53[00:40:59] <cws> unixbsd: Your original claim is that cat is better. So, show me in the source code why that's the case.
54[00:41:27] <cws> The official documentation refers to cp. One advantage is that it doesn't require shell redirection. cp is THE utility that exists for this function.
55[00:41:51] <cws> Use the right tool for the job and stop acting superior.
56[00:42:27] <cws> If you want to use cat, use cat. cat ALSO works. But cp is better. </discussion>
57[00:42:56] <unixbsd> sorry but you can do cat in about 15 lines of C code to put the image on the dev
58[00:43:17] *** Quits: msh07 (~msh07@replaced-ip) (Quit: "Las loterias son el impuesto de los gobiernos al desconocimiento de las matemáticas")
59[00:43:55] <unixbsd> fd open and it is done. cp is intended way more for files
60[00:44:11] <fling> busybox cp is superior to util-linux one on gentoo
219[03:42:29] <maroloccio> hi. i am on a deb server using ssh. need to search for packages to install based on description. which CLI | TUI tool is best? (I can install a tool if not installed already)
222[03:43:56] <Unit193> maroloccio: Are you looking for, eg `apt-cache search streaming server`, then `apt-cache show $pkgname`?
223[03:44:06] <abstractednoah> or maybe apt-file?
224[03:44:11] <abrotman> dpkg: search
225[03:44:11] <dpkg> Please search for things rather than just asking, e.g. «apt-cache search regex», or with <aptitude search>. The following ways can locate packages that own a file (installed, or candidate for install): «dpkg -S /bin/foo», <apt-file>, <pdo> (online) and «/msg judd help file» (bot). <search dpkg> for factoids. See replaced-url
226[03:44:27] <sney> apt search is pretty capable, but aptitude is more powerful and may give you better results depending on your search terms.
227[03:44:45] <sney> aptitude can also search tags:
228[03:44:46] <sney> !debtags
229[03:44:46] <dpkg> well, debtags is a way of categorising packages with tags (facets) that describe how a package can be used, what data it can read/write, what interfaces it has etc. This adds another dimension to package searching. Install the debtags package, then search either using "debtags" or "aptitude". For a GUI install packagesearch. Search (or add new tags) at replaced-url
230[03:45:26] <maroloccio> Unit193: those 2 commands seem like the do the job but let me review the other answers too.. (thanks)
240[03:52:33] <maroloccio> the GUI version of aptitude TUI is.. synaptic? or something else?
241[03:52:45] <Urk> I'm wondering if it is ok to install grub2 if I have already formated the partition as a linux partition? I'm at step 4.3.3.1 in the following URL instruction, and not sure what to do next replaced-url
242[03:53:19] <sney> maroloccio: aptitude has a gtk frontend itself. synaptic and discover and gnome-software are separate
243[03:53:25] <Urk> I can't install syslinux /dev/sdc because I am not running FAT 16
244[03:53:47] *** Quits: catman370 (~catman@replaced-ip) (Quit: See you later..)
245[03:54:04] *** Quits: Adbray (~Adbray@replaced-ip) (Quit: Ah! By Brain!)
246[03:54:16] <maroloccio> sney: what is "discover"?
267[04:08:26] <sney> Urk: as the error says, you would have to use a different filesystem. and nothing here is similar to the instructions in step 4.3.3.1. in fact, in the following step, it tells you to copy files from inside the installer iso, rather than writing the iso directly to the partition.
268[04:08:39] <sney> Urk: for best results, read the words on the page, and don't substitute from your imagination.
270[04:10:52] *** Quits: citypw (~citypw@replaced-ip) (Remote host closed the connection)
271[04:11:12] <Urk> sney: What i want to know is what can be used in place of syslinux since I am inside a Linux partition, and syslinux is only written for FAT16
276[04:12:14] <sney> and it's probably not going to work anyway. have you tried following the real instructions? or was the link to the install guide just a fakeout?
277[04:12:30] <Urk> oxek: I'm just following instructions found at 4.3.3.1 on replaced-url
278[04:12:43] *** Quits: twobitsprite (~twobitspr@replaced-ip) (Quit: Lost terminal)
279[04:12:54] <Urk> oxek: I used cfdisk to create my partitions so I don't have the option of installing FAT16
286[04:14:42] <oxek> unfortunately it appears that /dev/sdc is not partitioned at all, it instead directly contains a filesystem
287[04:14:44] <Urk> I would think there wouldn't be a problem with GRUB2, but received a different error when trying to install grub
288[04:14:56] <oxek> hence you can't install grub or anything else on it
289[04:15:08] <jmcnaught> Urk: you only need to do the steps in 4.3.1. Did you not see the big *Important* notice about the below options are for specialised needs?
290[04:15:49] <oxek> what's the end goal here anyway? And why wouldn't debootstick work?
292[04:16:09] <Urk> oxek: Its possible, but I did select the "new" option in cfdisk, and the "write" option so I assume new partitions were created. They look like they were created when running lsblk
294[04:17:24] <Urk> jmcnaught: I will take your advice and use the whole disk for I can revisit the issue later. Howeve, it would be handy to store missing drivers on a separate partition on the same USB device.
295[04:17:54] <jmcnaught> Urk: if you read 4.3.1 in its entirety it tells you that you can make an extra partition after copying the ISO
304[04:23:45] <dvs> Urk: At the start of section 4.3.3, it clearly says: If you like more flexibility or just want to know what's going on, you should use the following method to put the files on your stick.
305[04:24:14] <Urk> oxek: I tried both grub2, and syslinux; both had errors.
306[04:24:38] <Urk> dvs: I do want more flexibility, but don't have to have it tonight.
307[04:26:11] <jmcnaught> The instructions in 4.3.3 will only be able to do BIOS/legacy booting, so it will not install with UEFI support
312[04:31:36] <Urk> oxek: I was hoping to create a separate partition for drivers not included with the 10.9 Debian iso
313[04:31:53] <oxek> Urk: that's exactly what 4.3.1 lets you do
314[04:32:33] <jmcnaught> "The hybrid image on the stick does not occupy all the storage space, so it may be worth considering using the free space to hold firmware files or packages or any other files of your choice. This could be useful if you have only one stick or just want to keep everything you need on one device. " -- 4.3.1
324[04:47:06] <Urk> jcmcnaught: Not sure zxvf /path/to/firmware.tar.gz The firmware is already bundled in the iso so this line appears to be unnecessary. Should I skip it for now?
325[04:47:32] <Urk> jmcnaught: Not sure zxvf /path/to/firmware.tar.gz The firmware is already bundled in the iso so this line appears to be unnecessary. Should I skip it for now?
326[04:48:29] <jmcnaught> Urk: that line is from an example, it's not necessary. The point is that after you "cp debian.iso /dev/sdX" you can create a new partition on /dev/sdX's unused space and put whatever you want there. After copying the ISO.
344[04:59:02] <Urk> oxek: It is a goal to get a hard wire connection going, and order Dell's tool for doing this. There is no jack for an RJ47 on this laptop.
345[04:59:15] <somiaj> Urk: do you have the non-free firmware installed?
346[04:59:21] <Urk> I would have been done by now if I had a hard wire connection.
347[04:59:42] <somiaj> Urk: for iwlwifi, firmware-iwlwifi is the package name from non-free
348[04:59:45] <oxek> Urk: we literally found someone the other day, who has the exact same laptop as you, who installed debian using the iso containing non-free firmware just fine
349[04:59:49] <Urk> somiaj: Supposedly the url to 10.9 provided earlier has firmware on it.
350[05:00:31] <Urk> oxek: I had no problems with 10.4 and earlier, nor on systems older than the one that I have now.
367[05:07:06] <somiaj> Unsure what you mean by that, the driver is part of the kernel. You can check if you are missing firmware with 'dmesg | grep -i firmware'
385[05:22:03] <Urk> oxek: I'm booted up, but noticed that the kernel is only 4.19 or something like that, and I recall that I need 5.10 in order to get the wifi drivers to work
397[05:47:08] <Urk> does the add-cdrom tool also cover usb? I need to update my kernel to 5.10 so cdrom add --cdrom /dev/sda1 so that I can install my wifi driver, but received the following error "No CD-ROM could be auto-detected or found using the default mount point".
398[05:47:56] <Urk> The command that i typed was apt-cdrom add --cdrom /dev/sda1 Did I do this right?
399[05:55:42] <somiaj> Urk: The way forward if you really need a newer kernel but want to install buster, is to do a base install (the netinstall can do this without the internet), then copy over the kernel and firmware from buster-backports, install those and reboot, then get your network up, afterwards you can finish installing the software you need
400[05:56:05] <somiaj> alternativally you could just install bullseye, which is going to be released soon (maybe a couple of months)
401[05:56:42] <somiaj> You aren't just going to be able to copy a newer kernel onto the install media and have it work though
403[06:03:18] <Urk> somiaj: I have already installed the base install, and was trying to install kernel 5.10 from a memory stick. Its quite recent, and was used for the previous isntallation. The wifi driver will not work unless the kernels is updated. Unfortunately that wasn't fixed with 10.9
404[06:03:42] <Urk> somiaj: Does the tool apt-cdrom with with usb memory devices?
405[06:03:52] <somiaj> Debian does not upgrade kernels in point releases.
406[06:04:06] <somiaj> I would just download and install the .deb (and its depends) from buster-backports
407[06:04:43] <somiaj> well doesn't upgrade the version, it will include minor updates to the long term stable kernel it uses, mostly security/bug fixes, sometimes new network hardware gets added.
408[06:04:50] <Urk> somiaj: I have already done that and have it on disk.
409[06:05:03] <Urk> or USB. I'm just trying to get it detected so it installs.
410[06:05:11] <sney> Urk: this is why you're reinstalling again? because you thought 10.9 meant a new kernel for your hardware?
451[06:40:55] <Urk> My command is "dpkg -i /media/pdq/4430-BE54/Linux-image-5.10.0.0.bpo.3-amd64_5.10.13-1-bpo10+1_amd64
452[06:41:37] <Urk> I get an error indicating that 2 expected programs not found in path or not executable. This suggest another package is missing that must be installed before the 5.10 kernel is installed.
453[06:42:49] <Urk> also dpkg says that the path should usually contain /usr/local/sbin, usr/sbin and /sbin
454[06:43:13] <klys> urk, set | grep PATH; should show you if it is
455[06:43:14] <Urk> I just tried installing firmware required for the previous installation, and received the same error.
459[06:44:22] <Urk> klys: It does show that as the path, but I am at a loss as to what I should be using for a command.
460[06:44:46] <Urk> I am trying to follow instructions on stack exchange, and it just says sudo apt install /path/to/package/name.deb
461[06:45:08] <klys> urk, apt-cache show prog1 prog2 | grep pool | while read x y; do wget replaced-url
462[06:45:32] <klys> where prog1 and prog2 are your missing binaries
463[06:46:17] <klys> that isn't as cut-and-dried as I say it, though you can find the package of any binary at packages.debian.org
464[06:46:23] <jmcnaught> !buster su
465[06:46:23] <dpkg> In buster, su no longer overrides PATH by default, requiring that you use "su -" or "su -l" for login shells (which is not really a new thing at all...). See replaced-url
466[06:46:40] <jmcnaught> Urk: you need to use "su -l" not "su" alone
467[06:48:01] *** debhelper sets mode: +l 1015
468[06:48:50] <Urk> klys: error now says no packages found, but I don't think I can run wget until the wifi driver is installed, and the wifi driver can't be installed until kernel 5.10 is installed. Whether or not there is enough firmware with 10.9 is unknown. There wasn't enough firmware with 10.8, and at least three different packages had to be installed with dpkg in an exact sequence in order to get the wifi working.
469[06:49:33] <klys> urk, what are the missing binaries it lists? or consult packages.debian.org and put them in the lower form
470[06:49:56] <somiaj> Urk: you should provide better info on the errors you are getting, the first error you shared had nothing to do with the path, hence why I was thinking some other issues.
473[06:51:57] <Urk> Doesn't list any missing binaries after running the last command you suggested. However, the needed packages are on a usb stick that is inserted in my laptop. 10.9 could be improved in the sense that it should be able to look for the packages and drivers once the pointer is sent to the memory deivce. Unfortunately it doesn't do this, and there is a lot of work involved to get it working on this laptop. I really need a hardwire
474[06:51:58] <Urk> connection, and I am ordering Dells proprietary device in the morning. This laptop doesn't use an RJ47, but something similar that is very rapid, but sold as a separate accessory.
475[06:52:04] <somiaj> Urk: though you might beninifit from just installing bullseye on that laptop, as opposed to messing with buster.
476[06:52:35] <Urk> somiaj: I installed 10.9 which was recommended this morning, and it has a lot of missing packages just like 10.8
477[06:53:11] <klys> urk, mount /dev/sdb1 /mnt; ls /mnt; for example to mount the second detected disk, first partition thereof, and list the files on it.
479[06:53:24] <Urk> My media device where the firmware and wifi drive resides is /media/pdq/4430-BE54
480[06:53:41] <klys> urk, then forget the mount command I just said
481[06:53:43] <somiaj> What do you mean by missing packages? Buster is a fully functional release with 50,000+ packages. If you need a newer kernel or want newer versions of software, installing bullseye might benifit you (it isn't quite released, but it is frozen, and will be released in the next couple months from what I hear)
482[06:54:03] <somiaj> Buster doesn't have missing packages, it may not have the software you want, but that is different.
483[06:54:22] <Urk> somiaj: Is that consider a stable release? I thought I already installed Buster?
484[06:54:45] <klys> urk, what prevents you from telling us about the missing programs (binaries)?
486[06:54:54] <somiaj> Bullseye is going to be the next stable, it isn't released yet, but it is frozen and mostly in its final state, and if you just do reguraly updates you'll get all the fixes/changes by release time.
488[06:57:14] <Urk> somiaj: Do you have a link you recommend for this? Its getting late so I may have to continue this in the morning, but would benefit by having a link.
489[06:58:27] <klys> urk please don't ignore my question again. what programs did it say it was missing?
491[07:00:01] *** Quits: riff-IRC (~riff2@replaced-ip) (Remote host closed the connection)
492[07:00:46] <somiaj> Though it would be nice to know, what packages were missing in 10.8/9? Was it actualy software, or just you need a newer kernel, etc
493[07:01:41] <Urk> klys: dpkg didn't indicate programs were missing, but instead indicated that the root's PATH should usually contain /usr/local/sbin, /usr/sbin and /sbin.
495[07:02:24] <Urk> I know from the past that there were at least three packages that had to be manually installed to get it to work, and that kernel 5.10 or greater had to be installed or the wifi wouldn't work.
496[07:02:36] <klys> urk, and you mentioned that your PATH was correct. this is from dpkg -i ...deb; on your kernel deb?
497[07:02:37] <Urk> I have all three of those packages on memory stick.
498[07:03:13] <Urk> klys: Yes, this is from dpkg -i ...deb running kernel 4.19
499[07:03:30] <klys> urk, it should be as simple as dpkg -i /media/pdq/4430-BE54/*.deb
501[07:04:59] <Urk> klys: Thats what I thought. However, I recall from the help that jcmnaught offered that suggest there was a specific order for the installation of the missing packages, and I am a bit foggy on. The only thing I recall was that kernel 5.10 had to be installed before trying to install the wifi driver.
503[07:05:40] <klys> dpkg should install them in the correct order automatically, or else you can specify dpkg --auto-deconfigure
504[07:05:40] <dpkg> You are person #1 to send an unparseable request, klys
505[07:06:15] <Urk> klys: what does the --auto-deconfigure parameter do?
506[07:06:41] <Urk> dpkg ignore order
507[07:06:41] * dpkg sticks his fingers in his ears. "La, la, la! I can't hear you, order"
508[07:07:34] <klys> urk, for example if you have a package that depends on another package you're installing at the same time that conflicts with something that is installed but will be upgraded
509[07:07:53] <Urk> klys: OK, I just learned something.
510[07:08:01] <Urk> I am going to take this up tomorrow. Its getting really late.
511[07:09:12] <Urk> Has anyone tried Bullseye with the Dell XPS 7590?
699[11:29:54] <EdePopede> i never had to reboot unless there were kernel updates. so flatpak is a kernel extension it seems
700[11:30:08] <TheBigK02> dob1: every tool has its pros and cons. u should read some stuff about it and decide urself on the information u gained
701[11:31:04] <TheBigK02> EdePopede: i doubt flatpack updates the kernel. but may be it does provide a module? which i doubt as well... even then u could probably get away with just manually loading it
702[11:31:23] <EdePopede> there's modprobe for modules, no reboot needed
703[11:31:27] <EdePopede> yeha
704[11:31:34] <dob1> TheBigK02, from my point of view I am going for ansible because all the review seems it is simple (and better to start with something simpler) and it use ssh, not agents
705[11:32:05] <oxek> EdePopede: doesn't need reboot
706[11:32:08] <TheBigK02> dob1: if u made ur mind already why asking :D
707[11:32:13] <EdePopede> unfortunately i don't see a tag entry for the package though i doubt there's something like needs-task::reboot
708[11:32:41] <dob1> TheBigK02, because I noted that these things change very fast, so I don't know what I read in a review from a year ago is still there :)
709[11:32:46] <oxek> they provide those instructions for newbies, because it makes it easier to understand than "Logout and login, so that icons can be set up in your launcher"
710[11:32:57] <EdePopede> oxek: thanks. good to hear. what about suid root and ressource usage? all i potentially may need it for (not even sure) would be a matrix client.
711[11:32:58] <oxek> but nothing inherently requires a reboot in flatpak
712[11:33:24] <oxek> EdePopede: you can remove the suid root if you upgrade to kernel from backports, or manually enable user namespaces
713[11:33:47] <EdePopede> web is no-go because ressources (i want IMs&co to be able to run constantly), electron the same plus suid root for chrome-sandbox
715[11:34:12] <oxek> resource usage: Doesn't inherently consume more than a native app, with the exception that newer apps/libraries take up more space and are usually slower.
717[11:34:25] <EdePopede> ugh. never did such a thing. and i'm not sure if i want to try it before this system is going to be nuked anyway for its successor.
718[11:34:32] <oxek> obviusly it does take some storage space, because it duplicates a lot of system provided stuff
725[11:35:46] <oxek> CPU is identical as native app, with the exception that newer apps are slower than older apps. The same will happen with newer versions in debian, so can't be avoided.
726[11:35:52] <oxek> RAM is the same problem
727[11:36:05] <EdePopede> jelly: even one is bad enough. tested Franz! (or was it Fritz!) some years ago. was like firefox in the beginning and then +15% RAM for each service i activated. ouch.
728[11:36:29] <oxek> flatpak apps can't share debian-provided libraries in RAM, so they will have to load their own
729[11:36:48] <oxek> but overall flatpak is a good thing
730[11:37:00] <oxek> lots of software is not in debian repos, but are in flatpak
732[11:38:02] <EdePopede> the situation here usually is firefox needing most of the physical RAM, sometimes the bar is nearly maxed out, which means lot of swapping to an old HDD. slowing... things.... down....... endlessly. sometimes (oh, you REALLY need JS? i'll grant you a private window w/o NoScript) it just goes wild on the CPU also. and then videos mpv. but that's fine, i don't use it as a telly 24/7.
734[11:39:35] <EdePopede> the one i already installed (well, i threw the binary into $PATH) was goluks. i think i'd be fine with it, but it didn't let me create an account, now that i have one it presents me with an empty window without even a Quit option. matrix seem to be the Wild Wild West for now.
735[11:40:03] <EdePopede> not that i wouldn't play with it, i just like to have things running first :D
736[11:40:28] <oxek> matrix clients are a mess
737[11:40:30] <EdePopede> and maybe i'd even look at some flatpak stuff when i first have that one installed. ok, i'll be brave.
738[11:41:21] <jelly> oxek, I don't "libraries sharing RAM" was true in decades on linux
741[11:41:53] <EdePopede> debian.map.fastlydns.net <-- lol i was reading nap. while apt looked like waiting for the server to react.
742[11:42:08] <oxek> jelly: how many times do I have libc loaded then?
743[11:42:14] <jelly> oxek, so many.
744[11:42:25] <oxek> well that's annoying
745[11:42:31] <jelly> ¯\_(ツ)_/¯
746[11:42:46] <oxek> no wonder linux ate my ram :)
747[11:42:51] <EdePopede> seems i really have routing problems to that server... but ping worked, hm
748[11:42:52] <jelly> each with a slightly different address space, too
749[11:43:27] <EdePopede> is it possible to find out if a server is available, not just sitting there and reacting on pings? i mean the service being reachable
750[11:43:38] <oxek> I think I confused it with flatpak having to store its own version of libraries on disk, and not being able to use debian provided ones
772[12:06:13] <ratrace> yeah that's definitely the case. you can check /proc/PID/smaps and see how much of code is Shared_Clean. that's physically shared memory, in virtually-remapped, randomized layout, PIC process space
799[12:33:11] <EdePopede> trying to install Fractal (some matrix client which at least looked useable and fulfills some of my prerequisites), that's why flatpak. and now while running the command they have on their website (something i'd compare to apt install) it comes up with this
800[12:34:00] <ratrace> EdePopede: you can get it from snaps too
802[12:34:24] <EdePopede> ofc they don't mention it on their hp. matrix eould send you to the client room (which is out of reach obviously until there is a working client available) and the client folks just don't mention all the nasty bits in advance.
804[12:34:38] <EdePopede> ratrace, i don't even have non-free on this system ;)
805[12:34:58] <ratrace> snapd is in main
806[12:35:12] <EdePopede> and isn't snap an ubuntu thing after all? and what i've seen here flatpak seem to be embraced more by debian than snap
807[12:35:14] <ratrace> the the snap is gpl3-0
808[12:35:41] <ratrace> ubuntu is upstream for snap, but it's intended to be distro agnostic. basically same as flatpak, the two are competing technologies.
809[12:35:52] <EdePopede> i wouldn't even have considered flatpak with a usable native matrix client available.
810[12:35:59] <ratrace> but their impementation is different. flatpak requires shared runtimes, snap handles that automagically
813[12:37:00] <EdePopede> the only non-distro thing i needed so far was static builds of opera mayne years ago. firefox not really being a thing yet, Qt coming with a bad license.
814[12:37:43] <EdePopede> and i wouldn't even consider matrix if mozilla wouldn't have decided to run away from their users into a beta stage environment
815[12:39:20] <EdePopede> and as i don't like the idea to keep half a dozen matrix clients available because each of them is missing out something relevant i also don't like to have flat and snap and what not installed on top of debian, especially with each of them only needed by a single "app".
816[12:39:28] <ratrace> maybe you can install Element on a local LNPP stack: replaced-url
823[12:40:45] <oxek> even despite running that command wrong
824[12:40:57] <oxek> since it prompted me whether I meant to install it from flathub
825[12:41:15] <oxek> the right way would be `flatpak install flathub org.gnome.Fractal`
826[12:41:20] <EdePopede> oxek: it asked me during Fractal install. only i like to get the info beforehand, not while i'm on the boat already. and the url presented ( replaced-url
827[12:41:33] <EdePopede> so i have no chance to even find out wth this installer is trying to do
839[12:43:47] <EdePopede> > The following NEW packages will be installed: bubblewrap flatpak libgpgme11 libostree-1-1 xdg-desktop-portal xdg-desktop-portal-gtk
840[12:43:53] <EdePopede> that's how it should look like
841[12:44:02] <oxek> there's #flatpak on freenode, and they are very helpful in there if you need any help
866[12:48:03] <EdePopede> btw, afaik the romans really did it. the girls got names, the boys only numbers. chances were not small after all they'd die before they would need it in one of their wars.
874[12:52:28] <ratrace> so the Rösch company should TM the name in category "Software" and Worldwide. probably 7 or 8 figured sum in euros. they ain't selling that much detergent lol
875[12:52:38] *** Quits: rustyshackleford (uid236774@replaced-ip) (Quit: Connection closed for inactivity)
876[12:53:34] <EdePopede> seems intel won replaced-url
877[12:54:05] <ratrace> (obviously, sine we're (and Intel is) still calling them Pentiums and not Specexechole-iums)
895[13:10:47] <EdePopede> ratrace, oxek, just to finalize it (finally remembered the name): shop owners were sued for advertizing intel's Triton chipset in their windows. there was a dutch company owning Tricon™ - DM1300 for each case. since you now may know *where* it happened you also could imagine *who* the responsible lawyer was - a certain Günter Freiherr von Gravenreuth. if you feel like it, give google translate a try
924[13:33:42] *** StathisA_ is now known as StathisA
925[13:34:09] <cosimone> hello, this may be a stupid question, but is there anyone here using qemu for virtual machines? how do i get the guest to automatically resize its resolution when resizing the window? it worked with the exact same image (with spice guest tools installed of course) on a newer version on other distributions, but it doesn't seem to work on the current version in debian stable (3.1.0 if i'm not mistaken)
926[13:34:30] <cosimone> am i missing some package on the host? it could be a spice related thing, but i'm not quite sure where to look at
953[13:54:35] <hid3> Hello everyone. I have specified SystemMaxUse=10G entry in /etc/systemd/journald.conf file. However, ' journalctl --disk-usage' still returns size much over that. How do I invoke the "housekeeping" mechanism to get rid of the overhead data?
1100[16:11:23] <EdePopede> NetTerminalGene: none. because nothing works as expected.
1101[16:11:50] <EdePopede> trying since last night to get matrix onto this box but there's not a single client available i could use
1102[16:12:02] <oxek> EdePopede: nothing works as expected? Explain.
1103[16:13:14] <EdePopede> oxek: the only thing i managed is an account using the website of Element. dis was it.
1104[16:14:43] <oxek> I tried it a couple hours ago, using flatpak, and had a working nheko, fractal, spectral, quaternion and element.
1105[16:14:49] <EdePopede> tried gomuks first or what it's called, terminal client, looked ok. can't do anything it seems. then that Fractal or what its name is, which means Flatpak. started, but didn't let me into my account. no idea what it didn't like.
1106[16:15:10] <ratrace> linux gud for servahs. dextop linux sucks.
1107[16:15:32] <oxek> all matrix clients suck right now
1108[16:15:35] <ratrace> install windows in vm, enjoy matrix. windows-sound-ta-daaaa.jpg
1109[16:16:40] <EdePopede> you're right. lienucks is for developers only after all
1123[16:24:22] <chocolate> Is there a package like ircd-hybrid to install in the server, but for making video calls when the connection is cone locally in the server?
1134[16:27:55] <EdePopede> i installed a debian on a former XP computer some years ago and waited for the problems. there was only one, a firefox bookmark disappeared. dragged away, so you can imagine his competence. used the system for a few years without even noticing a difference.
1137[16:29:18] <EdePopede> there were 2 initial reasons: first he couldn't listen and HAD TO click "register your windows", and then it was not localized for him. turned out this all wasn't an issue anymore with squeeze or what it was.
1140[16:35:14] <EdePopede> the xdg-portal things are still running as children of user systemd, so does flatpak-session-helper - though they are all deleted ofc because purge. what's going on?
1157[16:56:18] *** Quits: idhugo (~idhugo@replaced-ip) (Remote host closed the connection)
1158[16:56:23] <urk> EdePopede: not sure. I got as far as set root=(hd0, msdos1), but not sure as to what ti do with rhe linux command line after that.
1159[16:58:01] *** debhelper sets mode: +l 1061
1160[16:58:06] <EdePopede> urk: i had this during my freezing era. fsck happened during boot, but i also had to do it manually on the shell a few times
1215[17:13:08] <drzacek> if the module is missing then it's bonkers, can't help with that. you'll need live system on pendrive, get some light live debian iso and boot from there, then fix grub from there
1246[17:43:40] <ZeroBeholder> I am trying to play with LDAP and kerberos (no need to tie into AD or Windows anything). It seems like LDAP will do centralized authentication by itself, but using kerberos with it is icing on the cake and incrementally more secure. Would you all recommend trying to both all at once or figure out LDAP then add kerberos functionality second? If you had to relearn the same stuff today, where would you start?
1247[17:44:37] *** Quits: odnes (~odnes@replaced-ip) (Remote host closed the connection)
1248[17:48:02] <trysten> usermod -aG <group> <user> is a simple pattern but I cannot remember it!
1255[17:51:01] <phogg> ZeroBeholder: I would suggest ldap first, mix in kerberos after. Or, freeIPA from the start
1256[17:51:32] <ZeroBeholder> Yeah, I hear you. As I get older, I look at commands that I should mostly know because I do use them regularly-- like usermod or mount (CD/DVD, VM disks, USB flash drives...). When I have choices of commands, I try to stick to the ones that work on more distros.
1257[17:51:40] <urk> Got bumped, but back. Cell ph disconnects channel if screen is minimized for too long
1258[17:52:03] <ZeroBeholder> phogg: Thanks, it seemed like that might be a strategy to minimize initial complexity.
1259[17:53:47] <ZeroBeholder> trysten: how could I forget systemctl and is it --list-units don't know. Use it all the time. Know that system --status-all is the parallel command I want systemctl to give me information for.
1264[17:55:28] *** Quits: mortderi1 (~mortderir@replaced-ip) (Remote host closed the connection)
1265[17:55:52] <urk> I am still trying to fix grub2, but stuck at the line for setting the kernel. I enter "linux /boot/vmlinux-4.4.9.210-generic and receive error unknown command linux
1267[17:56:35] <ZeroBeholder> phogg: I really like FreeIPA when I can give it its own Public IP address-- setting up FreeIPA with DNS behind NAT has been less fun. Also, I wanted to learn some of the behind the scenes magic FreeIPA was doing for me.
1268[17:57:16] <urk> Can someone again send me the link to the page on fixing grub? Page got closed accidentally on my cell phone.
1269[17:57:38] <phogg> ZeroBeholder: I mentioned it in case you were more interested in turn-key results than learning the bits and pieces. For learning it's less useful.
1271[17:59:48] <urk> I found the page in my cell phone browser history so ignore my last request.
1272[17:59:58] <ZeroBeholder> phogg: Agreed. I wanted to use what I implemented at work at home-- FreeIPA is such a nice tool for managing your own low level centralized authentication, authorization, DNS, and certificates for machine to machine data communication security. Turns out it isn't great for the home gamer, so I thought it was time to grow a little.
1287[18:14:03] <urk> And get: error symbol "grub_calloc" not found. Dicymentation I found onli e indicates to reinstall grub, but not sure how to do that.
1300[18:21:04] <ZeroBeholder> urk: worse, maybe do a hardware checkup to make sure your RAM is good and your disk are in okay shape. Bad bits are equally great at causing similar problems to human error.
1301[18:21:18] <urk> Somiaj: Need to first fix the stsyem that I was planning on downloading bullseye to. Its an ancient 32-bit system that periodically locks up.
1306[18:23:42] <urk> ZeroBeholder: System was not properly shut down which likely caused the problem. First need to figure out how to reinstall grub when in grub rescue>
1317[18:27:50] <urk> ZeroBeholder: I have been runnng the same 32-bit OS for awhile so not sure how relevant it would be to run a checksum of the install media.
1318[18:28:04] *** Quits: LeoTh3o (~th3o@replaced-ip) (Remote host closed the connection)
1393[19:49:17] <foormea> hi. i have an issue with ipv6 and systemd-networkd. i want my http services and my ssh service to run on separate, static ipv6 addresses. i've added 2 "Address=<a:b:c:d:e:f:0:1>" in my /etc/systemd/network/<interface>.conf. addresses are created on boot. but... it seems like the system stops using ipv6 if i do that. if i don't add those 2 "Address=" lines and if i manually add those 2 ipv6 addresses with "ip addr add ... dev ...",
1426[20:04:43] <cws> In the future, if you're going to redact IPv6 addresses, only redact the third and fourth segments (out of 8 total).
1427[20:05:07] <foormea> now, silly question. i'm a total ipv6 noob. how do i find my router/box's ipv6 address? i have its link-local address, do i just replace the first 4 fields with my public ipv6 prefix, and keep the last 4 fields?
1428[20:05:32] <cws> foormea: Actually, you can use its link-scope address as the gateway.
1438[20:09:56] <m4lvin> How/where should I report about upgrade-bugs from buster to bullseye? I updated three machines and each time KDE was unusable afterwards. I do have a workaround now but do not understand what happens.
1439[20:10:00] *** Quits: mthe878 (~mthe@replaced-ip) (Quit: Lost terminal)
1440[20:10:12] *** Quits: D4rk2020 (~darkangel@replaced-ip) (Remote host closed the connection)
1448[20:11:52] <foormea> cws: so it seems to work better now, i'll still need to run some tests. although sshd doesn't work until i manually reload the config (it's already configured to listen to its specific ipv6 address)
1453[20:13:22] <sney> m4lvin: there is a known upgrade issue that gets stuck due to the gcc10 transition, and results in a system that is half upgraded with multiple things broken. did you see a message like "libc6-dev: Breaks: libgcc-8-dev" during your upgrade?
1454[20:14:06] <foormea> cws: :D my home NAS has some http services running, it has sshd (which i want to be able to access from the outside), and possibly more to come. i understand it's good practice to use a temporary ipv6 address for outbound connections? so if i make my http/ssh services listen on that temporary address, every once in a while the connection will drop, + i need to update dns at every change
1455[20:14:31] <cws> foormea: temp addresses aren't any particular best practice, no.
1456[20:14:43] *** Quits: Vizva (~Vizva@replaced-ip) (Remote host closed the connection)
1457[20:15:44] <foormea> cws: okay. isn't it a good idea though to make ssh run on a separate address, to avoid risk of scanning once someone already has my http address? (well i use cloudflare's proxying but i'm not sure how exposed my real ip is)
1458[20:15:50] <m4lvin> sney: thanks, I will look at that metapackage. I did not have that message in my transcript and besides KDE plasma everything else seems fine.
1459[20:16:02] <m4lvin> sney: I do have "libgcc-8-dev:amd64: dependency problems, but removing anyway as you requested" in the log, but that looks normal.
1460[20:16:58] <sney> m4lvin: ok, good to hear. must be a new issue. you may want to skim the plasma source package changelogs for clues as well.
1461[20:17:14] <mutante> foormea: scanning happens all the time, it's like natural background noise, it shouldn't be very concerning if you avoid using passwords in favor of keys
1463[20:18:01] <foormea> cws: also, if i don't use the temp address, i'm not sure how to make networkd NOT expose my MAC address. this is my sysctl file: replaced-url
1465[20:18:31] <foormea> the `scope global dynamic mngtmpaddr noprefixroute` address always ends up with my MAC in it
1466[20:18:34] <mutante> foormea: your MAC address is only there until the next hop to a router
1467[20:18:34] <cws> Why would you care about your mac address?
1468[20:18:53] <cws> mutante: Not entirely true, its used to construct an EUI64-format address.
1469[20:19:05] <cws> So someone remotely could reverse your ipv6 address to get your mac address, and... who cares?
1470[20:19:28] <foormea> cws: well, in a way i don't, but i don't understand how NetworkManager can completely hide it, but i can't manage to achieve the same via networkd
1471[20:19:34] <mutante> ok, you can also set the MAC to something random
1472[20:19:37] <cws> foormea: Why do you *care*?
1473[20:20:51] <mutante> security-wise it's more concerning what sshd and version that NAS is running
1474[20:20:52] <foormea> cws: i'd like to understand how to make networkd use a random address or an address of my choice :) i'd like to better understand networkd. i understand the "fear" of exposing a MAC address is largely unfounded
1476[20:21:13] *** Quits: sauvin (sauvin@replaced-ip) (Remote host closed the connection)
1477[20:21:30] <cws> foormea: its not largely unfounded, its completely unfounded. So again, why do you care? This is not something you NEED to mess with.
1480[20:22:14] <foormea> cws: let's say, i'll need to change the network card and would like to retain some sort of stability in my setup, i wouldn't like to edit my DNS entries when i do
1481[20:22:39] <mutante> foormea: I wonder if you can change the configuration of the sshd on the NAS, you can do thing like disable the banner entirely, disallow insecure ciphers, disable password auth, maybe change port...
1494[20:28:39] <foormea> what gets me confused (and i realise that hiding a MAC is not useful), is how to set the "main" global address to something else than MAC-based via systemd-networkd.
1495[20:29:17] <mutante> edit /etc/network/interfaces and restart networking?
1533[21:03:31] *** ZeroBeholder is now known as _zerobeholder
1534[21:03:47] *** akp55_ is now known as akp55
1535[21:04:37] <m4lvin> sney: okay, I tried to report it. Hopefully I did it correctly, this was my first time ;-) reportbug did not manage to reach my smtp, so I copied the reuslt to my normal mail client...
1536[21:06:13] <sney> m4lvin: you'll get an email with a link to the bug within the next hour if it worked, which it probably did.
1601[22:03:47] *** Lord_of_Life_ is now known as Lord_of_Life
1602[22:04:04] <XLS202> Hey, I have a problem that my network slows way done and gets unstable as soon as start a few applications (minecraft server) on my server. The load is in the most times below 3 and I have a lot of memory free. If I have nothing running, I get a speed of around 950mbit/s, but if the applications are running it drops to 150 to 500mbit/s. Could
1614[22:07:41] <XLS202> sney: no it is not a VM. But there are running VMs on the server, but they dont make problems. And I only find tools to troubleshoot HDD bottlenecks
1615[22:09:47] <karlpinc> XLS202: vmstat, iostat (in the systat package) for cpu and IO, respectively. netstat is the classic for network throughput.
1616[22:11:06] <karlpinc> XLS202: Never hurts to look in the syslog and dmesg output.
1617[22:13:26] <Stx> Start applications one by one and run your diagnostics; you will see which one causes problems. Looking at the overall data/output from diagnostic tools when everything is already running is not gonna give you any valuable information. Use the exclusion method.
1618[22:14:18] <sney> rule out java first, there's a good chance it's the culprit somehow
1620[22:16:57] <XLS202> Also i just used vmstat while doing a test with iperf. Maybe you see my problem. replaced-url
1621[22:19:00] <XLS202> Stx and sney the problem starts with the first application. But its sadly not possible to not use Java, I rent out the servers.
1623[22:20:21] <karlpinc> XLS202: Is that a lot of interrupts? Maybe something is hitting the nic?
1624[22:25:42] <XLS202> karlpinc I dont know how much a lot is. I am new to troubleshooting such deep problems. But I recorded them with dstat, after the line "327680-edge xhci_hcd 157" i startet iperf. replaced-url
1649[22:57:19] <karlpinc> I'm wondering if "segs_out" and "segs_in" of `ss -i` is a packet count (and if there's any way to get this per-second with ss).
1664[23:09:15] <th0r> karlpinc, like I said, it has been years (like, maybe 20) since I was collecting detailed performance on networks. But at that time snmp was the only way to go for detailed data.
1665[23:09:33] <XLS202> karlpinc: On my server is telegraf collecting the data. I had it already setup, I only forgot it..
1666[23:10:16] <Stx> XLS202: Are you certain its not a hardware bottleneck? server/switches/router issues?
1667[23:11:59] <XLS202> Stx: I am 99% certain it has something todo with the server or the software running on it. Because if I stop every application the network runs fine.
1668[23:14:29] <XLS202> Stx: I just tested it again to be sure. It has something todo with the java applications that they are creating somewhere a bottleneck
1678[23:27:24] <XLS202> Stx: I hate networking! Now I am completly confuesd. Also the speed to another node in the network (via the same Switch) is at maximum speed. But the confusing thing is that from that node the speed to the Server is also way better.
1679[23:29:09] *** Quits: madbytes (uid207298@replaced-ip) (Quit: Connection closed for inactivity)
1682[23:31:13] <Stx> XLS202: Hah. Alright, aslong as you´re certain that the /network/ isnt affected - then you´ve done another exclusion, and it seems you might be right - its a server issue. I just wanted to confirm its not a network issue.
1685[23:31:48] <Stx> XLS202: Did you optimize the network card settings in the server for RX/TX? What kind of server is it?
1686[23:36:06] <XLS202> I didnt optimize anything. And it has a Supermicro X11SPi-TF mainboard, Xeon Silver 4210 CPU, 96GB RAM and I use the build in X722 NIC at gigabit speed.
1687[23:36:45] <XLS202> I also updated the driver today to version 2.15.9
1694[23:42:15] *** Quits: diogenes_ (~diogenes_@replaced-ip) (Remote host closed the connection)
1695[23:42:16] <karlpinc> XLS202: That's a clue. Does the problem go away if you "un-update the driver"?
1696[23:42:59] *** Quits: Jerrynicki (~niklas@replaced-ip) (Remote host closed the connection)
1697[23:43:12] <XLS202> karlpinc: No, I updated it as I attempt. I am struggling since months with this problem.
1698[23:43:36] <XLS202> *as a attempt to fix it
1699[23:45:00] <SpeakerToMeat> Question, I fetched the source for ffmpeg because I want to build enabling opencl use to get some opencl filter(s). Since this is for internal use I don't really need to do this, but just in case, does anybody know what the right build dep package for opencl library is to add it to build deps? I know I can compile opencl stuff so I have the right packages, but not sure, I have nvidia-opencl-dev
1700[23:45:01] <SpeakerToMeat> which is nvidia specific, but I don't see any "provides" with a generic name, it depends on ocl-icd-opencl-dev but not sure that's the right one either.... there's opencl-headers but that sounds too specific to be a generic "dev" (meta)package
1701[23:45:22] *** Quits: idhugo (~idhugo@replaced-ip) (Remote host closed the connection)
1703[23:45:57] <SpeakerToMeat> ... nevermind, the control package already has the right one built in, only commented out... duh
1704[23:46:04] <Stx> XLS202: if you do watch -n2 ifconfig <optional interface> - do you see any dropped packets? If so - you could always try to raise the RX/TX buffer using ethtool -G $iface rx/rx 1024/2048/4096 ...(ethtool -g $iface will give you the current settings). Doing this might be good, or bad, you should do your own reading and decide upon it.
1710[23:49:53] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
1711[23:50:18] <XLS202> Stx: the number of dropped packets does not increase. And a information i forgot: I use a bridge on the interface, could this cause this problem?