8[00:07:43] <wald0> folti7777: i think we have a serious
problem here, because we (soft devs, maintainers, distro makers,
etc) are putting our efforts everyday making this a better world (in
software level), and we should not let allow the effort making
things good to become bad because of human stupidity (to call it in
some form, no offense)
13[00:08:41] *** Quits: cosimone (~user@replaced-ip) (Remote host closed the connection)
14[00:08:57] <wald0> it is of course perfectly understandable
that a bunch of students discover how to do hello world in atom will
be amazed on that and write things in social media and become the
new fashion, than a pro telling about the optimizations that has
writting a hello world using specific flags compilers lol
17[00:09:48] <wald0> i have seen also some (very good) projects
to die because of the lack of X when they were better / more
valuable than other things that survived better just because
"they sell better" or by fashion
18[00:11:16] <folti7777> hahaha, yes, the shit always go up. It
doesn't matter. As an old slackware guy, I'll just change
distro when Debian will run only for the newer machines
19[00:12:19] <wald0> i dont want to see that day
20[00:13:19] <wald0> i still have some screenshots of years ago
with a 200 MB ram laptop where I ran SO MANY applications on the
desktop, on these days you can literally have anything because all
of them consume nothing
21[00:13:30] <cws> !offtopic
22[00:13:30] <dpkg> #debian is primarily a support channel for
Debian users. Please keep the discussions in #debian on-topic and
take longer discussions and non-support questions to
#debian-offtopic. Imagine the chaos if each of the hundreds of
people in the channel felt the need to wander off topic for a few
minutes every day.
23[00:13:31] <folti7777> me neither, but what to do? Attack
kids? Lowering ourselves to their level? They beat us with
"experience".
27[00:14:17] <wald0> you have now a bloated browser (any of
htem) that turns on entire trash your computer... thats the worst,
WM are quite bloated too, but if you go deep inside you see many
things that was not needed (pulseaudio, between 30 or 50mb for
nothing, so alsa always did a good job)
336[06:24:07] <judd> Search for libkImageAnnotator.so.0.3.2 in
buster-backports/amd64: libkimageannotator0:
usr/lib/x86_64-linux-gnu/libkImageAnnotator.so.0.3.2
337[06:24:18] <sney> you will have to install libkimagannotator0
^
340[06:25:45] <sney> the ksnip package depends on this, so idk
how you managed to install one without the other
341[06:25:46] <themill> that's a dependency of the ksnip
package already
342[06:25:54] <themill> *snap*
343[06:25:56] <themill> I win.
344[06:26:11] <john_rambo> sney: judd Will installing packages
from backports compromise stability
345[06:26:26] <themill> john_rambo: where did ksnip come from?
346[06:27:11] <john_rambo> themill: Ksnip was working just fine
....I dont know why this error came up later ....I installed using
#apt install ksnip ... No snaps
347[06:27:24] <sney> installing one library from backports will
not affect the system's stability in general.
348[06:27:38] <sney> ksnip is only in buster-backports (and
testing/sid) so you must already have backports enabled.
349[06:28:01] *** debhelper sets mode: +l 1025
350[06:28:30] <john_rambo> sney: I don't know if I have
backports enabled ....What's the command to check that ?
351[06:28:32] * themill tries out ksnip
352[06:29:08] <sney> john_rambo: see if it is listed in the
output of 'apt policy'
353[06:29:49] <themill> seems like ksnip from buster-backports
is buggy; needs recompiling against the newer library
357[06:31:36] <sney> I see you also have linux mint sources
configured.
358[06:32:32] <john_rambo> sney: I have no idea how Debian works
... I installed it like a month back ...You see Mint sources coz I
am using LMDE which as you know is based on Debian stable
359[06:32:40] <sney> SIGH
360[06:32:44] <sney> !frankendebian
361[06:32:44] <dpkg> When you get random packages from random
repositories, mix multiple releases of Debian, or mix Debian and
derived distributions, you have a mess. There's no way anyone
can support this "distribution of Frankenstein" and
#debian certainly doesn't want to even try. Ask me about
<reinstall>
362[06:33:01] <themill> john_rambo: you're not running
Debian (but that's also not the actual problem here)
363[06:33:37] <sney> "will installing these packages
compromise stability?" asks the madman who has already mixed
two separate operating systems
364[06:34:26] <john_rambo> sney: themill Is LMDE is good distro
? Should I install vanilla Debian instead ?
365[06:34:51] <sney> john_rambo: most distros are
"good" one way or another. But you should pick 1 and stick
with it, rather than trying to combine them.
366[06:34:53] <themill> #debian tends to recommend Debian. The
hint is in the name.
367[06:36:42] <john_rambo> sney: The only reason I chose LMDE is
coz I read that vanilla Debian by default doen't include non
free packages & one needs to add the non free repos....LMDE
includes non free repos by default
368[06:37:05] <sney> john_rambo: and then you needed debian
sources for a package you wanted anyway. there's a lesson here,
somewhere.
369[06:37:34] <john_rambo> sney: I did not combine anything
myself ....This is the default config of LMDE
370[06:38:25] <sney> john_rambo: if you need non-free packages
to support something in your system, it's easy to enable the
sources in debian and install them. but if you like mint, nobody is
stopping you. we just won't support it in #debian.
374[06:41:36] <themill> ftr, ksnip in buster-backports is buggy
and rebuilding its needed; there's an upload already there
(waiting to be accepted) that should fix it
375[06:45:06] <sney> good to know, never heard of it before
today
376[06:46:01] <themill> nor had I, it looks quite good
377[06:46:31] <themill> Unpacking ksnip (1.8.0-2~themill10+1)
over (1.7.3-3~bpo10+1) ...
451[09:10:52] <concordia> Hello, I upgraded to bullseye (early I
know) and discovered that libgtk2-perl is not available yet. Is
there any way to install it ?
452[09:12:21] <cws> concordia: Your best bet is to ask the folks
over on OFTC network in #debian-next.
453[09:12:23] <cws> !testing
454[09:12:23] <dpkg> Testing is a continuously updated release
between <stable> and <unstable>, currently codenamed
<bullseye>. See
replaced-url
455[09:12:29] <cws> !bullseye
456[09:12:29] <dpkg> The release following Debian 10
"Buster" is codenamed "Bullseye" (Woody's
horse in Toy Story 2) and will be Debian 11. It is the current
"testing" release. Remember that testing is called testing
for a reason; good bug reports with patches are greatly appreciated!
replaced-url
457[09:12:33] <cws> !next
458[09:12:34] <dpkg> Another happy customer leaves the building.
459[09:12:36] <cws> Oh, come on.
460[09:12:39] <cws> !debian-next
461[09:12:39] <dpkg> #debian-next is the channel for
testing/unstable support on the OFTC network (irc.oftc.net), *not*
on freenode. If you get "Cannot join #debian-next (Channel is
invite only)." it means you did not read it's on
irc.oftc.net. See also
replaced-url
462[09:12:42] <cws> That one.
463[09:12:54] * abff snickers
464[09:13:07] * cws is not an expert on the bot's commands,
evidently
474[09:19:55] <concordia> abff, I think I'll take one extra
step and go with sid indeed. I will make a copy of that bullseye /
and will use it later when it is stable. thansk cws, didn't
know there was a dedicated channel for testing.
476[09:21:54] <craigevi1> RoyK, which kpano package there are
lots of them
477[09:22:17] <craigevi1> apt search kopano
478[09:22:32] <abff> just remember, sid is constantly rolling,
so learn to manage your packages well. Having to roll back to fix
problems will happen from time to time, but this isn't the
place to be discussing it. The people in the OFTC channels are great
479[09:24:09] <concordia> abff, sure, I already use sid on my
desktop so no big deal.
481[09:25:05] <RoyK> craigevi1: I was at bullseye - misread -
thought I needed bullseye for this - I didn't - and kopano
isn't even in bullseye and there's no way I'll
install sid on anything in the vicinity of sid
550[10:40:35] <ratrace> alexrelis[m]: there are certain security
implications with console access, but if you understand those and
accept them, then it's fine.
551[10:42:44] <jelly> ratrace, there are SO many apps still
using gtk2 and not ported to either gtk3 or anything else, even if
unmaintained that perl binding should probably find its place in
oldlibs for a release or three
552[10:43:38] <ratrace> jelly: agreed. it's too soon.
557[10:49:37] <jelly> alexrelis[m], it's as dangerous as
leaving Xorg's Ctrl-Alt-Backspace enabled. Someone with 5
seconds of physical access might press it and ruin your day
558[10:51:16] <ratrace> cltr-alt-bkspc wasn't so bad, even
with sysvinit's inittab xorg restart
570[10:55:22] <dpkg> Since Debian 6.0 "Squeeze",
terminating an X server via Ctrl+Alt+Backspace ("zapping")
is disabled by default. To enable, run «dpkg-reconfigure
-pmedium keyboard-configuration» and answer Yes when asked if
Ctrl+Alt+Backspace can terminate the X server (or «setxkbmap
-option terminate:ctrl_alt_bksp»).
571[10:56:09] <ratrace> I'm vim binding everything in
sight. it's unnatural for me to twist around to ctrl-alt-bkspc
575[10:58:13] <hejux> always display pid and command name
576[10:58:19] <hejux> but i only need command name
577[10:58:26] <EdePopede> alexrelis[m]: some time in the past
(on wheezy) i had cpu issues, heating up, system freezing. REISUB
was my only chance to cleanly unmount my disks.
578[10:59:25] <EdePopede> and someone with physical access might
also be able to press the power button as well.
580[11:00:03] <EdePopede> C-A-Esc was nice, i liked the cursor
581[11:00:05] *** Quits: gpxlnx (gpxlnxmatr@replaced-ip) (Quit: Idle for 30+ days)
582[11:00:21] <jelly> xkill is one thing I still use regularly
583[11:00:55] <EdePopede> i go with htop now, even killing a
specific firefox thread seems to work. if the system still would let
me switch the desktop :>
587[11:03:01] <EdePopede> another argument btw against these
stupid wireless keyboard, better said their broken design. i often
end up on mine pressing INS instead of BSP, doesn't happen
here, but this one also has a gap between them i can put my thumb
into
645[11:38:07] *** Quits: toorop (~toorop@replaced-ip) (Remote host closed the connection)
646[11:40:03] <ratrace> now oyu want the pid :)
647[11:40:49] <ratrace> you asked for lsof initially. but.... if
you're not bound to lsof specific functionality, you can awk
the output of ss. neat, columns, no extraneous data
723[13:03:18] <HelloShitty> Hello peeps. When a program is
running, there is usually a pid file associated with that
application that is running, right?
727[13:04:42] <HelloShitty> abrotman: and when that application
is stopped for some reason, that file is kept or it is always
renoved until the application starts again?
735[13:06:20] <abrotman> HelloShitty: oh sorry, misread .. There
is at least one pid associated with a running program/script/daemon.
however, not all will create a pid file
738[13:06:37] <abrotman> but if a program creates that file, and
properly exits, it should remove the file
739[13:06:46] <abrotman> Look in /var/run/ for examples
740[13:06:54] <HelloShitty> but if one creates, then, it is
removed while the aapplication is not running, right?
741[13:07:03] <abrotman> if it exits properly
742[13:07:22] <abrotman> if it crashes or is killed, no promises
743[13:07:26] <HelloShitty> I want to write a small script and
check if an application is running by checking the existance of this
file
744[13:08:05] <ksk> HelloShitty: have you heard of systemd?
745[13:08:11] <HelloShitty> yes
746[13:08:19] <ksk> How about using it then? :)
747[13:08:19] <HelloShitty> I can check it with systemctl
748[13:08:20] <abrotman> might be overkill in this case .. need
more details
749[13:08:42] <HelloShitty> but systemctl can be werid
750[13:08:50] <HelloShitty> and I don't understand
it's output many times
751[13:08:57] <ksk> What are you really trying to do here? :)
752[13:09:25] <HelloShitty> for instance with Tor, the output of
systemctl status is not very straight forward
753[13:09:43] <HelloShitty> ksk: I want to write a script that
start an application incase it is not running
754[13:10:13] <ksk> whats "application" here? a
systemd service? a users desktop app?
755[13:10:15] <HelloShitty> wiat some time and then start some
other application, only after this first one is running and
completely sync'ed with aa database
756[13:10:23] <HelloShitty> which can take several minutes
757[13:10:51] <HelloShitty> it's not a systemd application
758[13:10:54] <ksk> "I want to write a script that start an
application incase it is not running" -- this is basicly
"monit", btw.
768[13:12:24] <judd> Package monit (admin, optional) in
buster-backports/amd64: utility for monitoring and managing daemons
or similar programs. Version: 1:5.27.1-1~bpo10+1; Size: 338.5k;
Installed: 995k; Homepage:
replaced-url
769[13:12:36] <abrotman> monit or runit ..
770[13:12:36] <ratrace> whoever wrote the dumb daemon never used
a *NIX CLI
773[13:13:36] <ratrace> meanwhile .. maybe HelloShitty should
explain _exactly_ what s/he's trying to do, as all this has a
feeble whiff of a XY :)
774[13:14:05] <abrotman> I suggested we needed more details five
minutes ago, you all jumped in with more recommendations :)
775[13:14:43] <ratrace> only if #define YOU_ALL ksk :)
776[13:14:48] <HelloShitty> I already said. I want to start an
application in case it is not already running, and then, wait for
this application to get sync'ed with a database an only then,
start one other applciation
782[13:15:51] <HelloShitty> You mean, how much time?
783[13:16:11] <HelloShitty> The second application will exit if
the first application in not fully sync'ed
784[13:16:13] <ratrace> no. I mean define teh
trigger/criteria/condition that is described as "synced with a
database"
785[13:16:24] <HelloShitty> So, I need to wait for the 1st
application to be 100% synced
786[13:16:40] <ratrace> that's a rather very high level
construct with no generic solution as there's many databases
and many meanings of "synced"
787[13:16:45] <ratrace> so again:
788[13:16:46] <ratrace> !xy
789[13:16:46] <dpkg> Slow down for a bit! Are you sure that you
need to jump through that particular hoop to achieve your goal? We
suspect you don't, so why don't you back up a bit and tell
us about the overall objective... We know that people often falsely
diagnose problems because they are too close to them -- it's
easy to miss that there is a better way to proceed. See
replaced-url
790[13:17:30] <HelloShitty> I'm not sure what else to say
791[13:17:41] <ratrace> so if you define "synced" as
"the process has exited with value N" ... then that's
a defined condition upon which you can act
792[13:18:13] <HelloShitty> It needs to check if the database is
complete, and if it is not, then it needs to fetch from other peers
with the complete database, the missing parts
793[13:18:20] <ratrace> if you need a magic fairy dust that will
consut a crystall ball if a process has issued a "COMMIT;"
SQL query and gotten a two-stage commit confirmed.... then...
that's a different condition.
794[13:18:42] <ratrace> "database is complete" does
not mean anything
795[13:19:08] <abrotman> ratrace: it is full, no more data
needed!
796[13:19:16] <ratrace> please understand: you're dealing
with software here, not HAL or other forms of AI that can read your
mind. you need to be operationally and computationally EXACT in
definig compputational state upon which the computer can act
797[13:19:47] <HelloShitty> I don't understand what else
you want me to say. A database is complete when it is complete...
lol. If my application is offline for let's say a day, then my
database is not complete
798[13:19:57] <HelloShitty> so it needs to catch up with other
peers databases
799[13:20:04] <ratrace> "database is compelete" does
not mean anything
800[13:20:22] <ratrace> does complete mean it has N rows
inserted into table XYZZY?
801[13:21:06] <HelloShitty> I don't know the exact
structure of the database. I only know that this database is
mantained my many peers and all peers needs to have the exact same
copy of the database
802[13:21:29] <HelloShitty> if my database is not an exact copy
of the others, then it must do its job to get the missing parts
803[13:21:36] <ratrace> HelloShitty: so how do you
operationalize that state?
804[13:22:07] <ratrace> or back to my previous question: you
launch a program and it exists. does its exit code define
"complete"?
805[13:22:12] <HelloShitty> The application does the reqired
tasks to ensure the database gets sync'ed with the others in
the network
806[13:22:24] <ratrace> ie, it exited with no error, that must
mean the task it performed is done, complete.
807[13:22:48] <HelloShitty> no, it exits with an error saying
the "first app" is not sync'ed
809[13:23:05] <ratrace> or in context of your original question:
you launch a program and ONLY and ONLY IF it exists with error code
0 (or some other number) can you proceed with something else. is
that what you seek?
812[13:24:53] <HelloShitty> application 1 will not exit. It will
keep running. applcation 2 checks if the first is sync'ed and
if it is not sync'ed, this application 2 exits with some error
code, and informing the user that application 1 was not fully
sync'ed
814[13:25:40] <ratrace> alright. and in context of this last
statement, what is your question?
815[13:26:36] <HelloShitty> I want to write a script that starts
application 1 if it is not running, wait for it to get fully
sync'ed, and then starts application 2
816[13:27:05] <ratrace> that's NOT what you last said. you
last said that app 2 is the one that does the checking
817[13:27:08] <HelloShitty> this script will be setup probably
with cron, so that if something goes wron, the script can restart
the applcations again
818[13:27:31] <HelloShitty> application 2 does the checking
819[13:27:55] <HelloShitty> I said the same before as now
820[13:27:57] <ratrace> so then how do you "wait for it to
get fully synced"
821[13:28:19] <HelloShitty> there are 2 checks being made
822[13:28:29] <HelloShitty> 1st, the one that checks if
application 1 is running
823[13:28:40] <ratrace> no you didn't. previously you said
APP 1 starts .... then APP 2 checks (presuming it's running
otherwise it can't check) .... then APP 2 returns an exit code
if there's error
824[13:28:43] <HelloShitty> and the other one is made by
applicaiton 2
825[13:28:51] <ratrace> NOW you're saying you start APP 1
.... then MAGICK .... and only THEN you start app 1
826[13:28:53] <HelloShitty> but that is the applicaiton 2 itself
827[13:28:54] <ratrace> app 2
828[13:28:55] <HelloShitty> it's not me
829[13:29:24] <HelloShitty> yes, APP2 checks for APP 1 when APP2
starts, obvisouly
830[13:29:32] <HelloShitty> it is not running prior to running
AAP 1
837[13:31:11] <HelloShitty> I start APP1, wait it to get fully
sync'ed. Then I start APP2 and this APPS checks for itself if
APP1 is runnning and sync'ed
838[13:31:32] <HelloShitty> But this is done by hand
839[13:31:32] <ratrace> alright. so where's the problme?
840[13:31:51] <HelloShitty> I visually check if APP1 is
sync'ed and only then I manually start APP2
841[13:32:06] <HelloShitty> in that case, APP2 sees APP1 running
and sync'ed, so no problems
842[13:32:28] <ratrace> alright ... and the question is?
843[13:32:32] <HelloShitty> My first question was if the pid
file disappears when APP1 shuts down for some reason
844[13:32:38] <HelloShitty> and the talk started there
845[13:33:06] <HelloShitty> because if the pid file disappears
for sure, I can check for the existance of this file to evaluate if
the app is running or not
846[13:33:19] <ratrace> the pid file disappears if something
removes it. there's no guarantee it'll disappear
847[13:33:42] <ratrace> pidfiles are ancient concepts. daemons
themselves (not the process manager!) create the pidfile and were
responsible for their cleanup on GRACEFUL shutdown.
848[13:33:50] <HelloShitty> ok, so I need a bette way to check
if the application is running or not
849[13:34:00] <ratrace> if the daemon exited abruptly,
nongracefully, it would NOT clear out the pidfile and that used to
be the source of many headaches
850[13:34:15] <HelloShitty> I can understand
851[13:34:24] <ratrace> so teh MODERN way with systemd is to
ignore pidfiles. systemd can know directly whether a process it
started is running or not
854[13:35:27] *** Quits: Iamahuman (~noname@replaced-ip) (Quit: to the sunshine!!)
855[13:35:36] <HelloShitty> ok, so I think I will try to create
a system unit file for these 2 apps and start from there
856[13:35:56] <ratrace> in other words, you can have a systemd
unit that starts a program and RESTARTS it in case it exits, with
various configurable conditions and timings
860[13:36:57] <ratrace> but systemd has no idea how to check if
"database is complete" . so I'm going back to my orig
question of how do you define that. and then you said you're
doing it VISUALLY .. which I'm afraid cannot be converted into
a shellscript at the moment. so you'll have to define another
API upon which a script or systemd or something else, can trigger an
action.
870[13:42:13] <ratrace> and has nothing to do with the problem
at hand, which is still undefined even that has to occur for program
2 to start, whether it starts at all, that's still unclear.
871[13:42:15] <hejux> just an example case of using $MAINPID
872[13:43:10] <HelloShitty> I can try to use a timmed fashion
way to wait for APP1 to get sync'ed
874[13:43:37] <ratrace> HelloShitty: or maybe your database has
a replication state you can query
875[13:43:41] <HelloShitty> In the screept I can have a spleep
5m thing or so, befora calling APP2 to start
876[13:44:02] <HelloShitty> ratrace: ok, I can ask for that
877[13:44:04] <ratrace> that sounds very fragile and will
definitely fail if the sync process decides to take longer than you
assumed it would
878[13:44:14] <HelloShitty> yes, that's true
879[13:44:40] <ratrace> HelloShitty: so THAT is the answer to my
previous question of how do you define "complete". what
programmatic state you can query, other than "visually"
881[13:45:43] <ratrace> in case of postgresql, for example,
there's clearly defined states for logical and streaming
replication, where an ongoing process will have a lag and WAL
identifiers to check how _far_ it came in syncing
882[13:45:45] <HelloShitty> ok, let me read first about the
options for unit files
883[13:46:30] <ratrace> another situation may be that the APP 1
exits with code 0 when it's complete, and non-zero when
it's not, which is another STATE you can query to know
what's going on and decide whether to launch APP 2
903[13:49:45] <ratrace> but APP1 exiting is NOT the condition to
trigger APP 2 to start, correct?
904[13:49:54] <HelloShitty> So I want to make regular checks,
and if APP1 is not running, I want to restart it, wait for the
database sync, and then, start APP2
905[13:50:00] <hejux> no, you just create a file in somewhere,
when app1's job is done
910[13:50:35] <ratrace> HelloShitty: 1) if APP 1 exits, it can
be easily restarted with systemd unit directives
911[13:50:46] <ratrace> 3) app2 can be made to start when
condition in #2 appears
912[13:50:54] <ratrace> 2) "database sync" is still
not defined
913[13:51:18] <ratrace> you need to define a programmatic way
for a script or service to QUERY, in order to know whether a sync is
complete or not
914[13:51:41] <hejux> yeah, that's a point. create somekind
of MARK
915[13:51:46] <ratrace> SO ... you can make APP2 service unit
start on TWO conditions. 1) "After" APP 1, and 2) you need
to define this condition
916[13:52:15] <HelloShitty> ok
917[13:52:26] <ratrace> HelloShitty: so you can use After=
directive in the unit for APP2 to start after APP 1
918[13:52:41] <HelloShitty> yes, I'm reading about unit
file options I can use
919[13:52:51] <ratrace> you can also use StartExecPre= to
check/query for a "database synced" condition, and if
complete, return with 0 so regular ExecStart can start APP 2
920[13:53:18] <ratrace> ExecStartPre ..... that's what I
mean.
921[13:54:06] <ratrace> critical thing here is to hang and wait,
not return error code or APP2 won't start, or will go into
restart dance and finally fail when the treshold is reached
922[13:54:12] <HelloShitty> ok, when I'm at tha tpoint,
I'll see what works best
930[13:58:28] <ratrace> HelloShitty: oh also, if as hejux
suggested, APP1 creates a file to mark completeness, you can use
ConditionPathExists= directive, that's explained in
systemd.unit(5) manpage along with some other conditions
931[13:59:55] <ratrace> these are all for app2.service.
app1.service only needs to Restart=on-failure (or some other value,
see systemd.service(5) manpage) unless you have some other condition
for it
934[14:05:03] <hejux> then in app2, you can do , until [[ -f
/path/to/the/file/created ]] ; do sleep 1 ; done ; run app2 command
next
935[14:05:43] <ratrace> no you don't need that with
ConditionPathExists
936[14:06:15] <hejux> e.g. i don't trust systemd's
network-online.target, i use until ping some ip ; do sleep 1 ;done
937[14:06:54] <ratrace> because network-online.target does NOT
do what you think it does
938[14:07:13] <hejux> yeah, i know that
939[14:07:20] <ratrace> network-online merely waits for the
local system to be come network enabled. there can still be upstream
conditions that ping may or may not even TRIGGER ...
940[14:07:26] <hejux> so that's a work around for my script
941[14:07:43] <ratrace> that said I do have a ping-test.service
myself and I made a bunch of network related services run AFTER that
one
942[14:08:02] <hejux> yeah, same here
943[14:08:55] <hejux> ratrace: be care of run AFTER.
951[14:12:32] <ratrace> if you use Type=oneshot without
RemainAfterExit (which you should), then dependencies won't
start until this one exits. and I've made my ping test exit
only after first successful ping
952[14:12:48] <hejux> ratrace: sometimes i use netcat ip port to
replace ping
960[14:16:59] <hejux> ratrace: as long as i can connect to
1.0.0.1:53, i am sure my network is online
961[14:17:11] <hejux> so, fk network-online.target
962[14:17:25] <ratrace> it's not its fault
963[14:18:09] <ratrace> it cannot know or care about network
states outside of 127.0.0.1. because your LAN could be working but
teh gateway is down. or the gateway is up but the nexthop is having
issues.
964[14:18:41] <ratrace> or you're in one of those networks
run by morons who thought 1/8 was a test subnet and are firewalling
it, so you couldn't query cloudflare :)
965[14:18:51] <ratrace> maybe google's 8.8.8.8 is more
reliable
966[14:20:16] <hejux> yeah, something like that
967[14:20:51] <hejux> so all the cost is the sleep command.
1073[16:58:13] <hejux> Aurora_v_kosmose: search debian.org for
backports
1074[16:58:16] <hejux> !backports
1075[16:58:16] <dpkg> A backport is a package from a newer Debian
branch, compiled from source for an older branch to avoid dependency
and <ABI> complications.
replaced-url
1076[16:58:32] <Aurora_v_kosmose> I know of backports. I'm
leading up to a question.
1077[16:59:24] <hejux> if a CVE is not published ... how the hell
Debian devs know what to fix
1078[16:59:58] <tomreyn> the linux-distros mailing list
1079[17:00:36] <Aurora_v_kosmose> Right, that is what I assumed
to be the case. So that leads me to wonder if keeping things to
latest releases introduces more exploitable bugs than it fixes, in
cases where the CVE submission & publishing system fails to work
appropriately.
1081[17:01:03] <Aurora_v_kosmose> I'm essentially wondering,
if CVE publishing falls apart, does the Debian model provide better
or worse security on Stable?
1082[17:01:17] <Aurora_v_kosmose> (I know that Stability is the
primary objective)
1098[17:14:51] <Aurora_v_kosmose> If CVEs work and are patched
sufficiently fast, then a stable & secure model works. But if
CVEs disappear, then how would maintainers know what to patch?
1099[17:15:21] <Aurora_v_kosmose> But I don't really know
how to estimate if new versions introduce more bugs than fixes.
1148[18:43:13] <Release_> greetingz gents, i was trying to figure
out an easy way to install dns but most tutorials ive seen seem long
1149[18:43:26] <Release_> on a debian 10 box
1150[18:44:01] <sney> 'apt install dnsmasq' and then
read how to set it up
1151[18:44:19] <Release_> thanx
1152[18:44:32] <sney> assuming by "install dns" you
mean you want a local dns server for your network, or for VMs or
containers or whatever
1153[18:45:09] <Release_> yes
1154[18:45:27] <Release_> all examples i come across seem to be
bind9
1155[18:45:38] <Release_> any reason why this is used most ?
1156[18:45:48] <sney> ok. dnsmasq is appropriate for that. and
it's pretty simple to use, iirc you can get most of the jist by
reading through the config file.
1158[18:46:39] <Release_> will dig into it some more thanks for
now sney
1159[18:47:03] <sney> bind is the dns server that's most
popular in large production environments. I guess the howto writers
don't know the size of your deployment so they just go with
what's popular.
1267[20:29:03] <sney> nvidia-driver in buster (debian 10) is
418.181, as seen above. for 460.39 you would need to enable the
optional buster-backports repo.
1268[20:29:12] <_0xbadc0de_> sney: I did that
1269[20:29:18] <_0xbadc0de_> I can re-try
1270[20:29:27] <sney> ok, then back to my other question: did you
verify that the driver was able to build? check for nvidia*.ko files
in /lib/modules/$(uname -r)/updates/dkms
1277[20:33:50] <_0xbadc0de_> search in that dir, instead?
1278[20:33:57] <sney> no
1279[20:34:23] <sney> the driver did not build. I think 460.39 is
incompatible with the 4.19 kernel. luckily the fix is easy
1280[20:34:42] <sney> install a newer kernel with 'apt -t
buster-backports install linux-image-amd64 linux-headers-amd64'
1281[20:34:57] <_0xbadc0de_> should I purge the recent driver
installation?
1282[20:35:00] <sney> no
1283[20:35:02] <_0xbadc0de_> ok
1284[20:36:51] <ratrace> no need, 460.39 works just fine with
4.19
1285[20:37:50] <_0xbadc0de_> OK I installed it regardless and I
am rebooting
1286[20:37:58] <sney> ratrace: hm, I swear I've seen that
one before. missing headers then?
1287[20:38:06] <sney> _0xbadc0de_: did you check if the driver
was present?
1288[20:38:35] <ratrace> the kernel from buster-backports is
lagging in security fixes a lot. might be it doesn't matter for
your desktop use but it's worth keeping in mind
1293[20:40:27] <ratrace> sney: don't know what's the
problem, scrolling back up I don't see much more detail.
1294[20:40:58] <ratrace> but I'm sure it works because
I'm currently reading this IRC using pixels generated by 460.39
on a 4.19.0-16 instance of buster's kernel :)
1295[20:41:59] <_0xbadc0de_> they seem to be loaded too, if I do
lsmod -l | grep -i nvidia I see many nvidia drivers loaded
1296[20:42:33] <_0xbadc0de_> but when I do ./hashcat.bin -b
--benchmark-all I still get no devices
1301[20:43:25] <sney> you probably need the additional cuda
packages, they don't come with the driver automatically, since
not everyone with a nvidia gpu is going to use that stuff
1302[20:44:11] <ratrace> _0xbadc0de_: back to square one. how did
you install the driver? apt install nvidia-driver is all you
should've done
1304[20:44:30] <sney> _0xbadc0de_: sounds right to me
1305[20:44:32] <_0xbadc0de_> since I am using backports?
1306[20:44:33] <_0xbadc0de_> ok.
1307[20:44:43] <sney> ratrace: that is what they initially
claimed to have done. you can get the next one
1308[20:46:18] <ratrace> well then, dkms can be retriggered.
something along the lines of this:
replaced-url
1309[20:46:52] <ratrace> or using `dkms` directly but I
don't know what all the modules are OTOH, as you need to name
them explicitely
1310[20:46:59] <sney> ratrace: anyway, as you can see from your
460.39/4.19.0-16 pixels, the driver is now built and loaded,
we're just in the adding cuda stage.
1311[20:47:00] <ratrace> (all nvidia related ones)
1312[20:47:13] <ratrace> ahhh.... sorry for the noise then.
1317[20:50:34] <somiaj> note, though probably too late, sometimes
it is worth installing the version from buster first, then upgrading
to backports, this can sometimes deal with issues of pulling in to
many libs from backports.
1318[20:50:37] <_0xbadc0de_> ok I am going to reboot
1319[20:50:59] *** Quits: matrixbot_bartab (~matrixbot@replaced-ip) (Remote host closed the connection)
1321[20:52:27] <_0xbadc0de_> sney: it seems like everything is
properly working, thank you very much. If you have a Dogecoin wallet
I can tip you?
1322[20:52:56] <sney> _0xbadc0de_: I don't support crypto.
donate some currency to SPI if you want, or just pay it forward and
help someone else in the future.
1343[21:27:48] <karlpinc> I'm using the User Agent Switcher
extension with Firefox (esr, from the debian repos). (And using
noscript.) When I click on the icon I never get a list of user
agents, I only get a little gear that spins forever. Any clues as to
what's going on?
1464[23:32:41] <Delf> Ah you mean login/greeter for both seats.
Yes, that part works. And color managing works on both seats too but
only for one user, and not the other. Trying to have night light
working for both users
1465[23:34:12] <jhutchins> Delf: Some hardwar can only be
controlled by one user at a time, like sound.
1466[23:34:23] *** Quits: wilson (~wilson@replaced-ip) (Ping timeout: 260 seconds)
1472[23:36:37] <jhutchins> I'm afraid I don't have any
current multi-seat experience, my knowedge of that is pretty
obsolete.
1473[23:37:04] <ruied> Hi. I would like to make the entire linux
distribution install from a single pen, is there any tutorial/info
to do that? I normally install it from a DVD iso and DD to a pen
disk. But I would like to have all the debian installation in a
single PEN. Is there any kind of iso or tutorial to do this?
1474[23:37:52] <ruied> By linux distribution, I mean Debian...
1476[23:38:38] <Delf> I'm not sure it's related to
multi-seat setup even. I've tried restarting and NOT logging in
with the user that can manage colors. Logged in with the user that
can't manage colors and still can't find devices to
manage. Not even as sudo, untill the user that CAN manage color
profiles logs in
1477[23:39:02] <cws> ruied: I'm not certain what you're
asking. What is it that you need that you're not getting from
the DVD image and copying it to a USB drive?
1478[23:39:39] <jhutchins> ruied: You usually don't need
anything but the netinstall or the first DVD
1481[23:43:40] <Delf> ruied: find out which /dev/sd? the usb
stick is and set 666 permissions on it then you can use wget -O
/dev/ url.to.iso
1482[23:44:16] <Delf> Before you do something I will regret, make
SURE it's the correct device
1483[23:45:07] <cws> Delf: Those are very strange
recommendations. Do not tell someone to change the modes on a device
node.
1484[23:45:20] <cws> How to write an iso to a usb medium is
well-documented. Stick to those.
1485[23:45:25] <ruied> I normally install from a pen. I download
the dvd iso file and 'dd' to a usb stick. Than start the
installation from the pen and the remaining staff debian gets from a
Internet debian repository. I would like to make a pen with all
debian repository, install every package from a single pen (without
the need of Internet)
1486[23:45:38] *** Quits: Vizva (~Vizva@replaced-ip) (Remote host closed the connection)
1487[23:45:46] <Delf> cws: Probably not the best way to do it
either. I'm a noob
1488[23:46:02] <cws> Delf: It's most certainly not the right
way to do it, to be a bit blunt about it. :/ Sorry.
1489[23:46:13] <cws> ruied: So you want to run a local mirror of
the ENTIRE debian repository?
1490[23:46:22] <Delf> cws: No, nothing to be sorry about
1491[23:47:08] *** Joins: wilson (~wilson@replaced-ip)
1497[23:50:15] *** Quits: Jerrynicki (~niklas@replaced-ip) (Remote host closed the connection)
1498[23:51:18] <ruied> Delf, I mean, if I have no access to
Internet at all, I still want to be able to install Debian from a
single pen as I did it years ago with several dvds or a bunch of
1.44 floppy disks...
1504[23:52:25] <sney> ruied: if that's your goal, you could
use jigdo to download the BD-DVD iso (I believe it's just one)
and write that to your flash "pen"
1506[23:53:13] <sney> sorry, not the bd-dvd, this one
replaced-url
1507[23:53:19] <_-Jon-_> Hi there, I have a silly question. I
have an IPv6 address provided by my ISP but I want to use a tunnel
broker. In /etc/network/interfaces I commented out the section about
the inet6 for that interface but it's still being assigned an
IPv6 address
1508[23:53:39] <ruied> sney, ok, nice! the procedure is the same?
after download just 'dd' to a usb pen?
1509[23:53:44] <cws> _-Jon-_: That's because IPv6
autoconfiguration kicks in automatically. Any reason why you want to
use the tunnelbroker instead of your ISP?
1510[23:54:05] <sney> ruied: yes, the iso installers all work the
same way
1514[23:54:43] <_-Jon-_> cws: More to test with on a VPS. My plan
is to use a tunnel broker at home eventually as well as my ISP
changes my address too often :(
1524[23:58:29] <jhutchins> ruied: Make sure you check the
checksum on the iso.
1525[23:58:45] <cws> Because a simpler option would be to get
yourself a domain, if you don't already have one, and just
update the DNS record for your IPv6 address whenever it changes.
1529[23:59:59] <_-Jon-_> cws: Glad you asked! I do that already,
but the problem with that is not only do I need to update all the
DNS records, but I have to update my firewall rules as well.