87[01:35:23] <Cobs> Hey, anyone here familiar with grub on the
PPC64 arch? Getting stuck at a Welcome to GRUB! screen with white
background. On an IBM POWER6 550. Trying to install
debian-10.0.0-ppc64-netinst
95[01:42:34] <Cobs> trying to use yaboot just complains it
can't find a config file at device NULL /etc/yaboot.conf. Have
confirmed that it was in /boot/yaboot.conf and /etc/yaboot.conf,
gave up and hoped grub would do the trck
96[01:43:09] <dvs> I'm just wondering why it isn't
the latest snapshot
97[01:43:10] <Cobs> unfortunately having to use a SID as there
doesn't seem to be many stable, if at all, PPC64 isos
98[01:43:42] <Cobs> I had the same issue with newer ones so
thought to try an older one.
99[01:43:50] <Cobs> I'll try reflashing a newer one again
110[01:50:56] <Cobs> i'm betting its probably an IBM
firmware problem as i'm stuck on one from 2010..annoyingly for
a 10 y/o server they still lock the firmware downloads behind a
paywall
134[02:03:38] *** Quits: YWH_1 (~foo_drive@replaced-ip) (Remote host closed the connection)
135[02:03:53] <Cobs> hmmm this latest image is stuck at
Loading... after pushing expert install. will leave it 10 mins more
but it usually only takes 10 minutes to load from usb
150[02:10:36] <dpkg> win32diskimager is much more reliable than
<unetbootin> for copying ISO images to USB sticks and you can
download it from
replaced-url
160[02:16:43] <Cobs> i flashed from windows so just used
diskpart to wipe everything on the usb. for the debian install i
created 512mb /boot ext2 (also tried ext3 and 4), 8.00mb PReP, 20GB
/swap and the rest allocated to /
161[02:17:04] <Cobs> waiting for the server to restart so i can
boot this new image
435[09:12:33] <shtrb> I just got an e-mail from cron daemon with
the title "Cron <user@machinename> script name", but
If I run crontab -e for that user I don't see any command at
all (but I remember setting them via crontab -e in the past). where
should I look for the config file ? tried already in
/var/spool/cron/crontabs (but it's gone from there already)
436[09:12:40] <jelly> b_100205: no, those are global/system
legacy init scripts
437[09:13:00] <shtrb> I must be missing something very basic (as
I didn't touch cron for ages)
445[09:16:54] <shtrb> jelly, I can see in syslog
"CRON[2676]: (user) CMD ($HOME/scriptname.pl magicvalue)"
and my subject "Cron <user@hostname> $HOME/scriptname.pl
magicvalue" with a content that is related to that script error
(the error is correct).
449[09:19:18] <shtrb> Checked already in /etc/cron.d
,/etc/cron.daily, /etc/cron.weekly , /etc/cron.monthly and
/etc/crontab (no mentioning of that script execution)
454[09:22:59] <jelly> shtrb: is it really user "user"
nane host "hostname"? Did you manipulate contents of
/var/spool/cron/crontabs manually at any point?
455[09:24:06] <shtrb> I replaced the user that exist on my
machine to "user" here , and hostname from my machine to
"hostname" , I doubt I had manually edited
/var/spool/cron/crontab manually ever (only used crontab -e)
510[10:05:35] <dob1> it's possible to know which flags were
used for compiling firefox?
511[10:05:46] <Deknos> did they remove stuff which the github
ceo requested?
512[10:07:11] <ratrace> the GH repo is still down. but the
package has been released and, I guess among other places, directly
to PyPI, thus it's pip installable into a virtualenv
513[10:07:34] <ratrace> I don't know if they removed those
links, but that was documentation only, no?
520[10:09:39] <azeem> plus there's the rolling cipher thing
which I haven't read up on, both things haven't been
addressed in the latest release according to the changelog
521[10:09:53] <ratrace> so then they should file a DMCA against
all the browsers and technically all the operating systems too
522[10:10:22] <shtrb> ratrace, isn't the version in the
debian repository safe from RIAA claims ?
523[10:10:36] <Deknos> ratrace, i thought the ceo requested that
they remove the rolling-cipher-circumvention?
524[10:10:46] <shtrb> ratrace, don't give them ideas
525[10:10:48] <ratrace> the version in debian is currently
broken and doesn't work
526[10:11:03] <ratrace> shtrb: oh I'd live to see them go
after chrome. google would eat them for breakfast.
527[10:11:05] <Deknos> but i do not think i have all the
information
528[10:11:07] <ratrace> *like
529[10:11:15] <azeem> Deknos: yeah
530[10:11:31] <azeem> where "the ceo" being the github
guy
531[10:11:44] <Deknos> yeah
532[10:11:45] <azeem> maybe they just don't feel like
continuing on github
563[10:38:02] <ychaouche> genr8_, I thought about putting
everything in /private, but since the certificate is public I was
thinking maube there's another place
564[10:38:09] *** dhgjkahsdg is now known as djkghasd
565[10:38:20] <genr8_> yea
566[10:38:26] <ychaouche> /private is where the private key
belongs
567[10:38:59] <genr8_> wherever it makes sense then
569[10:39:50] *** djkghasd is now known as rkeyren
570[10:39:55] <ratrace> ychaouche: I put both the key and cert
into /etc/ssl/local, owned by root:ssl-cert, and all the daemons
that should read them are in the ssl-cert group
591[10:48:40] <ychaouche> only the public certificate would be
in /etc/ssl/
592[10:48:45] <ychaouche> here's my namei :
replaced-url
593[10:48:49] <fireba11> jup, that#s one of the biggest issues
if you have 2 services read the same certificate ... you can usually
work around that with symlinks ;-)
594[10:48:49] <ratrace> well if that works for you...
595[10:49:08] <ratrace> fireba11: why tho, just put the daemon
that needs the keys into ssl-cert group. it's there for that
reason
596[10:49:25] <genr8_> i didnt learn that until now either
597[10:50:18] <genr8_> i didnt know about the namei -l command
either
636[11:11:57] <ratrace> but anyway, you must investigate why and
where that segfault occurs. maybe you can use strace
637[11:12:17] <jelly> gvvg: 1) are there any leftover apache
processes running as
replaced-url
638[11:13:18] <genr8_> looks like systemd is mad at him not
using it to restart it
639[11:14:08] <gvvg> lol - I'll stop using /etc/init.d ...
:)
640[11:14:13] <ratrace> or apache is for not being started via
systemd :)
641[11:14:32] <genr8_> both
642[11:15:04] <genr8_> apache2.service: Service hold-off time
over, scheduling restart. apache2.service: Unit entered failed
state. apache2.service: Failed with result 'signal'.
apache2.service: Service hold-off time over, scheduling restart.
Stopped The Apache HTTP Server. Starting The Apache HTTP Server...
643[11:15:29] <genr8_> bah i messed up a paste. the first paste
should have been: "httpd (no pid file) not running "
644[11:15:33] <ratrace> genr8_: this is the most relevant entry:
"Main process exited, code=killed, status=11/SEGV"
645[11:15:56] <ratrace> by default the service unit is
type=forking, and the master forker has segfaulted
646[11:16:44] <gvvg> hmm not sure what that means - I
didn't do any changes other than apt update apt upgrade
647[11:17:03] <ratrace> gvvg: you also have some configuration
warnings that you should address
648[11:17:15] <genr8_> did PHP get upgraded ?
649[11:17:30] <ratrace> (is mod_php even used)
650[11:17:39] *** Quits: Grldfrdom (uid391113@replaced-ip) (Quit: Connection closed for inactivity)
651[11:17:45] <gvvg> genr8_: not sure - this is an open source
billing system written in perl - freeside.biz
652[11:17:52] <gvvg> it uses mod_perl
653[11:18:12] <ratrace> gvvg: what if you used apachectl
directly? apachectl start
654[11:18:43] <genr8_> Have you tried turning it off and then
back on again ?
661[11:20:25] <ratrace> gvvg: eh... try this way:
APACHE_STARTED_BY_SYSTEMD=true apachectl start
662[11:21:02] <ratrace> but first make sure there's no
runnign apache process already like jelly suggested. ss -4lnp must
not show anything listening on ports 80 or 443
663[11:21:17] *** Quits: JohnML (~john1@replaced-ip) (Remote host closed the connection)
691[11:27:57] <gvvg> FS::part_event::Condition::payby is
disabled; skipping
692[11:27:57] <gvvg> [2545] [Mon Nov 2 10:27:38 2020] [warning]:
Configuration option WebExternalAuth is deprecated, and will be
removed in RT 4.4. You should use WebRemoteUserAuth instead.
(/opt/rt3/lib/RT.pm:964)
693[11:27:57] <gvvg> [2545] [Mon Nov 2 10:27:38 2020] [warning]:
Configuration option WebExternalAuto is deprecated, and will be
removed in RT 4.4. You should use WebRemoteUserAutocreate instead.
(/opt/rt3/lib/RT.pm:964)
694[11:27:57] <gvvg> [2545] [Mon Nov 2 10:27:39 2020] [warning]:
[notice] Apache::DB initialized in child 2545
(/usr/lib/x86_64-linux-gnu/perl5/5.24/Apache/DB.pm:31)
695[11:27:57] *** gvvg was kicked by debhelper (flood)
696[11:28:04] <ratrace> I'm pretty sure that change did not
affect mid-stable release , and apache wouldn't segfault. also,
the stretch package contains the prefork mpm
706[11:28:51] <ratrace> gvvg: (repost) I'm pretty sure that
change did not affect mid-stable release , and apache wouldn't
segfault. also, the stretch package contains the prefork mpm
707[11:29:01] <gvvg> ratrace: sorry
708[11:29:31] *** Quits: JohnML (~john1@replaced-ip) (Remote host closed the connection)
726[11:36:01] <ratrace> gvvg: I reckon something in that perl
script is causing this. can you try disable the vhost and/or
mod_perl and see if apache starts normally (using systemctl start) ?
727[11:36:12] <genr8_> make sure this file isnt bugged, or
loading something bugged: /etc/perl/sitecustomize.pl
730[11:37:36] <gvvg> not sure how to disable mod_perl
731[11:38:02] <ratrace> genr8_: "sneaky" .. welp,
that's what apache normally does in the service unit. it's
terrible. whoever is maintaining that has no clue how systemd is
working and is trying to hack around systemctl using apachetl using
systemctl to start itself in forking mode.... instead of being its
process manager directly, or optimally via notify type
732[11:38:35] <ratrace> gvvg: is there like ONE config file that
does all the setup for your billing app?
733[11:40:01] <gvvg> I feel like I need a SA for hire to help -
:(
734[11:40:19] <ratrace> gvvg: is that a public server/service?
735[11:40:29] <gvvg> ratrace: not sure where to find the config
file
736[11:40:33] <ratrace> (ie. accessible via public internet)
737[11:40:37] <gvvg> is it an apache2 config
738[11:40:46] <ratrace> gvvg: how did you even install that
application to begin with
739[11:40:49] <gvvg> ratrace: yes it is
740[11:41:05] <ratrace> gvvg: then definitely you should not
maintain a public server until you know what you're doing
741[11:41:09] <gvvg> freeside.biz but they are not responsive
742[11:41:18] <ratrace> gvvg: THEY set up your app?
743[11:41:22] <gvvg> I was paying them to maintain it
751[11:42:24] <gvvg> they were taking payments but not providing
support for over 1 year
752[11:42:36] <ratrace> no, I mean people runnin public servers
with no clue how to do it, and no skilled sysadmin on hand.
753[11:42:52] <gvvg> well I was paying for that service
754[11:43:06] <ratrace> managed hosting?
755[11:43:08] <gvvg> I'm surprised apt update and upgrade
broke my system
756[11:43:12] <gvvg> nope it's my vm
757[11:43:29] <gvvg> I guess I can try and hire an system
administrator now
758[11:43:36] <genr8_> DynaLoader - Dynamically load C libraries
into Perl code
759[11:43:39] <ratrace> yeah it shouldn't happen mid
release like that but who knows what ti is...... I have a suspicion
they installed custom perl thingy there and it didn't upgrade
along with apache
760[11:43:46] <themill> gvvg: what packages actually got
upgraded? can you pastebin /var/log/apt/history.log
761[11:44:11] <genr8_> This can also be set via the
PERL_DL_DEBUG environment variable.
778[11:48:10] <gvvg> is there anyway to see if it's a perl
thing that is causing the problem - is there any one error message
that could help
779[11:48:19] <ratrace> gvvg: what else did you do today?
reboot?
780[11:48:19] <gvvg> just this morning at 4am
781[11:48:28] <gvvg> apt update/upgrade reboot
782[11:48:30] <gvvg> nothing else
783[11:48:33] <genr8_> as long as that list is, that just says
it was a minor update: perl:amd64 (5.24.1-3+deb9u6, 5.24.1-3+deb9u7)
784[11:48:40] <ratrace> gvvg: I believe it's not todays
upgrade, but reboot just brought out whatever broke earlier
785[11:48:58] <ratrace> you have relevant, perl updates from oct
20th. that wouldn't restart the apache process.
786[11:49:02] <themill> I wonder what Upgrade:
libparams-classify-perl:amd64 (0.013-5.3, 0.013-6+b1),
liblocale-subcountry-perl:amd64 (1.63-1, 2.01-1) was from
787[11:49:03] <ratrace> so reboot restarted it and boom problem
798[11:50:32] <ratrace> either you have a panel, or someone else
playing root
799[11:50:43] <gvvg> no as I said I did it manually
800[11:50:50] <gvvg> they were not upgraded automatically
801[11:51:01] <gvvg> can I just remove them both?
802[11:51:03] <genr8_> maybe that was held back for a reason? :)
803[11:51:10] <gvvg> lol - yeah I guess
804[11:51:18] <ratrace> you didn't say you did it manually,
you said "not specifically" assuming in response to my
question of if you installed those two perl scripts
811[11:52:22] <genr8_> im gonna assume A) it was held back for a
reason, B) on the second held package, going from version 1.6 to 2.0
broke something since its a major version update, and C) you likely
need them and cant just uninstall them, youd wanna revert/downgrade
if anything
812[11:52:27] <themill> well it *should* have been July but was
only installed in Oct
842[11:59:09] <gvvg> I'm going to call the company - thank
you guys - this is way beyond me - I shouldn't have touched
anything
843[11:59:21] <gvvg> I really appreciate everything you guys
have suggested
844[11:59:27] <ratrace> tnR: hi, you'll need to be more
patient and maybe repeat your question sometime later. EFI
installations on raid0 are not something done that often
850[12:01:12] <ratrace> tnR: for starters grub couldn't
work with /boot on raid0 in the past, I don't know if that
changed with recent versions, or you still need a separate,
non-raid0, /boot
851[12:01:14] <genr8_> tnR, well, think about it right, if its
software RAID-0, how are you gonna boot
854[12:02:20] <ratrace> gvvg: as I thought, there's a
free.biz repo. I'm also guessing they install packages that
might've conflicted with regular debian ones, hence that
package held thing. but that's a guess. definitely DO call them
and have them sort this out, if you're paying them to do so
855[12:02:41] <gvvg> ratrace: thank you
856[12:03:13] <genr8_> yea the logic of it is off. /boot comes
first, software RAID-0 comes next, thus grub is like HUHH. its
perhaps possible but its probably beyond that installers internal
logic to figure out
858[12:03:53] <H-var> does anyone know how to enable error
logging in steam?
859[12:04:19] <ratrace> H-var: did you check if
~/.steam/debian-installation/error.log is there? also, for which OS
you're asking? I thought you asked the same for gentoo
yesterday
860[12:04:29] <ratrace> oh no, wait, confused you with someone
else :)
861[12:04:36] <genr8_> you could have a small seperate /boot
partition on either drive, and use LVM striping with the remainder
862[12:04:43] <ratrace> (y'all having same issues lol, and
I'm in too many chans)
863[12:04:46] *** Quits: isostatic (uid224824@replaced-ip) (Quit: Connection closed for inactivity)
864[12:05:05] <genr8_> or boot off USB.
865[12:05:22] <ratrace> OR use raid1 for /boot, 0 for whatever
needs 0
866[12:05:30] <ratrace> grub can work with /boot off of raid1
867[12:05:45] <genr8_> yea
868[12:06:39] <genr8_> at any rate its a "your current
config does not compute" error
869[12:07:21] <genr8_> ArchWiki might have info on the concepts.
*hides*
870[12:08:12] <ratrace> no shame, archwiki IS a good source of
info, even for debian things. I've set up gpu passthrough on my
debian consulting the archwiki primarily
890[12:19:03] <ratrace> tnR: use those two disks as non-raid PVs
in LVM
891[12:19:56] <ratrace> ie. like linear volumes, not striped
892[12:20:49] <nkuttler> is there a way to mark a package as
"don't remove"? still allow upgrades, so not hold
893[12:20:52] <genr8_> I would consider using LVM either way
894[12:21:00] <ratrace> that said... I'm not sure if the
installer itself can do all that. BUT... you can install regularly
on ONE drive with LVM with the installer, and then later just add
the other PV
895[12:21:03] <H-var> hi ratrace here I found the error log. I
renamed the old one, restarted steam, and tried running an app, and
it didn't run. Can you guys help me to make sense of these
errors, and make my games work on my debian please?
replaced-url
897[12:22:02] <H-var> please guys help me to be able to play
games on my debian. That would be awesome, if I could play the games
on linux!
898[12:22:16] <ratrace> H-var: proton or linux games? btw those
LD_PRELOAD errors due to wrong ELF class are safe to ignore
899[12:22:35] <ratrace> H-var: also, which gpu?
900[12:23:02] <H-var> I have gp102 (1080ti)
901[12:23:12] <H-var> on 8700k cpu
902[12:23:37] <ratrace> H-var: and using the proprietary
nvidia-driver?
903[12:23:53] <H-var> I use the drivers from official deb repos
904[12:23:57] <ratrace> first of all, I'd recommend you
install the driver from buster-backports.
905[12:24:01] <H-var> no backports
906[12:24:28] <ratrace> you wanna play games on debian,
you'll use backported driver or have issues :) a newer kernel
would help too, also from backports.
907[12:24:35] <H-var> do I have to uninstall the old ones first,
or how does it work?
908[12:24:57] <ratrace> H-var: apt install -t buster-backports
nvidia-driver will replace the stable one with backported one. just
enable the backports repo in sources.list
909[12:25:00] <H-var> do I have to uninstall xserver and
everything else?
913[12:25:31] <ratrace> as root of course, in whichever way you
want to become root. don't use sudo blindly.
914[12:25:52] <ratrace> also, if you will bump the kernel to
backported one, make sure you do it _independently_ of nvidia
upgrade, with a reboot between two steps.
915[12:26:00] <H-var> ratrace how do I install newer kernel from
backports?
916[12:26:12] <H-var> is it gonna create issues with currently
installed software?
918[12:26:32] <ratrace> H-var: normally it should not unless
you're running something that doesn't work with 5.8
919[12:26:44] <H-var> I'll go backup my linux partition
first, ratrace
920[12:26:54] <ratrace> most wise choice.
921[12:27:04] <genr8_> if you use any other DKMS related
packages youll possibly need to upgrade get those from backports
also
922[12:27:24] <H-var> I'll screenshot the commands. Ok,
I'll go backup it first. See you later, guys!
923[12:27:39] <H-var> what is DKMS genr8_?
924[12:27:49] <H-var> how do I know, if I have those packages
installed?
925[12:28:04] <genr8_> dynamic kernel module system. it would
say something about "building modules failed" when you put
the new kernel in
926[12:28:18] <H-var> roger
927[12:28:20] <ratrace> DKMS is system used to dynamically build
kernel modules/drivers because they're not part of regura
kernel. those are very rare. nvidia is a DKMS thing. ZFS is another.
as example
928[12:28:26] <ratrace> *regular
929[12:28:27] <genr8_> its not common but its not uncommon
either
930[12:28:53] <H-var> I only have standard soft that comes with
debian xfce
931[12:28:56] <ratrace> usually it's stuff you must
_deliberately_ install, and usually it's non-free due to
license (which is why they're DKMS and not part of regular
kernel)
932[12:29:23] <H-var> and then steam, double commander, and
veracrypt basically
933[12:29:26] <H-var> and that's all
934[12:29:34] <ratrace> ZFS being exception, it's in
contrib, not non-free, and yet the license prevents it to be part of
regular kernel source
935[12:29:44] <ratrace> eh veracrypt.... is that DKMS?
936[12:30:02] <jelly> well it sure is not in-kernel
937[12:30:16] <genr8_> its something.
938[12:30:20] <ratrace> but is it a kernel module, or purely
userland thing
939[12:30:30] <ratrace> if latter, all is good. if former, check
if it works with 5.8
952[12:32:54] <ratrace> genr8_: 4.19 is too old for zen2. I must
have five point eights!
953[12:33:37] <genr8_> what does that get you
954[12:34:00] <jelly> cake
955[12:34:04] <ratrace> a sane kernel behavior
956[12:34:18] <ratrace> the cake is a lie.
957[12:34:21] <genr8_> how is a CPU too old for a kernel. unless
its a GPU too.
958[12:34:29] <jelly> too new.
959[12:34:34] <ratrace> ask teh kernel that goes berserk about
io stats for zen2
960[12:34:35] <genr8_> that
961[12:34:46] <jelly> cpu is too new. kernel is too old.
962[12:35:00] <genr8_> that doesnt sit right with me
963[12:35:31] <ratrace> but it's important that GPL
POLITICKING pathces get backported to LTS, that's very too
important. to hell with working IO subset, unless GPL politicking is
DONE!
964[12:35:32] <jelly> new cpus have new features. To make them
work properly, OS needs to have code to support the features.
965[12:35:45] * ratrace flips a bird toward kernel devs and LTS
maintainers general direction
966[12:36:12] <jelly> ratrace: you can always pay those
devs' bills instead of RH
967[12:36:18] <genr8_> im not convinced
968[12:36:21] <jelly> then you get to choose what they'll
focus on
969[12:36:25] <ratrace> jelly: they're paid VERY MUCH AND
FATLY by the foundation and other corps
970[12:36:43] <ratrace> I'd like to ask THEM to pay ME for
time wasted due to their POLITICKING.
971[12:36:45] <jelly> foundation, whose members care about their
own interests.
972[12:37:03] <ratrace> and also they're paid THROUH my
payments to RH.
975[12:37:14] <ratrace> so no, argument invalid. segfault.
reboot that.
976[12:37:19] <jelly> !pal ratrace
977[12:37:19] * dpkg points at ratrace and laughs
978[12:37:57] <rk4> how much is RH costing you
979[12:38:23] <ratrace> !bad boy jelly
980[12:38:55] <genr8_> reddit says something about C6 power
state and sleep
981[12:39:19] <ratrace> rk4: ZERO. I donate elsewhere :))))
jokes aside, the kernel devs are very well paid. Linus is very well
paid. GKH is vey well paid. They're paid far more than I am to
clean up the mess GKH causes by backporting political patches that
aren't bugfixes or secvuln fixes.
982[12:39:21] <genr8_> that makes more sense. power management
was always terrible
994[12:43:07] <ratrace> I donate to postgres, python, gentoo,
debian, and a few other projects, yearly. I did many one time
payments for little programs I used here and there. So please,
don't tell me to waste money on devs who already are paid
nicely, only so they woulnd't do stupid things like backport
political patches that have no other value.
995[12:44:45] <genr8_> i wish they would backport that zen2
patch you spoke of, whatever it was, to 4.19 for political reasons
997[12:44:57] <ratrace> how 'bout technical reasons?
998[12:45:08] <genr8_> those too
999[12:45:14] <ratrace> maybe that's too much to ask? :)
1000[12:45:34] <ratrace> it's more important to screw out
ZFS in old kernels, than backport IO fixes for zen2.
1001[12:45:47] <genr8_> :/
1002[12:45:57] <ratrace> get your priorities straight, genr8_,
you're asking for technical improvements which are in the
backseat.
1003[12:46:10] <genr8_> well
1004[12:47:01] <genr8_> i saw it as political cause if i wanna
choose to use a well stabilized kernel, I should. they did the
technical fix and its only on the 5.x which I basically regard as a
beta, since 5.9 is still fixing things from 5.4
1005[12:47:20] <genr8_> i guess you're right too though
1006[12:48:01] <ratrace> genr8_: using stabilized kernel is
technical reasons. *backporting* GPL condom skullduggery (and I
don't mean nvidia) ... and note, it's about *backporting*
that, not "okay, yeah, you wanna improve GPL compliance in
future releases"....
1010[12:49:27] <genr8_> i was under the impression that GPL
condom situation was not going to happen anymore
1011[12:50:04] *** Quits: frgo (~frgo@replaced-ip) (Remote host closed the connection)
1012[12:50:08] <genr8_> i dont remember what I read but i feel
like it was GKH saying "that aint even technically legal
anyway"
1013[12:50:18] <genr8_> could be wrong tho
1014[12:50:26] <ratrace> personally I think thta removing those
patches from the kernel does NOT violate the GPL license. but IANAL.
in fact if anything, additing additional constraints to the kernel
is in violation of section 6 of the GPL2
1015[12:50:39] <ratrace> genr8_: it's not illegal. GPL is
about _redistribution_
1017[12:51:07] <ratrace> so ZFS, for example, can't be part
of the regular kernel due to license incompatibilities regarding to
behavior of future code distribution
1018[12:51:19] <ratrace> they may NOT prevent your or I how we
use the kernel we do not redistribute
1022[12:51:44] <ratrace> if I wanna link with ZFS or most vile
proprietary code the earth has seen (hint: WSL), it's my
business and not in violation of GPL
1023[12:51:53] <ratrace> they may NOT tell me if I may or may not
link non-GPL code with the kernel.
1024[12:52:09] <ratrace> genr8_: debian doesn't
1025[12:52:26] <ratrace> ZFS is DKMS in Debian, thus there's
no redist issues
1029[12:53:06] <genr8_> so who caused it. Ubuntu? Redhat ?
1030[12:53:24] <ratrace> and what they did, they said "Nope,
you may not build ZFS kernel module because your code is not GPL and
we make our APIs work only for GPL'd code.... yeah NO,
that's in violation of section 6 of GPL
1031[12:53:28] <ratrace> genr8_: caused what?
1032[12:53:42] <genr8_> idk, I dont get the argument then
1033[12:54:15] <ratrace> Linux kernel devs are blocking ZFS from
using kernel APIs because they said "these APIs maybe be used
only be GPL licensed code"
1034[12:54:21] <ratrace> that's in violation of GPL itself.
1039[12:54:43] <oiaohm> ratrace: china ruling in linking against
GPL gets horrible
1040[12:54:45] *** Quits: Hokedli (~laslie@replaced-ip) (Remote host closed the connection)
1041[12:55:02] <ratrace> ANYHOO... GPL is about distribution.
They may not prevent me to link and use the code as I see fit
without redistributing it further.
1043[12:55:55] <oiaohm> China ruled that GPL is between those
writing the code and end user. contract law not copyright law so
redistributing is not as cleared.
1047[12:56:38] <oiaohm> Turns out GPL holds up in china contract
courts.
1048[12:56:52] <genr8_> interesting
1049[12:57:09] <ratrace> killing people on the streets with tanks
is also held up in china courts. so..... xcuse me if I don't
care what they think :)
1050[12:57:16] <genr8_> gonna have to read up on that then
1051[12:57:59] <ratrace> GPL holds up in western courts too, but
nobody so far even attempted to argue that GPL restricts how you USE
the code, if you don't redistribute it
1052[12:58:12] <ratrace> or otherwise: Google would be the worst
offender on the planet
1061[12:59:46] <oiaohm> DKMS with the china rulings would still
be legal if you are using a GPL compadible license.
1062[12:59:50] <ratrace> you can kiss corporate backing goodbye
with that license. and i'm telling you that as a corp that uses
GPL, supports it, donates to it.
1063[13:00:06] *** Quits: alpernebbi (~quassel@replaced-ip) (Remote host closed the connection)
1064[13:00:14] <genr8_> good. they can make their own software.
1065[13:00:29] <ratrace> GoodLuckWithThat(tm) :)
1066[13:00:36] *** Quits: met (~met@replaced-ip) (Ping timeout: 256 seconds)
1067[13:01:01] <oiaohm> some of the other fun parts is some of
the patent grants to the Linux kernel are only for use with GPL
compadible works.
1069[13:01:24] <ratrace> oiaohm: section 6 contradicts itself
then
1070[13:01:29] <genr8_> they were not using GPL out of the
kindness of their hearts to be nice to us and contribute back, they
did it because it was cheaper
1071[13:01:45] <genr8_> and then they abuse it through legal
loopholes on the side
1073[13:02:06] <ratrace> true it's not about kindness of
their hearts. but they DO contribute becuse they see the value in
doing so: less work to maintain their custom fork
1081[13:03:50] <oiaohm> ratrace: you overlooked "subject to
these terms and conditions. " in section 6
1082[13:04:10] <ratrace> And then you have Netflix that uses BSD
_and_ very much contributes back because they know the value in
doing so, value for themselves, even if it's the most selfish
reason: it works both ways.
1083[13:04:11] <oiaohm> So section 6 only covers you ass if you
are using stuff that is GPL compadible.
1084[13:04:53] <ratrace> oiaohm: which the Linux kernel is, and
yet it's putting additional constraints that are NOT covered by
GPL: forcing other code to be GPL compliant to use the APIs
1089[13:05:36] <oiaohm> No section 6 states by GPLv2 terms.
1090[13:05:37] <ratrace> but nvm this... until it's tested
in court, this is just pure speculation.
1091[13:05:39] <genr8_> i thought that always was the constraint,
thats why we've had linking exceptions like OpenSSL
1092[13:05:42] <oiaohm> Not by GPLv2 incompadible terms.
1093[13:06:00] <gvvg> Hi again - I'm trying to install the
package but it has unmet dependencies - I have installed the
packages via cpan - is there a way to get the apt install to ignore
unmet dependencies?
1094[13:06:54] <ratrace> genr8_: by that logic, WSL is illegal
1095[13:06:59] <oiaohm> ratrace: basically the Linux kernel is
just enforcing constraints saying GPL compadible code that is in
fact in section 6 gplv2 if you read it carefuly.
1096[13:07:11] <ratrace> WSL is code that's definitely not
GPL using GPL code APIs.
1097[13:07:24] <oiaohm> WSLv1 you mean.
1098[13:07:31] <oiaohm> WSLv2 uses a full blown Linux kernel.
1099[13:07:38] <ratrace> in dom0
1100[13:07:38] <klys> gvvg, short answer: no, you shouldn't.
long answer: it may be possible with approptiate edits to
/var/lib/dpkg/status.
1102[13:07:49] <genr8_> likely for that reason. microsoft knows
what they did.
1103[13:08:05] <gvvg> an example is freeside-lib : Depends:
libgeo-streetaddress-us-perl but it is not installable
1104[13:08:13] <gvvg> I have installed that perl module using
cpan
1105[13:08:17] <jelly> !equivs
1106[13:08:17] <dpkg> equivs is a package that enables you to
create dummy packages that tell <apt> you really have
installed (through some other means) the package. apt install
equivs, and read /usr/share/doc/equivs/*, see also <usrlocal>.
A better plan is often to adapt the Debian packages to your needs,
ask me about <package recompile> <uupdate> <ssb>.
1107[13:08:20] <jelly> gvvg: ^
1108[13:08:26] <ratrace> klys: gvvg: what about installing
specific packages via apt, instead of cpan. it's not automatic,
but you should be able to install -perl stuffs you need, via apt
1109[13:08:30] *** Quits: de-facto (~de-facto@replaced-ip) (Remote host closed the connection)
1110[13:08:31] <oiaohm> There is work to allow Linux kernel as
hyper-v dom0
1113[13:08:44] <judd> No package named
'libgeo-streetaddress-us-perl' was found in amd64.
1114[13:08:54] <gvvg> ratrace: right that's what I did
1115[13:08:54] <ratrace> oiaohm: I know, that's why I
mention it. that's SURELY in violation then. non-gpl code using
gpl APIs
1116[13:08:57] <jelly> ratrace: going to be a bit hard to install
that via apt
1117[13:09:26] <ratrace> jelly: what's "that"?
1118[13:09:49] <gvvg> if I did install it by cpan I'd like
to by pass the apt warning
1119[13:09:49] <jelly> gvvg: if you care about things being neat,
you can use dh-make-perl to make local .deb from CPAN modules
1120[13:09:51] <ratrace> from what I understood, packages were
installed via cpan. if the same packages are covered by apt, they
can be (manually) replaced, no
1121[13:09:54] <ratrace> ?
1122[13:10:06] <genr8_> microsoft is top on my list for "not
invited to the linux club"
1133[13:11:48] <gvvg> most of the time I can make stupidly easy
stuff :)
1134[13:11:51] <gvvg> thank you
1135[13:11:53] <oiaohm> ratrace: Microsoft hyper-v drivers in the
Linux kernel are all GPL or GPL compadible license. Hyper-v
interface to hypervisor was used for windows first and still
contains the special value that is bigboobs in hex to work.
1136[13:12:21] <genr8_> howd that pass CoC :P CANCEL!
1137[13:12:21] <ratrace> oiaohm: long story short, GKH explicitly
stated on the mailing list he's deliberately screwing ZFS over.
Because the GPL symbols export work by stating which licenses are
okay for the export. CDDL is not on the list, EVEN THOUGH, it's
a libre/open source license.
1138[13:12:34] <jelly> gvvg: whoops, that's old and
incomplete syntax. dh-make-perl --cpan Geo::StreetAddress::US
1139[13:12:46] <jelly> no more --build apparently
1140[13:12:49] <ratrace> oiaohm: so yeah I can understand them
not wanting _proprietary_ modules abuse the GPL APIs. that's
fine. but ZFS is not proprietary, it's open AND libre, just not
complatible with GPL
1141[13:13:11] <oiaohm> Mixing CDDL and GPLv2 has been legally
questionable thinking that CDDL was designed exactly to prevent
that.
1143[13:13:58] <ratrace> oiaohm: which is the reason ZFS is not
merged. but **NOT** the reason they should prevent exports for API
usage by DKMS'd code
1144[13:14:23] <oiaohm> China ruling DKMS the module like ZFS may
not be legal either.
1145[13:14:27] <genr8_> are you sure that ZFS was the issue at
hand
1146[13:14:30] <ratrace> we're talking about two different
things: merging code to main tree vs using the kernel's APIs by
out of tree modules
1147[13:14:47] <ratrace> oiaohm: I really do not care what China
is ruling
1148[13:14:49] <oiaohm> The big thing that caused the recent
crack down was in fact Nvidia.
1149[13:14:51] <genr8_> it could be nvidia as well
1150[13:14:51] <ratrace> nobody in the west should either.
1151[13:15:24] <ratrace> oiaohm: AND THAT IS FINE :) nvidia
driver's license is not on the list of licenses allowed by GPL
exports
1152[13:15:28] <oiaohm> the netgpu patch for the Linux that was
linking against nvidia cuda stuff binary only.
1153[13:15:38] <ratrace> CDDL is not either but CAN BE because
it's not proprietary, it's libre and open.
1155[13:15:49] <oiaohm> No CDDL is not that simple.
1156[13:15:52] <ratrace> GKH and friends deliberelately do not
want to add the license
1157[13:16:05] <ratrace> oiaohm: it's about DKMS AND NOT
MERGING UPSTREAM, GAAAH
1158[13:16:12] <oiaohm> reason why ZFS don;t want to change from
CDDL to a new license is if they do they don't have Orcales
patent grant.
1159[13:16:20] <oiaohm> if they change license.
1160[13:16:20] <genr8_> ^
1161[13:16:26] <jelly> gvvg: sorry, I really should verify a
command line actually WORKS... Try this: dh-make-perl --build --cpan
Geo::StreetAddress::US
1162[13:16:33] <ratrace> it's very simple. they're not
violating the GPL nature of the main source by allowing CDDL to use
the APIs
1163[13:16:33] <oiaohm> Now are those Orcale patents compadible
to be in a GPL work.
1164[13:16:48] <ratrace> are you listening......
1165[13:17:10] * jelly points ratrace and oiaohm gently in the
direction of #debian-offtopic
1166[13:17:11] <ratrace> I'm not talking about merging the
potentially patented code (and btew OpenZFS is fork and rewirte
before Oracle's acquisition)
1167[13:17:12] <oiaohm> When you license patents as only X
licenses only things get horrible,
1169[13:17:50] <gvvg> I got
/root/Geo-StreetAddress-US-1.04/../libgeo-streetaddress-us-perl
already exists, skipping rename at
/usr/share/perl5/DhMakePerl/Command/make.pm line 599.
1170[13:18:00] <ratrace> that affects only MERGING code with
upstream and redistributing it further. it has zero applicability to
locally built binaries and DKMS linked code. and indeed, offtopic,
so I'm done :)
1171[13:18:24] <oiaohm> ratrace: Ihave moved over to
#debian-offtopic
1172[13:18:33] <gvvg> worked!
1173[13:18:36] <jelly> gvvg: don't run it as root!
1205[13:37:03] <jelly> ratrace: it's a metapackage. I'm
not sure a dkms module deps are smart enough to pull the actual
latest headers that match the latest kernel image
1207[13:37:29] <jelly> H-var: are you asking how to see which
firmware packages are installed?
1208[13:37:38] <ratrace> jelly: I don't know, but if the
headers were installed to begin with (say, as needed for dkms) then
upgrading to backported kernel should've upgraded the headers
package too
1215[13:38:06] <H-var> is this how you do it in linux?
1216[13:38:10] <ratrace> will someone FINALLY fix dpkg NOT to
respond to anything other than ! ?
1217[13:38:13] <H-var> I searched for firmware-
1218[13:38:24] <H-var> I found firmware-misc-nonfree and
firmware-free
1219[13:38:51] <jelly> so install those from buster-backports now
1220[13:38:54] <ratrace> jelly: well, just checked a system I
bumped to 5.8 and I'm 100% sure I never installed headers
explicitly and there's both nvidia and zfs
1221[13:38:59] <jelly> thereby upgrading them
1222[13:39:02] <H-var> installed. I ran sudo apt install -t
buster-backports firmware-linux-free but got "
irmware-linux-free is already the newest version (3.4)."
1223[13:39:12] <n4dir> i wouldn't have known that installing
the kernel from backports installs the according headers too. Must
be new
1224[13:39:14] <ratrace> another system bumped to 5.8 wihtou
dkms, no headers installed, and none were additionally pulled in
1225[13:39:17] <H-var> it should be ....bpo10~" in the end,
right?
1245[13:48:11] <judd> Package redshift (x11, optional) in
buster/amd64: Adjusts the color temperature of your screen. Version:
1.12-2; Size: 105.9k; Installed: 586k; Homepage:
replaced-url
1253[13:53:09] <jelly> ws2k3: see if your daemon supports setting
SSL/TLS protocol version somewhere
1254[13:54:02] <ws2k3> jelly: i already put this
ALL:!aNULL:!SSLv3!TLSv1.3 into my TLSCipherSuite but logs show that
it still uses TLS 1.3
1255[13:55:40] <jim> hi, trying to help someone with this:
[8086:15f3] is 'Ethernet Controller I225-V' from
'Intel Corporation' with kernel modules
'snd-hda-intel', 'ata-generic' in stretch. See
also
replaced-url
1256[13:56:23] <jelly> ws2k3: Ciphers are a different thing from
Protocols.
1257[13:56:42] <jim> I wanted to see if a netinst image would see
the card and work with it; he says the card doesn't work at
all, except under windows
1258[13:56:48] <jelly> sorry. Cipher suites are a third different
thing, too.
1267[13:59:21] <genr8_> standard intel Ethernet controller, but
you mention a sound card and a hard drive controller module ? and
then stretch but no stretch ? none of this makes sense
1277[14:03:58] <genr8_> hmm i guess its not standard then.
"igc" is the module tho
1278[14:04:54] <jim> he has installed another dist; I want him to
download a debian netinst to see if it works with his card
1279[14:05:18] <n4dir> if anyone is trying to help or "give
back", that should be the basics he knows. But that might well
be me.
1280[14:05:27] *** Quits: Haudegen (~quassel@replaced-ip) (Quit: Bin weg.)
1281[14:06:05] <jelly> jim: that bug report suggests a recent
bullseye with 5.8 or 5.9 kernel might work. Or installing buster,
then upgrading to buster-backports kernel.
1282[14:06:36] <jelly> jim: but I am not thrilled at all with you
proxying questions from non-debian users in ##linux to here
1287[14:09:19] <jim> not asking any questions about the dist he
currently has installed... in fact, if this works, maybe he'll
install the debian, and then it should be all good
1288[14:11:25] <jim> I have no idea if he's actually a user
of that dist or not, I'm given to understand he tried to
install it but couldn't get net working, and about that, he
felt that the installation of the other dist may have no point
1290[14:11:57] <jim> but I'll find out what he wants to do
1291[14:12:18] <jelly> jim: you're helping someone
who's not running debian, and have avoided saying which distro
it was so far.
1292[14:12:24] <ychaouche> guys, I would like to do an unattended
dpkg-reconfigure locales on distant machines, is there a way to do
that ?
1293[14:12:39] <ychaouche> unattended or merely just not
interactive
1294[14:12:53] <ychaouche> I would like to set the system locales
to en_US.UTF-8
1295[14:13:07] <jim> do you really want to know? maybe he
won't be running it long, if this experiment works
1296[14:13:21] <jim> I'll tell you if you want to know
1297[14:13:52] <jelly> jim: no. I don't want that. I want to
keep helping Debian users, and not random Linux users.
1298[14:13:58] <ratrace> ychaouche: set them in /etc/locale.gen
and run locale-gen
1299[14:14:06] <jim> but you're right, I avoided saying what
dist it is so far, because I'm given to understand it's
off topic here
1300[14:14:10] <jelly> jim: which is why I'm active in here,
and not in ##linux
1301[14:14:37] <ychaouche> ratrace, ok thanks
1302[14:14:42] <jelly> jim: that means the issue is offtopic in
here as well, unless you're SURE the user wants to switch to
Debian
1303[14:14:52] <jim> well let me find out what he wants to do
1304[14:15:12] <jelly> jim: so I'll ask you to stop
screening questions from other channels and coming in here for
expert help
1305[14:15:14] <jim> I have no idea yet, so far, I just want to
see if it works
1306[14:15:35] <ratrace> ychaouche: also there's
DEBIAN_FRONTEND=noninteractive for non-interactive use of apt and
dpkg reconfigure but that then uses defaults, basically for
automation you're better off using files directly.
dpkg-reconfigure is really just a user-facing convenience tool
1307[14:15:55] <genr8_> it only works if you know how to get it
to work. if you don't, it doesnt.
1308[14:15:57] <genr8_> thats the TLDR
1309[14:16:01] <jelly> ychaouche: locales package has debconf
values to set, use preseed
1310[14:16:04] <jim> I'm pretty sure that if it does work,
he
1311[14:16:06] <jelly> !preseed
1312[14:16:06] <dpkg> [preseed] Preseeding provides a way to set
answers to questions asked during the Debian installation process,
to allow for a streamlined or completely automated installation. The
<install guide> provides an extensive appendix on preseeding
with a full example. See
replaced-url
1313[14:16:11] <jim> would switch
1314[14:16:22] <jelly> jim: no excuses please, it's not the
first time you've done this
1315[14:16:42] <ratrace> jelly: can preseed be used to alter
locale post-hoc?
1316[14:16:54] <jim> well let's see what he wants to do
1317[14:17:14] <jim> you're of course welcome to observe
1318[14:17:21] <jelly> ratrace: typically you seed debconf
answers, then do a dpkg-reconfigure
1319[14:17:52] <jelly> with a noninteractive frontend
1320[14:18:16] <ratrace> I'll look into that. so far I
prefered the automation did what dpkg-reconfigure would do, for a
more robust solution
1321[14:18:37] <jelly> dpkg-reconfigure itself tends to do what
dpkg-reconfigure does
1322[14:19:19] <jelly> I don't remember when we set
locales...
1323[14:19:22] <ratrace> yes but it's abstraction
1324[14:19:37] <ratrace> btw the second link doesn't work
1334[14:25:18] <ratrace> no, the idea here is to use
debconf-set-selections, I think, to set dpkg-reconfigure defaults
and let dpkg-reconfigure then work normally via interactive frontend
1335[14:25:37] <ychaouche> interactive or non-interactive ?
1336[14:25:38] <jelly> ychaouche: no, you can seed answers to
debconf in advance at any point
1338[14:25:57] <ratrace> I just have to figure out how to
actually do that :)
1339[14:26:30] <ratrace> debconf-get-selections for example shows
exactly the things for locales: the locales enabled multiselect, and
default locale. corresponding to /etc/locale.gen and
/etc/defaults/locale
1340[14:26:58] <ratrace> so by this example, I'm tryna
figure out how to use debconf-set-selections, or whatever, to preset
what I want for dpkg-reconfigure to do it automatically
1341[14:27:49] <ratrace> and yay ther's ansible module for
it..... but I wanna do it manually
1345[14:32:27] <ws2k3> i would like to configure ssh to it can
authenticate users against an external source.(php script or mysql
or something). what options do i have for this?
1346[14:32:37] <jelly> ratrace: you can dpkg-reconfigure locales
(or whatever) manually, then grab lines for that package from
debconf-get-selections output
1347[14:32:59] <jelly> and feed them to debconf on another system
1381[14:59:04] <ychaouche> PATH IS
/home/ychaouche/.pyenv/plugins/pyenv-virtualenv/shims:/home/ychaouche/.pyenv/shims:/home/ychaouche/.pyenv/plugins/pyenv-virtualenv/shims:/home/ychaouche/.pyenv/shims:/home/ychaouche/.nvm/versions/node/v8.7.0/bin:/bin:/home/ychaouche/.pyenv/bin:/home/ychaouche/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin:/usr/games:/usr/local/games:
1389[15:05:18] <ychaouche> no, I smell a rat, it's not
asking for my password and then executes next command in the script
(another scp, then an ssh command). It seems to be stepping to the
next command after a timeout
1390[15:05:58] <ychaouche> ah there it is
1391[15:06:00] *** Quits: DodgeThis (~DodgeThis@replaced-ip) (Remote host closed the connection)
1392[15:06:02] <ychaouche> maybe a network problem
1551[17:32:20] <sney> maybe the maintainer reorganized the
package split but didn't set the Breaks/Replaces in d/control
to account for it
1552[17:32:26] <jelly> TomyWork: if they come from a PPA, report
to the owner of said PPA. Their build system might be broken and
they might have debs from multiple builds in the repo with the same
version (which Should Not Happen)
1553[17:32:55] <jelly> if they're from debian,
1554[17:32:58] <jelly> !bat
1555[17:32:58] <dpkg> [Basic Apt* Troubleshooting]. To diagnose
your problem, we need ALL OF THE FOLLOWING information: 1. complete
output of your apt-get/apt/aptitude run (including the command used)
2. output from "apt-cache policy pkg1 pkg2..." for ALL
packages mentioned ANYWHERE in the problem, and 3. "apt-cache
policy". Use
replaced-url
1556[17:34:12] <TomyWork> i know what to do
1557[17:34:34] <TomyWork> I'm just confused about why this
worked in the first place
1566[17:39:03] <jelly> TomyWork: sounds like the binary packages
probably ought to split more, and keep the config file in a
foo-common deb, then depend on it in both of those
1567[17:39:37] <TomyWork> one of them is a foo-common type
package
1568[17:39:55] <TomyWork> but apparently, they messed up
1569[17:40:00] <jelly> then the other probably shouldn't
ship the same fucking file path
1572[17:41:42] <TomyWork> but like i said i know what to do.
I'll just build this myself (doing that anyway for unrelated
reasons) so I can use the software again.
1651[18:52:14] <dpkg> README.Debian (or README.Debian.gz) is a
document found in the /usr/share/doc/$packagename/ which explains
any Debian specific details in the package's operation or
configuration.
1653[18:53:32] <karlpinc> Exagone313: I don't know about
cloud-init, but the readme.debian might help. (or not) Anyway,
I'm unclear on daily images. I don't know anything about
them but thought that they were for testing, not stable. Could be
entirely wrong.
1674[19:08:42] <Exagone313> karlpinc: unfortunately I don't
see anything specific to Debian in that directory, I'll take
that cloud-init documentation is incomplete and that I need to
change the password differently (probably with chpasswd -e)
1705[19:15:59] <dpkg> The deb-multimedia.org repository was once
the only way of getting good multimedia support. By the release of
Debian 6.0 "Squeeze" this was mostly not true and for
Debian 7 "Wheezy" this situation is even better. Libraries
from dmm are known to cause incompatibilities with packages from
Debian; packages may not meet the <DFSG> so are not allowed in
Debian. Use dmm as a last resort; ask me about <dmm pinning>
and <dmm remove>.
1706[19:16:12] <LuxInterior> oh never heard of that either
1707[19:17:04] <LuxInterior> so add buster DMO to repo?
1708[19:17:13] <LuxInterior> repo to sources
1709[19:17:36] <rgr> whats the correct wz to configure a keyboard
in debian_ Google chucks up loads of conflicitng indo. Ive tried
locales, console-data/setup. contents of /etc/default/kezboard sazs
de but when I login its a british keyboard.
1718[19:20:33] <sney> LuxInterior: if venv seems too complicated,
you're definitely going to break something by enabling dmo. in
this case, I recommend option #3: wait for youtube-dl 2020-11-01 to
migrate to buster-backports
1747[19:37:30] <sney> LuxInterior: pip doesn't require it, I
said maybe venv so that the app you install is in an isolated
environment and can't cause any weird conflicts with system
python
1748[19:37:36] <sney> this has been known to happen with pip on
debian
1810[20:29:08] <karlpinc> LuxInterior: You can run the very
latest youtube_dl in a venv. It does not mess with anything
installed anywhere else but within the venv. Venvs are designed to
be discarded; if you need to you can have several laying around with
different versions in them.
1811[20:29:59] <karlpinc> LuxInterior: That's the long
answer. The short answer is: Don't mix package managers. Pip is
a package manager and shouldn't be installing anything anywhere
to do with apt.
1812[20:31:19] <karlpinc> LuxInterior: Installing with pip into
~/.local/ is just a un-flexible corner case of installing into
virtual environments.
1885[21:26:55] <dpkg> If you have a question, just ask! For
example: "I have a problem with ___; I'm running Debian
version ___. When I try to do ___ I get the following output ___. I
expected it to do ___." Don't ask if you can ask, if
anyone uses it, or pick one person to ask. We're all
volunteers; make it easy for us to help you. If you don't get
an answer try a few hours later or on
replaced-url
1887[21:29:42] <Brigo> hyiltiz, programs get pids when launched,
so you have to monitorice when a program is launched and check the
pid.
1888[21:30:02] *** Quits: Vizva (~Vizva@replaced-ip) (Remote host closed the connection)
1889[21:30:20] <hyiltiz> yeah; that monitoring should either
happen at the process that spawnes everything else (like systemd) or
at the kernel level, right?
1890[21:30:42] <greycat> To be pedantic, a new PID is born when
an existing process forks itself. The child will often then exec a
new program. So, the PID comes first, and the new program second.
1908[21:36:40] <karlpinc> hyiltiz: You might be interested in
wazuh for generating the graphical output you're looking for.
1909[21:36:50] <hyiltiz> Hmm thx! That seems to be logging ever
since I ran auditctl -a task,always; now checking which config file
to edit how so full logging starts since startup
1940[21:49:21] <hyiltiz> karlpinc: since wazuh isn't on
debian repos, and seems too commercial, I am hesitant; do you know
if they have a community support version? Their installation is
talking about a manager and an agent.
1941[21:49:40] <hyiltiz> If I could do it with auditd, I'd
rather not install yet another monitoring program into my system tho
1983[22:23:55] <karlpinc> hyiltiz: Wazuh is completely open
source. It uses auditd. It has an agent to "report back",
and do polls when it can't get the data real-time.
2085[23:35:37] <karlpinc> hyiltiz: You can also install from
github.... I don't know what the pre-requsites are. I know it
uses elastisearch under the covers.
2094[23:41:24] <hyiltiz> Thx! Suppose I did install it, do u k if
the PID histogram comes out of the box from next reboot, or I need
to configure it somehow?
2095[23:42:09] <hyiltiz> Sorry I am asking a bi too much (not
quite sold for a solution that a) requires adding external repo b)
based on java and xml)
2096[23:42:18] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
2097[23:42:23] <hyiltiz> I am willing to try it out if not too
much committment for configs and package dependencies tho
2099[23:42:30] <karlpinc> hyiltiz: I don't know. I've
logged into wazuh, and it does all sorts of graphing. And I'm
sure you can configure it to graph your histogram if it does not.
But beyond that I don't know much.
2100[23:43:15] <karlpinc> (I bet it has dependencies. I've
never setup the stack.)
2120[23:56:21] <hyiltiz> any idea how to use this? It's
documentation is too hard to dig thru the fluf of
orchestration/devop and it didn't register anything named wazuh
to my systemctl
2121[23:56:41] <hyiltiz> and it installed a full python distro
under /var/ ...
2122[23:57:02] <karlpinc> hyiltiz: No clue. It seems like a
"big project".
2123[23:57:24] <hyiltiz> welp so long then (and thanks for all
the fish)
2124[23:57:37] <karlpinc> hyiltiz: It also puts stuff in /var/
that seems like a wonky place.