1[00:01:05] <ndroftheline> yeah the diff b/w openbox and xfce
looks like about 10mb and a half dozen packages...i think i'll
go with xfce for now. thanks
2[00:03:45] *** Quits: erle- (~stephan@replaced-ip) (Remote host closed the connection)
10[00:15:56] <deadrom> my NAS box just crashed and I have no
idea why. syslog has nothing. any idea where to start investigating?
I ran memtest86+ after recetnly bad ram bit me, temperaturewise the
system should be ok, power supply is stable... no clue where to look
11[00:17:28] <ratrace> retroactively nowhere. but you can prime
the system for a future crash. export the log to another machine,
re-run memtest for at least a day if you haven't, check disks
41[00:49:17] <kingsley> Can you suggest a nice, user friendly
package, with a GUI tool, that reveals the contents of all sessions
recorded to a multisession DVD?
77[01:28:37] <deadrom> kingsley: I was pulling your leg. I am
surprised that soemone still uses them at all, and then such a fancy
features such as multisession. I wasn't even aware DVD does
muti-session, thoguht that was for DVD-RAM. But to aid you in your
quest, I'd expect writer programs for optical media to give you
the information on the sessions
78[01:29:12] <deadrom> last time I wrote a DVD is years ago so
I don't know what's hot right now
79[01:30:27] <deadrom> (btw it's Digital Versatile Disc,
but that's just me being fun at parties :) )
80[01:32:44] <deadrom> seriously though, a 64GB thumbdrive is
under 10 bucks, why?
87[01:43:17] <deadrom> ratrace: machine capsized again.
I'm starting to suspect a power supply issue, but no telling
like that. will hook a monitor on and see.
149[02:43:24] *** Quits: craigevil (craig@replaced-ip) (Quit: Lost terminal)
150[02:43:26] *** Seto_Kaiba is now known as Hunterkll
151[02:44:09] <cybrNaut> you would think they would want
Bullseye to be easy to get so that the whole world doesn't wait
until release then all download it at the same time
152[02:45:28] <cybrNaut> found it (had to search outside of
Debian to find it)
replaced-url
153[02:46:12] *** Quits: rustyshackleford (uid236774@replaced-ip) (Quit: Connection closed for inactivity)
154[02:46:49] <epsilon> you only want DVD if the machine has no
or slow internet
243[04:51:31] <Grobo> I'm trying to set sshd options for a
specific user. I have a .conf file in /etc/ssh/sshd_config.d but it
doesn't seem to be working.
244[04:52:12] <Grobo> Specifically, I have:
PasswordAuthentication no (newline) AuthenticationMethods publickey
245[04:52:18] <Grobo> But the user can still log in with their
password
251[05:03:49] <alkisg> Grobo: I've noticed that when I put
some entries under sshd_config.d/blah.conf, they don't work
(while others do), yet when I put them directly in sshd_config, they
do work. Can you test if the same settings in the main configuration
file work? (do restart sshd to test)
259[05:09:05] <somiaj> could this be the sshd_config is read
after sshd_config.d, thus things there would overwrite previous
settings?
260[05:09:54] <somiaj> hmm, I notice here that I don't have
sshd_config.d by default, did you have to add a read line to
sshd_config for that to work?
261[05:10:46] <somiaj> Grobo: in sshd_config, do you have the
line, "Include /etc/ssh/sshd_config.d/*.conf", if so where
is it in relation to other settings?
262[05:11:29] <somiaj> I just notice that none of my systems
here have that
263[05:12:05] <alkisg> somiaj: e.g. this is on top of
bullseye's config: Include /etc/ssh/sshd_config.d/*.conf
264[05:12:21] <alkisg> Maybe it wasn't there in buster, and
it's in testing only
265[05:13:29] <Grobo> Include /etc/ssh/sshd_config.d/*.conf --
it's at the beginning
266[05:15:53] <somiaj> so that means configuration options in
sshd_config will override options in sshd_config.d
267[05:16:05] <Grobo> oh?
268[05:16:08] <somiaj> alkisg: ahh nice, yea it isn't here
in buster
269[05:16:30] <somiaj> Grobo: yea, since it is at the top, those
are all read first, then sshd_config is read second, so if there is
ever a conflict, the second read will take presedence (my guess)
270[05:16:55] <Grobo> hmm
271[05:16:55] <alkisg> Good guess, Grobo if you want, try to put
that include at the bottom...
272[05:17:25] <alkisg> (or put the settings at the end of the
main file to check that one first)
273[05:17:32] <somiaj> this would also explain alkisg
observation that some things work and others don't
280[05:19:33] <somiaj> Grobo: if you put the Include at the
bottom of the sshd_config file does it then work?
281[05:19:57] <Grobo> No
282[05:20:26] <somiaj> Okay, guess my guess was wrong, but look
at that bug, aperentally this was a bug in upstream,
replaced-url
283[05:21:37] <somiaj> hmm, well that bug report says this was
supose to be fixed in 8.4 which is the bullseye version, but it
appears the issue is still somewhat present.
284[05:21:45] <somiaj> There really isn't much info on the
debian bug though.
285[05:21:56] <alkisg> Grobo: which sshd version do you have?
286[05:22:17] <somiaj> well 8.4p1, wonder what the p1 means in
the upstream version
287[05:23:01] <Grobo> OpenSSH_8.2p1
288[05:23:11] <somiaj> Grobo: what version of debian are you
running?
314[05:27:36] <somiaj> anyways, your system is out dated, It
appears the bug you have should be fixed in 8.4
315[05:28:10] <somiaj> alkisg: well hopefully on your 8.4 system
you don't run into that bug anymore either.
316[05:28:23] <somiaj> at least if the upstream bugreport is
correct that it was fixed
317[05:29:28] <alkisg> somiaj: yeah possibly; unfortunately I
reported the bug for a software I develop, ltsp, not for my local
configuration, so I can't use it in ltsp until it's
supported in all debian releases, i.e. in 6 years or so :)
318[05:30:02] <alkisg> I wanted to drop a conffile at
sshd_config.d from the ltsp package
420[09:24:14] <shtrb> Is there a better option to check the
current loaded nvidia driver info than "grep -i "NVIDIA
GLX Module" /var/log/Xorg.0.log " ? tried modinfo nvidia
but it failed
421[09:24:14] <jim> that worked fine :)
422[09:24:53] <jim> shtrb, maybe grep the output of lsmod?
432[09:28:05] <shtrb> is modinfo simply don't work for non
GPL modules ?
433[09:28:40] <somiaj> I don't see why it won't work,
but it more likely gives you info about the module on the disk vs
the one loaded in memeory (if there is a version difference)
441[09:41:04] <somiaj> it is probably depmod (or some other
thing) that puts all the module information in a location modinfo
can find the module without a full path isn't being updated for
that module
444[09:42:41] <somiaj> though that only gives you info about the
module on the disk, not the one loaded in memeory, so I think dmesg
is probably better for that
451[09:46:08] <antto> i can move the mouse cursor around, but
the graphics are frozen including the clock on the panel, cpu
meters..
452[09:46:28] <shtrb> antto , try to move to tty1 and back to
tty7
453[09:46:47] <antto> ctrl+alt+F1?
454[09:47:02] <shtrb> followed by by ctrl+alt+f7
455[09:47:10] <antto> no reaction
456[09:47:17] <antto> also no reaction to pressing numlock
457[09:47:53] *** Quits: Z4CHe (uid496478@replaced-ip) (Quit: Connection closed for inactivity)
458[09:47:54] <shtrb> Can you ssh into it or do you see the hdd
led blinking ?
459[09:48:02] *** debhelper sets mode: +l 1008
460[09:48:35] <somiaj> and if ssh isn't running, you can
try ping, almost sounds like xorg has crashed, or it could be that
you are running low on memeory and everythign is beign swapped, that
can make things appear to come to a halt
461[09:49:08] <antto> just before it froze, it started thinking
intensively for a bit (i heard fans spinning up) and then it stopped
thinking, now the red LED periodically blinks on
462[09:49:30] <shtrb> could be X missbehaving
463[09:49:38] *** Quits: shtrb (~shtrb@replaced-ip) (Remote host closed the connection)
464[09:50:38] <antto> i *think* there's ping to it
478[10:00:03] <somiaj> okay, yea memeory is fine, guess
something else has caused your xorg to crash. What type of upgrade
where you doing?
479[10:01:06] <shtrb> how much free disk space do you have ?
xorg can crash or lock if there is no disk space for the current
logged in user
480[10:01:18] <antto> hadn't booted this debian since a
long time, i Reloaded and Marked all upgrades.. there were many new
versions of things, firefox v68 to 78 or so, for example, also, new
versions of the kernel, grub, etc..
481[10:01:33] <antto> not sure about disk space :/
482[10:01:51] <somiaj> antto: what version of debian do you run?
483[10:01:58] <antto> 10
484[10:02:09] <somiaj> do you mean firefox-esr?
485[10:02:15] <antto> yes
486[10:02:24] <antto> the one from synaptic
487[10:02:36] <somiaj> I guess there have been a point release,
just making sure you weren't trying to say mix stable/testing
488[10:02:59] <antto> no, it's debian stable
489[10:03:16] <somiaj> anyways, you might just need to reboot
the machine and then see if you can fix the upgrade using apt
(generally better to use for this anyways)
490[10:04:04] <antto> somiaj, from last time i had upgrades to
be interrupted, i remembered that this will break my grub at least
491[10:04:09] <somiaj> It could be that maybe some xorg packages
and various things were upgraded, and the upgrade caused a problem
while running xorg. It is often suggested you at a minimum run
upgrades in screen/tmux just in case xorg crashes during the upgrade
(as you see it can happen)
492[10:04:10] <antto> i wanna avoid this
493[10:04:29] <antto> hm
494[10:04:58] <somiaj> since you didn't run the upgrade in
screen/tmux and it is stuck in a crashed xorg, I don't know of
a way to recover it. I guess you could try to kill xorg and all the
offending processes then see if you can finish the upgrade from over
ssh (I just suggested a reboot since it was the easiet to get back
to a nice clean state)
496[10:06:09] <antto> i don't know about this screen/tmux,
isn't that h4x0r-grade stuff?
497[10:06:17] <somiaj> !screen
498[10:06:17] <dpkg> screen is a terminal multiplexor, allowing
you to have n terminals and easily switch between them, add key
bindings, scroll back, copy/paste and lets you resume all your
sessions if your terminal dies or are disconnected. See
/usr/share/doc/screen/README.Debian and ask about <screen
tutorial>, <byobu>. See
replaced-url
499[10:07:01] <antto> i'm more a GUI click-clackz0r :/
500[10:07:08] <somiaj> basically it just lets you run a terminal
that isn't attached to a specifc tty/pty, so if say in this
case xorg crashed, you could reattach yourself to the running
screen/tmux session in the same state.
501[10:07:42] <antto> but i upgraded via synaptic
502[10:07:43] <somiaj> I would suggest in the future trying to
learn a little bit about the cli, as doing these types of upgrades
from apt inside a screen/tmux session could avoid some of these
issues in the future.
504[10:08:02] <somiaj> yea in this case, since it appears your
xorg has crashed, it took synaptic with it.
505[10:08:19] <somiaj> I don't know of any way to recover
that (it might be possible but I don't think so)
506[10:08:48] <antto> if i just press the reset button, grub is
gonna be broken, isn't it?
507[10:08:56] *** Grldfrdom_ is now known as Grldfrdom
508[10:09:14] <somiaj> hard to say, it really depends on where
during the upgrade it crashed.
509[10:09:24] *** Grldfrdom is now known as Grldfrdom_
510[10:09:26] <somiaj> there is a good chance it will be fine,
but again it depends.
511[10:09:28] <antto> it froze in the middle of "reading
changelogs"
512[10:09:54] <somiaj> It might just have been waiting for you
to hit okay, so that is probably a decent place as it wasn't in
the middle of doing things, but waiting to start doing things.
513[10:09:54] <antto> with progress bar at about 60-70%
514[10:11:04] <antto> i think this dialog comes before the
actual unpacking and such
515[10:11:37] <antto> usually there are no messages in it, so
after it "thinks" it goes away.. sometimes i've seen
text come up in it which i have to read first
580[12:13:21] <shtrb> drop the default gw and dns records , have
a static route to your vpn
581[12:13:29] <shtrb> Botyuba, ^
582[12:14:35] <Botyuba> shtrb: so i can only connect to one of
the vpn's servers?
583[12:14:39] <shtrb> yep
584[12:15:10] <shtrb> you can even add a rule in your firwall to
drop any outgoing connection to anything other than your vpn server
ip
585[12:17:25] <Botyuba> so i could in theory put a DROP policy
for in and outgoing, and then only have exceptions for each and
every vpn server as well as my lan?
586[12:18:12] <Botyuba> i think lan is just 0.0.0.0 and then i
add additional exceptions for each and every vpn server i want to
add
587[12:18:32] <Botyuba> i am using the classical iptables
directly as a firewall
595[12:26:34] <shtrb> If your vpn provide you routes and a tap
access it's even easier to block everything and allow traffic
only via the tap device
596[12:27:44] <Botyuba> Ah, i understand now, so tap, so i i
understand the penultimate statement correctly, i should use
192.168.1.0/24 to refere to lan instead
597[12:28:35] <shtrb> (*) If that's your network, you can
also ask at ##networking if you don't get answers here fast
enough
598[12:28:40] <Botyuba> wouldn't 192.168.1.0/24 still refer
to EVERYTHING as the router is incluided there too?
660[14:13:52] <EdePopede> couldn't use youtube-dl for maybe
an hour now (the open programs all are running on a .fritz.box host
though), lsof showed me it was running on a weird numbered hostname?
661[14:14:20] <kirk781> maybe youtube changed the API again and
youtube-dl needs to be updated
664[14:15:59] <EdePopede> i also tried different versions i
still have here and some other sites (which were working yesterday),
and nothing changed since
665[14:16:09] <EdePopede> it's definitely the dns thing
681[14:39:53] <EdePopede> what *could* add "-5" (like
it did before) or maybe some other number to the hostname? and who
decides what hosts a program uses when it's started?
702[14:52:50] <ratrace> however, avahi wasn't LP's
original project, he joined his mdns implementation with it later
703[14:53:26] <phogg> he works on other stuff, but those are the
big headline items
704[14:53:35] <ratrace> pulseaudio is dead. long live pipewire
and 10 more years of half-assed half-buggy audio server until it
matures and we can kill it and start writing NodeAudio from scratch!
719[14:56:05] <EdePopede> technicalle i even have 2, i just
don't use them
720[14:56:13] <ratrace> been there, done that. two box of
tissues later, I switched to i3-wm and carefully crafted UI
environment.
721[14:56:13] <phogg> EdePopede: Does this sound familiar?
"$a was working, then it stopped. I found that $b was running,
so I removed it. Now $a works again." You can use "name
resolution" or "playing audio" for $a, and
"avahi" or "pulse" for $b.
722[14:56:23] <EdePopede> don't print for just 2 weeks and
you can buy new ink
724[14:56:36] <phogg> For his third attempt he was cleverer and
made the fix much harder.
725[14:57:49] <EdePopede> the thing is firefox/hexchat are using
host.fritz.box for their connections, youtube-dl suddenly was
running on host-5.fritz.box. then suddenly it was ok again. i
suppose that's the 2 blocks of entries i found in syslog.
726[14:57:56] <ratrace> good thing IBM nixed it and is now
reinventing all the bugs via PipeWire
727[14:58:07] <phogg> In all seriousness I think pipewire is
here to stay, unless something fundamentally changes in the demands
of the audio world. I don't see any fundamental problems with
it that can't be solved.
728[14:58:21] <ratrace> that was said for pulseaudio
729[14:58:23] <EdePopede> ping and traceroute on host-5
didn't even work btw.
730[14:58:28] <phogg> ratrace: never by me
731[14:58:34] <ratrace> I was there when lenny said it, on a
conference. and all I got was a lousy t-shirt!
732[14:58:43] <ratrace> *at a conference
733[14:58:49] <phogg> ratrace: pulseaudio had some pretty
obvious deficiencies from the very start
734[14:59:05] <ratrace> yeah
735[14:59:15] <ratrace> it got all the hate because distros
pushed it out before it was ready
736[14:59:18] <phogg> ratrace: you have to take what he says as
marketing speak. I think he gets paid for every install or
something.
744[15:00:27] <EdePopede> i understand that it doesn't work
on my 80s hifi, since it's a simple mechanical solution there.
so i was surprised to see it's possible on linux. and then
there came pulseaudio.
745[15:00:35] <ratrace> when the BadThings start overtaking
GoodThings in systemd and people will have had enough.
746[15:00:50] <phogg> ratrace: there are without a doubt
fundamental flaws and an appetite for a better-designed system which
accomplishes some of the same technical feats without the politics
and drama. In that way it is much like pulse.
747[15:01:04] <EdePopede> like with dropping MS Office in the
gov after 20 years?
748[15:01:14] <ratrace> EdePopede: the joke is that Alsa later
got the ability to multichannel so (one of the) initial premise for
PA was busted shortly later
749[15:01:27] <EdePopede> nice
750[15:01:31] <phogg> EdePopede: which government is doing that?
751[15:01:54] *** Quits: kristijonas (~kristijon@replaced-ip) (Remote host closed the connection)
752[15:01:58] <phogg> ratrace: it's hard to say that
wasn't hastened by the competition
753[15:02:08] <EdePopede> phogg: just a situation. won't
work. though Munich seems to return to at least OSS, even if
ignoring the F for now.
754[15:02:37] <ratrace> EdePopede: Munich is flipping between
Win and FOSS so frequently, you can't even use those news as
PRNG any more :)
755[15:02:49] <phogg> EdePopede: it's still good to hear.
Governments certainly should be striving in that direction by
policy, however little they succeed.
756[15:03:04] <EdePopede> their excuse to return to MS was
"our office workers were not able to use LibreOffice" or
something that stupid
757[15:03:20] <EdePopede> phogg: data protection. MS never will
be able to guarantee it.
758[15:04:02] <EdePopede> and then the problems using old files.
i should still have something around from word 4. i doubt their
newest product will be able to open them :)
759[15:04:34] <EdePopede> ratrace: don't forget, their
prefered place to do even politics is in a beer tent ;)
760[15:04:36] <phogg> EdePopede: MS Office is actually a lot
like Unix. A collection of generic utilities that the end user can
combine in useful ways to solve a much larger number of problems.
762[15:04:46] <ratrace> EdePopede: har har har! :)
763[15:04:57] <shtrb> phogg, flipping to FOSS is not always
really in FOSS favor, I know of a certain municipally department
that had Open Source activists code for them to develop a system
that was used only by that department and the activists didn't
get payed (the code should still be at github or something like
that)
764[15:04:59] <phogg> EdePopede: It's just too bad that
they forgot "do one thing well" and "operate on
streams of text."
765[15:05:46] <ratrace> but it wasn't a tent originally,
'twas a beer hall. Bürgerbräukeller. :)
766[15:05:46] <phogg> shtrb: I still call it a good thing. In
the long run governments must move away from closed software, or
nationalize the companies they buy it from.
767[15:05:56] <shtrb> Also a curse on docx and xlsx
768[15:06:01] <EdePopede> a lot of bug fixes which never
happened, a lot of features (read: code) implemented in the core.
771[15:06:43] <shtrb> phogg, the problem is it's a project
that will not be used by anyone and is no more than a hobby project
that the devs did not get payed for it
772[15:07:26] <EdePopede> important FOSS projects should
definitely get tax support.
773[15:07:41] <shtrb> tax support ? people need to pay rent and
pay for food
774[15:07:47] <phogg> shtrb: sounds like most software to me
776[15:08:16] <EdePopede> there's a *need* for software
meeting certain criteria and commercial software won't be able
to fulfill them all. never ever.
777[15:08:40] <EdePopede> like proved GDPR compliance and no
hidden costs and backdoors and what not.
778[15:09:27] <shtrb> let start they all drop all links to CDNs
,google ,amazon,micrsoft and facebook from all sites
782[15:10:35] <EdePopede> how about a button in the browser
which sends a mail complaining about such connections whenever an
official site implements something shady?
785[15:11:12] <EdePopede> phogg: even more if as a government
you want to keep backdoors alive to get around privacy tools
786[15:11:53] <EdePopede> phogg: backdoors which work only for
yourself ofc. because the dealer on the darknet promised you to not
sell it to anybody else.
790[15:17:47] <Na_Klar> I have to do some piping. I mkfifo a
pipe and then "tail -n 1 -f /pipe | someprog". which works
fine. then I tried "someprog | otherprog" which worked
fine as well. but when I "tail -n 1 -f /pipe | someprog |
otherprog" and then write something to /pipe, the
"otherprog" does not catch any stdout from
"someprog". what am I missing?
815[15:34:03] <chilversc> Is there a limit of 255 block devices
for a single hard drive? When running partprobe I get "Error:
Partition(s) 256 on /dev/sdb have been written, but we have been
unable to inform the kernel of the change"
871[16:35:11] <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>.
935[17:36:32] <Peyam> petn-randall, I want to be able to block
ads without using adblock extensions in FF
936[17:37:39] <petn-randall> Peyam: You're essentially
trusting your DNS to a 3rd party, so I'd rather install adblock
in FF than allow someone else to control my name resolution of my
whole system.
937[17:38:01] *** debhelper sets mode: +l 1033
938[17:38:10] <Peyam> petn-randall, no . I fear nothing
947[17:43:23] *** Quits: sig_9 (~sig@replaced-ip) (Killed (Sigyn (Spam is off topic on freenode.)))
948[17:44:39] <EdePopede> Peyam: once i had some /etc/hosts
redirects which blocked already a lot of ads. and then a script
blocker, never used a "real" adblocker.
958[18:02:20] <unixbsd> somiaj: it seems that he is not there
today, anyhow, here the amd64 edition that would give him a fix for
his LVM:
replaced-url
959[18:03:14] <blackroot> i chmod'ed a file to 444, not i
cant chmod, chown or delete the file. what do i do?
960[18:03:32] <Peyam> petn-randall, I was not satisfied with
that thing and now I can not remove it
963[18:04:04] <jelly> blackroot, what is the complete path to
that file and does the user have write permissions for the
_directory_ where that file is?
992[18:19:34] <jelly> Peyam, if they don't have
uninstallation instructions, you'll probably have to restore
the original /etc/hosts file manually and remove the directory with
the code, and the files their installation method placed on your
system
1009[18:28:49] <somiaj> though unsure if it adds your domain name
if you ahve internet durning the install, but fairly minimial and
easy to edit if you nee to change it
1065[19:51:29] <cybrNaut> the filename slightly differs. I
downloaded "debian-edu-testing-amd64-BD-1.iso", but the
SHA256SUMS file gives the hash for
"debian-edu-bullseye-DI-rc1-amd64-BD-1.iso"
1066[19:51:48] <petn-randall> They're not the same contents,
either.
1067[19:52:09] <petn-randall> cybrNaut: Doesn't the torrent
come with it's own SHA256SUM and signature?
1068[19:52:18] <cybrNaut> obviously. this is the bad file =>
replaced-url
1069[19:52:51] <cybrNaut> petn-randall: the torrent is broken
1071[19:54:35] <somiaj> why are you using the weekly builds and
not rc1?
1072[19:56:29] <cybrNaut> not sure.. i think i was unable to find
the rc1 file. it was quite a search just to find a Bullseye ISO.
1073[19:56:46] <sney> !testing installer
1074[19:56:47] <dpkg> Repeat after me: the testing installer is
for testing the <installer>, not for installing
<testing>. To install testing, do a minimal installation using
the <stable> installer and ask me about
<stable->testing>. Installer bugs should be filed against
the debian-installer pseudopackage.
replaced-url
1075[19:57:04] <sney> it's pretty reliable atm, but
expecting release quality from a pre-release installer is only
setting yourself up
1076[19:57:06] <somiaj> You might want to wait for the release,
it could be the issue you are running into is just due to these
weekly builds
1079[19:57:39] <sney> if you can reproduce the same checksum
issue with a buster installer then please file a bug, otherwise
it'll most likely come out in the wash
1102[20:25:10] <Reviziis> Hey guys.. Just installed Debian 10
XFCE using the non-free iso. I noticed in synaptic that there's
a ton of hunspell stuff installed. What's the easiest way to
completely remove all hunspell stuff without having to manually
select everything in synaptic....
1112[20:27:55] <sney> if you aren't short on disk space, it
may be wiser to just disable spell checking in whatever application
you're using where it bothers you
1115[20:29:40] <Reviziis> Thanks for the tips :) Might just leave
it installed. Was only going to uninstall it because the system will
mainly be used for gaming, YouTube & discord....
1116[20:29:48] <RoyK> ratrace: I just wondered why that
wasn't in the manual, really
1117[20:30:08] <ratrace> RoyK: it should be. Are you reffering to
the new, 2.0+ manual, or the old one?
1118[20:30:13] <RoyK> lz4 has been default for years, so I
thought it would still be that, but after splitting up the manpages,
I can't find it
1119[20:30:27] <RoyK> this is buster
1120[20:30:34] <RoyK> nothing fancy
1121[20:31:05] <ratrace> RoyK: yes the split up manpages are a
damned shame, nothing can be found now
1122[20:31:30] <ratrace> I have ZERO clue where the properties
are documented, and iirc the default should be mentioned for the
dataset property "compression"
1124[20:32:00] <RoyK> I guess some smartass found that they were
too long and just split them without adding hints on where to find
the stuff
1125[20:32:16] <ratrace> indeed. systemd at least has the
grep'able systemd.directives(5)
1126[20:32:42] <ratrace> and to think ZFS originated in the BSD
land that laughed at linux for lack of good manpages. TROLOLO merry
had a little lamb and what goes around comes around.
1127[20:32:55] <ratrace> (yea yeah, solaris ain't bsd, I
know)
1130[20:33:44] <RoyK> ratrace: zfs came from opensolaris and was
adopted by freebsd and is now more or less the same across platforms
1131[20:33:50] <ratrace> AH! here we go. zfsprops(8)
1132[20:34:00] <ratrace> RoyK: I know, I've been there at
the beginning ;)
1133[20:34:09] *** Quits: ov3rmind (~over0-07@replaced-ip) (Remote host closed the connection)
1134[20:34:37] <RoyK> hm current default compression algorithm is
either lzjb or, if the lz4_compress feature is enabled, lz4.
1135[20:34:50] <ratrace> and that one's enabled since v5000
iirc
1136[20:35:24] <RoyK> but zfs get compression only shows on,
without saying anything about whether lz4 is enabled or not
1137[20:35:32] <RoyK> I guess it's lz4, though
1138[20:35:34] <ratrace> check zpool properties
1139[20:36:49] <ratrace> zpool get feature@lz4_compress
<pool>
1140[20:36:58] <RoyK> extern feature@lz4_compress active local
1141[20:36:59] <RoyK> extern feature@zstd_compress enabled local
1142[20:37:11] <RoyK> that's from zpool get all
1143[20:37:12] <ratrace> yah zstd is now a thing with zfs
1144[20:37:36] <ratrace> BUT it's underwhelming. ZSTD is
good only if you want to beat GZ compressratio at the price of some
speed
1145[20:37:43] <ratrace> if speed is neeed, lz4 still is the king
1146[20:37:53] <ratrace> zstd is at lz4 speed but at much lower
compressratio
1147[20:38:13] <RoyK> I've used lz4 for about a decade and I
know it works ;)
1148[20:38:18] <ratrace> same here
1149[20:38:52] <RoyK> I guess 'active' is the one
activated and 'enabled' is just 'hanging arund'?
1150[20:38:59] <unixbsd> cd
1151[20:39:09] <RoyK> ~>
1152[20:40:10] <ratrace> RoyK: yes
1153[20:40:49] <ratrace> RoyK: enabled also implies versioning.
eg, a pool with that enabled cannot be imported with the .ko versio
that doesn't support it
1262[23:00:04] <Diagon> Anyone have experience with btrbk? In
particular privilege separation with btrfs-progs-btrbk?
1263[23:00:54] <Diagon> This is a patch-set for btrfs-progs.
I'm wondering why not just write a bunch of scripts for eg
btrfs-subvolume-show and the like.
1264[23:01:13] *** Quits: Gerowen (~Gerowen@replaced-ip) (Remote host closed the connection)
1265[23:02:09] <sney> binary tools can be faster and have fewer
runtime dependencies than scripts.
1266[23:02:33] <sney> but for btrfs internals/development
discussion, you might have better results in #btrfs. #debian is more
of an existing usage/support channel.
1267[23:03:11] <Diagon> Well, btrbk is actually just a script, so
... ?
1268[23:03:45] <Diagon> Not sure if #btrfs has much connection to
btrbk, but ... I suppose I could try there, thanks.
1311[23:50:46] <Mister00X> at this point in time probably by
upgrading to buster..... but maybe someone else here has a better
idea. The switch from oldstable to stable should be save though