107[01:11:26] <plt> The developer should have done it ine
reverse order write and distrubute on the stable version first m
wait 6 months then do the next release
201[02:59:43] <Numero-6> Hi all, i have make a service systemd
for show if the router(getaway) is up or down, so i have make log
too with StandardOutput=file:/home/user/log1.log (i would make color
error or warning, but i dont't how) My question is: Is i use ss
-natup i see the connection for test the getaway but in process/pid
i see nothing
259[03:46:36] *** Quits: Numero-6 (~Numero-6@replaced-ip) (Quit: << - Qui etes vous ? - Je suis le nouveau numero 2 -
Qui est le numero 1 ? - Vous etes le numero 6 - Je ne suis pas un
numero ! Je suis un homme libre!! >>)
480[09:05:28] <ratrace> my nginx logs stopped rotating properly.
eg. access.log is rotated into access.log.1 but nginx keeps posting
into it, as if it's not signaled to reopen logs. and looking
into the logrotate.d hook I see invoke-rc.d nginx rotate ; rotate ?!
481[09:06:09] <ratrace> and LO and behold what happened.... of
course.... rc-policy.d exits 101 to prevent service restarts...
FFFFFFFFFFUUUUUU
488[09:07:43] <ratrace> yea I remember messing with policy-rc.d
causing a metric ton of other problems... but this is plain moronic.
"rotate" never was a service command. we're in
systemd era now. why is this shiz still acting like it's sysv
bash scripts in the backend.
489[09:08:24] <ratrace> and morons voted to exacerbate the
problem even more instead of adopting systemd like it's 2020
and not 1980.
497[09:19:32] <ratrace> themill: remember who was it we
supported few days ago about using policy-rc.d to prevent service
(re)starts on install/upgrade? it's out of my backlog...
505[09:24:27] <themill> ratrace: you were talking policy-rc.d
with AquaL1te (there may have been others; I was only half paying
attention!)
506[09:25:39] <ratrace> themill: that's it, AquaL1te, much
obliged.
507[09:27:12] <ratrace> AquaL1te: heads up, messing with
policy-rc.d will cause nasty, silent, unlogged side effects as dumb
things rely on invoke-rc.d for things other than just starting or
stopping a service.
508[09:29:17] *** Quits: Night-Shade (~TimF@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
509[09:30:17] <themill> That's probably also why that demo
script was specifically filtering on actions from dpkg
512[09:31:25] <ratrace> the second positional parameter is
supposed to be action, so I'll see if filtering on that is
better than looking up dpkg in user's ps tree
513[09:31:32] <themill> Using a denylist of actions to prevent
actions you don't want (start/stop/restart) rather than an
allowlist of actions to permit would have potentially let
'rotate' through for you
514[09:32:15] <ratrace> themill: frankly, what I want to do
there is find all the nonsense like this and fix it _properly_, like
in this case, not use the init script to rotate a systemd service,
but `systemctl kill`
515[09:32:46] <themill> you'd then have the "boohoo I
don't use systemd" crowd chasing you
516[09:32:51] <ratrace> I know I can't open a bug report
for this because debian recently voted for other init support to
stay and it'll be refused as the logrotate hook has to support
BOTH inits... my rage intensifies.
517[09:33:19] <themill> When the options are "you can do A
or B" Debian always tries to do both.
518[09:33:24] <ratrace> themill: yes. so we need another
fork...... devTWOan. the Debian fork for TRUE init freedom,
systemd-only.
519[09:34:54] <ratrace> ubuntu looks more and more appealing as
dual-init problems on Debian rear their ugly heads ever more as
months pass by. but then tweakin 'buntu to remove all the nasty
THEY do (resolved, netplan, zsys, ...) is much more work.
520[09:35:24] <ratrace> maybe I should start and maintain a wiki
page. How to remove the sysv cancer from your Debian. Subscribe
today.
524[09:38:15] <ratrace> (and chances are the same is done on
ubuntu and they won't accept the change because they want to
change as few packages from debian as possible)
544[09:55:07] <dreamer> logging is pure hell now. how many flags
do I need to read a simple log?
545[09:55:10] *** Quits: Eduard_Munteanu (~Eduard_Mu@replaced-ip) (Remote host closed the connection)
546[09:55:20] <ksk> You either do not want to understand it, or
have not worked with it extensivly enough imho :>
547[09:55:55] <ksk> logging is awesome. And you only need one
flag, to specificy the unit journalctl should show you
548[09:56:08] <dreamer> every time I have to do anything other
than restarting a service I have to look up how to do things. imo
it's highly unintuitive
549[09:56:15] <dreamer> not being able to grep in a simple
log-file is really shit.
550[09:56:37] <dreamer> if you don't understand this, then
you haven't been using computers long enough
551[09:56:54] <dreamer> the only big advantage I see is readable
and easy to write init files
554[09:59:25] <ratrace> I use journald as in-memory only short
buffer I can search with journalctl's variety of filters and
options. The rest goes to rsyslog where I change the defaults etup
(which writes same data to multiple files and is very inefficient),
to have one main catchall, and one extra for errors so my log watch
script can keep an eye on those. and there's the traditional
mail logs because postfix is dumb and wants
555[09:59:31] <ratrace> to use rsyslog only.
556[09:59:35] <ksk> "jouralctl | grep foo" does also
work.
557[09:59:47] <ratrace> so for that purpose, journalctl is
awesome as I can quickly look up things. for the rest and grep, I
have the log.
559[10:00:28] <ratrace> however I often find grepping VERY
inefficient. different date formats between applications. different
languages used, eg. we have some localized WP installations and look
and behold PHP error logs in non-english. good luck parsing THAT.
561[10:00:58] <ratrace> ksk: it does but it's a bit
inefficient, and journalctl is supposedly getting a --grep flag or
something too, to make it more efficient.
564[10:02:02] <teus> windows update replaced GRUB by the windows
loader again, and i broke the bootloader by performing bcdedit (to
set it to grub again)
565[10:02:04] <ksk> I am not totally against acknowledging some
of your points, but service handling has greatly improved since
systemd, if you asked me.
566[10:02:11] *** fling_ is now known as fling
567[10:02:14] <ksk> !grubinstall
568[10:02:18] <teus> im trying to reinstall grub simply, but
everything is going wrong
569[10:02:21] <ksk> eh.
570[10:02:37] <teus> im on debian live, chrooted, but im
probably on legacy boot so i cant write to uefi
571[10:02:39] <ksk> !fixmbr
572[10:02:39] <dpkg> To reinstall <GRUB> boot to your
Debian install disk/live CD, switch to the other console (Alt-F2),
mount your root filesystem (mount -t ext4 /dev/whatever /target ;
mount --bind /dev /target/dev ; mount -t proc none /target/proc ;
mount -t sysfs none /target/sys), chroot into it (chroot /target),
run "mount /boot/efi" on EFI and "update-grub
&& grub-install /dev/whatever". See also <rescue
mode>, <dual boot guide>, <supergrub>.
603[10:25:37] <shtrb> anyone have a clue why sudo hcitool -i
hci0 lescan , get me a "Set scan parameters failed:
Input/output error" while regular scan work as expected (AC9560
, should have LE)
606[10:29:39] <AquaL1te> ratrace: thanks for the headsup! but i
don't do the 101 exit code, i do the 106 and then go through
this list of actions "reload force-reload stop
force-stop". so if the service is already running, a reload is
done. if the service is stopped, then the reload will fail, as well
as the force reload. and then as a final option a stop is done. if
the service is not running (e.g. when installing it the first
607[10:29:45] <AquaL1te> time) then a a reload is done, but
won't do much else. if all of that fails a stop is done. i
haven't seen a situation where stop is ran. so maybe i'll
remove it to ensure best predictable behaviour. in which stage does
invoke-rc.d trigger a logrotate? and by the way, i only do the exit
106 when i update or install software with ansible. i then set the
POLICYRCD environment variable to point to my script
608[10:29:51] <AquaL1te> that does that exti 106 and provides
those options.
609[10:31:06] <ratrace> AquaL1te: how would that handle the
"rotate" action that I found this morning was busted due
to my messing with policy-rc.d ?
610[10:31:40] <ratrace> `invoke-rc.d rotate nginx >/dev/null
2>&1` is the atrocity in nginx' logrotate log reopen
hook
611[10:32:12] <ratrace> ... nginx rotate ... typo
612[10:32:23] <AquaL1te> ratrace: that depends on when and how
that's triggered. ah i see. well in that case it will just
work. because that's not triggered by ansible, so that
environment variable is not set. so normal operation is then done
613[10:33:21] <ratrace> AquaL1te: can you pastebin your
policy-rc.d perhaps?
625[10:39:38] <AquaL1te> ratrace: added the extra info:
replaced-url
626[10:41:10] <ratrace> AquaL1te: I see. so with 106 exit code,
the script must output a list of allowed actions, and the invoker
decides which one to use
627[10:41:12] *** Quits: pvdp66556 (~pvdp@replaced-ip) (Remote host closed the connection)
628[10:41:28] <ratrace> AquaL1te: so you'll have to add
"rotate" to that list, for logrotated services that use
that hook, like nginx
629[10:41:46] <ratrace> and who knows what other non-standard
"service" shell script command is used by the packages.
630[10:42:07] <AquaL1te> ratrace: yes, 106 is fallback mode, and
the stdout list provided is done in that sequence. if one fails, the
next one is tried
632[10:43:07] <AquaL1te> ratrace: no ;) because this POLICYRCD
environment variable, pointint to my custom script which does the
exit 106, is only set and ran by ansible. native system stuff, like
crons, don't do that
649[10:53:13] <Guest52883> Hey guys, I'm trying to create a
new directory from root: mkdir /root/Desktop/file name...It says
"mkdir:cannot create directory mkdir /root/Desktop/file name...
So any help what must I do?
650[10:53:49] *** Guest52883 is now known as Baddaz
658[11:02:07] <Baddaz> ratrace: I'm trying to create a
directory here bro... Isn't it possible to?? (ratrace) Baddaz:
/root/Desktop .... seriously? you want to run GUI applications as
root!!
659[11:02:34] <Baddaz> From root?
660[11:02:53] <ratrace> first of all, I'm not your bro.
second, what's the _exact_ error message you get from mkdir.
third, you shouldn't run GUI applications as root.
665[11:07:01] *** Quits: endstille (~endstille@replaced-ip) (Quit: I'll be back.)
666[11:07:40] <Baddaz> I said error message :: mkdir: cannot
create directory '/root/Desktop/File_name' : Not a
directory... I'm trying to create one in my Desktop
667[11:08:43] <ratrace> Baddaz: apparently "Desktop"
is not a directory at all. what's the output of ls -l
/root/Desktop ?
668[11:09:40] <diogenes_> sudo ls i guess.
669[11:10:31] <ratrace> well if THAT has to be specified....
then they probably shouldn't touch anything in /root/ to begin
with
682[11:16:39] <ratrace> where's "Desktop" in that
ouput?
683[11:17:02] <ratrace> how about you don't type it
manually but actually copy&paste the actual output of ls -l ?
684[11:18:01] <nlpqda> 3 things piss me off the most about
debian.. My Dell laptop 1) fails to give any indication before
battery completely dains out and shut off suddenly on me. 2) adding
one more entry of recovery mode permanently for each suden shut off!
3) being unable to recover opened not saved documents e.g. gedit ...
please let me know if anyone think there are fixes to these things
as I'm using debian as my daily bases
685[11:18:07] <nlpqda> os.
686[11:18:17] *** coruja__ is now known as coruja
687[11:19:34] <Baddaz> (ratrace) Nope there isn't desktop
688[11:20:03] <ratrace> Baddaz: so you need to create it first,
or use the -p flag for mkdir if you're creating multiple
subdirectories at once
689[11:20:17] <nlpqda> photo for number 2
replaced-url
690[11:21:04] <Baddaz> (ratrace) okay lemme try it I'll let
you know
691[11:21:07] <ratrace> nlpqda: apt autoremove to remove old
kernels. it's adding one entry per kernel found in /boot
692[11:21:18] <diogenes_> nlpqda, try a different DE, for me
Debian is the smoothest OS for my hardware off all other distros.
693[11:21:29] <themill> nlpqda: looks a lot like you need to
uninstall kernel packages you're no longer using
694[11:21:32] <ratrace> nlpqda: as for 1) it is desktop
environment specific, so what diogenes_ said
700[11:24:58] <nlpqda> diogenes_: ratrace: from wandering around
many distro(s) I found out debian is the most reliable daily OS with
the least bugs in general even though it lacks many packages or
newer releases but stable
701[11:25:44] <nlpqda> Fedora is a RUN TO HILLS both as OS and
support team atittude
707[11:29:27] <KOLANICH> Hello. How can I manually create a
trigger that is executed after every package installation and gets
the name of an installed package?
708[11:30:11] <ratrace> KOLANICH: maybe via DPkg::Post-Invoke
hook in /etc/apt/apt.conf.d/... ?
709[11:30:31] <ratrace> KOLANICH: but eh, what's the final
outcome you want to achieve?
710[11:31:27] <KOLANICH> ratrace: I need to update an own
inverted index of files (mapping path in fs -> package name).
711[11:31:55] <ratrace> can you rephrase that? I don't
understand what you're saying
715[11:33:36] <KOLANICH> a just an SQLite DB with package names
+ a trie for paths mapping them to primary keys in that DB
716[11:33:48] <KOLANICH> but it has to be kept up to date
717[11:36:05] <ratrace> KOLANICH: well, you can define dpkg
triggers via /etc/apt/apt.conf.d/ maybe Post-Invoke is not what you
want, that's invoked once per transaction. so look up
/usr/share/doc/dpkg-dev/triggers.txt.gz for more info
722[11:42:06] <KOLANICH> ratrace: I have looked into them. They
have not helped much/ Files triggers look promising, but it seems
they require a deb package to be installed
752[12:08:44] <KOLANICH> themill: do hooks in
/etc/dpkg/dpkg.cfg.d/ execute even with dpkg -i? The docs says like
DPkg::Post-Invoke is an apt hook, not dpkg
753[12:09:01] *** debhelper sets mode: +l 1153
754[12:09:30] <themill> /etc/dpkg/dpkg.cfg.d/ are for dpkg;
DPkg::Post-Invoke is the dpkg step within apt
827[12:48:53] <NoneDone> I've got some issues with systemd
828[12:49:01] <ratrace> JyZyXEL: sorry, it's been a while
since I KDE'd, but it _should_ work if enabled unless
there's some bug. iirc there was a startup service for it, so
check if that's enabled too
839[12:52:36] <NoneDone> So, after reboot, i cat start it
manualy
840[12:52:41] *** cdown_ is now known as cdown
841[12:53:14] <NoneDone> but i have no idea why doesn't
start automatic after reboot
842[12:53:56] <NoneDone> There is two interfaces. First one is a
USB LAN card. Second one is a OpenVPN interface
843[12:54:53] <ratrace> I'd guess the devices become
visible in the system too late for those services. nftables is
running Before=network-pre.target, for example
844[12:55:02] <NoneDone> I thank that is the reason nftables
wont start. There is ^^^^^ under those two names of the interfaces
845[12:55:20] <ratrace> for unbound, see if it has a log of its
own with more details
846[12:56:01] <ratrace> NoneDone: how do you setup your
networking? via /etc/network/interfaces ?
847[12:56:15] <NoneDone> Ok. And how to fix nftables?
852[12:57:22] <ratrace> frankly, I'm not sure. I have a
similar problem running Bind, even though it's
After=network-online.target it'll have issues on startup and
not fetch proper DNSSEC keys, and goes bell-yup on that. So I
injected a service before bind to actually check and wait for
availability of WAN (simple ping to our external server) before
proceeding.
855[12:58:00] <tarzeau> i'm surprised after 10 years of
systemd, to still see such questions - but then i'm glad
i'm not alone with problems
856[12:58:23] <ratrace> tarzeau: I don't think it's
systemd specific issue. systemd just exposed the race conditions
that would plague any init
857[12:58:34] <themill> JyZyXEL: the plasma notifier explicitly
says it's only for the home directory, so if you've got
media/downloads in other places then it's not looking at them
858[12:58:37] * tarzeau is using cron with @reboot /usr/bin/screen
-dmS name software
859[12:58:41] <ratrace> NoneDone: btw, can you pastebin your
interfaces file? how do you bring them up? with "auto" ?
876[13:01:08] <ratrace> tarzeau: you can always switch to Devuan
877[13:01:16] <tarzeau> ratrace: having multiple people work on
systems, it's now all gotten messed up, people going for
systemd replacing cron and how hostname (the command) and dhcp
client/dns(resolv.conf) should work
878[13:01:27] <tarzeau> ratrace: i can't, there's
other people having a say
879[13:01:35] <ratrace> oh, well.
880[13:02:39] <ratrace> NoneDone: can you pastebin entire
journalctl -b or at least determine when the interface becomes
available in the boot timeline, related to startup of
nftables.service ?
881[13:03:09] *** Lord_of_Life_ is now known as Lord_of_Life
882[13:03:22] <ratrace> NoneDone: one quick remedy would be to
set the service After=network.target (be sure to comment out current
Before= as well), but that'll mean you'll have a short gap
during which the NIC is up but not netfiltered
883[13:04:32] <ratrace> tarzeau: then again you can still use
/etc with favourite $EDITOR, systemd didn't take away that.
907[13:17:32] <ratrace> NoneDone: as I suspected, nftables is
logging before the link becomes ready. One fix would be, as I
suggested earlier, you change when nftables is run, like
After=network.target (remove current After= and Before= rules).
908[13:17:59] <ratrace> Another fix would be to switch to
systemd-networkd for network configuration and make nftables
dependant on the interface .link
917[13:21:30] <NoneDone> Okey, i'm going offline for a
while
918[13:21:34] <ratrace> NoneDone: wait
919[13:21:37] <NoneDone> to do the test
920[13:21:51] <ratrace> NoneDone: I think you should also change
WantedBy to multi-user.target, I'm not sure
921[13:22:12] <ratrace> btw, I'm assuming you're doing
this to a COPY of the service unit in /etc/systemd/system/ ?
922[13:22:26] <NoneDone> Yes
923[13:22:30] <ratrace> mkay
924[13:23:41] <NoneDone> Thank you! I'm going to reconnect
925[13:23:50] <ratrace> tarzeau: btw, this whole mess is because
debian is using interfaces(5) and not systemd-networkd. networkd can
do proper signals, and services can explicitly wait on devices to
appear. right now it's a oneshot shell script for network.
pfft. :)
930[13:27:19] <tarzeau> ratrace: i get that yes, the reason why
we turn on systemd-networkd for our machines @work (that's easy
to do): but imagine that:
replaced-url
951[13:32:44] <tarzeau> yes, people play that DAYS! and systemd
kills the ascension
952[13:33:04] <ratrace> well, easy to configure journald.conf
not to output anything to console :)
953[13:33:39] <tarzeau> you only got ONE life, gameover and you
start at dungeon level 1 again... and they're random, and you
need to go down 30-45 to find the amulet of yendor (read backwards),
then bring it back to dlvl:1 and then 0,-1..-3 or so
955[13:33:57] <tarzeau> ratrace: should be default, not to be
configured
956[13:34:21] <ratrace> ForwardToConsole=no; ForwardToWall=no;
ForwardtoWall is Yes by default. and that's DISTRO defaults
fault, not really systemd's ;)
957[13:34:31] <abff_> Is there a simple package I can download
from the repositories to install debian from my existing insta
959[13:34:56] <ratrace> tarzeau: systemd internals grind my
gears in many places, but vast majority of issues I've had with
it, is due to bad distro defaults
960[13:34:57] <abff_> install onto another disk, one that used
in the livecd installer or netinst
961[13:35:19] <abff_> Or do I have to install the old fashioned
way
962[13:35:37] <tarzeau> ratrace: so biebl should set a sane
distro default... try telling him
963[13:35:58] <ratrace> NoneDone: try
After=network-online.target or After=networking.service (note:
networking not network)
964[13:36:27] <ratrace> actually... not sure networking.service
is correct condition, it's an oneshot with no readiness
notification, so try network-online.target
967[13:38:02] *** Quits: NoneDone (5218664c@replaced-ip) (Remote host closed the connection)
968[13:38:36] <ratrace> NoneDone: currently the problem is still
that "Jul 22 12:24:57 debian kernel: ax88179_178a 3-1:1.0
enx8cae4ce9bb0c: renamed from eth0" happens after
nftables.service is running; that's probably due to
interfaces(5) running concurrently and not activating the interface
fast enough
986[13:49:50] <wingedrhino> Can someone tell me if there's
a .deb powered distro that's somewhat stable, but rolling
release? I used to use Sidux back in 2008 but that project is dead
now.
989[13:51:20] <ratrace> NoneDone: same problem really, but
openvpn@.service is also running After=network-online.target ; so
basically take the common denominator and make nftables.service run
After=openvpn.service ; the openvpn@.service is
PartOf=openvpn.service, so that one target should suffice
990[13:52:01] <wingedrhino> aaro: does it aggressively track the
mainline kernel?
993[13:52:25] <ratrace> wingedrhino: with an added caveat emptor
that Debian Testing is not receiving security fixes until
they're part of regular upgrade of a package in Testing
1010[13:57:27] <ratrace> this
After=sys-devices-virtual-net-tun0.device was my next suggestion,
just make nftables wait on BOTH devices too (the problem is also
with an USB NIC)
1011[13:57:40] <NoneDone> So, i just added another line under
network-target. The line is "After=openvpn.service"
1012[13:57:55] <ratrace> so, After=network-online.target
openvpn@yourservice.service sys-devices-virtual-net-tun0.device
sys-devices-virtual-net-enx8cae4ce9bb0c.device
1040[14:11:36] <KOLANICH> ratrace: one of its features is
autodiscovery of deps. It inspects the files in the package for
deps. I mewn it scans all elfs and resolves the shared libs. Then
the shared libs paths are converted to package names using the
inverted index
1043[14:13:49] <KOLANICH> ratrace: it scans the newly built
packages. You just drops a link to a github repo with a supported
build system + additional args to it, my tool automatically builds
the package. Metadata is partially generated automatically.
1044[14:14:00] <KOLANICH> *drop
1045[14:14:11] <tarzeau> KOLANICH: where is it?
1046[14:14:45] <tarzeau> KOLANICH: does it handle font packages
that need be built using fontmake? (input files .glyphs)?
1047[14:15:10] <tarzeau> KOLANICH: scons? cmake? make? what build
systems do you support?
1048[14:15:27] <KOLANICH> tarzeau: it is called prebuilder.py but
it is very unmature and unfinished, pretty a lot of work. It
doesn't handle font packages. For now only elfs. other langs
are also planned, but not now.
1049[14:15:56] <tarzeau> KOLANICH: your nick is .pl? is your
prebuilder.py on github?
1052[14:17:43] <tarzeau> KOLANICH: are you aware of repology.org
?
1053[14:18:05] *** Quits: martian67 (~martian67@replaced-ip) (Remote host closed the connection)
1054[14:18:06] <tarzeau> KOLANICH: when i built debian packages
and solaris packages (i had some scripts to transfer meta data), and
later freebsd ports, as well
1055[14:18:08] <KOLANICH> tarzeau: my nickname is KOLANICH
1060[14:21:06] <KOLANICH> tarzeau: and yes, I am aware of
repology and I'd like to utilize it. The tool uses global
distro-agnostic package identifiers and transfoms them into
distro-specific ones according to the rules
1061[14:21:24] <tarzeau> KOLANICH: strange, i like the
debian/rules makefiles, hardly ever did i find something not doable
1063[14:23:18] <KOLANICH> tarzeau: you likely haven't used
them for more complex tasks, like repackaging fedora packages into
debian ones with fixing paths or repackaging other prebuilt packages
distributed as tar.gz.
1064[14:23:31] <KOLANICH> tarzeau: it requires custom logic
1067[14:24:01] <tarzeau> KOLANICH: i've packaged many
packages not even in fedora yet, or in fedora-but-not-yet-in-debian
for debian...
1068[14:24:29] *** Quits: gavlee (~gav@replaced-ip) (Remote host closed the connection)
1069[14:24:39] <tarzeau> KOLANICH: my starting point has always
been the upstream source repo though, and the fedora one only maybe
to look at it, how they screw around with broken things, if they do
(seldom did i find something useful)
1070[14:25:15] <tarzeau> KOLANICH: care to name a few complex
tasked software packgages you're referring to, so i can get an
idea?
1085[14:31:53] <ratrace> NoneDone: Works now? Reason I
didn't want to recommend depending on specific devices first is
to make a generic dependency case instead, so if you add or remove
NICs later, your nftables is still working properly; so I'd
rather try with After=openvpn.service, which includes the openvpn @
templated services, which includes running after
network-online.target, which includes having the static NIC brought
1086[14:31:59] <ratrace> up by interfaces(5), ie. the common
denominator of all possible cases without specific (fragile) device
dependency.
1093[14:32:58] <ratrace> NoneDone: but adding .device
dependencies SHOULD've worked. you sure you made it correctly?
can you pastebin nftables.service you have?
1099[14:34:49] <tarzeau> KOLANICH: tensorflow 1.x or 2.x?
1100[14:34:56] <KOLANICH> tarzeau: 2
1101[14:35:08] <tarzeau> KOLANICH: with cuda 9.x? 10.x? or 11.x?
1102[14:35:19] <KOLANICH> tarzeau: no cuda, cpu
1103[14:35:41] <KOLANICH> tarzeau: I have an AMD GPU
1104[14:35:44] <tarzeau> KOLANICH: we also have these problems,
and reprepro repos for like bionic, focal, xenial, stretch,
buster... with like ~550 packages
1105[14:36:37] <tarzeau> KOLANICH: we have ~200 machines, a 64
node slurm cluster, 50 machines with cuda (incl. cudnn)
pytorch/tensorflow (up to 8 gpus in one node) all of them are
workstations (some in the serverroom with 10gbit links)
1199[15:10:14] <ratrace> NoneDone: well, the device names _are_
correct. Not sure why it's not working then. I'd like to
help you more now, but I have to leave the office, be back later.
1200[15:10:51] <NoneDone> Thank you for support me
1238[15:51:52] <nbusrone> Anyone here using nvidia driver on
Desktop ? The brightness get reset for every restart , logout or
wake up on the monitor , opening a web browser with youtube also
affected.
1293[16:40:12] <dreamer> nbusrone: display brightness (laptop?)
is afaik not related to the gpu
1294[16:40:25] <dreamer> which window manager are you using?
1295[16:41:47] <nbusrone> dreamer : as I mention nvidia driver on
Desktop 440
1296[16:42:42] <nbusrone> dreamer : I am using ubuntu debian base
and I went to the #ubuntu and #nvidia channel but no one able to
answer my question
1297[16:43:27] <nbusrone> dreamer : I really hope I can seek some
information regarding with the brightness reset before i fill up the
bug report to confirm which is actually the culprit .
1298[16:44:01] <nbusrone> dreamer : can you help me ?
1308[16:51:39] <nbusrone> annadane : no one will response on my
question because I am newbie.
1309[16:51:44] <somiaj> !ubuntu
1310[16:51:44] <dpkg> Ubuntu is based on Debian, but it is not
Debian. Only Debian is supported on #debian. Use #ubuntu on
chat.freenode.net instead. Even if the channel happens to be less
helpful, support for distributions other than Debian is offtopic on
#debian. See also <based on debian> and <ubuntuirc>.
1313[16:52:08] <somiaj> nbusrone: ubuntu support is better done
over their forums, always see good stuff there. But part of a distro
is its support community. This community is only for debian pure
blends.
1320[16:53:30] <somiaj> nbusrone: I see lots of good info from
ubuntu on their askubuntur forums, that might get better support
(seaching through it too) than irc
1323[16:54:00] <somiaj> nbusrone: debian and ubuntu use the same
package manager and optional/extra package sources, but the core
systems are quite different.
1324[16:54:07] <dreamer> nbusrone: go to #ubuntu
1325[16:54:13] <dreamer> (and you didn't answer my question
at all)
1326[16:54:16] <somiaj> so outside of ubuntu can use .deb's,
there are actually tons of differences (and quite important ones)
1329[16:55:15] <dreamer> somiaj: honestly, the ubuntu forums are
usually dead ends and lots of noobs trying to "anwser" by
giving weird suggestions
1330[16:55:17] <nbusrone> dreamer : which question ?
1331[16:55:23] <dreamer> (but maybe I'm looking for too
specific things)
1332[16:55:42] <dreamer> nbusrone: I asked you what window
manager you are using. but since you are not even using debian. you
should go to the support channels of your distribution, as
suggested.
1333[16:55:47] <somiaj> dreamer: I just mention them because I
always find good answers to questions there (never asked any though)
1335[16:56:06] <dreamer> somiaj: ok, I usually find non-answers
1336[16:57:02] <nbusrone> dreamer : Wayland
1337[16:57:20] <somiaj> It could also be the only place I get
there is from google, so only find the good answers and didn't
realize there are lots of bad ones
1339[16:57:40] <dreamer> nbusrone: not a window manager. and
again -> go to #ubuntu with your questions
1340[16:58:20] <annadane> also as a general cautionary advice,
don't just google a question for ubuntu and blindly take advice
from the ubuntu forums, many of those results are years old and may
not apply
1341[16:58:21] <nbusrone> somiaj : I did a lot of google search
but no result on my question.
1342[16:58:35] *** Quits: NoneDone (5218664c@replaced-ip) (Remote host closed the connection)
1343[16:58:55] <nbusrone> dreamer : ok i will go and ask there
thanks
1344[16:59:29] <nbusrone> annadane : my question actually quite
complicated since I am newbie , I don't know how to express the
condition.
1355[17:09:57] <nbusrone> dreamer : Stacking window managers -
Metacity and compositor window managers - Wayland .I found out ,
both also the having the same issue.
1356[17:11:17] <dreamer> again, we are not here to give support
to Ubuntu
1357[17:11:29] <dreamer> please direct all your questions and
remarks to the kind folks in #ubuntu
1367[17:18:15] <epictetus2> if I want to send email from my
little personal debian server just for simple notifications, which
software am I looking for?
1372[17:23:48] <petn-randall> epictetus2: It depends to a great
amount on the details. I'd either go with postfix or
nullmailer, but depends on your use case how to configure them.
1373[17:24:04] *** Quits: zapatista (~zapatista@replaced-ip) (Remote host closed the connection)
1374[17:24:07] <petn-randall> epictetus2: What kind of
notifications do you want to send, and to where?
1375[17:25:03] <epictetus2> I have a bash script that checks
uptime of servers and I want an email when they go down.
1376[17:25:13] <epictetus2> To my personal email at google
1377[17:25:45] *** Joins: Prints (~333@replaced-ip)
1396[17:41:45] <somiaj> it allow syou to ssh to the system and
you can do most things that way, and it has x-forwarding setup, if
you want to run a xorg app
1414[17:54:54] <Mat63> Hi, A netinst CD allows me to install the
complete system ... Debian. Now it's okay that when I start I
only have tty on the screen?
1415[17:56:25] <sney> that sounds like you didn't have a
network connection during the install, and instead only got the
small base system included on the netinst iso
1416[17:56:30] <Mat63> How to access the graphical environment
from the console (tty1)?
1417[17:56:36] <sney> this is common especially if you were
trying to install over wifi.
1418[17:56:50] <Mat63> yes
1419[17:57:05] <petn-randall> !firmware installer
1420[17:57:05] <dpkg> Debian-Installer is able to load additional
<firmware>, by including it within installation media or
supplying on removable media (e.g. USB stick, floppy). See
replaced-url
1421[17:57:17] <petn-randall> Mat63: You likely want to try
installing with this. ^^^
1422[17:57:33] <Mat63> I have the option of automatic network
configuration, but failed due to ignorance
1423[17:57:36] <tinga> What's the best way to set up remote
file access (security and convenience are important, performance is
not) between two Debian boxes?
1424[17:57:37] <petn-randall> This will allow you to use you wifi
device and download any GUI you select during installation.
1425[17:58:01] <petn-randall> tinga: sshfs, I guess.
1426[17:58:02] <sney> tinga: nfs if you have a *very* reliable
network connection between them. otherwise, samba
1427[17:58:21] <sney> oh, I misread. not nfs if you care about
security
1428[17:58:49] <tinga> petn-randall, sney, sshfs is nice for
techies like me but this is for mum, so not sure; I've never
used nfs, maybe; samba I've used, still kinda painful.
1431[17:59:31] <tinga> I was thinking maybe WebDAV actually works
nowadays? Maybe not?
1432[18:00:20] <sney> webdav? er.. what's the purpose of
this file share anyway?
1433[18:00:25] <petn-randall> tinga: Are there any other details?
If it's for your non-techie mum you might want to use something
like nextcloud for file sharing and web frontend access.
1434[18:00:42] <tinga> Just exchanging files for printing or from
scanning and the like.
1435[18:01:05] <tinga> The occasional copied DVD maybe.
1436[18:01:39] <tinga> The two computers are both in the same
local network.
1449[18:05:09] <sney> Mat63: you can try to install the firmware
for your wifi card manually, but it's easier to start over with
the firmware-included netinst that dpkg linked above.
1450[18:06:01] <Mat63> I am signed to that link, so ask
1461[18:08:28] <sney> most linux wifi drivers need an extra
binary component that is not in the regular debian netinst CD image.
they are in the firmware image instead.
1462[18:08:35] <sney> oh, I see petn-randall sent the wrong link.
1463[18:08:37] <sney> !firmware image
1464[18:08:37] <dpkg> Unofficial <netinst> and DVD images
containing non-free Debian <firmware> packages are available
from
replaced-url
1465[18:08:41] *** Quits: factor (~factor@replaced-ip) (Remote host closed the connection)
1466[18:08:43] <annadane> i guess a decent general rule is free
image for desktops, nonfree for laptops
1467[18:08:51] <Mat63> I'm going to finish reading and then
asked what's up
1468[18:08:53] <annadane> or nonfree for wifi in general
1524[18:48:01] <nebunez> Hey, I'm running Debain bullseye
(testing) but I wonder if people on stable are being affected by a
bug in xserver-xorg-video-intel involving a segfault with
iris_dri.so when running GL applications, e.g. glxgears.
1527[18:50:46] <somiaj> nebunez: what intel card do you have?
1528[18:50:59] <sney> I *think* I've only seen that intel
video issue referenced with bullseye but a bug number would help
make sure
1529[18:51:28] <nebunez> Bug#959400
1530[18:51:34] <somiaj> although not able to verify if the bug is
there are not, I think it is more proper to use the modesetting
driver and not xserver-xorg-video-intel with most intel chipsets
1535[18:55:30] <nebunez> So, how do I interepret that (this is my
first time tracking a bug like this in testing)? I see it says,
"in src:mesa (closed, upstream)", does that mean it was
fixed? I've been checking the issue on mail-archive.com for
weeks for an update but I'm not sure what to look for.
1536[18:56:39] <sney> yes, as the last message in the thread
says, "We believe that the bug you reported is fixed in the
latest version of
1537[18:56:40] <sney> xserver-xorg-video-intel, which is due to
be installed in the Debian FTP archive."
1538[18:57:19] <nebunez> oh wow, I've been following the
wrong site all along it seems
1539[18:57:43] <sney> it would be uploaded to sid first, then if
there aren't any rc bugs after 10 days it'll also hit
testing. (sid packages can be installed directly on testing if
you'd prefer not to wait) so the earliest it's likely to
be in bullseye would be sunday
1551[19:01:00] <nebunez> I have been using the modesetting driver
since that bug affected me, but video playback is not as smooth as
when using the intel driver
1552[19:02:26] *** Quits: dionysus69 (~Thunderbi@replaced-ip) (Remote host closed the connection)
1573[19:13:00] *** Quits: d3vnull (devnull@replaced-ip) (Quit: Lost terminal)
1574[19:13:44] <nebunez> neither, I run just X with i3
1575[19:14:00] <NetTerminalGene> anyone use thunderbird irc
client here?
1576[19:14:13] <nevivurn> !anyone
1577[19:14:13] <dpkg> Please do not ask if anyone can help you,
knows 'something' or uses 'some_program'.
Instead, ask your real question. (If the real question _was_
"does anyone use 'some_program'?" ask me about
<popcon> instead.) See <ask> <ask to ask>
<polls> <search> <sicco> <smart questions>.
1607[19:50:56] <nebunez> I just reinstalled the intel driver (it
works), did some back and forth comparisons with the modesetting
driver by running glxgears and streaming video via mpv, and I can
confirm that I'm getting smoother video playback with the intel
driver on my HDMI output; I see no difference on the laptop display
itself.
1636[20:31:53] <ZitZ> hi, has anyone had any luck with install
nvidia proprietary drivers on bullseye. The nvidia module will not
load with modprobe nvidia
1637[20:32:21] <ZitZ> i've successfuly blacklisted nouveau
1672[21:04:35] <diogenes_> nebunez, CPU usage and temperature are
very important because on those two factors depends how long your
CPU will serve you.
1673[21:05:48] <nebunez> do you have any recommendations for
benchmarking ?
1712[22:01:12] <kittonian> hi all. i've got an hp 1810-24G
switch and in order to keep the time/date correct, I enabled SNTP
and pointed it towards my debian box running NTP. It successfully
got the date but the time is 1 hour off. The debian box's
time/date are perfect. Any ideas of what to check?
1713[22:01:25] *** Quits: IsThatAPlanet (~dawgy@replaced-ip) (Remote host closed the connection)
1729[22:06:08] <kittonian> there is a daylight savings time
option that is disabled right now. i wonder if that's
what's doing it
1730[22:06:11] <greycat> from google search result: Oct 25, 2010
- NTP does not handle time zones. All time data handled by NTP is in
UTC; your local time zone setting determines the offset from there.
1736[22:08:40] *** Quits: mezzo (~mezzo@replaced-ip) (Quit: leaving)
1737[22:09:03] <epictetus2> are all software in apt main repo
free software? considering installing the telegram-desktop client
but a bit sceptic about malware/privacy
1747[22:12:44] <greycat> You've been incredibly vague this
whole time. Just to be absolutely clear: when you run "date
-u" on the Debian system, what does it say?
1748[22:13:41] <kittonian> don't mean to be vague. happy to
check whatever is asked
1778[22:32:44] *** Quits: slabber (~slabber@replaced-ip) (Remote host closed the connection)
1779[22:33:03] <kittonian> greycat, fixed it. had to manually
setup daylight savings time on the switch and put in the 2nd sunday
in march and the 1st sunday in november with a 60 minute offset for
it to calculate properly. seems like this is something that should
be automatic in 2020, lol
1780[22:33:07] <kittonian> thanks for helping
1781[22:34:31] *** Quits: rgr (~rgr@replaced-ip) (Remote host closed the connection)
1814[23:07:29] <brenlae> hello, i was browsing the internet
(youtube) and i heard a quick system beep, i can't seem to find
anything running in the task manager that shouldn't be - i am
on Debian 10.04
1815[23:08:35] <brenlae> perhaps it's just a minor bug,
i'm thinking
1816[23:08:49] <brenlae> i have an AMD apu, and i use chrome
1817[23:08:55] <diogenes_> brenlae, maybe it was in the video.
1818[23:09:06] <brenlae> diogenes_: no i rewatched just to be
sure
1826[23:12:07] <jhutchins> I can't get the -depth command to
work in find. find . -type d -depth 1 gives me the generic
"Pathss must preceed,,," error.
1875[23:26:53] <sney> f8e4: sd-pam is a systemd component for
pam, which handles authentication. it's a normal part of your
system. I believe it has an occasional memory leak bug, but if you
don't see it going nuts with memory usage, then it's doing
what it's supposed to.
1884[23:29:36] <tom_> I have multiple internal ipv4 address, and
one of them is floating. However I dont want the system to assign
the floating ip itself that will be managed by keepalived
1885[23:29:44] <tarzeau> sney: and if you don't have
ifconfig?
1886[23:29:54] <sney> I believe there's an ip link command
too but I don't remember the syntax. man ip
1887[23:30:08] <tom_> I tried adding auto eth0:1
1888[23:30:08] <tom_> allow-hotplug eth0:1 in
/etc/network/interfaces but upon rc-service networking restart, i
get
1911[23:33:16] <sney> when you're troubleshooting your
/etc/network/interfaces config, use ifup/ifdown eth0:1 instead of
restarting the whole networking service, so you only get relevant
error messages
1913[23:33:44] <sney> I haven't dealt with subifs in a
minute, but could you pastebin your interfaces file?
1914[23:34:00] <tom_> ok
1915[23:34:24] <tom_> but how do i setup an interface alias
without any inet or inet6 addresses bound to it
1916[23:34:39] <tom_> not how do i do that at runtime, but how do
i configure /etc/network/interfaces to do that at boot
1917[23:35:04] <tom_> just iface eth0:1 ?
1918[23:35:12] <tom_> no inet static
1919[23:35:28] <dvs> manual?
1920[23:35:43] <mutante> auto eth0:1
1921[23:35:48] <tom_> if i do that i get
1922[23:35:49] <sney> 'man 5 interfaces' should have
that information. I *think* that when you're dealing with
subifs, if you assign an address to the main interface (eth0) and
you create a subinterface (eth0:1) without an address, it will just
have the address from eth0. so addresses should only be assigned to
the subifs
1923[23:35:49] <tom_> ifup: /etc/network/interfaces:31: unknown
or no address type and no inherits keyword specified
1924[23:36:56] <tom_> interfaces from manual 5 is missing. Is
there a package that installs that documentation?
1942[23:40:56] <sney> there's probably a way to not have it
on a debian system, like if you followed a howto that said to always
disable recommends, but that file should be present on any sane
install
1955[23:46:07] <tom_> when i try # ifup eth0:1 all i get is an
error basiclly telling me the syntax is wrong ifup:
/etc/network/interfaces:30: unknown or no address type and no
inherits keyword specified
1956[23:46:09] <mutante> tom_: is it missing the "inet"
part for eth0:1 ?
1957[23:46:11] <sney> if you put an ifup command in
/etc/network/interfaces it will probably error out, or maybe get
stuck in a loop, since ifupdown is already running ifup for each
interface named there.
1958[23:46:19] <mutante> tom_: i think "inet" is the
"address type" that is missing
1959[23:46:45] <sney> but yes it should be iface eth0:1 inet
manual, and the 'up' command should not call ifup.
1960[23:46:45] <tom_> mutante, yes, I don't want to assign
anyIP address to eth0:1 yet, I want keepalived to manage doing that
not debian
1961[23:46:55] <tom_> but debian just needs to bring up the
interface alias
1962[23:47:14] <plt> Does anyone know if this package name direct
cli with vertical pinstripes is named tw-cli?
1963[23:47:19] *** Quits: ska (~ska@replaced-ip) (Quit: Leaving)
1964[23:47:32] <sney> ,v tw-cli
1965[23:47:33] <judd> No package named 'tw-cli' was
found in amd64.
1966[23:47:40] <tom_> sney, just putting up there still results
in ifup: /etc/network/interfaces:30: unknown or no address type and
no inherits keyword specified
1967[23:47:42] <sney> plt: probably not! what's your real
question
1968[23:48:00] *** Quits: wonderer (~quakeroat@replaced-ip) (Quit: Famous quotes #3: "A lie gets halfway around the
world before the truth has a chance to get its pants on." - Sir
Winston Churchill (1874-1965))
1969[23:48:06] <sney> tom_: did you change the line to read
'iface eth0:1 inet manual'?
1970[23:48:24] <mutante> tom_: it does not mean you have to add
an "address" line as well
1971[23:48:29] <plt> I need the denian package that does direct
cli with vertical pinstripes is named tw-cli?
1972[23:48:41] <tom_> yes
1973[23:48:57] <sney> plt: your question does not make sense.
1974[23:49:02] *** debhelper sets mode: +l 1165
1975[23:49:17] <plt> I need the denian package that does direct
cli with vertical pinstripes?
1976[23:49:48] <sney> plt: what is your native language? there
may be a channel where you don't have to try english. :)
1977[23:50:18] <tom_> mutante, sney this is exactly what it looks
like
replaced-url
1978[23:50:44] <sney> tom_: one more time, with feeling: iface
eth0:1 inet manual
2005[23:56:50] <sney> yes obviously, it's on linode.
2006[23:56:52] <tom_> and yes i'm sure linode allows
floating ips
2007[23:56:56] <sney> are you sure they... ok.
2008[23:56:58] <tom_> if that's what you mean
2009[23:57:07] <mutante> tom_: hotplug is only needed if the
interface is not already present during boot. the example from the
manpage says "USB network card" as an example
2012[23:57:28] <mutante> (nothing to do with detecting whether a
cable is plugged in)
2013[23:57:44] <sney> because the 'ip link set dev eth0:1
up' run by ifup should have brought the link up, and the fact
that it's not may indicate a more fundamental problem.
2014[23:57:58] <tom_> i can run it by hand
2015[23:58:24] <mutante> can you manually bring it up _WITH_ the
IP?