03 < AWSCUBER> Server Info.......: Intel(R) Xeon(R) CPU E3-1240 v5 @ 3.50GHz, 4 cores 11:03 < AWSCUBER> OS Info...........: Debian Linux 7 11:03 < AWSCUBER> Host Name.........: dubaichannel 11:04 < AWSCUBER> nai 11:04 < Marikus> nicholasBPM Maybe use a SSH Tunnel 11:04 < nai> AWSCUBER: try sudo apt-get install znc 11:04 < nicholasBPM> Marikus, thank you, i will google ssh tunnel 11:05 < AWSCUBER> $ sudo apt-get 11:05 < AWSCUBER> We trust you have received the usual lecture from the local System 11:05 < AWSCUBER> Administrator. It usually boils down to these three things: 11:05 < AWSCUBER> #1) Respect the privacy of others. 11:05 < AWSCUBER> #2) Think before you type. 11:05 < AWSCUBER> #3) With great power comes great responsibility. 11:05 < AWSCUBER> [sudo] password for Break: 11:05 < AWSCUBER> [sudo] password for Break: 11:05 < AWSCUBER> Break is not in the sudoers file. This incident will be reported. 11:05 < AWSCUBER> $ 11:06 < nicholasBPM> Marikus, that was what im looking for, thanks bro! 11:06 < well_laid_lawn> use a pastebbin dude 11:06 < AWSCUBER> so nai 11:06 < nai> AWSCUBER: don't paste multiple lines in here. use a paste service like pastebin 11:06 < jim> yep, [[pastebiN]] 11:06 < sud0x3> AWSCUBER: ask in #znc for a start, second there are full instructions on znc wiki https://wiki.znc.in/Installation 11:06 < nai> AWSCUBER: it seems that you don't have the right to install packages on that server 11:06 < jim> yep, you can pastebin the output of an arbitrary command by running "anArbitraryCommand | nc termbin.com 9999", and to include error messages, "anArbitraryCommand 2>&1 | nc termbin.com 9999" 11:07 < learningc> kurahaupo, ok, thanks for the info 11:09 < BluesKaj> 'Morning folks 11:09 < supernovah> To run a service as a chrooted user, do I have to specify a user in 50-udev.rules or something 11:13 < learningc> nai, what is printf -v x? 11:14 < nai> it outputs text directly in the variable "x" instead of the standard output. it's a bash feature 11:15 < Triffid_Hunter> so more like sprintf? 11:15 < nai> exactly 11:19 < learningc> I see 11:19 < learningc> I used myvar=$(printf ...) but I cannot get it to work 11:20 < learningc> http://termbin.com/yol2 11:20 < supernovah> is there a way I can mount -t proc proc /home/chrooted/proc, to hide other user's processes? 11:20 < learningc> Any idea why? 11:20 < nai> learningc: no spaces around =. move to #bash for shell questions. 11:21 < learningc> nai, ok, thanks 11:23 < learningc> nai, working :) 11:25 < supernovah> oh so remounting with ro works 11:25 < supernovah> mount -o remount,ro,hidepid=2 /home/chrooted/proc 11:30 < supernovah> can you make entries in fstab with $(getent passwd | cut -d: -f6) ? 11:32 < nai> supernovah: why not use ~user ? 11:35 < supernovah> nai: I just did umount /home/chrooted/proc, and sono after all my ssh sessions dropped 11:36 < supernovah> AndI can't access the server from the root account any more... 11:36 < nai> too bad, you lost the linux game. try again next week 11:37 < luna> Live release of OpenSUSE 15: https://streaming.media.ccc.de/osc18/main 11:38 < supernovah> why would unmounting a copy of proc cause that? ssh is still active, it just drops me every time I connect probably due to this: pam_loginuid(sshd:session): Cannot open /proc/self/loginuid: Read-only file system 11:40 < nai> since it was a bind mount and you modified the mount options, i'm guessing you also modified the mount options for /proc 11:41 < Neighbour> Does someone know if it is possible to specify a non-local identity in a running ssh-agent to use when ssh'ing to another machine? `man ssh_config` hints that this is possible with the options 'IdentityFile' and 'IdentitiesOnly' combined, but I'm having trouble getting it to work 11:44 < Tech_8> hi 11:45 < supernovah> nai: I ran `mount -o proc proc /home/chrooted/proc` 11:45 < supernovah> so it wasn't a bind as far as I know 11:47 < supernovah> before I got booted I saw `proc on /proc type proc (rw,noexec,nosuid,nodev)` and ` 11:47 < supernovah> `proc on /home/chrooted/proc type proc (ro,hidepid=2)` 11:48 < Tech_8> ... 11:56 < supernovah> is there a way to issue a reboot through ssh if it's failing due to /proc/self being read only? 11:58 < luna> Leap 15 is soon out online waiting for Ludwig Nussel to fix some last stuff 11:59 < Alexander-47u> are there legit tools that work like c2 ? 11:59 < Alexander-47u> :p 12:00 < Alexander-47u> like for normal business purposes, remote administration from a firewalled environment 12:02 < Alexander-47u> bbl gtg 12:10 < luna> http://download.opensuse.org/distribution/leap/15.0/iso/openSUSE-Leap-15.0-DVD-x86_64.iso 12:10 < Tech_8> hi 12:32 < supernovah> lol mxtoolbox removed their portscan option 12:33 < Armand> supernovah: nmap 12:37 < jnor> an app that can ring a bell/alarm/reminder at continuous intervals? 12:38 < jnor> two times every hour for example 12:41 < dianni> cron? 12:42 < supernovah> questions as statements? 12:42 < jnor> ayyy im silly thanks lol 12:42 < jnor> cron 13:02 < supernovah> hey so would `mount --bind /proc /home/chrooted/proc && mount -o remount,bind,ro,hidepid=2 /home/chrooted/proc` prevent the real /proc from becoming read only? 13:05 < bartmon> hi. can i achieve serving the same multicast addresses on several networks? maybe with source-based routing (per iface routing tables)? 13:06 < pingfloyd> I love how this comcast box said to me, "We'll need to go ahead and start your update". Makes you wonder if Bill Lumberg was the project manager of its software team. 13:07 < supernovah> pingfloyd: lol 13:13 < kurahaupo> supernovah: last time I tried anything complicated with bind mounts, I quickly found that flags such as ro & noexec don't actually have any effect; the new mount point for the same flags as the original mount point. That might have changed, but I'm sure they wouldn't have regressed to altering the original mount point. 13:13 < supernovah> kurahaupo: yea, the reason changing the flags doesn't work is that the bind mount occurs prior without any errors being given, that's why I remount straight after 13:15 < kurahaupo> supernovah: I guess I'd classify this under TIAS 13:16 < supernovah> what's tias, I've seen it before... 13:16 < supernovah> !tias 13:16 < kurahaupo> Try It And See 13:16 < supernovah> Ah 13:16 < supernovah> well I just did before, but with readonly because it wasn't a bind mount, and I destroyed my server, I just called up the tech guys in fiji to reboot it for us 13:17 < supernovah> it's not a production server thankfully lol 13:17 < kurahaupo> Ouch. 13:18 < supernovah> the guy who answered isn't at work and will do it in two hours, sounds like he's out drinnking 13:18 < supernovah> heh 13:18 < kurahaupo> Fiji is like 00:18 ? 13:18 < supernovah> 11:18 I think 13:18 < supernovah> depends on daylight savings 13:19 < kurahaupo> 23:18 now 13:19 < supernovah> Yea same time as here 13:20 < kurahaupo> Lazy so-and-so's who want to quit work before midnight! And sleep in until after 4am! 13:21 * kurahaupo waves to the Kiwis from West Island 13:24 < supernovah> kurahaupo: ^_^ 13:26 < supernovah> It's so weird having a server effectively offline because pam can't open /proc/self/loginuid 13:29 < jelly> which pam module is that 14:04 < sn00bie> i cant overwrite an file (executable) but delete and copy is possible, why? 14:04 < sn00bie> i use winscp, txt file is busy is the message 14:08 < kurahaupo> sn00bie: do you remember the three scenarios I said before? 14:10 < kurahaupo> sn00bie: the first thing to understand is that a file and a filename are not synonymous; a file can have any number of names including zero. If it has zero names, it will be deleted when there are no more open references to it 14:11 < kurahaupo> "Being executed" and "current directory of a process" both count as "open" 14:12 < kurahaupo> sn00bie: so it should be obvious that removing a name and creating it again makes a new file separate from the old one 14:13 < sn00bie> kurahaupo, sry im a n00b :-( 14:14 < kurahaupo> sn00bie: do these explanations make sense? 14:15 < fub> Hi. Im using this trick to run 3 processes in parallel: https://stackoverflow.com/a/41762802 -- this works 14:15 < fub> but when I do a ctrl+c to stop them, the script is stopped, but a ps tells me the processes still run. 14:15 < fub> Is there any way to ensure they also get the ctrl+c sent? 14:15 < kurahaupo> fub: put () around the whole thing 14:16 < sn00bie> i ve a file with the name test and its 755, the file cant have different names or? how? 14:16 < kurahaupo> fub: So that it makes a single process group 14:16 < sn00bie> the name of the file is test 14:17 < kurahaupo> sn00bie: "test" is the name of a command that is built into the shell 14:17 < kurahaupo> Which may take priority over any program by that name 14:18 < supernovah> readlink -m $(which test) 14:18 < fub> kurahaupo: the commands are long, so this would make the script ugly, any other way? http://dpaste.com/1HMCZDH 14:18 < fub> kurahaupo: maybe a trap 'kill $P1 $p2 $p2' INT ? 14:18 < sn00bie> k sry the filename isnt test the file name is another only an stupid example its i_host 14:19 < supernovah> you can change the file's name with mv command, mv i_host new_name ... 14:19 < kurahaupo> fub: if you put that text in a script it will work as you want. But when you do so on the command line, each line gets a separate process group - a separate job 14:19 < fub> kurahaupo: okay 14:19 < fub> so the trap thing worked :) 14:20 < fub> added it just before the "wait" 14:20 < kurahaupo> fub: you shouldn't need the trap in a script 14:21 < kurahaupo> Unless you have set -m in your script, which is seriously discouraged. 14:22 < kurahaupo> fub: you can put the ( and ) on separate lines by themselves; you don't need to fold the lines together 14:22 < sn00bie> mv is working, but copy over winsco not 14:23 < sn00bie> -o +p 14:24 < fub> kurahaupo: oh, I did not know this with the separate lines 14:24 < supernovah> what is winsco? 14:25 < kurahaupo> Winscp 14:25 < supernovah> Is it windows software? 14:25 < sn00bie> copy over ssh files 14:25 < interrobangd> linux question: are all drivers made as modules? 14:25 < sn00bie> yes windows 14:25 < fendur> supernovah: yes 14:26 < interrobangd> intel_pstate is a driver, but not listed with lsmod 14:26 < interrobangd> why? 14:27 < kurahaupo> sn00bie: try scp to a different filename and then mv. Or mv the old one, then scp 14:27 < steinex> interrobangd: it's compiled into the kernel, not as a module 14:28 < steinex> interrobangd: probably for reasons, it's a decision done by the linux-people 14:28 < interrobangd> steinex, but intel_pstate is not mentioned in my kernel config 14:28 < interrobangd> where we can decide if module or not 14:28 < steinex> CONFIG_X86_INTEL_PSTATE=y 14:29 < steinex> if it would be a module, there would be an m 14:29 < sn00bie> kurahaupo, thats working 14:29 < interrobangd> steinex, i am sorry, ... grep search case sensitive without -i, i've forgotten thsi 14:35 < atif5> Hi 14:37 < supernovah> hi 14:37 < Pentode> hi 14:37 < sn00bie> hi 14:38 < turkeyhand> I have CUPS installed and the printer shows up via LSUSB 14:38 < turkeyhand> but it tries to open "gnome software centre" to find the driver 14:38 < dTal> hi 14:38 < turkeyhand> I found the exact driver in the AUR for this thing, but it didn't seem to work 14:38 < turkeyhand> it says "canon pixma 490 series" 14:39 < turkeyhand> in LSUSB but the driver I found was for canon pixma 495, in windows it shows up as 490 series too 14:42 < Pentode> so it's not the proper driver? 14:43 < Pentode> im not that familiar with cups, but i wonder if the driver identifies the printer via it's unique device ID like other operating systems. maybe if the devices are similar enough you could modify the drivers somehow. that is, if the proper ones arent available.. 14:44 < Pentode> just a suggestion. it may not even be remotely possible, but i've gotten drivers working with devices in windows using that method. especially printers. 14:45 < Psi-Jack> Nope. Completely different. 14:46 < Pentode> figures ;{ 14:46 < turkeyhand> you mean modify the windows driver? 14:46 < Pentode> well at the very least he can print ascii through the line printer device file. 14:46 < turkeyhand> man it's a canon pixma MX495 and surely it's not that hard to do 14:46 < Psi-Jack> He means to modify the .inf file for the driver, not the actual driver itself. LOL 14:46 < turkeyhand> the fucking canon site says there is no driver or none required 14:46 < turkeyhand> I'm using arch 14:46 < Psi-Jack> Ahem! 14:47 < turkeyhand> swearing? sorry 14:47 < Psi-Jack> Yeah, you should know better already. 14:47 < turkeyhand> it's a habit I need to break 14:47 < turkeyhand> especially since I moved to germany 14:47 < turkeyhand> luckily I don't know how to swear yet. 14:47 < A4L> 3rd party yt dl like wget example.com/?yt= 14:48 < turkeyhand> how do I make this printer work 14:48 < Psi-Jack> What /actual/ specific model is it? 14:48 < turkeyhand> canon pixma mx495 14:48 < turkeyhand> or "490 series" 14:49 < turkeyhand> https://www.usa.canon.com/internet/portal/us/home/support/details/printers/support-inkjet-printer/mx-series/pixma-mx490?tab=drivers_downloads 14:49 < Psi-Jack> mx or mp? 14:49 < A4L> i alswo need help w/a printer HP LaserJet M1005 using CUPS 14:49 < turkeyhand> printer says mx 14:49 < Psi-Jack> Hmmm... openprinting.org claims that the mp495 is a paperweight. Common with many models of Canon in Linux land. 14:50 < Langley> Hello, is it possible to see somewhere what protocol a connected printer uses (LPD, SLP, IPP etc.)? 14:50 < turkeyhand> Bus 001 Device 007: ID 04a9:1787 Canon, Inc. PIXMA MX490 Series 14:50 < turkeyhand> a paperweight 14:50 < turkeyhand> GREAT 14:50 < rx7a> Hi all ! .o/ 14:50 < A4L> how to setup hp laserjat m1005 14:50 < turkeyhand> it has paper and it has weight 14:50 < turkeyhand> so I have to boot to windows to print something 14:50 < Pentode> i guess most cheap printers now-a-days are "software" driven, analogous to the old "winmodems." 14:50 < Psi-Jack> In fact, the MX435 is also a paperweight. 14:51 < turkeyhand> it prints fine in windows 14:51 < turkeyhand> I wish I didn't have to use windows though 14:51 < turkeyhand> why is it showing up in lsusb? 14:51 < pankaj_> I am building installing packages from source. But many times the basic files like INSTALL or README are missing. So, how to get instructions for that particular package for installing it correctly. Or that it means that just follow standard build options -> ./configure -> make -> make install 14:51 < Psi-Jack> Blame Canon for being the problem in this case, not making proper cups drivers. 14:51 < turkeyhand> I didn't buy this thing 14:51 < Psi-Jack> Because it's a USB device. 14:51 < turkeyhand> so linux will never print with it? 14:51 < A4L> turkeyhand: just get driver file for cups? 14:52 < turkeyhand> canon doesn't supply one 14:52 < Psi-Jack> turkeyhand: Never say never, just it's not supported today. 14:52 < supernovah> Could always write one 14:52 < turkeyhand> I would like to print today 14:52 < turkeyhand> not next year 14:52 < Psi-Jack> Many Canon's aren't. I generally avoid Canon for printers because a good lot of their printers won't work in Linux. 14:52 < turkeyhand> well I didn't buy it 14:52 < turkeyhand> it's the only one I have at hand 14:52 < A4L> turkeyhand: get a non-official driver, or other model driver, that works with ur model 14:52 < turkeyhand> windows it is >:| 14:52 < turkeyhand> which one works though? 14:53 < Psi-Jack> HP usually works great. 14:53 < turkeyhand> no which driver works 14:53 < Psi-Jack> None. 14:53 < supernovah> type /join ##windows and have fun 14:53 < A4L> turkeyhand: idk trial and error some simmilar models and google 14:53 < turkeyhand> it just means I have to reboot into windows and do it there, piece of crap 14:53 < turkeyhand> I hate this 14:54 < turkeyhand> how can I make arch hibernate? 14:55 < DrDonkey> I am confused about this command xrandr --addmode VGA-0 1368x1024_60.00 14:55 < Psi-Jack> https://wiki.archlinux.org/index.php/Power_management/Suspend_and_hibernate#Hibernation 14:55 < Pentode> commenter's on openprinters appear to indicate that it works on some distros and not others for whatever reason. so apparently it _can_ work 14:55 < pankaj_> How to get details of a package installation when it does not contain basic files like README or INSTALL that contains user build instructions? 14:55 < DrDonkey> noy that sorry but this: 14:55 < DrDonkey> xrandr --output VGA-0 --mode 1368x10 24_60.00 --right-of LVDS 14:56 < DrDonkey> LVDS is my laptop monitor 14:56 < geirha> pankaj_: check its homepage 14:56 < Pentode> whats confusing about it? 14:56 < DrDonkey> What exactly it does? 14:57 < Pentode> output to vga 0 at 1366x1024 at 60hz (low voltage differential signaling) 14:57 < DrDonkey> it adds another monitor of the stated resolution right of my laptop monitor?? 14:58 < Pentode> uh huh 14:59 < etaleo> Hey there! Is there a way to append the output of "date" to every line while reading from a pipe? 14:59 < DrDonkey> Let me explain a bit, I am using i3-gaps, and have 10 workspaces, will the right monitor also have 10 workspaces?? 14:59 < noodlepie> etaleo, yep! 15:00 < etaleo> So far I've come up with: cat /my/pipe | awk -v date="$(date)" '{print date, $0}' 15:00 < etaleo> But it initializes date at the start and then uses the same date for every line 15:00 < DrDonkey> (It's really confusing how it works) 15:00 < DrDonkey> also do you guys have a discord group?? 15:01 < Psi-Jack> No. 15:01 < noodlepie> etaleo, read the printf docs (manual page etc) for details of date insertion into a string variable 15:01 < noodlepie> WE ARE ORDER! 15:01 < pingfloyd> turkeyhand: or set up a windows print server and have linux send jobs to it. 15:01 < DrDonkey> sure. 15:02 < tsarompy> hi 15:02 < A4L> btw cups doesnt work with hp lj m1005. or does it?? 15:02 < A4L> help 15:03 < A4L> printer 15:03 < Psi-Jack> A4L: Works /mostly/. 15:03 < Psi-Jack> http://www.openprinting.org/printer/HP/HP-LaserJet_M1005_MFP 15:03 < Pentode> DrDonkey, something like that is up to the window manager's virtual desktop implementation afaik 15:03 < A4L> on raspberry pi it doesnt 15:05 < DrDonkey> Pentode ok got it. Whatever they did implement, it is so confusing. Now I am going to talk with i3wm guys. Thank you. 15:06 < Pentode> np, good luck. 15:07 < kubast2> Hey ,can parallel read a custom ssh port? 15:07 < A4L> kubast2: ?? 15:11 < A4L> kubast: please explain situation 15:11 < A4L> kubast2: look up 15:14 < A4L> how to install foo2xqx 15:14 < A4L> ?? 15:14 < A4L> ? 15:15 < Pentode> love it when they ask a question then they don't pay attention, leave, or flat out ignore you... 15:15 < A4L> they... 15:16 < oneko> Can someone suggest for me a used Apple devices lol 15:16 < Armand> Nope 15:16 < oneko> -a 15:16 < A4L> Ctrl+C 15:16 < oneko> I've never used Apple things before and I think I need them for this project 15:17 < Armand> oneko: Off-topic. 15:17 < Armand> kthx 15:17 < A4L> oneko: describe situation 15:17 < Armand> A4L: This isn't ##apple 15:17 < oneko> Do we have anything like ##apple on freenode ? 15:17 < A4L> probably ##apple 15:18 < Armand> /msg alis list *apple* 15:18 < Armand> You're welcoome 15:18 < Armand> -o 15:18 < A4L> is Ctrl+7 bell still working?? 15:18 < oneko> I'm working on something that I need to work on almost all platforms, A4L - I figured going for a Macbook or something would be the most reasonable choice 15:19 < A4L> letgo (-; 15:19 < oneko> And, also, my current PC is a bit old 15:19 < A4L> maybe macbook emulator?? 15:19 < A4L> iosx emulator 15:20 < oneko> Yeah, that's indeed one of the options 15:21 < A4L> why not? 15:22 < oneko> So, what macbook emulator works on Linux ? 15:22 < oneko> Or, iosx emulator ? 15:22 < Psi-Jack> None 15:23 < A4L> i dnt even know if macbook emulator exist. if it does, u probably have 2 use VirtualBox... maybe wine to emulate windows to emulate macbook :-P 15:23 < oneko> lmao 15:24 < sn00bie> to emulate ios? 15:24 < oneko> Is it a crime ? :-P 15:24 < revel> I've had minimal success with emulating OSX on VirtualBox. 15:25 < revel> They can't legally develop and/or distribute guest additions for that platform though, as far as I know, so the performance is subpar. 15:25 < revel> "Performance" being a very broad term. 15:25 < supernovah> server finally back online 15:26 < A4L> another option is to get a chair and work in Apple store 15:27 < etaleo> noodlepie: I've now tried with the following, unfortunately it's still the same date for every line: cat /tmp/mypipe | xargs -I %line printf "%s: %line\n" "$(date)" 15:28 < noodlepie> etaleo, you don't need to specify the date command variable if you use the data formmating subtexts from printf 15:28 < hellyeah> hey 15:28 < oneko> I do have a chair, A4L except I live in the middle of nowhere and we don't have Apple stores lol 15:28 < noodlepie> Hi hellyeah, wonna "do" Linux? 15:29 < hellyeah> i try to add service i wrote this file. Is there anything wrong here http://dpaste.com/2VWGEPW 15:29 < sn00bie> what about rent a device? 15:29 < hellyeah> ofc 15:29 < sn00bie> or rent a device online, only "virutal" 15:30 < etaleo> noodlepie: Is it possible to use the output of another commands? In the end I'd like to append the current milliseconds to each line 15:31 < noodlepie> "%d %h %s" will automatically put the parts of date into those variables. 15:31 < rypervenche> hellyeah: Remove EOF, use full paths. 15:31 < noodlepie> so "Hello World %d %h %s" does what you want! 15:31 < noodlepie> :P 15:33 < oneko> So, revel, elaborate on performance ? 15:35 < revel> Well, in the literal sense, it's slow as hell (haven't tested bare-metal OSX on this hardware, so, yymv) and it has no guest additions, so that's a minus (stuck in some terrible resolution, pretty sure it also generally adds some graphics drivers to improve performance) 15:36 < revel> s/yymv/ymmv/ 15:38 < oneko> Well, I also had also set up some hackintosh setup before for kicks and it was indeed slow 15:38 < etaleo> noodlepie: What could I do if I'll want to use the output of another command and append it to each line? 15:39 < revel> Also, I had to use some command-line tools to change some EFI variables or somesuch to get it to boot in the first place, but there's guides up online for that. But I don't really think it's worth it. 15:40 < alexandre9099_> hi, i need help troubleshooting my openvpn server, i get connected but i cannot even ping local pcs/routers on the server's network, with pptpd i have a local+internet connection without problem 15:41 < oneko> Thanks for the information, revel :-) 15:42 < revel> Well, you'd probably have found out about that before you even downloaded an image to boot from :P 15:42 < etaleo> noodlepie: I got it! cat /tmp/mypipe | xargs -I %line sh -c 'printf "%s: %line\n" "$(my_command)"' 15:44 < rypervenche> alexandre9099_: Did you set up an SNAT or masquerading on your server? 15:44 < oneko> Hmm, I'd guess so too :-P 15:52 < supernovah> how can you use cat Excellent stuff etaleo, Linux is worth using then? :P 15:53 < noodlepie> supernovah, install wireshark tcp/ip monitoring :P 15:54 < rypervenche> Or use tcpdump like a boss. 15:55 < noodlepie> rypervenche, yesh! 15:55 < supernovah> I'm trying to see how a chroot user can potentially expose itself out of the jail, apparently /dev/tcp// is provided by bash 15:55 < supernovah> perhaps I need a bash built without it 15:58 < JeffATL> if i use dmcrypt to encrypt a real physical disk partition, making /dev/mapper/encrypted, and i want to cut that up into pieces so i can use LVM to make LVs out of the pieces, what's my next step...run fdisk on /dev/mapper/encrypted? 16:01 < JeffATL> and if i do that, won't it try to make a new /dev/sda1 etc.? 16:02 < toad_poloer> I want to make a `dd` backup of a partition I have, but most of that space is unused, so I think I can save myself a bunch of time by only copying a subset of the image. 16:03 < toad_poloer> The problem is that I don't want to resize the partition in gparted without having a backup in case things go wrong, and getting a backup is the problem I'm trying to solve. 16:03 < toad_poloer> For ext4 if I just copy as much as the size it says is "used", will that work? Is there no worry about disk fragmentation or metadata? 16:04 < JeffATL> toad_poloer: i'd be wanting to back up the contents of the partition, not make an image of it 16:04 < toad_poloer> I would presumably copy more than is "used, like it looks like it's maybe 34 GiB so I could copy 50GiB. 16:04 < toad_poloer> JeffATL: Yeah I did that as well. 16:04 < revel> Shrinking an ext4 partition from the right side should be fairly safe. 16:04 < revel> Note: Don't sue me, please. 16:04 < compdoc> Im getting a lawyer 16:05 < fattredd> I would try rsync, but that's just me 16:05 < revel> Also, resizing from the right side should take, like, seconds. 16:05 < rypervenche> supernovah: You might want a more restricted/simple shell. 16:05 < toad_poloer> JeffATL: But I am planning on put a new distro on here, and if that drags on for whatever reason I want to be able to just wipe the slate clean by restoring the partition image. 16:06 < toad_poloer> revel: Yeah, I think I've done that before, though it seems like that's consistent with the idea that I should be able to just copy the first N bytes of the image and that's equivalent to having done the shrink first, right? 16:06 < toad_poloer> Since the shrink is just changing the GPT table or whatever. 16:07 < JeffATL> toad_poloer: might do this in this order: 1) defrag 2) dd if=/dev/zero to a file on the partition until it runs out of space 3) delete that file 4) dd the partition to stdout and pipe through gzip to write another file 16:08 < toad_poloer> JeffATL: How do you defrag an ext4 partition? I've always heard "it's not necessary" 16:08 < revel> If you move a partition or grow/shrink left, then I think you have to basically copy the whole partition to a new place, chunk by chunk, meanwhile when you shrink/grow right, it doesn't. 16:09 < revel> I think it's supposed to defragment itself and should be fine with regular use. 16:09 < fattredd> Swap #1 and #2 16:09 < revel> Lemme check. 16:09 < fattredd> defrag the large blank file 16:09 < toad_poloer> Also I think that dd if=/dev/zero would cause problems if I put the file in the wrong place. 16:10 < S_Gautam> I tried Ubuntu and Linux Mint on a Core 2 Duo (no dedicated GPU) and the result was a pretty sluggish system (the UI felt like a viscous fluid), have anyone of your tried Manjaro on such a system? 16:10 < toad_poloer> The home directory is encrypted, so if I filled it with a huge empty file I'm guessing it would fill the drive with opaque encrypted files. 16:10 < revel> Yeah, it shouldn't need defragging. 16:10 < toad_poloer> Though maybe the home directory encryption stuff works differently. 16:10 < toad_poloer> I think I'll just sorta risk it. 16:10 < amosbird> Hi, how can I find out which process generates high sys cpu usage ? 16:10 < toad_poloer> Dump more than the used space to a file, shrink from the right, dump the whole thing to a file. 16:11 < JeffATL> toad_poloer: ordinarily it's not necessary but you're trying to do a weird thing 16:11 < tsarompy> thats what u get for usin ubuntu and linux mint on a core 2 duo 16:11 < alexandre9099> rypervenche, if you mean this command "iptables -t nat -A POSTROUTING -s 192.168.5.x/24 -o enp7s0 -j MASQUERADE" yes (is the x relevant?) 16:12 < S_Gautam> tsarompy, I'm currently using Lubuntu although I kind of hate LXDE 16:12 < rypervenche> alexandre9099: It should not be an x :) 16:12 < tsarompy> you oughta be usin AWESOME 16:12 < alexandre9099> well, i thought that i wrote 1 or 0, not sure, let me check 16:12 < tsarompy> minimal ubuntu or arch install + xorg + awesome = perfection 16:12 < tsarompy> with kitty terminal 16:12 < tsarompy> best workstation evar 16:12 < rypervenche> alexandre9099: You could do: 192.168.5.0/24 16:13 < alexandre9099> rypervenche, would it make any difference? 16:13 < tsarompy> i have an old quadcoar toshiba with 4gb ram it runs perfectly with that setup 16:13 < rypervenche> alexandre9099: Well, the x would not be proper syntax, that I know of. If you did a 1 or 2, I don't know what source IP you're using 16:14 < mawk> just use .0 16:14 < alexandre9099> i haven't wrote actually x, i just put it there because i was not sure which number i wrote there ;) 16:14 < alexandre9099> on openvpn i have "server 192.168.5.0 255.255.255.0" 16:14 < supernovah> rypervenche: yeah I want to build bash without tcp/udp device hijacking 16:15 < supernovah> or to apply tcp wrappers to the jail 16:16 < rypervenche> alexandre9099: You should use a .1 or something, as 192.168.5.0 would be the network address. 16:16 < alexandre9099> just .1 without the netmask(? that \24 thing) 16:16 < alexandre9099> */ 16:17 < rypervenche> alexandre9099: With the netmask. 16:17 < alexandre9099> i'll try again 16:17 < rypervenche> server 192.168.5.1 255.255.255.0 16:18 < alexandre9099> i'll try to connect now (maybe i'll lose conection) 16:18 < mawk> what ? 16:18 < mawk> why rypervenche ? 16:18 < alexandre9099> ? 16:18 < mawk> it's a mask, the .1 will be masked out anyway 16:21 < rypervenche> mawk: Oops, sorry. Just checked my config. Yes, .0 is correct. 16:21 < rypervenche> alexandre9099: In your iptables rule, use -s 192.168.5.0/24 16:22 < Trel> Can anyone help me understand what this line is doing in a script with a /bin/sh hashbang? https://pastebin.com/nDVeUPH2 16:24 < solidfox> Trel, the context is missing 16:24 < Trel> Top of a node.js binary, it looks like it's somehow continuing the hashbang, but I don't get how 16:25 < supernovah> sweet, that was easy, download bash source, ./configure --disable-net-directions, make and copy it into the jail, chown it and good to go 16:26 < Trel> supernovah: here's my whole test file http://termbin.com/91e9 so context is pretty non-existant 16:26 < Toadisattva> using Ubuntu 18.04 for some reason the file menu and sidebar just completely disappeared from my file manager. alt+m/e/v andsuch don't do anything, I've tried reinstalling ubuntu desktop and nautilus, can anyone ponit me the right direction to fix this? 16:27 < supernovah> OH that's a javascrpt file with a shebang at the top, you use that when you want to run a nodejs script as a service 16:27 < supernovah> Trel: 16:27 < Trel> supernovah: I know what it is, I can't figure out how that second line is causing it to execute in node 16:28 < supernovah> ':' is a javascript string with a comment following it, technically the syntax is incorrect 16:28 < Trel> Yes, but I'm trying to figure out how that line works at all 16:29 < supernovah> Trel: that won't be causing it to execute in node, generally to execute a nodejs script as a service, people place arule in /etc/udev/rules.d/50-udev.rules 16:30 < supernovah> and when they have output like console.log('test') there, it will output to /var/log/messages or journal or whatever your distro is using 16:30 < Trel> The problem is that it IS working 16:30 < solidfox> Trel, where did you get this code? 16:31 < Trel> I compiled node.js and installed (locally), and installed a node module (vtop) and looked at the executable. I copied that line from there, and put the console.log myself. 16:31 < Trel> I'm trying to figure out WHY it's working, but if I take out the line starting with ':' I get a syntax error on the console.log part. 16:31 < Trel> So I know that line is somehow getting it to run in node 16:31 < solidfox> command is a command that runs a command 16:32 < solidfox> $0 is argument 0 16:32 < solidfox> $@ is the rest of the arguments 16:32 < revel> But what's $* ? 16:32 < Trel> What is "':' //;" doing 16:32 < solidfox> those are the arguments passed to nodejs or node 16:33 < supernovah> where is that file that contains it 16:33 < Trel> I'm still not getting how the three line script I pastebinned worked 16:33 < solidfox> so it kinda evaluates to something like: nodejs /bin/sh ':' // 16:34 < compdoc> igp 16:34 < supernovah> Trel: what file contains the odd syntax, what's it called, and is there an entry for it as a service 16:34 < seven-eleven> hi 16:35 < solidfox> or perhaps nodejs thefilename ':' // 16:35 < seven-eleven> hi i want to run this command from a python script: `find /some/dir -mtime +60 -exec rm {} \;` should I rewrite this with python or should i execute the linux command from python because the linux command is more efficient? 16:36 < Trel> solidfox: if THAT's what it's doing, I think I see how it works now 16:36 < benjwadams> Is there some kind of "Rosetta stone"/cheatsheet for correspondences between the `netstat` and `ss` commands 16:37 < supernovah> Trel: what's the filename of it 16:37 < revel> seven-eleven: If you're going to use Linux coreutils, then you might as well write a shell script, right? :P 16:37 < Trel> test3.sh 16:37 < supernovah> and is there an entry for it as a service 16:37 < benjwadams> I've grown used to `ip` over `ifconfig` these days, but i still haven't really grokked `ss` 16:38 < Trel> supernovah: 100% no 16:39 < Trel> if I'm understanding that line, I THINK I may get what's happening, I need to test something 16:39 < alexandre9099_> rypervenche, nice, i changed the server ip to 192.168.251.0 and now it works, maybe 192.168.5.0 was already used on the network i'm on and was making some conflict 16:39 < solidfox> I'm trying to test it out but sh is confusing me 16:39 < solidfox> I'm used to bash 16:40 < rypervenche> alexandre9099_: Grats :) 16:40 < alexandre9099_> :) thanks 16:40 < Trel> solidfox: got what it's doing 16:41 < alexandre9099_> now let's see if i can connect trough proxy, i already set openvpn to listen on 443 (actually it is on the default, but i'm using sslh :D) 16:41 < Trel> "':' " just returns true 16:41 < solidfox> Trel, I still don't get it 16:42 < Trel> so "':' //" = true, so nothing happens shell-wise. The semi colon is to begin a second statement which is to pass the whole file to node. 16:42 < solidfox> why is it so cryptic lol 16:42 < Trel> When it goes through node, it sees ':' (true), and then // the rest is commented out 16:43 < Trel> and then onto the javascript lines which node can handle 16:43 < solidfox> this doubly indirected way of executing a file in sh to execute itself in nodejs is black magic sourcery 16:43 < supernovah> Trel: ':' will be true in javascript just because its a string that doesn't contain the ascii for 0 16:43 < Trel> yes, and in shell it's a built-in for true apparently 16:44 < Trel> or rather a null command that always exits successfully 16:44 < Trel> so for shell // is a paramter to :, and in node, // comments out the rest of the line 16:44 < supernovah> I think its so cryptic for backwards compatability on older systems 16:45 < Trel> I'm going with black magic that I need to memorize for the future. 16:47 < supernovah> Trel: man bash, look for command [-pVv] also 16:47 < solidfox> here's an example: https://searchcode.com/codesearch/view/98389438/ 16:47 < supernovah> unless you know it 16:49 < seven-eleven> revel, i want to run the command from within a python routine 16:49 < supernovah> its basically a type or which for a command 16:49 < revel> Well, nobody's stopping you, it's really up to you and what the goals/limits/whatever are for the Python script. 16:52 < alexandre9099_> well. not sure if this is the best place to ask, anyway, how can i bypass a proxy server to connect to my openvpn server? 16:53 < Psi-Jack> Why do you want to bypass a proxy server? 16:53 < alexandre9099_> well, to connect to my openvpn server ;) 16:53 < alexandre9099_> or go trough it 16:53 < Psi-Jack> So company/school/etc blocking you? 16:53 < alexandre9099_> yes 16:53 < Psi-Jack> Yeah, not going to help you then. 16:54 < alexandre9099_> i tought so :D 16:54 < supernovah> Trel: also the comment // doesn't do anything, it's passed as an arg to ':' and you are right, comments uot the rest of the line if the file is read as javascript, for this reason - what the // achieves is making this file work if the file is invoked in a node environment, but it will create a node environment if not 16:54 < Trel> Yep, looks like a neat magic trick :D 16:54 < alexandre9099_> Psi-Jack, interesting thing is that bitmask (on my android device) connects fine trough the proxy, they might do some black magic XD 16:55 < alexandre9099_> afaik they use openvpn 16:55 < supernovah> its like a comment that precedes a terminator (the colon for the bash) depending on what reads the file, definitely neat 16:55 < Psi-Jack> Stil not going to help you. 16:55 < alexandre9099_> ok, understandable, thanks anyway ;) 17:00 < rypervenche> lol 17:08 < etaleo> How come busybox adjtimex gives me a 9 digit number for time.tv_usec when microseconds should be 6 digits long? 17:15 < bls> how are you determining that it's giving you a 9 digit number? what format specifier are you using to print it? 17:16 < Siecje> How do you set the format of the syslog? I want to add the year to the date. 17:16 < etaleo> bls: I am running "busybox adjtimex" in the shell 17:17 < supernovah> etaleo: are the first 3 whole seconds, or perhaps the last 3 are sub-micro based on the ULP of the floating point architecture used, or perhaps the accurancy of the system clock 17:19 < mawk> in all my open chrome tabs, my own website is the only one to have DNSSEC 17:19 < etaleo> supernovah: They seem to be nanoseconds 17:19 < mawk> when even TLD servers won't implement it, it's the sign of a dead standard I guess 17:19 < supernovah> can you tell what the minimum difference is between them 17:20 < cluelessperson> question, how do you set the sudoer's file to allow a user to run a command as ANOTHER user, WITH THEIR HOME? 17:20 < supernovah> you can run sudo -Siu to achieve that... maybe sudoers does it also though 17:20 < mawk> what do you mean by with their home cluelessperson 17:21 < cluelessperson> mawk: I'd like to cd to their home I suppose 17:21 < cluelessperson> that would be good enough 17:21 < mawk> ah 17:21 < mawk> yeah sudo -i will do that maybe 17:21 < mawk> but it's not the intended use 17:21 < mawk> just cd yourself 17:21 < bls> cd is a command internal to the shell, not something you can use with sudo 17:21 < mawk> yeah with sh -c I mean 17:22 < supernovah> $(getent passwd | cut -d: -f6) for a users home directory if you want to be obtuse like me 17:22 < cluelessperson> mawk: how do I allow that in the sudoer's file? 17:22 < mawk> it's already allowed 17:23 < bls> oh, yeah, you just need to give sudo the right args to reinit the environment as the other user so HOME gets reset 17:23 < supernovah> and the path variable doesn't get set with env, woop 17:23 < mawk> it's already the case by default no ? 17:23 < mawk> sudo -E will keep the old env, but just sudo will set $HOME to the right thing 17:24 < bls> one of sudo -H or sudo -i should accomplish this 17:24 < bls> depending one what's really being asked for 17:25 < supernovah> pretty sure $PATH doesn't get set with -E as well 17:28 < supernovah> sudo -Eu env PATH=$PATH sh -c 'whatever' 17:32 < supernovah> if you have custom paths set up in .bashrc etc you probably want to use -i though 17:36 < adrian_1908> `cat some_file | wc>` The angle bracket is bullshit, right? 17:36 < stevenm> can someone tell me a proper way to handle zombie/defunct processes without rebooting - it's really ***king STUPID. 17:37 < stevenm> e.g. (this a repeat offender) my file manager (caja) will hang... forever... so I think to kill it... which works - but one last caja process/thread goes defunct 17:37 < stevenm> can't be killed - but launching another instance of caja won't work - presumably because it sees the first one as defunct 17:38 < stevenm> so the parent process is mate-session - i.e. my whole desktop environment... don't want to kill that 17:38 < stevenm> is there some way I can just REMOVE the defunct process entry so things can go on as normal? 17:38 < stevenm> and don't say "Oh MATE should write their software better to handle child processes etc... etc..." - because I know that ... but every now and then you encounter this, not just with caja. 17:39 < stevenm> there has to be a workaround to some of this alleged "bad programming" than to just reboot the whole damned system or kill a potentially important parent 17:39 < triceratux> not that exactly, but what were you doing that causes it to hang, & how do you ordinarily kill it ? 17:40 < stevenm> it got stuck on accessing a mount - it wouldn't close normally... so I killed it with xkill (I have a keyboard shortcut that allows me to run xkill and then I just click the program) 17:42 < storge> that might be just killing a gui and not an underlying process 17:42 < triceratux> tbh ive never heard of xkill, but then im a n00b https://www.thegeekstuff.com/2009/12/4-ways-to-kill-a-process-kill-killall-pkill-xkill/ 17:42 < stevenm> storge, no it actually does kill the program - i'm very familiar with it 17:42 < bls> there's not really a workaround or a solution. that's just how the system was designed 17:42 * triceratux would never rely on an x11 program like that 17:42 < stevenm> it wouldn't matter if I'd have done kill -9 17:42 < storge> triceratux: me neither 17:42 < stevenm> which sometimes i do from a terminal instead of xkill 17:43 < triceratux> nuke it from orbit 17:43 < storge> its the only way to be sure 17:43 < stevenm> bls, then it's flawed - when it forces the user to reboot or kill a potentially more important parent process - to go on with using the system properly... there should be *some* was of overriding it! 17:46 < stevenm> if anyone has a better answer . .please PM me... as i have to go home now and my scrollback won't last the weekend 17:46 < stevenm> thanks 17:46 < bls> I understand there's a group of people out there that don't care or believe otherwise, but there are very good reasons to not allow that 17:46 < triceratux> install thunar ftw :) 17:48 < bls> hence the "don't make us add in an unsafe workaround just for people that don't care about data integrity to their flakey mounts/hardware/network connections, fix your system/program" 17:50 < storge> heh 17:51 < triceratux> as part of my linux best practices i dont use the gui FMs for anything other than filesystem mounting. i do all file operations including umounts from mc. the gui FMs are too prone to unrecoverable hangs because theyre so reliant on udev & udisks2 & dbus & the rest 17:52 < Pentode> i love xfce and everything but thunar is poopoo :| 17:52 < benjwadams> /quit 17:54 < vktec> Anyone know a way to make xmodmap automatically work for newly-plugged in keyboards? 17:54 < epicmetal> vktec: probably some udev config 17:55 < vktec> Fair enough 17:55 < bls> vktec: https://superuser.com/questions/249064/udev-rule-to-auto-load-keyboard-layout-when-usb-keyboard-plugged-in 17:56 < S_Gautam> will manjaro linux run good on a Core 2 Duo with 5gigs of RAM 17:56 < S_Gautam> i just want to do some programming and web browsing 17:56 < vktec> S_Gautam: Depends what software you run on it 17:56 < bls> S_Gautam: any distro will run fine 17:56 < vktec> What DE, etc 17:56 < vktec> But most likely anything will run fine 17:56 < bls> yeah, the DE/WM and associated tools will determine if your system is fit 17:56 < S_Gautam> I'm asking because I tried Ubuntu GNOME and Linux Mint Cinnamon and the system felt very sluggish 17:57 < bls> then try a different DE and not a different distro 17:58 < bls> try XFCE or LX**, try it with just a window manager 17:59 < darkxploit> Hello how do we log the “execution time” of all mysql queries. Example execution time of all queries in General Log ? 18:00 < bls> darkxploit: mysqlbench or try asking in #maria 18:01 < S_Gautam> How's Deepin, compared to Cinnamon and GNOME? 18:02 < bls> S_Gautam: deepin is a distro. cinnamon a gnome are desktop environments. 18:02 < catalase> hi, does anyone have instructions on how to downgrade from transmission 2.94 to 2.93 on ubuntu and then pin the package 18:02 < S_Gautam> Deepin is a DE, no? 18:03 < bls> ah, looks like they spun off their own 18:05 < hipp> deepin is a distro and a de 18:08 < triceratux> S_Gautam: the extix version of deepin looks pretty interesting. he creates some pretty good liveisos http://www.extix.se/?p=407 18:15 < yakiza> https://www.linuxquestions.org/questions/slackware-14/uk-london-slackware-meet-4175630377/ 18:15 < alexandre9099__> hi, i'm using sslh to multiplex my 443 port, but my https website is not working as expectected (cloudflare gives "SSL handshake failed"), what can i do? 18:17 < bookworm> check why the handshake fails? 18:18 < alexandre9099__> well, how? 18:18 < bookworm> Is it valid if you connect directly without cloudflare? 18:18 < alexandre9099__> i'll try (but the domain is set trough cloudflare, i need to reconfigure it) 18:23 < alexandre9099__> bookworm, hmm firefox gives me "SSL_ERROR_RX_RECORD_TOO_LONG" 18:24 < alexandre9099__> oh it might be because of the listening port not being set for https 18:26 < alexandre9099__> hmm, whenever i get home i'll try to set the virtual host listening to _default_:443 18:28 < Gnjurac> hi can i somhow run script on hid device connect 18:28 < Gnjurac> ? 18:28 < Gnjurac> even better if i can sort per device id 18:49 < smallville7123> how is memory optimized in standard up to date distributions (not limited to distributions specifically designed to optimize as much memory as possible) 18:51 < Psi-Jack> The same way/ 18:52 < lnnb> memory? optimized? bahaha 18:52 < jelly> smallville7123: what do you mean by optimizing memory? 18:53 < smallville7123> like memory consumption 18:53 < smallville7123> or memory usage 18:53 < jelly> short answer is "it's not" 18:53 < smallville7123> why (◉-◉) 18:54 < lnnb> it's optimized by you deciding to not install bloatware 18:54 < Psi-Jack> Nothing special is going on. 18:54 < jelly> because it requires expert knowledge which is a lot more expensive than RAM :-) 18:54 < Psi-Jack> Now... What are you REALLY trying to do/ask about? 18:54 < linuxconformer> guys how do i use my public key if i'm using an image of an instance that already has a key? 18:55 < Psi-Jack> linuxconformer: Your question makes no sense. 18:55 < jelly> linuxconformer: ssh public key? 18:55 < jelly> linuxconformer: you don't use the public part, you use the private part to log in to a remote user that lists the public part in authorized_keys 18:56 < linuxconformer> yeah i'm not really sure how to explain it. Basically i had a vps instance which i am able to ssh into, then i created a new instance based off that original instance, and now when i try to ssh into this new vps i get "WARNING: REMOTE HOST IDENTIFICATION HAS CHANGED!" 18:56 < cxc99> can a krb5p nfs share work with another local-only user besides root? 18:56 < jelly> linuxconformer: and that's expected 18:56 < linuxconformer> how do i resolve it? 18:57 < Psi-Jack> linuxconformer: Well, yeah. That's a host key... 18:57 < smallville7123> When an application or library is loaded is its memory layout modified at all or is the entire file and its corresponding segments loaded into memory 18:57 < Psi-Jack> Have you bothered to read wthe full message? 18:57 < Psi-Jack> the* 18:57 < jelly> linuxconformer: the remote HOST key is different than a previous host with the same name or the same IP address 18:58 < linuxconformer> my bad, figured it out 18:58 < linuxconformer> thanks guys 18:58 < Psi-Jack> heh 18:58 < jelly> linuxconformer: if you are 100% certain this is okay, you can do what ssh client tells you to do to remove the old remembered host key 18:58 < Psi-Jack> Yeah, the message even tells you what to do to clear the problem. 18:59 < smallville7123> and could that memory layout be optimized in any way, for example, by possible attempting to trim unneeded data from it while still maintaining stable execution 19:00 < Psi-Jack> smallville7123: LOL. Seriously. What are you on about? Linux memory management is linux memory management. It's the same in any distro. 19:01 < jelly> smallville7123: size of code is usually a lot smaller than size of data the code works on 19:02 < jelly> smallville7123: the right questions would be how to choose data structures to achieve your goal (which might be optimizing for size, or optimizing for speed, or a combination thereof) 19:02 < jelly> and this is a topic for a (couple) programming class(es) 19:04 < rypervenche> linuxconformer: ssh-keygen -R ipaddress (put the ip address) 19:05 < rypervenche> linuxconformer: Or if you are using a domain, put that there. Essentially, whatever it showing up in the error. 19:05 < Psi-Jack> rypervenche: You're late. 19:05 < rypervenche> So I am. 19:05 < Psi-Jack> By over 5min. :p 19:06 < rypervenche> Might still be useful information to someone else :) I only learned about ssh-keygen -R recently. 19:06 < Psi-Jack> Really? How long you been using Linux? 19:06 < jelly> smallville7123: then when you get to learn how actual current hardware works, you get to read about tiered memory architectures and data locality and code locality and TLBs and cache lines and all sorts of interesting bottlenecks 19:07 < EvilRoey> "then when you get to learn how actual current hardware wor" aww that's easy! The power plug's connected to the--so-cket ♫ 19:07 < smallville7123> like.... could it, in theory, be optimized by using jit techniques (like monitoring for hot spots then if one is detected replace it accordingly) (except it will attempt to do so on compiled code instead of source code) and dynamically modifying the applications code/data to be as small as possible than what strip can produce 19:08 < Psi-Jack> smallville7123: ... 19:08 < jelly> smallville7123: no, your definition of "it" is not good enough 19:08 < Psi-Jack> heh 19:08 < smallville7123> it == any linux application 19:08 < lnnb> why though 19:09 < smallville7123> because why not 19:09 < lnnb> you'll only encourage more horrible programming practices 19:09 < smallville7123> (◉-◉) 19:09 < djph> ^ 19:09 < smallville7123> how 19:09 < Psi-Jack> Because you make 0% sense. 19:09 < jelly> smallville7123: start with basic "data structures and algorithms", you want to learn about the bigger picture first 19:09 < Psi-Jack> Because you obviously know near to nothing about memory and applications. 19:09 < djph> you wanna see good code, go get a master's in mathematics, and read Knuth 19:10 < jelly> there's no such thing as "any linux application", and Linux kernel itself can run on tiny appliances to large 4096-cpu systems 19:10 < djph> I tried reading Knuth (not a math major) ... good god the math ... 19:11 < jelly> if you want to learn about bottlenecks on x86, you have to learn about x86 19:12 < jelly> and x86 has grown to be very complex during its lifetime 19:12 < dave_grol> hi, qq - I've a list of IDs followed by a "release version" e.g.: 55555_1.2.6_2 - 55555_1.2.6_3 - 55555_1.2.6_4, that's just a single ID but the list is 700 lines, many duplicated IDs but diff version numbers, is there any command (not aware) to mathc only the latest release number? 19:13 < smallville7123> 03:10 jelly: there's no such thing as "any linux application", and Linux kernel itself can run on tiny appliances to large 4096-cpu systems. Basically i mean ANY executable application that can be ran in linux, and thus any dynamic shared object files that may or may not be loaded 19:13 < jelly> dave_grol: no, but you can script/program one with awk, python or perl (or even shell) 19:13 < smallville7123> by that application 19:14 < jelly> smallville7123: it doesn't matter 19:14 < dave_grol> yea , thought that too... 19:14 < smallville7123> for example when the dynamic linker loads a shared object does it load the entire file or is it possible to load only the required parts of it 19:15 < EvilRoey> oh that's a good question 19:15 < jelly> smallville7123: apps are tiny. datasets are orders of magnitude larger 19:16 < blasen> hello. i have copied linux partition from usb onto a harddrive. i have marked that partition as 'active'. trying to boot from that disk, it failed, obviously unable to find the proper bootsequence. now it's not the 'first' partition on that disk, could it be the reason it fails? or what else would i need to care about in order to make it work? 19:16 < jelly> smallville7123: I _think_ linux dlopen()/dlsym() map the whole library but the kernel only loads what's needed when it's needed. 19:16 < djph> a shared lib is (more or less) always in memory 19:16 < jelly> smallville7123: but again, it's probably the wrong place to optimize anything 19:17 < djph> but only *once* 19:17 < jfe> hello 19:17 < smallville7123> for example if an application only calls printf, could the dynamic linker load libc in such a way so that only the required information for the successful execution of the printf function is loaded into memory and everything else that is not needed would be discarde 19:17 * jelly gives djph a COW 19:17 < smallville7123> discarded* 19:17 < djph> jelly: cowsay moo? 19:18 < jelly> djph: that thing about a library loaded only once is either a halftruth or a big fat lie. I'm not sure which. 19:18 < smallville7123> obviously taking into account for recursive function calls and initialization functions required and what not 19:19 < lnnb> and global variables 19:19 < jelly> given ASLR is a thing, I'd guess probably more the latter 19:19 < smallville7123> for example, vprintf i think is a function directly called by printf and so on for any other function 19:19 < djph> if you call to a shared library, that *specific* shared object is loaded only once into memory, and all programs linked against that shared object make calls to the relevant portions of RAM that that shared object resides in. 19:19 < jfe> i've noticed that on systems where 1000's of TCP connections are made per second, a single connect() call to a non-existant unix socket will hang the system temporarily. 19:19 < lnnb> its nightmare tier work with almost no payout 19:20 < smallville7123> Wonder if a statically compiled application has a lower memory usage than a dynamically compiled application 19:20 < jfe> does anyone have a clue as to why that might happen? 19:21 < smallville7123> Since the statically compiled application is compiled with only what is needed from each library it may depend on 19:21 < jfe> my devops guy has noticed this happening on haproxy nodes. when we run one of our programs that connect()s to a non-existant socket our grafana graphs (including cpu usage) drops to zero for up to a minute 19:21 < djph> this is, of course, the polar opposite to a static library (which is compiled into the binary) 19:22 < djph> shared objects get mapped at runtime 19:23 < djph> smallville7123: no, the statically compiled executable is larger - it pulls in the entire library 19:24 < smallville7123> (◉-◉) 19:24 < smallville7123> so even if it only contains printf, it will ALSO contain the enirety of all other libc functions aswell as printf itself? 19:25 < djph> but it's faster 19:25 < Psi-Jack> Yes 19:25 < smallville7123> why? 19:25 < Psi-Jack> Do you have a Linux question? 19:27 < blasen> smallville7123: but dynamically linked progs wouldn't share memory for the functions/libs both make use of? or would they? 19:27 < lnnb> might be able to mess around with linker options to remove code 19:28 < lnnb> unused code, confusing stuff but something like -ffunction-sections -Wl,--gc-sections will clean up unused symbols from a c file 19:29 < bls> it depends on how the library is coded. if you put all your functions in a single object file, yes, it'll link in the entire thing. but most static libraries break code out into separate object files, and only the ones you need are included 19:30 < Frith> I'd also venture that static libraries are slightly less safe as well, in that the position of the code is fixed at link time. 19:30 < Frith> As in, "ever so slightly". 19:32 < blasen> smallville7123: .net applications eventually would .. there is sth. called global assembly cache, i can imagine it could manage loading and sharing of libs among processes, but that's only a guess. 19:34 < djph> Frith: why would static linking be 'unsafe'? 19:34 < smallville7123> i guess that i am asking is could it be possible for shared objects to be broken down into segments of functions and global data, then attempt to disregard as much unneeded segments/data as possible 19:35 < djph> no, you can only break them down as far as the library is itself broken down. 19:35 < smallville7123> like what bts said except applied to compiled objects instead 19:35 < djph> Once you've finalized the *.so file, it's as big as it is. 19:35 < smallville7123> bls* 19:37 < lnnb> maybe you can load it, then force the whole lib to be swapped out of memory 19:37 < lnnb> would it only load in pages as needed? 19:37 < djph> er, what? 19:38 < smallville7123> for example, via attempting to disassemble and identify functions and global data, then construct a new file consisting of what is needed based on the disassembly and global data, in relation to what code possible to be executed 19:38 < bls> what would be the point of thinning out a shared library even more? 19:39 < bls> you'd spend a lot of time and effort for no real value other than saving a couple pages of memory 19:39 < smallville7123> wich is optimized for the specific binary being loaded 19:40 < djph> smallville7123: that makes no damn sense 19:41 * Pentode puts some dressing on that word salad 19:42 < smallville7123> In which if applied to all of linux's existing running applications, could save upto half memory tho obviously the actual amount saved varies greatly depending on the application complexity 19:42 < bls> you'd probably end up wasting more memory than you saved, because you'd likely have to isolate methods on a page by page basis 19:42 < djph> a random garbage program that includes stdio.h links against these three shared libs (memory addresses removed, because space) -> linux-vdso.so.1; libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 ; /lib64/ld-linux-x86-64.so.2 19:43 < bls> that assumes that 50% of all the methods of all the libraries on the system have zero callers 19:43 < djph> ^ 19:43 < Pentode> it would just become a huge mess really fast. 19:44 < CrazyTux> Opensuse Leap 15 is now available in live versions. 19:44 < lnnb> wooooooooooooooohooooooooo! 19:45 < triceratux> yes but only in gnome & kde. ill stick to my tumbleweed. xfce is the only hope 19:45 < smallville7123> obviously it would depend on how optimized the application itself that is optimizing the other applications is 19:45 < djph> smallville7123: you do understand the concept that *shared* libraries get loaded into memory *once* and then everything that needs something out of one (or more) of those libraries just use a pointer to the relevant memory location(s), right? 19:45 < CrazyTux> triceratux, is tumbleweed stable enough for an end user? 19:46 < smallville7123> yes 19:46 < bls> so you'd either go with a naive/performant/wasteful version that puts a single method on a page, or the dynamic linker is having to constantly pull individual methods in into memory, intelligently organize them, eject them, etc 19:46 < triceratux> CrazyTux: nope. its more broken than arch even 19:46 < CrazyTux> triceratux, then better to stick to leap 19:47 < triceratux> CrazyTux: lubuntu lxde is no more. its been converted to lxqt in the cosmic daily builds. there is no going back 19:47 < CrazyTux> I must say, Opensuse looks more professional. 19:47 < triceratux> theyve been doing it a long time 19:47 < CrazyTux> triceratux, great. I wanted to test lxqt. 19:47 < smallville7123> tho my definition of shared memory probably is not the same as yours 19:47 < djph> smallville7123: okay, so since a shared library is in memory *one time*, regardless of how many programs are sharing the resource; how do you figure you can save 50% of your RAM? 19:47 < bls> the complexity of the management overhead would far outweight the small memory savings. if you're that memory constrained, you're not going to be using things like glibc, boost, etc 19:48 < triceratux> CrazyTux: im running it right now from the 18.10 nightly. it makes lubuntu look better & gets most of the gtk out of the way. its quite cool actually 19:49 < bls> we used to separate out libsocket, libm, libresolve, etc, but nearly everything is using them constantly, so we stopped 19:50 < triceratux> CrazyTux: its not like an ordinary ubuntu nightly. its been good to go for several releases. its actually quite finished & stable http://cdimage.ubuntu.com/lubuntu/daily-live/ 19:50 < CrazyTux> triceratux, so, the next lts release is going to be only with lxqt? 19:50 < bls> put them in glibc, load them once, enjoy the efficiencies of not having to go to disk for them over and over, let swap eject the unused pages 19:51 < CrazyTux> triceratux, is lxqt more userfriendly than lxde? 19:51 < sloshy> lxqt is very new 19:52 < sloshy> it is not as mature as lxde is 19:52 < sloshy> i find lxde easier 19:52 < triceratux> CrazyTux: not just that, the last LXDE release was 18.04. so its lts & itll be around. & folks still need LXDE for the raspis. but LXQt is actually maintained & theres reasons why its absurd to think LXDE will be converted for GTK3. its not going to get any more maintenance 19:52 < sloshy> but really they are pretty similar, the difference is if you prefer qt or gtk apps 19:52 < CrazyTux> ok 19:53 < sloshy> i just stick with xfce 19:53 < sloshy> even though it is not actively developed at all 19:53 < sloshy> that can be a good thing sometimes 19:53 < TRS-80> afternoon fine people 19:53 < CrazyTux> I think KDE is the easiest for windows users. 19:53 < triceratux> CrazyTux: theyve put some work into the panel & the filemanager & it almost seems as robust as xfce. & their sessionmanager is written to make it trivial to change windowmanagers. formally, openbox isnt part of LXQt in fact. its just a widely observed convention 19:53 < sloshy> gnome is actively developed, so you get new bugs all the time 19:53 < qrvpzvb> does anyone play overwatch? on linux? with wine? 19:54 < alexandre9099> bookworm, oh, i'm dumb :D virtualhost had :443 instead of the new port :8443 19:54 < sloshy> kde imo offers too much customization 19:54 < sloshy> and seems to crash a lot 19:54 < Pentode> ditto 19:54 < hexnewbie> KDE offers the full Windows experience, with crashing and bugs and all. :) 19:54 < CrazyTux> sloshy, so which DE for windows users then? 19:54 < sloshy> it takes a long time for me to set up kde the way i like it 19:54 < sloshy> i recommend xfce 19:54 < Pentode> i've been testing KDE off and on for years and years and I still feel that way 19:54 < CrazyTux> and I think Zorin makes linux a lot easier to newbies. 19:54 < sloshy> it is easy to make it like windows 19:55 < sloshy> just drag the taskbar to the bottom 19:55 < TRS-80> trying to compile encfs 1.7.4 from source (https://github.com/vgough/encfs/tree/v1.7.4), looks like it needs GNU autotools which I have installed but first time I have done this and I seem to be stuck 19:55 < triceratux> CrazyTux: on lubuntu 18.10 lxqt i was able to install kstars in only 7 debs. thats a yuugh kde app running on a desktop as light as lxde. you cant touch that with trinity 19:55 < TRS-80> I love KDE but then again I came from Windows. LOL 19:55 < CrazyTux> triceratux, ok 19:55 < sloshy> xfce is like classic windows, kde is more like windows 7 and newer 19:55 < Pentode> sloshy, im using the Chicago themes with xfce on my thinkpad. once and a while I get a "my god, are you still using windows98?" lol 19:56 < TRS-80> I actually love all the customization options, wouldn't have it any other way, I personally find Gnome, etc. very limiting like iPhone with 1 button 19:56 < CrazyTux> triceratux, the current version of Zorin OS is not based on Ubuntu 18.04. 19:56 < CrazyTux> sloshy, I use redmond theme on ubuntu mate. 19:57 < sloshy> TRS-80: https://www.youtube.com/watch?v=BGGOn-H7s3Q 19:57 < CrazyTux> redmond panel. 19:58 < cxc99> can a krb5p nfs share work with another local-only user besides root or does that defeat the purpose? 19:58 < triceratux> CrazyTux: nope. zorin 12.3 updated to 16.04.4 which for now is a better move. almost noone is making their clones of 18.04 yet, the notable exception being voyager 19:58 < TRS-80> sloshy: that's great XD 19:58 < Pentode> lol 19:58 < Pentode> cute 19:59 < CrazyTux> triceratux, OK 19:59 < cxc99> never mind read the man page 19:59 < smallville7123> So what if you have two different libraries that have the some of the same functions except named differently for each library 20:00 < bls> if they're different libraries, they're not the same functions 20:00 < djph> then you did "shared objects" wrong ) 20:00 < djph> :) 20:00 < smallville7123> for example, a.so has foo() and b.so has bar() and foo() and bar() are 100% identical to eachother except for there function names 20:01 < smallville7123> but could be for entirely different applications without knowing it 20:01 < CrazyTux> triceratux, have you used Zorin before? 20:02 < smallville7123> if so wouldnt you have two of the exact same function in memory? 20:02 < bls> they're in no way 100% identical. they're from different libraries and have different names. you have absolutely no way to affirm that they're 100% the same 20:02 < smallville7123> shared memory* 20:03 < djph> smallville7123: no, but then they're also *not* using a shared library. 20:03 < triceratux> CrazyTux: yep its on my shortlist of xfce clones i keep bootable. its a mint wannabe like lite or voyager. & thats alot easier these days with the state mint is in 20:03 < djph> s/using// 20:03 < danieldg> smallville7123: yes, you can have the same code more than once in memory in that case 20:03 < TRS-80> OK so I did automake and autoconf and they generated a configure, I do ./configure and it appears to do some stuff but when I do make I get: make: *** No targets specified and no makefile found. Stop. 20:03 < smallville7123> for example if both foo() and bar() where just "int foo() { return 0; }" and "int bar() { return 0; }" 20:03 < djph> I mean, they themselves are each "shared objects"; but ultimately they're not the same. 20:03 < CrazyTux> triceratux, why? what happened to mint? 20:04 < danieldg> we're quibbling about the definition of "same" here 20:04 < smallville7123> and both compiled in the exact same way 20:04 < TRS-80> what the definition of "is" is 20:04 < danieldg> smallville7123: why do you care? What's the reason for asking? 20:05 < djph> whoever wrote the shared objects is at fault. However, that doesn't really happen that often (more likely you'd have a.so and b.a - one being a shared lib, and the other being compiled into the program it's acting as a library for) 20:05 < sloshy> 3rd person indicative of "to be" 20:05 < sloshy> but try telling bill clinton that 20:05 < triceratux> CrazyTux: its not just their security, but they dont ship the codecs on the iso anymore now that the rest of the distros do. its now an extra admittedly trivial download 20:05 < TRS-80> sloshy: that's the joke XD 20:05 < bls> how often do you suppose something like that happens? where the code is so dead simple there's no possible way the compiler can't produce the exact same machine code that has the exact same behavoir and side-effect, and is using enough memory where it's worth it to figure that out and eject one of them? 20:06 < FManTropyx> dpkg failed 20:06 < CrazyTux> triceratux, ok. Do you consider Zorin to be better than Mint? for beginners? 20:06 < smallville7123> it doesnt NEED to be that simple, it just needs to be the same in rare cases 20:06 < CrazyTux> triceratux, has Mint not sorted out the security issues yet? 20:06 < TRS-80> OK I guess I'll try the GNU channel... 20:07 < bls> if it's rare, then at what ratio is the rarity of it happening worth the work required to detect it? 20:07 < triceratux> CrazyTux: theres an argument to be made for running base distros instead of their derivatives, considering all the derivatives do is put everything together a certain way which you can do yourself anyway. my vies is not merely to use a preconfigured downstream derivative, but to use the best one because a lot of them are indeed lousy 20:07 < djph> smallville7123: in terms of the *standard* libraries (e.g. glibc); the chances of that happening are ... well, not that high. 20:07 < bls> at that point, you're essentially asking for deduplication of memory 20:07 < triceratux> CrazyTux: mint & zorin are both fine at this time. theyre right in the same ballpark 20:07 < danieldg> smallville7123: you might save a few KB that way, but it comes at a cost of having to do a lot more fetching from disk for each program 20:08 < triceratux> *view 20:08 < CrazyTux> triceratux, ok 20:08 < CrazyTux> triceratux, so, you prefer Ubuntu to mint or zorin. 20:08 < nixfreak> anyone using tahoe-lafs for file sync 20:09 < triceratux> CrazyTux: but also the days of mints primacy are nearing their end with the ascension of manjaro to the top of the distrowatch 100. it may seem merely symbolic but it means something about the arch worldview has prevailed over the deb worldview 20:09 < smallville7123> Or (taking this one step further) assuming all function names are irrelivant, it could be applied at the instructional level and, tho then it ultimately depends on what the size of a single instruction is in memory 20:10 < CrazyTux> triceratux, I agree. I am using Manajaro also. 20:10 < TRS-80> nixfreak: I just came across that the other day, actually... 20:10 < danieldg> smallville7123: instructions are usually smaller than memory addresses, so you can't deduplicate them 20:11 < smallville7123> Tho then again it also depends on how many of that instruction there is 20:11 < triceratux> CrazyTux: in a threeway between mint, zorin & ubuntu i wind up saying its all the same crap. which is good because when something special like mx-17 comes along you notice ;) 20:11 < bls> it makes way more sense to just eject unused pages to swap than to try to deduplicate executable memory, which would only be safe if the pages were marked W^X 20:11 < smallville7123> Total across all shared memory 20:11 < CrazyTux> It has been exceptionally stable so far. But, I keep it updated always. I am not sure what would happen if I don't update it say for more than a month. 20:11 < danieldg> smallville7123: no, a deduplication itself consumes memory 20:12 < smallville7123> How 20:12 < triceratux> CrazyTux: your manjaro ? we are curious too. let us know 20:12 < danieldg> smallville7123: more pointers to where in memory the shared object is 20:12 < TRS-80> nixfreak: I came across it while reading about I2P, but right now it's just "on my radar" 20:12 < CrazyTux> triceratux, but, why MX 17 doesn't have its default kernel patched for spectre and meltdown? 20:12 < danieldg> so you have one "add" instruction and 10000 pointers to it 20:12 < danieldg> that's not less memory than 10000 adds 20:13 < smallville7123> lol 20:13 < mattfly> software recomendation, is there any simple fast ocr program / gui front end , that makes easy the process of copying the processed text of a image into text? 20:13 < smallville7123> ;-; 20:13 < mattfly> to the clipboard 20:13 < CrazyTux> triceratux, MX is my favorite too. 20:13 < bls> danieldg: haha, a 1:1 compression algorithm! 20:14 < triceratux> CrazyTux: cause they release sound stuff once a year & the likelihood of being exploited by spectre / meltdown is relatively small despite the hype 20:14 < TRS-80> mattfly: I heard good things about OCRFeeder 20:14 < CrazyTux> triceratux, ok 20:15 < triceratux> CrazyTux: its not like every server in the cloud & every 2.4 box in redhat shops is going to suddenly be running 4.17 either 20:15 < mattfly> i would like a practical thing, like select and its copied to clipboard 20:15 < djph> danieldg: smallville7123 however, 10000 pointers to "int foo (var1, var2) {...}" is considerably smaller than 10,000 copies of that function. hence shared objects. 20:15 < CrazyTux> triceratux, ok 20:15 < djph> oops, didn't mean to tag you danieldg :| 20:16 < bls> was going to say pdfsandwhich or tesseract, but that's not close 20:19 < TRS-80> mattfly: I think OCRFeeder is about as good as it gets on GNU/Linux 20:19 < mattfly> im testing it 20:19 < TRS-80> with F/LOSS I mean 20:19 < mattfly> thanks for telling me 20:20 < TRS-80> bls mattfly it's front end GUI for Tesseract (among other things) 20:29 < pankaj> I googled this topic but could not get a complete and most important picture -> 'What is the most important use of header file and why to use it?' 20:29 < TRS-80> pankaj: study your textbook 20:31 < pankaj> TRS-80: You are right. But I think I forgot that. 20:32 < pankaj> TRS-80: I know it is used for function declaration but the article said that now it has all changed due to advancement but still it has some important function. 20:35 < FManTropyx> how does ? turn into f? 20:35 < lnnb> ? turn into f? 20:36 < barometz> FManTropyx: that question will need some context 20:36 < FManTropyx> now I have a fatal error: no input files!!! 20:36 < barometz> aha 20:36 * barometz scrolls 20:38 < FManTropyx> ahh... there's a directory called f! thanks for the help 20:38 < bls> :| 20:45 < alexandre9099> i got unmetered internet connectin, tell me distros to seed :) (i am already seeding ubuntu/ubuntu server, kali, all linux mint distros, kali and arch) 20:48 < bls> the top ones on distrowatch.com? 20:50 < alexandre9099> bls, oh, how haven't i tought about distro watch :) 21:01 < vutsuak> hey /var/log/somefolder/somefile.... always require user permission to write something in this location. Is there a way to prevent this.... even if somebody else clone a repo a runs on his machine.? 21:03 < Elec_A> Hi, I have an ethernet loopback adapter, how can I test it ? 21:03 < mad_hatter> trying to test a gdpr ip from the us. is there anyway i can change my ip to one coming from the EU? 21:03 < bls> mad_hatter: get a VPN 21:04 < bls> vutsuak: you can set it up like /tmp 21:04 < mad_hatter> bls: do you know of any free vpns that are EU based? 21:04 < bls> vutsuak: where everyone can write into it, you just can't delete other people's stuff 21:04 < bls> mad_hatter: no, don't use one and not really on topic for this channel 21:04 < vutsuak> @bls thanks I will try that.. but can it be overwritten too..? 21:05 < bls> vutsuak: overwritten by whom? you can't overwrite someone else's files in /tmp if it's configured properly 21:07 < vutsuak> @bls okay thank you.. 21:11 < saran> hey guys, I have some errors in my journal when trying to use optirun/primusrun, any ideas why? http://dpaste.com/2X2QPES 21:11 < RayTracer> alexandre9099: last time I tried the download of raspbian images with torrent was much slower than direct dl 21:12 < alexandre9099> i'll also see that one, thanks (I also have an RPi, but haven't used it in a while) 21:16 < sari> hello people, i am having an issue with pip3 on ubuntu 18.04 21:16 < sari> server... i keep getting ImportError: cannot import name 'sysconfig' 21:16 < sari> what is going on? 21:16 < lnnb> pip is that python package manager? 21:17 < stmiller> sari: sudo apt install python3-distutils 21:19 < sari> stmiller: thanks, but now when i install something i get this error: Command "python setup.py egg_info" failed with error code 1 in /tmp/pip-build-qozik_lv/ebaysdk/ 21:20 < smallville7123> "04:15 djph: danieldg: smallville7123 however, 10000 pointers to "int foo (var1, var2) {...}" is considerably smaller than 10,000 copies of that function. hence shared objects." For example, converting instruction sequences into functions should there be more then x of them 21:21 < smallville7123> then replacing them with calls to that function 21:22 < smallville7123> tho that is only applicable if the multiple sequences match eatchother 21:22 < danieldg> that can have performance impacts due to decreased locality 21:22 < danieldg> function calls aren't free 21:22 < smallville7123> like pattern matching 21:22 < danieldg> if "memory use" is your only performance metric, then yes, that'd be a win 21:23 < solidfox> danieldg, decreased locality? 21:23 < smallville7123> like shared functions lol 21:23 < lnnb> instruction cache 21:24 < smallville7123> lnnb: i doubt every single instruction is cached 21:24 < danieldg> it's a tradeoff 21:24 < solidfox> you can't cache instructions where pointers to functions are used repeatedly instead of copies? 21:24 < danieldg> sometimes you'll win, sometimes it's not worth it 21:24 < lnnb> smallville7123: all memory is cachced 21:24 < danieldg> if you can manage to unify a 10KB function, it's more likely worth it than a 300-byte loop 21:24 < smallville7123> Like cache 5 million instructions 21:25 < lnnb> i mean, if it goes through the main CPU 21:25 < smallville7123> As the cache can only hold x amount 21:25 < lnnb> yeah that's why locality is important 21:25 < smallville7123> I cant remeber exactly how much 21:25 < solidfox> I fail to see any advantage to having a copy of a function, but I have no idea 21:25 < solidfox> I'm a n00b 21:27 < solidfox> I guess it's kinda nice if your program runs on its own, instead of using .dll or anything like that 21:27 < smallville7123> solidfox: the advantage would be instead of having say 480 occurences of a 12 line instruction sequence, u could only have 480 calls to a single 12 line instruction sequence 21:27 < solidfox> plus dll can be created to inject arbitraty code 21:28 < jim> please spell out u as you, it helps people (particularly new english speakers) to understand, at least, most of what's going on 21:28 < smallville7123> you* 21:28 < solidfox> smallville7123, that seems backwards 21:29 < solidfox> 480 occurences of 12 instructions sounds like copies of functions to me. 21:29 < jim> thanks 21:30 < solidfox> 480 calls to a single 12 line instruction sequence seems like using a single function, maybe pointers to it too 21:31 < smallville7123> It can be optimized at compile time but if you have multiple files that share more or less of the same code, if they where to be merged into a single file that code would probably be optimized into just a single section of code instead of multiple sections of the same code, however since they where compiled separately you now have duplicate code being loaded from different files and functions as t 21:31 < smallville7123> here could be millions of duplicate code segments spread about the shared memory originating from different functions and files 21:31 < danieldg> smallville7123: -flto 21:31 < lnnb> smallville7123: calling a function is not free, look up cdecl calling convention 21:33 < solidfox> totally lost now. Sorry I even joined this convo 21:33 < solidfox> conversation* 21:33 < lnnb> and it probably screws up the cache depending on how big it is and where it's located, but maybe not 21:33 < danieldg> calling a function can take a good 6-12 instructions on its own, due to register saving/restoring 21:34 < danieldg> usually that's not obvious because the compiler will spread those around 21:34 < lnnb> so you save a few instructions maybe, but it doesn't mean it's going to happen faster 21:34 < solidfox> danieldg, what is the alternative to calling a function? inline? 21:34 < danieldg> yep 21:34 < solidfox> ahhhhh I get it now 21:35 < danieldg> of course, that has its own cost... extra code size, you fit less in memory/cache 21:35 < danieldg> everything's a tradeoff 21:35 < bls> and the justification seems to be 1. libraries are packed with functions no one calls and getting rid of them would save magnitudes more memory than the use ones and 2. a typical linux distro is rife with massive copy-pasted functions 21:35 < smallville7123> how is inlining the code any smaller then having multiple copies of it 21:36 < danieldg> inlining can drop register spills that aren't required, or avoid moves to common arguments 21:36 < solidfox> smallville7123, the function prologue takes up instruction time, and also maybe (I think) executing the function prologue uses space on the stack 21:36 < danieldg> or constant propagation 21:36 < lnnb> yeah you have to store return pointer on stack and parameters 21:36 < solidfox> as well as local args 21:36 < lnnb> err parameters on i386 if it's more than some number i forget 21:36 < solidfox> sorry. as well as local variables 21:37 < lnnb> arch specifics make it annoying 21:37 < smallville7123> So would it even be possible to optimize it? 21:37 < solidfox> this is why I write everything in main 21:37 < danieldg> linux x86_64 passes 6 integer/pointer args in registers, and clobbers 8-ish registers 21:38 < smallville7123> since your basically saying any attempt at optimizing it will be bad 21:38 < misternumberone> Hi I get segmentation fault when running grub-mkrescue 21:38 < misternumberone> specifically sudo grub-mkrescue --modules="lvm cryptodisk scsi ahci disk luks" -o floppy.img ./img 21:39 < misternumberone> sorry wrong paste it's sudo grub-mkrescue --modules="lvm cryptodisk scsi ahci disk luks" --compress xz -o floppy.img ./img 21:39 < danieldg> smallville7123: no, I'm saying that anything could be bad, and you'd need to test to see benefit/cost 21:39 < misternumberone> i.e. with compression, and when I use it without the compression it works 21:40 < solidfox> misternumberone, that's a bit odd for xz to seg fault. 21:41 < solidfox> misternumberone, is xz installed 21:41 < smallville7123> Ok 21:42 < misternumberone> I just tried it with gz instead, and then with just nothing after --compress, and then with just --compress for the same results 21:42 < bls> looks at it like zram. you're trading CPU cycles for memory capacity. 21:43 < misternumberone> "grub-mkrescue --compress" returns "Segmentation fault" 21:43 < nixfreak> anyone know if there is a way to get a signed ssl cert of ddns ? 21:43 < bls> is free memory more important than having CPU capacity? use zram. if you've got plenty of ram and need to keep your cores spinning, zram is a bad deal 21:46 < solidfox> I didn't realize C programs can use inline as well. Thought it was only a c++ thing 21:47 < djph> nixfreak: you mean like a ssl cert for youdomanin.dyndns.net or whatever? 21:48 < Elec_A> Hi, are eth adpaters anywhere in /dev ? I want the address of e.g eth2 to use with "i2c" command 21:48 < sari> people i am in a root in terminal (on a server) how do i exit root and into a spesific user? 21:48 < smallville7123> Could the function call be replaced with a jmp instead 21:48 < bls> Elec_A: no 21:49 < bls> sari: sudo -u theuser 21:49 < smallville7123> and then end with a dynamic jump 21:49 < misternumberone> solidfox: xz is installed also gzip also lzop, which are all the encryption options grub-mkrescue --help lists 21:49 < oleo> sari: sudo su 21:50 < sari> bls: sudo -u USER from ROOT? 21:50 < smallville7123> that is determined during the execution of the code 21:50 < oleo> sari: or just su depending on if you have to use sudo or not.... 21:51 < sari> oleo: i keep getting no passwd for user 21:51 < oleo> sari: well you have to have one 21:51 < sari> ok, i used a different user, but now nothing happens\ 21:51 < sari> i remain in root 21:52 < oleo> sari: which user do you try ? 21:52 < sari> odoo 21:52 < sari> sudo su odoo 21:52 < oleo> sari: it has to be a user with <= ID 1000 and a home 21:53 < solidfox> smallville7123, call is an instruction that jumps. but to use jump instead, you'd need to manually push the return address as well as arguments 21:53 < sari> oleo: can i get a lisy of users to make sure mine is there or not? 21:53 < solidfox> smallville7123, so I don't think any time is saved using jmp instead of call 21:54 < Gnjurac> hi anyone know how can i make a script run on hid usb device connect 21:55 < Gnjurac> like if 28bd:0042 device connect run script 21:55 < oleo> sari: cat /etc/passwd 21:55 < smallville7123> solidfox: it depends if it takes arguments or not 21:56 < solidfox> smallville7123, do you know algebra 21:56 < kurahaupo> Gnjurac: add a config line in udev rules 21:56 < oleo> sari: you should have something like odoo:x:1000:1000::/home/odoo:/bin/bash 21:56 < smallville7123> it mainly depends on if the register contents are preserved before and after the jump 21:57 < oleo> sari: and a /home/odoo folder, if you used the useradd commandline utility 21:57 < Gnjurac> kurahaupo, ok will reasearch how to do that 21:57 < Gnjurac> ty 21:57 < smallville7123> as it wont actually be a function, it will only mimic one but have absolutely no return value of any kind 21:58 < solidfox> push eip; jmp func; is the same as call func 21:58 < smallville7123> eg jmp xxx .... do stuff ... jmp yyy 21:58 < solidfox> smallville7123, yyy is the return address 21:58 < solidfox> push eip 21:58 < solidfox> sets yyy 21:59 < oleo> sari: useradd -d /home/odoo -u 1000 -g 1000 -G audio,video,users -p something like that 21:59 < solidfox> smallville7123, if you add arguments it's the same. just like an equation in algebra. since you add the arguments to both sides of an equation 21:59 < smallville7123> even if it takes no arguments? 21:59 < oleo> sari: you can also use the -m there to let it create the home folder 22:00 < oleo> sari: just check useradd --help 22:00 < solidfox> smallville7123, the first assmbly I sent just now is no arguments 22:00 < oleo> sari: or man useradd 22:00 < solidfox> push argv; push argc; push eip; jmp func; is the same as push argv; push argc; call func; 22:00 < solidfox> again no arguments would be: 22:00 < solidfox> push eip; jmp func; is the same as call func 22:01 < solidfox> since call pushes eip 22:01 < solidfox> then jmp 22:01 < smallville7123> why push eip? 22:01 < solidfox> smallville7123, it's the place to return to 22:01 < solidfox> smallville7123, your yyy 22:01 < solidfox> aka a return address 22:01 < smallville7123> ok 22:02 < solidfox> saving 2 or 3 instructions would be some very embeded device software :) 22:02 < solidfox> I'd hate to work on that I tihnk 22:02 < smallville7123> "06:02 solidfox: saving 2 or 3 instructions would be some very embeded device software :)" per what? 22:03 < solidfox> smallville7123, saving 2 or 3 instructions per would-be function call 22:03 < smallville7123> xD 22:03 < kurahaupo> It's push eip+(length of the following jump instruction) 22:04 < solidfox> kurahaupo, thanks. forgot about that 22:04 < solidfox> my assembly was an infinite loop 22:04 < smallville7123> Lol 22:04 < solidfox> no instructions were saved :P 22:05 < smallville7123> Wonder how much memory would be saved if this got impliented 22:05 < smallville7123> Like in a working usable state 22:05 < solidfox> smallville7123, I think some ncurses functions use shared object files 22:05 < solidfox> some ncurses programs I mean 22:05 < solidfox> nah I should stop talking. this goes beyond what I know 22:07 < smallville7123> It could save over half the memory considering the sheer complexity and length of assembly in comparison to mid-high level languages 22:08 < smallville7123> As a simple 50 line C file could be like 270 lines of assembly lmao 22:08 < solidfox> so that's what. 1kb? 22:09 < ayecee> a lot more than that 22:09 < ayecee> a lot more lines than that, that is 22:09 < smallville7123> Considering average is around 1 GB to 3 GB memory usage 22:10 < solidfox> depends on what kind of data you're loading too 22:10 < solidfox> there are other segments to a program 22:10 < smallville7123> True 22:11 < solidfox> printf(user_input) instead of puts(user_input) and printf("%s\n", user_input) 22:12 < solidfox> this will save stack space, and instruction space >:) 22:12 < solidfox> highly recomend 22:12 < ayecee> citation needed 22:12 < smallville7123> At the very least id say about 3 quarters of all total application code can be greatly reduced 22:12 < ayecee> seems like one of those "premature optimizations" that i keep hearing about. 22:13 < ayecee> the ones that don't really change performance but make it harder to read and maintain 22:13 < lnnb> shared library isn't application code though 22:13 < djph> ayecee: more "it's a simple matter of programming" 22:13 < ayecee> shared libraries gotta be maintained too 22:13 < ayecee> djph: heh 22:14 < smallville7123> leaving more room for resource space 22:14 < djph> ayecee: and two hours ago, he was on that two different (standard?) shared libs would both implement the same code, just one named the function 'int foo() {...}' vs. 22:14 < djph> *vs. 'int bar() {...}' 22:15 < solidfox> lnnb, do shared libraries load a copy of the function for each program that uses them? 22:15 < solidfox> lnnb, that's what I wonder about 22:15 < solidfox> lnnb, it's shared, maybe only on disk. maybe 22:15 < lnnb> it only copies if you try to modify the address space i think 22:15 < lnnb> CoW 22:15 < djph> solidfox: no, that's not how it works. 22:15 < solidfox> ok 22:16 < solidfox> gtg! have a good weekend everyone 22:16 < solidfox> whoops 22:16 < twainwek> don't tell me what to do 22:16 < solidfox> gtg is "got to go" 22:16 < smallville7123> "06:13 lnnb: shared library isn't application code though" technically it is 22:16 < djph> if you ink to libc.so.6 (for example), you get some memory address at runtime (i.e. where libc6 happens to be loaded). Your program, and every other program linking against libc6 points to there. 22:16 < lnnb> it does not belong to your application, it's a system resource 22:17 < smallville7123> still has executable code 22:17 < kubast2> Can I do "find something -exec ./test.sh {} \; 22:17 < lnnb> sure you can split hairs if you want to be ignored 22:17 < djph> kubast2: you're missing a closing quote, but otherwise it seems to check out. 22:17 < longxia> lnnb: shared libraries are not exclusive to system libs. Any program can compile its own shared objects. 22:18 < kubast2> I see thx 22:18 < lnnb> and any program can load symbols from it 22:18 < smallville7123> Yup 22:18 < djph> kubast2: you may want to provide more help (e.g. find . -name "someglob.*" [...]) 22:19 < smallville7123> As long as it can be executed in some way it counts as executable code 22:19 < lnnb> smallville7123: how big is your kernel in memory? 22:20 < smallville7123> i dont know 22:23 < sari> ls 22:23 < jim> sari: ls: command not found 22:28 < FlorianBd> Hi there! If I make a read-only system and I want all logs (syslog, messages, etc) to be sent to a remote server, what would be the best approach? I am thinking about writing to ramfs and then having a script that empties these files right away and sends their content. But is that clean? .... 22:29 < ayecee> the normal way would be to log to syslog, and have syslog send logs elsewhere 22:30 < FlorianBd> so absolutely all kernel logs can be logged via syslog? Even btmp, wtmp, and messages? 22:31 < ayecee> i don't know what btmp is. wtmp isn't normally considered a log file. messages is normally created by syslog in the first place. 22:31 < FlorianBd> Ok. How does the kernel communicates with syslog by the way? Is there a socket or something? 22:32 < ayecee> klogd reads the kernel ring buffer and sends the messages to syslog 22:32 < FlorianBd> so it's it's something in /proc ? 22:33 < FlorianBd> (that ring buffer) 22:34 < Gnjurac> kurahaupo, hmm i aded this to my udev to test, ATTRS{idProduct}=="0042", ATTRS{idVendor}=="28bd" , RUN+="/home/gvozdje/Documents/APPs/runme.sh" 22:34 < Gnjurac> it works but isse is it runs my scrip like 10 times when i put it in and 10 times when i unplug 22:34 < Gnjurac> do you meybe know why meybe i need to make match better 22:34 < Gnjurac> and how can i run it only on plugin 22:35 < jim> do you have more than one device that matches that vendor and product? 22:36 < Gnjurac> hmm not butt cuz this is tablet for drawing it simulates mouse , buttons so probbaly more hids , will metch by lowest chidren i guess 22:37 < Gnjurac> tought still duno why it runs on uplug too 22:39 < smallville7123> Solidfox: take this for example, https://en.wikipedia.org/wiki/Kernel_same-page_merging, for simplicity say they are 1111, 1112, and 1113, and each is on a page, the 3 pages could be valid as all contain "111" but are consider invalid as they contain 1, 2, and 3, at the end respectively and thus cannot be merged 22:41 < jim> so then you have more than one device that works like a mouse? 22:43 < smallville7123> however if applied byte by byte then this approach then becomes feasable on thi scale as it is not limited to a certain length in order to be considered valid 22:46 < ayecee> if only pages were byte-sized 22:46 < lnnb> i have an *entire system* sitting in ramfs that i use just to run bash script to load modules wastefully idling there because i'm lazy and didn't delete them from initramfs, the whole thing only accounts for like 220MB 22:47 < ayecee> imagine the massive page table you'd need for byte-sized pages XD 22:47 < djph> except now you need to do a lotta work to follow 1 -> 1 -> 1 -> {1,2,3} 22:47 < djph> ^ said it better than me ayecee 22:48 < smallville7123> .lmfao 22:48 < smallville7123> lmfao* 22:49 < rud0lf> .lmao - hidden laugh? 22:49 < smallville7123> xD 22:50 < lnnb> some day i'll get around to compressing that into a 50K c program 22:52 < smallville7123> ^-^ 22:53 < smallville7123> the day when everything is super compressed t 1 MB 22:53 < smallville7123> to* 22:53 < smallville7123> With advanced technology 22:53 * ayecee compresses smallville7123 to 1 MB 22:54 < smallville7123> (◉-◉) 22:54 < smallville7123> ( ▀ ͜͞ʖ▀) 22:55 * smallville7123 self compresses to 1 byte and becomes malware 22:55 < smallville7123> Imma 1 byte virus xP 22:58 < smallville7123> 06:54 Fatalnix: https://i.redd.it/sbr4veiar0011.jpg 22:58 < xx2> is there a program for linux like shazam or soundhound? 22:59 < Bl4ckC0r3> hello 22:59 < Bl4ckC0r3> need some help here please 22:59 < Bl4ckC0r3> i just finished to scan with rkhunter and it found 6 possible rootkits 22:59 < revel> Good to know. 22:59 < revel> Probably false positives. 22:59 < Bl4ckC0r3> can you please read the log file and tell me if it is real warning or false 23:00 < Bl4ckC0r3> there are the log file https://pastebin.com/uRr5UfPX 23:00 < Bl4ckC0r3> please help 23:01 < ayecee> this file is long. what was found? 23:01 < Bl4ckC0r3> wait 2 sec 23:01 < Bl4ckC0r3> this /usr/bin/egrep [ Warning ] 23:01 < ayecee> which line is that 23:01 < Bl4ckC0r3> Warning: The command '/usr/bin/egrep' has been replaced by a script: /usr/bin/egrep: POSIX shell script, ASCII text executable 23:01 < ayecee> line number 23:02 < Bl4ckC0r3> ok 23:02 < Bl4ckC0r3> 219 23:02 < Bl4ckC0r3> 221 23:02 < Bl4ckC0r3> 222 23:02 < Bl4ckC0r3> 244 23:02 < ayecee> open the file, verify that it's a script 23:02 < Bl4ckC0r3> 319 23:02 < Brainspackle> if egrep was replaced by a script it sounds you got rekt 23:03 < revel> ldd and egrep are supposed to be scripts. 23:03 < Brainspackle> oh, then rkhunter is on the spectrum 23:03 < ayecee> ain't that the pot calling the kettle black 23:04 < revel> And besides, it's not like replacing them with a small compiled wrapper that runs some script in /var/log/everything/.some/hidden/path/script.sh would be difficult. 23:04 < Bl4ckC0r3> Checking if SSH root access is allowed [ Warning ] 23:04 < Bl4ckC0r3> Checking if SSH protocol v1 is allowed [Warning ] 23:05 < Bl4ckC0r3> Warning: Hidden file found: /usr/share/man/man5/.k5identity.5.gz: gzip compressed data, max compression, from Unix, original size 22 23:05 < ayecee> okay, you can stop now 23:05 < Bl4ckC0r3> Warning: Hidden file found: /usr/share/man/man5/.k5login.5.gz: gzip compressed data, max compression, from Unix, original size 19 23:05 < Bl4ckC0r3> ok i stop 23:05 < Bl4ckC0r3> what are that 23:05 < revel> The SSH ones are saying some options are not explicitly set and the defaults *may* be bad. 23:06 < Bl4ckC0r3> k5identity.5.gz and k5login.5.gz 23:06 < steinex> rkhunter really is a joke these days 23:06 < Bl4ckC0r3> what are that 23:06 < revel> gzipped data. 23:06 < ayecee> manpages 23:06 < Pentode> kerberos 23:06 < Pentode> its just a false positive 23:07 < Bl4ckC0r3> oh...thank you 23:07 < Bl4ckC0r3> God bless you all 23:07 < revel> I don't think I ever gathered any useful data from rkhunter... 23:07 < CompanionCube> ...does it really alert on all hidden files like that? 23:07 < CompanionCube> damn that's sensitive 23:07 < ayecee> spent a lot of time tuning it to avoid false positives though. 23:08 < Bl4ckC0r3> i have lynis too 23:08 < Bl4ckC0r3> just lynis audit system 23:08 < ayecee> sounds like a medical condition 23:08 < Bl4ckC0r3> and will scann your system 23:09 < revel> Lynis Dorwelds. 23:09 < Bl4ckC0r3> ? 23:09 < revel> Nothing. 23:09 < ayecee> a rare but serious complication of vendor lockin 23:09 < Brainspackle> does it itch? sounds like you might have that new STD called ayecee 23:10 < Bl4ckC0r3> there it is https://github.com/CISOfy/lynis 23:10 < ayecee> nice try, brain spackle 23:13 < Gnjurac> best udev i cen do is to 6 time to run , this sux 23:15 < Gnjurac> and my logic is its cuz its like that reported 6 times in dseg 23:28 < torncolours> hello 23:29 < iflema> o/ 23:29 < Gnjurac> is there some way to limit udev to run only 1 time even if you have devace that creates 4 inputs whit same id vendors.. 23:46 < dslegend> hello 23:47 < dslegend> I'm looking for an area to discuss PCIe driver development help 23:47 < kurahaupo> Hello dslegend 23:47 < kurahaupo> #kernel 23:47 < dslegend> Thank you 23:48 < ice9> dslegend, also #kernelnewbies on OFTC server 23:48 < Psi-Jack> And, of course, the LKML --- Log closed Sat May 26 00:00:33 2018