15[00:10:39] <DDDD> I tried installing python 3.9 as altinstall
from source, but getting the error "import _tkinter # If this
fails your Python may not be configured for Tk ModuleNotFoundError:
No module named '_tkinter'"
16[00:11:06] <DDDD> As far as i know turtle comes with python
per default so what am i missing?
17[00:11:21] <DDDD> Is there any deb package for python 3.9
that one could install instead?
18[00:12:04] <sney> python 3.9 is available in bullseye, and if
this is a desktop/development system then it may be smart to upgrade
anyway
24[00:13:09] *** Quits: Lupricon (~Lupricon@replaced-ip) (Remote host closed the connection)
25[00:13:12] <DDDD> when did bullseye get released?
26[00:13:42] *** Quits: LucaTM (~LucaTM@replaced-ip) (Quit: To infinity and beyond...)
27[00:13:43] <sney> it'll likely be released this fall.
but the freeze process started today, so it'll be pretty
consistent from now on
28[00:13:51] <sney> dpkg: buster->bullseye
29[00:13:52] <dpkg> In /etc/apt/sources.list, change
"buster" to "bullseye", remove lines like
buster-backports, debian-multimedia <dmo>, and other 3rd party
repos as they are known to cause issues then do: apt update
&& apt upgrade && apt full-upgrade. Note that
testing is a <moving target> and may be buggy, and read the
sid FAQ:
replaced-url
36[00:17:21] <sney> I've been using it regularly for over
a year and it's very solid now. there may be some changes to
how your applications behave after upgrades, but debian testing at
this point in the dev cycle is more stable than some distros'
actual releases
37[00:19:20] <DDDD> Guessing it's not possible to just
install that python .deb package on current buster?
45[00:24:26] <jhutchins> DDDD: One way to do it is to set up a
dual-boot so you can still use the computer while they're
fixing stuff. You just chroot to the testing install, run apt
update, and see if it works.
46[00:25:10] <jhutchins> I did that for the intel video
upgrade, took a LOONG time.
47[00:26:21] <DDDD> Reason for why being so careful is because
i do some work from home which requires me to encrypt all my disks.
I do have backups for my most important files but if something goes
wrong i don't have the linux technical knowledge of
reinstalling debian while setting up the LUKS to work again without
formating everything.
48[00:26:42] <petn-randall> sney: Now new transitions can be
started from today, and buildessential packages are frozen. There
are still a lot migrations going on right now in bullseye.
54[00:28:40] <sney> petn-randall: it's still pretty ok for
a desktop, and the most straightforward way to get a debian system
with python 3.9 without jumping through extra hoops.
123[01:21:20] <sney> you can have multiple DEs installed as long
as you have disk space
124[01:22:34] <sney> I have lxqt installed as a backup on this
system in case of plasma shenanigans (haven't needed it in a
while, but you never know), same user, no conflicts.
125[01:22:56] <DDDD> KDE is not but it's a bit to blingy at
times
137[01:26:05] <sney> Mister00X: ~/.config is pretty well
separated in the current era. in the past, there was some chance of
things being overwritten, but modern stuff keeps to itself
138[01:26:36] <Mister00X> good to know!
139[01:26:38] <sney> it might get hairy if you're using the
same program in both DEs, and trying to keep different settings for
it, but that's really an edge case
140[01:27:18] *** Quits: Tobbi (~Tobbi@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
146[01:29:15] <DDDD> sney: same program such as something
included in the DE's? Does this mean changing settings on for
example firefox on one DE will overwrite the changes on the other
one?
147[01:29:26] <DDDD> Not that i care as i'd want the same
settings on both anyways but :P
148[01:30:18] <sney> like if I wanted hexchat to use a different
theme in lxqt than it does in plasma, I'd probably have to
script something to run at login, because hexchat's settings
are always in ~/.config/hexchat regardless of DE
149[01:31:15] <sney> same idea with firefox, yeah
150[01:31:57] <DDDD> ahh okay
151[01:32:12] <DDDD> I'd want same settings for everything
probably so sounds like i don't need to care.
152[01:32:13] <DDDD> Thanks sney!
153[01:32:16] <sney> np
154[01:32:21] <DDDD> Also got bullseye working on my VM now :D
159[01:34:25] <Xacyllum> Hi, a question regarding static
routing:
160[01:35:22] *** Prints is now known as VolatileEmpanada
161[01:41:19] <DDDD> sney apparently i have two display managers
installed, lightdm and sddm. xfce installation is asking which i
want to use ask default
162[01:42:34] <sney> it makes no practical difference. I find
sddm looks nicer, but it's your choice
166[01:47:40] <DDDD> Strange getting import tkinter as TK
ModuleNotFoundError: No module named 'tkinter' even on
Bullseye. Python 3.9 is installed by default but seems like
it's not compiled with tkinter
167[01:48:19] <Xacyllum> ratrace: have to anonymize it and write
it down at first on kwrite ;)
168[01:48:24] <Xacyllum> not very much
169[01:48:43] <sney> DDDD: you are probably just missing the
python3-tk package
189[02:02:12] <Xacyllum> Ok, now I'm done, only five lines
:D had to get an overview to don't get muddled.
190[02:02:32] <Xacyllum> we have the networks 192.168.88.0/22,
192.168.40.0/23, 192.168.10.0/24.
191[02:02:33] <Xacyllum> On 192.168.80.73 is a router to route
from 192.168.40.0 to 192.168.10.0, it has a interface on
192.168.10.0, too. It works already.
192[02:02:54] <Xacyllum> Which static route do I need on
192.168.40.131 to make it work?
194[02:03:16] <Xacyllum> I tried this static route, but it
doesn't work, because I also need to add a second route route
from 192.168.40.0/23 to 192.168.88.0/24.
283[03:21:32] <qr3461> anyone know if virtual box is available
on debian in the repo?
284[03:21:42] *** jotaxpe_ is now known as jotaxpe
285[03:21:47] <sney> !virtualbox
286[03:21:47] <dpkg> Oracle VM VirtualBox is
<virtualization> software. Not in buster and unlikely to be in
any future debian stable releases due to #794466. Unofficial
backports are available as well as 3rd party packages from Oracle,
see
replaced-url
287[03:22:46] <sney> if you don't *need*
virtualbox-brand-virtualbox, and just want to virtualize some stuff
on the desktop to test, take a few minutes to learn virt-manager.
less headache in the long run.
288[03:23:12] <qr3461> i actually have virt-manager
289[03:23:26] <qr3461> and i like it quite a bit, but was having
some trouble with the networking
290[03:23:35] <qr3461> and the documentation was a bit difficult
to parse at times
292[03:26:04] <sney> virt-manager's documentation can be a
little vague because it's just a frontend to libvirt/qemu
tools. anything more complex than a basic nat or bridge probably
requires understanding how to do it with qemu, and then applying
that to the gui; but at that point you might have outgrown
libvirt/virtualbox anyway and need a more dedicated hypervisor
294[03:29:45] <sway> I use libvirt/qemu with the Virtual Machine
Manager (VMM) and find it to be rock solid. The only non-linux VM I
have problems building with it is Win10.
295[03:30:07] *** Quits: sway (~Milo@replaced-ip) (Remote host closed the connection)
297[03:31:18] <qr3461> sney, are you good with
virt-manager/libvert/qemu? i actually have a networking question
that I haven't been able to figure out from the docs
298[03:31:59] <qr3461> basically, I set the networking to user
networking, but would like to create a firewall on my host that will
block some IP addresses that the guest is trying to connect to
299[03:32:38] <qr3461> ive tried with firewalld on the libvirt
group but it did nothing unfortunately
306[03:36:03] <sney> if you want to firewall traffic from your
vm guests then applying it to the virbr interface is the correct
approach, you probably just need to prioritize your firewall rules
better, and make sure you understand the distinction between the
'input' and 'forward' tables
307[03:38:16] <sney> er, chains. I rarely use host firewalls so
sometimes I lose the terms.
368[04:44:19] <qr3461> sney, apologies, was away for a bit. If
youre still here and can give me a quick rundown on the difference,
that would be very helpful
369[04:44:34] <qr3461> glad to know i was on track with applying
the rules on virbr
371[04:45:24] <qr3461> like for example, if I sent a ping to
1.1.1.1 on the guest, it would get sent from the vnet to the virbr
and then to the physical devices on the host if i understand
correctly
372[04:45:52] <qr3461> how would I then, on the host, check that
the packet is going to 1.1.1.1, and then create a rule that will
deny this packet?
374[04:46:18] <qr3461> maybe i need a forwarding deny rule or
something? im not too sure, not an expert on this stuff
375[04:46:30] <sney> I'm here but I can't really
commit the time to help with firewall basics. luckily, there's
youtube, and a lot of text manuals. and ##networking on this network
is ok if you have thick skin.
376[04:46:44] *** Quits: krkini (~kini@replaced-ip) (Remote host closed the connection)
377[04:47:17] <sney> networking/firewalls seem like something
that linux people should just be able to pick up, but it's
really a whole different field. netfilter.org has some good diagrams
to give you a leg up iirc
407[05:08:53] <galex-713> what about debugging symbols? do they
come necessarily from the same build process?
408[05:09:03] <sney> yes, and in a separate section.
409[05:09:06] <sney> !dbgsym
410[05:09:06] <dpkg> Packages that end in '*-dbgsym'
contain the symbols required for debugging executables and
libraries. The dbgsym packages are automatically generated packages
that are in a separate archive; add a line like "deb
replaced-url
412[05:09:37] <sney> security.debian.org doesn't have a
dbgsym section but those packages can often be found in
stable-proposed-updates too, which does
413[05:10:30] *** Quits: BrianG61UK_ (~BrianG61U@replaced-ip) (Remote host closed the connection)
433[05:14:57] <galex-713> but why isn’t the gcc version
the same as I have now then?
434[05:15:18] <sney> !your screen
435[05:15:19] <dpkg> Your screen is in front of your face.
#debian cannot see it. Please give as many details as you can about
the problem. It wastes our time and yours when we have to guess what
is on your screen. See <context>, <what>
436[05:15:30] <galex-713> it seems to say
“gcc-8_8.2.0-15” instead of “gcc-8_8.3.0-6”
there
444[05:18:58] <sney> ok, so telegram-desktop was uploaded to sid
(and then built, with the log you see there) on 03 Feb 2019. so it
was built with the sid/buster toolchain that we had at that time.
replaced-url
519[07:06:26] <asarch> One stupid question: how do yo make a
specific kernel your default kernel? E.g.: buster-backports:
5.9.0-0.bpo.2-686 (5.9.6-1~bpo10+1)
520[07:07:59] *** Quits: kini (~kini@replaced-ip) (Remote host closed the connection)
521[07:08:32] <asarch> I remember there is a dpgk make-default
linux-image-5.9.0-0.bpo.2-686 command for that
567[07:34:41] <Voltum> hey i installed ejabberd from
debian-backports, and this was missing (erlang-p1-pgsql). debian
splits support for mysql and psql into different packages..
568[07:35:00] <Voltum> installing the deb from their site, never
had that issue
569[07:37:10] <oxek> Voltum: 'apt -t buster-backports
depends ejabberd | grep sql' tells me they are Suggests
packages
584[07:59:34] <unixbsd> badblocks: hello, mkfs.ext2 -c /dev/sda1
is readyonly, mkfs.ext2 -cc /dev/sda1 is fine but takes several days
/ It is possible to have it faster? Which possible method would
exclude but in less time than -cc ?
623[08:51:55] <unixbsd> ratrace: I havent found any other
alternative. I noticed that ntfs bad block is little bit faster. xfs
does not have any bad block.
630[08:54:01] <dpkg> The hash of a file is a string of
numbers/digits calculated from its contents in such a way as to mean
that if the hash is correct then the file is OK. Debian uses hashes
extensively to ensure that apt downloads verified packages for you.
Ask me about <secure apt> or see
replaced-url
631[08:54:47] <Mister00X> gholinbrown: you can verify the hash
with the sign file
632[08:55:19] <gholinbrown> how do i know if the sign file is
authentic
648[08:58:15] <gholinbrown> how do i start a new game?
649[08:58:19] <rudi_s> If you started with the right iso then
all further updates are authenticated (assuming nobobody compromised
the build architecture, signing architecture, developer machine,
etc.)
650[08:59:03] <rudi_s> Well, use a trusted system (that might be
a problem). Find the proper GPG key (e.g. from colleagues), download
ISO, SHASUM, signature, verify everything and install it.
651[08:59:31] <rudi_s> However, there are many additional attack
vectors if you're really paranoid (usb stick, disk, BIOS/UEFI,
CPU backdoor, whatever).
652[09:00:14] <rudi_s> And you have to start somewhere. But
checking the signature using gpg and the hash is a good thing to do.
655[09:00:44] <rudi_s> That protects against malicious mirrors
which I think is the most likely attack (if "most" is even
realistic, doubt it happens often).
656[09:01:17] *** Joins: mezzo (~mezzo@replaced-ip)
668[09:03:42] <rudi_s> How do you know that your brain
wasn't manipulated and you actually verified the checksums and
it wasn't just a fake memory (e.g. as shown in Ghost in Shell,
fake memories that is).
669[09:03:46] <steven> and hope the mail man won't
manipulate the image :D
671[09:04:37] <pasiz> or if your switch have malware to
manipulate just debian iso:s
672[09:04:46] <rudi_s> But if you're up against resourceful
attackers they will most likely just enter your house and install
malware/cameras/whatever. Or manipulate the BIOS of the computer you
just ordered ...
697[09:16:08] <ratrace> unixbsd: have you actually tried
badblocks(8)?
698[09:17:29] <ratrace> unixbsd: also, have you measured the
throughput to your drive with iostat or vmstat? is it saturated? too
much iops, not enough MB/s? increaes block size for badblocks(8) and
reduce block count tested at a time so (-b times -c ~ 64k).
699[09:18:04] <ratrace> (so that -b X -c ~= 64k)
700[09:18:18] <rudi_s> gholinbrown: Just to increase your
paranoia. A USB stick (or a hard disk/SSD for that matter) is just a
tiny computer. So it can have malware as well and behave differently
when it's booting on your computer then when it's being
read from another computer.
701[09:18:43] <rudi_s> (I'm not trying to troll you. I just
want to point out, that you have to draw the line somewhere.)
709[09:20:46] <rudi_s> gholinbrown: "new usbs"? Why
not just use an existing one.
710[09:21:08] <rudi_s> It should be a few directory up, they are
called 'dvds' in the mirror hierarchy I think.
711[09:21:09] <gholinbrown> rudi_s i've plugged some of
them in to the machine. and i don't know when the potential
infection/whatever may have started
712[09:21:41] <rudi_s> gholinbrown: As I said, you have to draw
the line somewhere. You can just erase overwrite it and it will
(most likely) be just like new.
713[09:22:05] <rudi_s> (And how can you assume that it
wasn't infected in the first place when you bought it?)
722[09:25:03] <dob1> I am connected via ssh on a debian server
and on htop I don't see the names of the user of the processes,
I read them if I highlight the line where the process is
728[09:26:51] <Mister00X> dob1: have you tried changing them
729[09:27:54] <Mister00X> wait... htop uses curses nevermind
that was stupid to say
730[09:27:57] <rudi_s> Or your TERM is incorrect (either locally
or on the remote side).
731[09:28:01] *** debhelper sets mode: +l 1191
732[09:28:07] <dob1> Mister00X, if I run htop on the same
terminal on the local pc it works perfect, it's on ssh that it
doesn't work. for example sudo htop I just see root as user not
the other ones until I move the lines
733[09:28:18] <rudi_s> (You should never set TERM in any shell
configuration file, only in your terminal emulator.)
748[09:33:37] <unixbsd> ratrace: wow, this is pro answer, I
neeed to study this area. thank you
749[09:33:50] <rudi_s> dob1: And locally? And do you use screen?
screen-256color terminfo installed on remote host? Does it work
without screen, etc.
750[09:34:31] <dob1> rudi_s, local it works, tried several
terminal emulator it's the same. I am using tmux. same even if
I connect without having tmux running
754[09:39:28] <oxek> my /etc/firewalld/firewalld.conf contains
the line 'FirewallBackend=iptables' yet says that
'nftables (default)'. Anyone know why debian changes this
default?
783[09:58:52] <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
784[10:00:04] *** Quits: rnm[m] (mambangmat@replaced-ip) (Quit: Idle for 30+ days)
786[10:02:29] <gildasio> thanks ratrace, but it isn't a xy
case, because I really need this version. It isn't a
troubleshooting issue or something like this. If it were, I would
count on your help to a better diagnose
787[10:02:54] <ratrace> gildasio: I guess there's no other
way than with jigdo:
replaced-url
828[10:31:42] *** Quits: towo` (~towo@replaced-ip) (Remote host closed the connection)
829[10:33:20] <mrjpaxton> Hey, I'm looking for a way to
configure the TTY size when picocom starts up without having to type
in "stty cols <x> rows <y>" every time. Anyone
have an idea about that?
830[10:33:59] <ratrace> Lope: dunno.... forgot to mount /boot ?
836[10:40:59] <gholinbrown> it's a shame writing to usb
takes forever. esp. if usb 2.0
837[10:43:23] <Lope> gholinbrown, 99% of USB flash drives are
absolute trash. I've stopped using them.
838[10:43:36] <Lope> gholinbrown, I just use SSD's inside
USB enclosures now.
839[10:44:18] <mrjpaxton> Oh yeah, for sure. I don't miss
USB 2.o in the slightest. I'm loving USB-C too much. :p
840[10:44:52] <gholinbrown> progress bar has been stuck at
halfway mark for like 5 minutes (or more)
841[10:45:03] <mrjpaxton> yeah, about the only good thumb drives
are the Sandisk Extreme Go.
842[10:45:13] <mrjpaxton> And maybe Samsung.
843[10:45:30] <Lope> USB 3 flash drives are also trash. USB-C is
just a connector.
844[10:45:58] <Lope> I had a samsung 128GB USB3 flash drive.
Worst piece of shit ever.
845[10:46:20] <gholinbrown> i ended the transfer. way way too
long
846[10:46:38] <mrjpaxton> I have an SD card adapter that can
take any micro-SD, and "convert" it into a thumb drive. It
has one USB 3.0 port, and another one as USB-C.
847[10:46:44] <Lope> Flash drives are worse than trash actually.
Trash doesn't waste time. Time is money and life. Time is
everything. USB Flash drives are Satan.
848[10:46:55] <mrjpaxton> I can post an Amazon link to it if I
can find it. Lol.
849[10:47:05] <Lope> If you see a USB flash drive, get a hammer.
854[10:47:52] <Lope> Amazingly I've found microSD is often
vastly better than USB flash drives.
855[10:48:09] <Lope> It's like they think they can get away
with putting the absolute lowest quality crap in a flash drive.
856[10:48:24] <Lope> Whereas microSD must at least handle mobile
demands.
857[10:48:36] <mrjpaxton> Yeah, here it is:
replaced-url
858[10:48:42] <mrjpaxton> This thing is amazing.
859[10:48:43] <oxek> quadrathoch2: so you think I should switch
it to nftables?
860[10:49:02] <gholinbrown> and now the usb report won't
respond
861[10:49:44] <ratrace> Lope: well I can't disagree with
that. _every_ USB stick I've bought in the past several years,
even from known brands like kingston.... would die within a year.
bad sector kinda deid. and I used them in write once, read seldom
cases, for ISOs only
889[10:59:13] <ratrace> gholinbrown: well if it hogs again, same
questions apply, before you remove the drive....
890[10:59:39] <gholinbrown> lol. so i assumed now it's
done. so i do the safe remove. now it's stuck there acting like
it won't remove.
891[11:00:10] <ratrace> probably because linux is lying about
data transfers if there's no sync involved and it isn't in
dumb tools like file manager GUIs
892[11:00:32] <ratrace> so it's waiting for the kernel to
actaully page out the data to the device. again, check with iostat
if any data is moving toward it.
896[11:03:45] <gholinbrown> most times i've written
gigabyte sized files things do this. lock up etc. and i've
gotten pretty screwed before. only way to know for sure if it worked
is hash check comparison
900[11:05:13] <mrjpaxton> Yeah, sync (or fsync) is a thing that
I really didn't understand for a while when using Linux, but it
is a good way to increase the life of flash media as it decreases
writes to them in general, depending on how much RAM you have, or
when the device is removed, or your comptuer shuts down.
901[11:05:39] <mrjpaxton> But it can cause slow flash drives to
hang for a while. Lol.
902[11:05:57] <ratrace> if anything, sync _increases_ writes
903[11:06:26] *** Joins: jerry (~jerry@replaced-ip)
904[11:06:29] <ratrace> data may change in page cache a few
times before the kernel actually pages it out. if you sync'ed
after each change, it'd be writing out that each change
905[11:06:57] <gholinbrown> also always unplug and replug the
usb. else the hash check lies. i don't know why this is.
907[11:07:13] <mrjpaxton> So, when you write to a file, those
changes are stored in RAM, not written to the disk, but then when
you remove the flash drive, those changes are finally written. So it
could also be a method of wear leveling by spreading writes to RAM
and the flash media.
908[11:07:22] <mrjpaxton> That's as far as I understand it.
909[11:07:35] <ratrace> problem in linux is that without syncing
regularly on (large) data transfers, linux will wait until certain
dirty page tresholds are met. that's why you see, eg. dd
without sync, blast out super fast and then the rate goes down
910[11:08:54] <mrjpaxton> For large data, you might also
consider wanting filesystem compression. That will also decrease
writes.
911[11:08:56] <ratrace> mrjpaxton: it has nothing to do with
wear leveling. it's just writeback cache.
928[11:13:14] <Lope> I was having problems with the realcrap, I
mean realtek ethernet adapter going dead, but I contacted the host
and they told me to disable TCP segmentation offload. I also
installed r8168 dkms, and from the journal, the ethernet adapter
comes up and looks good (logs wise anyway)
929[11:13:29] <Lope> ah, thanks for reminding me to setup a root
pass
930[11:13:42] <Lope> could not having a root pass prevent the
system booting properly?
932[11:14:01] <ratrace> I'm not aware of any such situation
933[11:14:03] <mrjpaxton> Lope: Just so you know, Debian's
interfaces is a legacy method at this point. Most people are
switching to systemd-networkd, NetworkManager (nmcli) or something
else, like netplan.
934[11:14:14] <ratrace> mrjpaxton: not true
935[11:14:29] <mrjpaxton> I mean interfaces is still useful!
936[11:14:35] <ratrace> ifupdown is STILL pretty much the
defacto and default network framework in Debian
938[11:14:41] *** Quits: lichito (~licho@replaced-ip) (Remote host closed the connection)
939[11:14:44] <Lope> mrjpaxton, I think that's only true on
distros like raspbian or whatever they call it.
940[11:14:48] <BCMM> "legacy" as in Debian plans to
stop supporting it at some point, or "legacy" is in you
personally have stopped using it?
941[11:14:52] <ratrace> you're confuzing it with Ubuntu and
the nonsense they do with netplan that nobody else would touch with
a 10ft pole
942[11:15:16] <mrjpaxton> Sorry, I guess netplan is a bad
example.
943[11:15:38] <ratrace> extremely bad example
944[11:16:01] <BCMM> stuff like network-manager is probably a
good idea if you have, like, a laptop that moves around and attaches
to different wireless networks, and maybe a cellular dongle and a
vpn too
945[11:16:49] <BCMM> but if you have a desktop or server that
just always plugs in to the network, and *you want it to keep
working*, ifupdown is pretty hard to beat
946[11:16:53] <mrjpaxton> BCMM: You can also roam with just
wpa-supplicant. I have been doing it for a while.
947[11:17:09] <BCMM> mrjpaxton: sure, and i've done it, but
it's kind of a pain
948[11:17:26] <mrjpaxton> It is, sometimes. Lol.
949[11:17:27] <BCMM> for just moving around wifi networks i
quite like wicd, though
950[11:18:20] <mrjpaxton> Oh man, I used to use wicd-curses.
That was interesting.
951[11:18:26] <gholinbrown> alright. time to see if the iso
works
953[11:18:57] <ratrace> Lope: installing NIC dkms doesn't
sound normal. which chipset is it exactly? also, TCPSO can be
configured via ethtool iirc
954[11:20:25] <quadrathoch2> oxek as long as you use nftables,
yes ;)
955[11:20:31] <mrjpaxton> Yeah, realtek should be well
supported. Maybe you should try backporting the new kernel and see
if that works better. Otherwise, I'm not sure....
956[11:20:44] <mrjpaxton> I think the newer kernelhas some
better Realtek firmware at this point.
957[11:21:04] <BCMM> <mrjpaxton> Yeah, realtek should be
well supported. <-- there are a *lot* of different Realtek NICs
973[11:23:36] <mrjpaxton> Hmm... There's three different
numbers. Lol.
974[11:23:38] <Lope> ratrace, yes, I've done the ethtool
thing.
975[11:23:46] <mrjpaxton> That's totally not confusing. /s
976[11:23:52] <BCMM> "This driver should only be used for
devices not yet supported by the in-kernel driver r8169. Please see
the README.Debian for instructions how to report bugs against r8169
that made it necessary to use r8168-dkms."
977[11:23:58] <BCMM> hmm interesting
978[11:24:11] <ratrace> r8169 supports a wide range of 'em
, yea
979[11:24:28] *** Quits: bolt (~r00t@replaced-ip) (Remote host closed the connection)
980[11:24:38] <Lope> BCMM, before 8168-dkms the ethernet adapter
was going offline upon boot
981[11:24:46] <Lope> but that was before I disabled the
offloading
982[11:24:52] <Lope> I kind of did them at the same time
983[11:24:54] <BCMM> then i reckon using the dkms package sounds
reasonable
984[11:24:56] <ratrace> and i'd be surprised a hosting
company would set up such hardware.... what company is that? perhaps
you should replace it with some other, wher techs know what
they're doing
985[11:25:33] <Lope> ratrace, yeah, not too impressed. I asked
them to install another nic and they told me to use workarounds haha
986[11:25:41] <ratrace> what company is that?
987[11:26:04] <Lope> I'll give you a clue, starts with h
and ends in er
988[11:26:05] <jim> hi... are there plans to have a debian debug
server?
989[11:26:14] <ratrace> Lope: Hetzner?!
990[11:26:28] <Lope> ratrace, well, I don't want to diss
any companies.
991[11:26:31] <quadrathoch2> i assume that's a dedicated?
992[11:26:38] <quadrathoch2> Lope ^
993[11:26:39] <Lope> Yes
994[11:26:51] <ratrace> Lope: that's extremely surprising.
are oyu using some of those old Serverbörse machines*
995[11:27:06] <Lope> ratrace, nah, a new ryzen
996[11:27:21] <Lope> must be consumer hardware
997[11:27:26] <ratrace> ssh mx-01rr
998[11:27:28] <quadrathoch2> ratrace sadly it's not that
surprising, as their normal offerings are just consumer hw
1002[11:28:41] <ratrace> Lope: all Intel NICs on our ryzens at
Hetzner
1003[11:29:13] <Lope> ratrace, not this one!!!!
1004[11:29:17] <Lope> it's a realcrap
1005[11:29:22] <ratrace> Lope: if Hetzner's known for one
thing, is that they do custom server chassis and components that
they extensively test and use only those that give them least amount
of trouble
1013[11:30:21] <Lope> ratrace, this is /etc/network/interfaces
replaced-url
1014[11:30:22] <mrjpaxton> I am lucky with Realtek Ethernet,
it's jsut the Realtek Wi-Fi I have a bigger issue with. I have
a hard time getting some of the newer Wireless ad ones working.
1015[11:30:26] <Lope> It should be working :/
1016[11:30:41] <Lope> The boot journal shows the bridge comes up
etc.
1017[11:30:48] <ratrace> Lope: uhhh who set that up?
1018[11:30:49] <mrjpaxton> And Wireless AC has alaways been iffy,
too.
1019[11:31:07] <Lope> The journal says everything is fine, shows
the SSH is listening with port number etc.
1020[11:31:35] <ratrace> Lope: also, if you're bridging with
eth0 ... did you make sure to include net.ifnames=0 in the kernel
command line?
1021[11:31:36] <Lope> ratrace, I set that up. for now I've
commented out everything else, just trying to get it accessible
1022[11:31:42] <mrjpaxton> Ew, that netmask though.
1023[11:31:50] <ratrace> Lope: be _very_ careful with bridges at
Hetzner
1024[11:32:04] <Lope> ratrace, well, it's not really called
eth0, it's enp blah blah blah
1025[11:32:05] <ratrace> one tiny packet going out with wrong
src-ip, you're dun goof'd
1026[11:32:10] <Lope> I just put eth0 for show
1027[11:32:35] <Lope> ratrace, okay cool. Well I've been
running bridges on dedicated for years. No issues.
1028[11:32:36] <jim> mrjpaxton, apparantly gdb has had a feature
added which looks at the thing you're debugging, and based on
its "build id", will import symbols and source (so with
this system, there might no longer be a need for compiling with -g
(not sure if this is true for all cases)
1029[11:32:56] <ratrace> Lope: also, cosmetics but .... use the
CIDR IP notation, don't use netmask
1030[11:33:17] <Lope> ratrace, ah, I didn't know you can
CIDR in /etc/network/interfaces
1031[11:33:25] <Lope> I thought it's too oldschool.
1032[11:33:49] <Lope> But I mean, there's nothing wrong with
my /etc/network/interfaces. Journal shows the bridge comes up.
1033[11:34:07] <Lope> (as far as I can tell)
1034[11:34:22] <ratrace> Lope: are you looking at this through
LARA?
1035[11:34:58] <Lope> ratrace, I dunno what LARA is. I've
been editing the files in nano, and then checking the journal after
the boot fails to be internet accessible.
1038[11:35:25] <ratrace> it's hetzner's KVM console.
you also have an option of vKVM. how are you checking the files
then... rebooted into rescue mode, mount, chroot ?
1045[11:36:29] <ratrace> Lope: open a ticket and request KVM.
they'll give you connection details. the java applet you can
download and exec directly on the command line with javaws, since
browser won't support the applet
1046[11:36:32] <Lope> ratrace, I asked for the console thing, but
they seemed to ignore my request.
1048[11:37:05] <ratrace> Lope: they won't ignore it.
they'll get back to you as soon as a device becomes available.
they have limited supply of them
1049[11:37:15] <Lope> ratrace, I suppose the KVM is the way to go
at this point. It's been such a monumental waste of time. With
the NIC unexpectedly being dead etc.
1050[11:37:21] <mrjpaxton> jim: Oh, so you're looking for
packages that you can debug? There are a lot of "-debug"
packages available inthe repos, but not a whole lot. Try looking up
"deb-src" and see if you can build in packages with debug
symbols.
1051[11:37:27] <ratrace> it's not OVH with OOB IPMI over VPN
:)
1052[11:37:33] <mrjpaxton> Because I'm not sure about that!
1053[11:37:36] <Lope> Naturally I debugged all my other
shenanigans first, all the layers of shit I'm running.
1054[11:37:51] <Lope> mdadm, luks, zfs, etc.
1055[11:38:10] <ratrace> Lope: make sure you have root pass set,
or unpriv user in wheel or with sudo capabilities, or else you
won't be able to log in on the KVM console :)
1056[11:38:12] <mrjpaxton> LUKS with ZFS? Interesting....
1057[11:38:27] <Lope> ratrace, thanks for the reminder :/
1058[11:38:35] <Lope> (poo emoji)
1059[11:38:47] <Lope> Anyway, I better do useful stuff today, not
just bang my head on a server.
1060[11:38:56] <Lope> Thanks for the amazing advice as always
ratrace
1061[11:38:57] <ratrace> ZFS atop of LUKS is neat-o. half our
prod servers are that. the other half is btrfs atop of LUKS :)
1064[11:39:28] <mrjpaxton> I am even having a hard time getting
AirPrint and AirScan to work with CUPS. So I also have more
networking projects at hand. Lmao.
1065[11:39:31] <Lope> haha, just kidding.
1066[11:39:37] <Lope> That would be a good way to troll the ZFS
people.
1067[11:39:52] <ratrace> Lope: i'll betja in advance,
you've got wrong enp* name . just shove that into net.ifnames=0
and make your life way easier ;)
1068[11:39:55] <mrjpaxton> It's realyl fun to do networking
projects... until they become frustrating for us not completely into
networking.
1069[11:40:17] <Lope> ratrace, I've quadruple checked the
interface names bud!
1070[11:40:23] <ksk> mrjpaxton: draw a picture, it helps :D
1071[11:40:37] <Lope> ratrace, but yes, it seems like eth0 would
be better for a remote dedi. Do you do it?
1072[11:41:05] <mrjpaxton> I'll stick to BTRFS with RAID 1
or 10 for my home NAS. :)
1073[11:41:06] *** Quits: JohnML (~john1@replaced-ip) (Remote host closed the connection)
1074[11:41:16] <ratrace> Lope: yup. there's always just one
NIC, but our ansible scripts bind them to their MACs anway
1075[11:41:21] <ratrace> "you never know"
1076[11:41:37] <ratrace> (OVH servers come with two, for example)
1089[11:44:04] <mrjpaxton> Well okay, they are modules. I know
ath9k-htc does that.
1090[11:44:08] <jelly> kernel gets the preferred name from... I
honestly don't know where, might be drivers, might be driver
type hardcoded, but userspace does re/naming
1091[11:44:19] <ratrace> mrjpaxton: does _what_ exactly?
1104[11:46:53] <jelly> I have a usb-c ethernet: enx340a46aa1e77
1105[11:47:29] <mrjpaxton> Yeah, I have an Armbian device that is
hosting my Wi-Fi with hostapd. I'm debating if I should just
turn interface renaming off or not. Does it really provide that much
more security?
1110[11:48:01] <mrjpaxton> It's okay, I still have the Wi-Fi
and Ethernet bridged as one LAN interface, so I just use the name of
the bridge for everything. Lol.
1111[11:48:11] <mrjpaxton> Problem solved.
1112[11:48:14] <jelly> it can provide help with provisioning. Or
hinder it
1113[11:49:34] <jelly> I guess it might be ince if you're
plugging and unplugging dozens of interfaces during the lifetime of
a machine
1114[11:49:43] <ratrace> maybe it's security by obscurity
because the haxxors can't .... uh...... predict....... the
predictable name in advance :))))
1115[11:49:57] <jelly> (we don't, give me back eth0 and eth1
any time of day)
1118[11:50:43] <mrjpaxton> I think having predictable names is
just to prevent user mistakes, more than anything. Because the
computer should know best. :p
1130[11:58:15] <quadrathoch2> I would say almost mrjpaxton, but I
like networkd
1131[11:58:59] <mrjpaxton> Yeah, just as a challenge I converted
a working interfaces to networkd, and it works just like it did
before, after fiddling. Lol. It's fun.
1132[11:59:13] <ratrace> mrjpaxton: you can rename them even with
ifupdown
1133[11:59:22] <mrjpaxton> I just use itnerfaces as a testbad at
this point, because I'm overly familiar with it.
1154[12:12:20] <mrjpaxton> Ah, is this another new-ish RedHat
thing that's been going on?
1155[12:13:02] <jim> dunno, I just heard about it 5 mins ago
1156[12:13:17] <jim> or maybe 15
1157[12:13:25] <mrjpaxton> Just for the record, Debian is very
slow to adopt new changes, and is a community/volunteer funded
project by SPI. Whereas RedHat and Fedora are corporately sponsored.
Don't expect new and shiny from this distro. ;)
1159[12:14:46] <jim> I understand, having run debian for about 20
years... I just figure it might be something that folks in debian
might want to look into
1160[12:15:35] <jim> maybe if people find out, we get a server
around the time current sid gets released stable
1161[12:15:52] <jim> (aka 2 releases from now)
1162[12:16:14] <mrjpaxton> Okay, well you are outclassing me
right now, since I've only used GNU/Linux for about 10 years
since switching off from Windows 7. Haha.
1163[12:17:18] <jim> yeah, I can tell stories about watching
debian upgrade itself from libc4 to libc5, and at the same time move
from a.out to elf
1165[12:17:41] <jim> (this would be before apt-get :)
1166[12:18:10] <mrjpaxton> Oh actually. Yes. I'm glad that
you brought that up, becuase that was pretty painful. AND I was
playign with Testing at the time, so even more pain. I would say it
was just as bad, if not worse, than the upgradefrom GRUB 1 to GRUB
2.
1167[12:18:58] <jim> the upgrade went very smoothly, which
impressed me
1168[12:19:31] <jim> I liked grub1 though, it had that grub shell
1169[12:19:40] *** Quits: JohnML (~john1@replaced-ip) (Remote host closed the connection)
1170[12:19:56] <mrjpaxton> I weep for my days when I was on
Unstable all the time. It turns out Ubuntu was really all I wanted
in a desktop. Debian makes a good stable distro, but they really
suck at doing rolling-release. I guess I just learn the hard way.
1171[12:20:31] <quadrathoch2> mrjpaxton so what is rolling in
ubuntu Oo
1173[12:20:47] <mrjpaxton> But I really do not like how bloated
Ubuntu feels compared to Debian. I am thinking of switching my
desktops back to it.
1174[12:21:21] <mrjpaxton> I'd miss how up to date the
software is, though.... Hmmm....
1175[12:21:29] <rk4> ubuntu is going down a little weird path,
packages being a mix of snap and deb
1176[12:21:35] <mrjpaxton> But yeah, for servers... Debian all
the way.
1177[12:22:09] <mrjpaxton> rk4: I know. It sucks. Chromium is no
longer in the repos anymore. It's only available as a snap.
1178[12:22:23] <jim> that was at the beginning of my linux use...
I went from sls, to slackware (which I destroyed by trying to
upgrade libc), and looking for a dist that could upgrade itself, so
trying redhat, it could not, then I went to a newish project called
debian, which did
1180[12:23:56] <mrjpaxton> Yeah, I don't think I will ever
use Debian Unstable/Testing again. I'd rather just use stable
with backports/Flatpak or even install new software with containers
or a chroot if I have to.
1181[12:24:16] <mrjpaxton> I just want a stable base system, but
with some up to date packages, if I need them (for a desktop).
1182[12:24:21] <rk4> debian definitely nailed the package
management side of distroing, took a very long time for redhat to
almost catchup [up2date, yum, dnf[?], but dnf is a slow user
experience]
1183[12:24:39] <mrjpaxton> APT will forever be king for me. :)
1184[12:25:16] <jim> much later, maybe 3 years ago, I compared a
ubuntu install to a debian install... and for speeds (which is
pretty much the only thing I could do), I found that debian was
faster, to download (but I cheated, got a major university with a
debian mirror 15 miles away), to install, and to use (sometimes
debian was about twice as fast) BUT! this was not a study that I
repeated, and is in no way scientific
1187[12:26:36] <mrjpaxton> Oh yeah, minimal Debian is very much a
thing. if you want such a thing as minimal Ubuntu, then you'd
have to install Ubuntu Server. They recommend not to even use it as
a desktop, but I don't see why not!
1188[12:27:07] <jim> but these were not minimal installs
1189[12:27:31] <jim> in debian's case, I told it to install
kde
1190[12:27:49] <mrjpaxton> Oh. Pff... I only use the netinstalls
or minimal ISOs. It's just so much easier to install exactly
what you want afterwards.
1193[12:28:22] <mrjpaxton> I tried to do the custom live CDs, but
figured it was way too much work. I'd rather just install to a
USB drive with persistancy... or something.
1194[12:28:49] <jim> for the debian install, I used a netinstall,
but when it got to the tasksel, I had it install kde, ssh server,
standard tools
1203[12:32:20] <Lope> ratrace, seems super dodgy to use?
1204[12:32:28] <Lope> ratrace, could be MITM?
1205[12:32:39] <mrjpaxton> jim: Ooh, what about options for
persistency, though? I know you can do the same with Ubuntu or
TAILS, but I'm not sure about Debian.
1206[12:32:44] <Lope> Do you have any mitigations for the risk?
1207[12:32:56] <ratrace> Lope: none, unfortunately that's
how their KVM is set up
1208[12:33:16] <Lope> haha, so u just login and pray?
1211[12:34:01] <ratrace> Lope: unfortunately. you could bind the
cert in your browser for future use. I don't know if they offer
a CA you could download
1212[12:34:06] <jim> mrjpaxton, well really, the only live image
I built so far, has borg backup, network support, I think xfce and
gparted
1213[12:34:33] <jim> I made it to have an image to restore from
backup
1214[12:34:53] <mrjpaxton> Oh yes. I desperately miss gparted
from the default live images. I always think what do I need without
an Internet connection, just in case....
1215[12:35:07] <mrjpaxton> And that's totally one of them.
Lol.
1216[12:35:17] *** Quits: eliotome3000 (~eliotime@replaced-ip) (Remote host closed the connection)
1217[12:35:33] <Lope> ratrace, it's this thing:
replaced-url
1224[12:38:20] <Lope> ratrace, you mean like tinyproxy or
whatever?
1225[12:38:35] <ratrace> Lope: no, ssh socks proxy
1226[12:38:50] <Lope> ok
1227[12:39:00] <jim> and/or... I can try to find my image config
files and you can try that, possibly modifying to lxqt instead
1228[12:39:38] <mrjpaxton> I remmeber the good old days of using
Parted Magic before they turned into payware. Lmao
1229[12:40:07] <jim> I think someone I knew bought one
1230[12:40:43] <mrjpaxton> Yeah, it used to be completely free.
1231[12:41:23] <ratrace> omg paying for someone's work! what
an atrocious thought!
1232[12:41:31] <jim> now you can use gparted, which is still free
1233[12:42:06] <mrjpaxton> Or parted. ;)
1234[12:42:15] <jim> right
1235[12:42:53] <jim> I couldn't get things right with
parted, (but I could before with fdisk)
1236[12:43:32] <mrjpaxton> Typically, you want to use parted
together with fdisk/gdisk
1237[12:43:58] <jim> that could have made sense when I was doing
it :)
1238[12:44:49] <jim> by luck, a good partition table was backed
up near the end of the drive
1239[12:45:39] <mrjpaxton> I think parted helps with the
partition table and creation, where fdisk/gdisk does the layout
more. They do seem to have options that clash, but yeah.... It is
interesting.
1240[12:45:53] *** Quits: nickodd (~nickodd@replaced-ip) (Remote host closed the connection)
1241[12:45:57] <mrjpaxton> And then you'd use
"mkfs" to create the actual filesystems for those
partitions.
1242[12:46:11] <mrjpaxton> Learning how to do all of that on the
CLI is a useful skill, but time consuming. Lol.
1243[12:46:33] <ratrace> parted can create, layout, remove,
rename and set flags on partitions just fine
1262[12:49:09] <EdePopede> that one too, yes. only it's
called FORMAT.COM only (FORMAT /mbr, remember?)
1263[12:49:29] <mrjpaxton> Oh yeah, lol.
1264[12:49:31] <Lope> ratrace, I've opened the HTML KVM
thing in firefox, it showed a small window, but when I tried to open
it up, it failed every time.
1273[12:50:55] <EdePopede> i think there were also some issues
with sfdisk/cfdisk or what they were called, not sure of fdisk used
to do everything i wanted, but i stayes away from the others.
1274[12:51:11] *** lovetolearn is now known as ltl3
1277[12:52:05] <mrjpaxton> Yeah, I don't like frontends to
basic utilities that work fine. For example, htop for top, or
aptitude for apt. The regular ones can work and be configued close
enough just fine!
1278[12:52:33] <mrjpaxton> So obviously not a fan of cfdisk.
1279[12:52:34] <ratrace> htop is not a frontend for top tho
1280[12:52:41] <ratrace> and it has a ton of features that top
doesn't
1281[12:52:52] <EdePopede> htop is nice, it lets me do things
with a few keypresses. or i just didn't find top's
interactive mode yes. only it messes up process names sometimes.
1282[12:53:08] <jim> I think aptitude (and apt) used libapt,
rather than apt-get
1283[12:53:08] <mrjpaxton> I do agree, but you are talking to a
hardcore top user here. :p
1284[12:53:27] <EdePopede> can top give you info on a particular
process?
1285[12:53:29] <mrjpaxton> I want my processes to ALWAYS BE
ACCURATE.
1286[12:53:33] <mrjpaxton> Yes, please.
1287[12:53:42] <mrjpaxton> Any bug in htop is no problem for me.
Lol.
1288[12:53:44] <ratrace> and htop info isn't accurate?
1290[12:54:05] <mrjpaxton> I'm just being silly now. I
don't hate on htop users.
1291[12:54:17] <mrjpaxton> I just don't need it.
1292[12:54:18] <EdePopede> or sort it / switch forest on and off
while it runs?
1293[12:55:00] <mrjpaxton> I'm just waiting for someone
who's like, bruh just use "watch ps aux"
1294[12:55:32] <ratrace> htop > top > ps in context of
features, but it's purely a preference which one you like
1295[12:55:32] <EdePopede> right now i have a process (sd-pam)
which is /lib/systemd/systemd. not sure if it's meant to be
like this, got the impression in the past that it happens around
upgrades when the original file disappears with the old package.
1296[12:55:58] <ratrace> and real admins grep the procfs :)
1297[12:56:08] <mrjpaxton> "top" is three letters, btw.
Not 4. :)
1298[12:56:19] <EdePopede> accidentally used top recently
(instead of ps, happens *shrug*). it cuts off process names,
couldn't find a switch to turn this off?
1299[12:56:36] <ratrace> (real-er admins awk th eprocfs and
real-er-er admins just launch htop in their emacs OS :) )
1300[12:57:18] <mrjpaxton> EdePopede: Could be the terminal
emulator/TTY?
1301[12:57:40] <oxek> I still need to learn the proper flags to
pgrep so that it shows what I want
1302[12:58:01] <mrjpaxton> I also have tree mode enabled, I use
kitty (terminal) and I've never had it cut off.
1303[12:58:08] <mrjpaxton> So I dunno.
1304[12:58:18] *** Quits: szorfein (~daggoth@replaced-ip) (Remote host closed the connection)
1305[12:58:33] <EdePopede> and right now i also have the
hierarchy SCREEN -R > SCREEN -R > /bin/bash, with the 2nd
SCREEN -R actually being bash, as the other direct child processes
of SCREEN are.
1306[12:58:35] <jim> if you like python, you might appreciate
xonsh
1309[12:59:38] <mrjpaxton> Oh, speaking of python... export
PYTHONDONTWRITEBYTECODE=x and export PYTHONSTARTUP=~/.pythonrc
1310[12:59:45] <mrjpaxton> You'll thank me later.
1311[12:59:59] <EdePopede> mrjpaxton: nope, they don't end
at the same column. besides that my maximized xterm is 212 chars
wide, it's mostly firefox and xfce-panel stuff reaching its
right border in htop.
1313[13:00:32] <EdePopede> they are cut to 15 chars and some to
another (still fixed) value. which is the next odd thing, why 2
values?
1314[13:02:17] <mrjpaxton> Hmmm.... I know you can adjust the
column size with "stty cols <X>", but I don't
think it can go above the limit, unless your terminal emulator has,
like, a horizontal scroll bar.
1315[13:02:33] <EdePopede> as an example (normal mode) >
panel-13-xfce4-
1316[13:02:52] <mrjpaxton> Have you tried top in a TTY?
1317[13:02:57] <mrjpaxton> I'm just curious.
1318[13:03:04] <EdePopede> and there's plenty of space next
to it
1333[13:05:40] <mrjpaxton> My columns on 1080p is 240, because 12
point font. I don't think I have a process that goes long
enough. I also have threaded mode turned off.... Maybe I can test by
turning it on.
1341[13:07:16] <oxek> ratrace: but I learned a lot. For example,
I learned that the situation regarding nftables and iptables is a
complete mess that will not be solved for years to come.
1342[13:07:21] <Lope> how can I run the jnlp "with
icedtea"
1349[13:08:33] <mrjpaxton> oxek: Yeah, I tend to stick with pure
nftables without firewalld. Probably because I'm crazy, but...
it seems to work better. Lol.
1350[13:08:55] <Lope> ksk, ah yes I remember.
1351[13:08:57] <ksk> Lope: might be you need to install said tool
first, but its kind of bundled to icedtea, iirc.
1352[13:09:06] <Lope> ksk, I don't have javaws, only java
openjdk
1353[13:09:34] <ratrace> Lope: no javaws after installing
icedtea-netx?
1357[13:09:45] <quadrathoch2> oxek oO buster is defaulting to
nftables, so let me read it thouroughly
1358[13:10:01] <oxek> mrjpaxton: but do you actually really have
pure nftables, or is it still going through all sorts of
compatibility layers and shims pretending to be iptables?
1359[13:10:13] <ratrace> you mean.... iptables in buster use the
nftables backed by default, unless you install iptables-legacy
1360[13:10:15] <mrjpaxton> Yes, and thank God. After learning to
use nftables for a week, I am very much a fan of it.
1363[13:10:41] <mrjpaxton> oxek: Nope! I have iptables completely
uninstalled from my system now. :p
1364[13:10:48] <mrjpaxton> Just nftables.
1365[13:11:40] <jim> quadrathoch2, maybe in later kernels, they
took iptables out, and replaced it with a compatibility layer, that
still lets you write iptables rules, but what the compatibility
layer produces, is nftables rules
1409[13:23:21] <ksk> ubuntu does also still ship jre-8 O_o
1410[13:23:32] <ratrace> but jre doesn't come with javaws
1411[13:23:46] <ratrace> or..... is there anotehr way to start
the jnlp thingy?
1412[13:24:40] <mrjpaxton> I feel like the amount of old shared
libraries Java 8 would need would render it uninstallable at this
point. If you can chroot or use a Linux container/VM with oldstable,
that might be your best chance to install Java 8.
1413[13:24:40] <ksk> icedtea-netx depends on
"default-jre", and as I understand it there is a
connection between "javaws" and the jre you have
installed.
1415[13:25:19] <ratrace> Lope: Hetzner also has vKVM ... which is
basically starting your rootfs inside a VM.... the network
won't work from ousside (different IP ranges tehy use iirc) but
at least you can check if your config brings out hte NIC with the
static IP properly
1416[13:25:21] <ksk> I once was supplied with a Windows XP VM
running Java6 ;)
1417[13:25:53] <mrjpaxton> Oh boy. I am planning to do an offline
Windows XP machine for old PC games.
1418[13:26:04] <ratrace> ksk: Lope: ah..... also look at
nvidia-openjdk-8-jre maybe that'll help
1419[13:26:05] <mrjpaxton> But I also want to use RDP over the
LAN. Hmmm....
1420[13:26:05] <Lope> ratrace, that's not a bad idea!
1421[13:26:22] <Lope> ratrace, I've actually used JRE
directly before for this purpose
1422[13:26:27] <Lope> But trying now I get this: Bad
installation: JAVAWS_HOME not set: No such file or directory
1423[13:26:38] <Lope> I've set the environment variable, but
it doesn't seem to want that.
1425[13:27:10] <ratrace> dunno.... now I'll be trying thir
KVM interface on one of our test servers, and if I find that I
can't use LARA with tools in Buster .... it's goodbye
Hetzner time.
1426[13:27:13] <mrjpaxton> Not that I dislike Wine so much, but
it does have regression issues with old games. Also, Windows XP is
nostalgic. :)
1427[13:27:31] <ratrace> as if plain http access to the core of
your server wasn't enough lololol....... OVH ain't that
bad any more.
1428[13:27:43] <mrjpaxton> Not as nostalgic as Windows 98 or
2000, but still nostalgic enough.
1429[13:27:48] <ratrace> mrjpaxton: why not win10. surely it can
run xp things
1430[13:27:56] <mrjpaxton> ratrace: Ew, no.
1431[13:28:01] <ratrace> ew why no
1432[13:28:05] <mrjpaxton> I'm saying for OLD games.
1433[13:28:25] <ratrace> you'd be surprised at the
compatibility maze win10 _has_ to support
1434[13:29:03] <mrjpaxton> So I'm trying to figure out if I
want to build a Windows XP machine, or just have a VM, and
PCI-passthrough an older GPU. But I also want RDP so I can play
those games on the go.
1435[13:29:14] <mrjpaxton> Yeah, I'm a wierdo when it comes
to tech stuff. Lol.
1436[13:29:33] <ratrace> I pci passthrough nvidia to my gaming
win10 VM
1502[14:28:40] <Lope> I wonder if I should try removing
r8168-dkms
1503[14:28:56] <ratrace> Lope: "a virtual MAC address needs
to be requested for each IP address via the Hetzner Robot and
assigned to the guest network card." ::
replaced-url
1504[14:29:01] <Lope> Going to give it a shot. Might regret it,
lol.
1553[15:20:09] <freem> Hello. acpid will not daemonize when it
receives something on stdin, which is to support systemd. But it
seems that asking it to go foreground, with --foreground, makes it
assuming stdin is open and have something to read on. Is that
normal?
1554[15:20:11] <Lope> ratrace, I'm doing a test to figure
out why it didn't work. I'm wondering if debian
didn't like that I had tabs (instead of spaces) in my file for
all lines from address and down.
1557[15:22:06] <azeem> freem: what's the problem with that?
1558[15:22:10] <Lope> ratrace, okay, it wasn't the tabs. I
re-did my bridge config with spaces, and it's not working
again.
1559[15:23:38] <freem> azeem: I use runit to run my daemons, and,
in a common file sourced everywhere, I 1) redirected stderr to
stdout and 2) close stdin. 2) makes acpid closing in loop, with no
reason, and that behavior is not explicit
1594[15:31:35] <JohnDohh> is it possible with partman to format a
drive without partition (using whole disk) ?
1595[15:31:48] <Hash> I am using debian sid on a test machine. I
have discovered something odd. minimal debian xorg setup,
xmonad/xmobar + terminal. I try to start a tray program called
'trayer' like so:
1610[15:33:58] <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
1646[15:40:17] <yanmaani> 1gb of stuff gets swapped into zram.
Then 1gb of zram gets swapped into zram, again. And so on.
1647[15:40:29] <Hash> yanmaani: Are you using it to have a ram
block device or using it for compressed swap?
1648[15:40:38] <yanmaani> Hash: compressed swap
1649[15:41:14] <Hash> yanmaani: I recommend ZSWAP instead as zram
meanages paging different and has much higher chance of stale pages
left and going to disk
1650[15:41:23] <Hash> I read something about that, hang on
1651[15:41:38] <ratrace> yanmaani: do you actually have on-disk
backing storage for swap?
1652[15:42:14] <Hash> It was someoen in #kernel... I forgot the
link now.
1653[15:42:36] <ratrace> Hash: I'd love to read it because I
use zswap too, and I'm being recommended to switch to zram
whenever I mention zswap
1654[15:42:48] <slop> Hi all, I am getting back into Debian -
havent used linux as my daily driver since like 2012. Hope this
finds you doing well
1657[15:44:42] <Hash> zram: Once a page is stored in zram it will
remain there until paged in or invalidated. The first pages to be
paged out will be the oldest pages (LRU list), these are
'cold' pages that are infrequently access. As the system
continues to swap it will move on to pages that are warmer (more
frequently accessed), these may not be able to be stored because of
the swap slots consumed by the cold pages.
1658[15:44:55] <Hash> What zram can not do (compcache had the
option to configure a block backing device) is to evict pages out to
physical disk. Ideally you want to age data out of the in-kernel
compressed swap space out to disk so that you can use kernel memory
for caching warm swap pages or free it for more productive use.
1660[15:46:02] <Hash> zswap: In-kernel cache is compressed, the
compression algorithm is pluggable using the CryptoAPI and the
storage for pages is dynamically allocated. Older pages can be
evicted to disk making this a sort of write-behind cache.
1661[15:46:06] <jelly> Hash, so zram backend is actual RAM.
That's kind of expected from the name
1662[15:46:15] <Hash> Needs a physical swap device (or swapfile).
1663[15:46:15] <ratrace> Hash: now that's weird, because
zram _can_ evict to disk
1664[15:46:24] <Hash> Hmm
1665[15:46:27] <jelly> and zswap does less compression, IIRC
1666[15:46:39] <jelly> I get 2.5:1 with swap on zram
1667[15:46:48] <Hash> Ditto
1668[15:47:17] <Hash> I just build a new 5800x cpu 64gb ram
thingy
1669[15:47:33] <Hash> Not expecting to use much swap at all for a
while.
1691[15:54:33] <ratrace> personally, the whole idea of
compressing it _without_ paging out is dumb. the whole idea is to
free RAM of unused pages. compressing them only half-asses that job.
if you're lucky to get half (50%) compressratio
1692[15:54:37] <jelly> the idea is to avoid certain classes of
bugs, not to avoid buying enough RAM
1712[15:59:47] <jelly> (also, I've already killed an SSD
swapping on it ;-)
1713[16:00:08] <Hash> I swap to a 10year old WD black, still 100%
healthy in smart
1714[16:00:09] <ratrace> what, 10 years ago?
1715[16:00:16] <koollman> I don't think zram is
single-threaded anymore
1716[16:00:17] <ratrace> jelly: ^^
1717[16:00:36] <jelly> let me see how old the replacement is
1718[16:00:45] <Hash> Regardless of the value passed to this
attribute, ZRAM will always allocate multiple compression streams -
one per online CPU - thus allowing several concurrent compression
operations. The number of allocated compression streams goes down
when some of the CPUs become offline. There is no
single-compression-stream mode anymore, unless you are running a UP
system or have only 1 CPU online.
1719[16:01:04] <koollman> zramctl shows a number of streams used,
by defaut equal to output of nproc I think
1721[16:01:17] <koollman> so you don't need separate devices
anymore
1722[16:01:55] <jelly> Power on time: 2757 days, 17 hours
1723[16:01:56] <ratrace> jelly: first gen SLC/MLC SSDs were
garbage. but if you bought an ssd int he past few years, unless you
bought some super cheap semi-anonymous carp, there's fat chance
in hell you can kill it with swapping alone, unless you had multi
MB/s swapstorms 24 hours a day, for many months.
1724[16:01:58] <Hash> I use 4 since I only have 4 cpu in my
script
1725[16:02:17] <jelly> ratrace, Vertex 2 was second gen.
1726[16:02:19] <oxek> !sources.list
1727[16:02:19] <dpkg> A suitable /etc/apt/sources.list for
"Buster" has the lines: "deb
replaced-url
1728[16:02:32] <ratrace> jelly: so, one of them oldie snowflakes
:)
1729[16:02:37] *** Quits: lesless (~lessless@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
1730[16:02:43] <Hash> oxek: you can also /msg dpkg !sources.list
1731[16:03:01] <jelly> ratrace, they didn't even track GB
Written in smart, then. The replacement Vertex 3 does.
1756[16:06:38] <ratrace> jelly: and on the servers, it's
mostly unused at all, with few dozen MB when ZFS is too slow to
release ARC and I have load peaks
1757[16:06:41] <jelly> ratrace, fun thing: kubernetes developers
think kernel doesn't have bugs. They disable and actively check
for swap being disabled.
1758[16:06:57] <ratrace> heh
1759[16:07:03] <jelly> predictably, k8s nodes hang BADLY under
pressure
1760[16:07:26] <ratrace> jelly: and nobody cares because
you're "supposed" to have many of them, failovers,
serverfarms to migrate between!
1761[16:07:33] <oxek> Hash: thanks
1762[16:08:04] <jelly> they have fugly workarounds like userspace
helpers doing "earlyoom"... that of course gets stuck like
all the other userspace
1763[16:08:18] <pasiz> is out of memory situation kernel bug?
1764[16:08:30] <ratrace> jelly: so that's why lenny invented
systemd-oomd :)
1765[16:08:44] <ksk> jelly: facebook begs to differ about that
assessment of yours :P
1766[16:09:04] <ratrace> pasiz: no, problably sysamind's bug
in the poor provisioning domain
1787[16:13:52] <Lope> I've got another hetzner server (with
an intel NIC) that has a bridge, doesn't do anything special
with MAC's and it's been fine for ages, never had a
problem.
1788[16:14:16] <ratrace> Lope: all my bridges on Buster share the
sae MAC with the eth in it.... but on Hetzner at least, that's
a must as it's the NIC's MAC they use to forwar packets to
you AND from you
1789[16:14:17] *** Quits: alexandros_c (~alexandro@replaced-ip) (Remote host closed the connection)
1822[16:20:36] <Lope> But oh so relieved to have figured this out
1823[16:20:45] <ratrace> Lope: but has that ACTUALLY solved the
issue?
1824[16:20:50] <Lope> I wouldn't have figured it out if you
didn't show me the stuff about the bridge
1825[16:20:54] <Lope> gonna test it now lolz
1826[16:20:56] <koollman> the main idea of oom userspace handling
is that it is limited by other ways first (cgroups), then once a
process is reaching the limits of the cgroup, rather than killing
it, something else may be done (like raising the limit a bit and
killing it properly)
1827[16:21:12] <ratrace> Lope: it was just my guess based on
knowing how hetzner routers packets.... did it actually solve the
issue?
1833[16:22:35] <ratrace> koollman: yeah the idea is not that bad,
to use cgroups and other variables for a bit more control over
usually haphazard kernel oom ... but ... one should not have oom to
beign with :)
1835[16:23:02] <koollman> ratrace: I mostly agree. but the
majority seems to disagree :)
1836[16:23:09] <ratrace> overcommitting your resources is not
unlike overselling shared hosting. you can cram thousands of users
based on a statistic that most of them use virtualy nothing, until
that one joker who does, and down goes your server.....
1838[16:23:56] <ratrace> koollman: that's not majority, that
the very same "lets cram thousnds of kates and then sell
support" RHEL crowd.
1839[16:23:56] <koollman> in general I prefer using
overcommit_memory=2, but tthere are some cases where overcommitting
is justified (or preferable, maybe). typically stuff doing large
mmap, or large programs that use fork
1840[16:24:41] <koollman> the influential crowd ? whoever defines
defaults, and whoever doesn't ask for them to be changed.
probably the majority :)
1841[16:25:01] <Lope> ratrace, the fucking bridge works.
1846[16:25:38] <koollman> like, sure, I don't care if a jvm
fails memory allocation, since it can handle it. But redis do really
like to avoid overcommit_memory=2, because the persistence model is
designed a specific way :)
1847[16:25:43] <jelly> that was apparently the thing I saw on k8s
hosts
1848[16:26:23] <koollman> (basically, fork the process then write
memory to disk. But the kernel thinks "wait, all that space may
be written later, so I need more space to duplicate pages. fork
fails)
1861[16:36:21] <Lope> ratrace, haha cry-laughing-emoji... now
that I uninstalled r8168-dkms it looks like NIC dies again upon
boot. So back to recovery to re-install r8168-dkms
1895[16:57:59] <ratrace> yanmaani: it's not about
performance but about kernel going belly up in high mem pressure
situations with no storage to page out stuff
1912[17:09:17] <Lope> ratrace, haha so I've finished my epic
investigation. I got fucked by 3 things: 1. Realtek, being a PoS
r8168-dkms is absolutely required for it to function. 2. MAC Address
filtering. 3. Bullseye changing the MAC.
1933[17:15:19] <Lope> I've got a buster and a bullseye
system. The behavior is different.
1934[17:15:32] <dermoth> Lope, I played with realtech wifi
dongles on my rpi lately... I find the new stock driver taht is
meant to replace all previous usb drivers works on a cold boot, but
has issue on warm boot - I came across a patch in this driver for
specific id's that performs a full init even if the dongle is
already initialized (or something like that, will have to dig). If
cold boot / reinserting dongle fixes it that may be the issue with
stock kernel driver.
1935[17:15:33] <dermoth> FWIW...
1936[17:15:40] <Lope> Both using the exact same style of etc
network interfaces
1938[17:16:33] <dermoth> haven't had time to test further,
but I'd like to add a module param to allow it on any dongle..
that is of course if your specific adapter is supported (not all of
them are yet)
1939[17:16:37] <Lope> dermoth, although sometimes I get confused
between realtek and ralink.
1940[17:16:56] <Lope> I'm not sure if I've used realtek
wifi or ralink. But one of them I had to disable hardware crypto and
then it worked.
1941[17:17:09] <Lope> With the crypto enabled, it transferred
data at snails pace.
1964[17:21:20] <Lope> ratrace, If they give you realcrap `apt
install r8168-dkms` and put this on your bridge and forever hold
your peace: hwaddress ether 00:11:22:33:44:55 \n pre-up
/usr/sbin/ethtool -K eth0 tso off
1966[17:23:10] <dermoth> yeah, the module in was rtl8xxxu - for
now I switch to another dongle with stock rtl8192cu driver (for the
chipset of the same name)... stable.... the symptoms were no auth
after a reboot, but cold boot worked (reinserting the dongle should
work too causes a power draw spike that resets the rpi anyway)
1967[17:23:18] *** Quits: tryte (~tryte@replaced-ip) (Remote host closed the connection)
1970[17:24:02] <ratrace> Lope: ah, no. if they give me realkrap,
I'll be asking for a different server
1971[17:24:47] <dermoth> a while back I also used one on my
computer, would fail on occasions. I wrote a network monitor daemon
that killed usb power for a second when router became unreachable.
different issue though.
1972[17:25:45] <ratrace> my workstation's r8169
WorksFine(tm) tho
1973[17:26:21] <Lope> ratrace, wise hahahaha
1974[17:26:43] <dermoth> So...back to what I came here for :)
I'm running two graphical desktops on my home computer (two
users... both me so security / separation between both isn't a
concern) - but I'm having audio issue on occasions - I
don't reboot often too. I feel like I may not have the ideal
setup as I've been though many os upgrades.
1975[17:27:01] <ratrace> Lope: btw, pretty sure you can ask for
another NIC, if you're not in the mood of changing the whole
server. for ~50€ they'll replace it
1976[17:27:05] <dermoth> is there any doc that describes exactly
what audio daemons/setup I should have in a fresh Buster install?
1977[17:27:23] <dermoth> before I install in in a vm and start
comparing...
1980[17:27:37] <Lope> ratrace, I've had problems with
realcrap in my laptop, and on my workstation's onboard (4 years
ago) and again a month or 2 ago, to the point where I said fuckit
and installed another realcrap NIC that's the SAME CHIPSET, but
different revision, that works.... LOL.
1981[17:27:54] <BCMM> dermoth: don't think so, and also it
will depend on what tasks you selected
1982[17:28:04] <Lope> But I've just tried the new tso
disable workaround in my workstation's init scrips. So
I'll test my onboard again some time :)
1987[17:28:45] <dermoth> I run pulseaudio - one daemon per user -
in the worst cases I have to likk all, unload and reload all audio
modules, restart pulseaudio and reapply the proler sink profile
1988[17:28:48] <Lope> ratrace, surely on a ryzen consumer CPU
they're using onboard shit?
1989[17:29:02] <Lope> ratrace, I asked them for another NIC, they
suggested I do workarounds!
1990[17:29:04] <Lope> hahahaha
1991[17:29:16] <dermoth> I think for my case tey recomment a
single system pulde daemon though
1992[17:29:55] <dermoth> BCMM, so on a fresh install I would have
a user-spawned pulse daemon correct?
2041[17:44:38] <dermoth> well tbh my vdi plugins is very old as
it has to match the old version I have to use at work, until they
upgrade it - I actually use it, but it's barely usable once you
know all the quirks and workaround to limit crashes to a minimum
2049[17:45:41] <dermoth> latest version started crashing
sometimes even without zoom plugin loaded, but before that it was
pretyt sble. There's also occasional keyboard layout annoyances
if you don' use a us layout
2057[17:47:36] <jelly> yanmaani, when were your /etc/hostname and
/etc/hosts files last changed?
2058[17:47:36] <dermoth> slop, fwiw been using it for quite some
time, probably ~5 years... before they renamed it from Citrrix
Reciver... so with previous Debian Stable versions too
2076[17:52:44] <jhutchins> dermoth: cron was the original
scheduling program. cronie and anacron are comparative newcomers.
2077[17:52:50] <yanmaani> jelly: /etc/hostname, a few months ago.
/etc/hosts, a few hours ago, when I changed it to add the new
hostname so sudo would stop whining.
2078[17:53:07] <dermoth> anacron uses cron to start hourly,
dayly, weekly, monthly jobs
2096[17:59:06] <dermoth> 25 6 * * * root test -x
/usr/sbin/anacron || ( cd / && run-parts --report
/etc/cron.daily )
2097[17:59:14] <Voltum> yeah mine too
2098[17:59:33] <dermoth> so anacrons overides run-parts dor daily
and breater intervals apparently
2099[17:59:38] <dermoth> for
2100[17:59:41] <Voltum> i haven't figured if cron and
anacron are complementary or replace one another
2101[18:00:23] <Voltum> from what i remember anacron has an extra
functionality, that if you miss your scheduled cron job, it tries to
run it as soon as possible
2102[18:00:26] <Voltum> instead of skipping it
2103[18:00:28] <yanmaani> jelly: /etc/hostname has the old
hostname (that I want).
2104[18:00:28] <dermoth> I think anacron spreads the jobs rather
than runinng then all at once...and/or ensures jobs are run after
downtime
2121[18:03:46] <dpkg> Use «hostname foo» to set the
hostname, $EDITOR /etc/hostname to set it for the next boot (create
/etc/hostname if it does not exist) and $EDITOR /etc/hosts to set up
local translations for a FQDN. See also 'man 1 hostname',
or ask me about <mailname>.
replaced-url
2122[18:04:01] <yanmaani> jelly: I just rebooted, and it was
after this the issue started. So my hostname was not what it said in
/etc/hostname.
2123[18:04:05] <dermoth> actually nowadays anacron also has a
systemd timer for hourly... interesting... why I still have the
hourly run-parts (regardless of that timer) and why other intervals
haven't been implemented with timers too
2124[18:04:10] <yanmaani> So do I have `hostname hostname` in
some autostart script
2152[18:15:55] *** Quits: kakaka (~koniu@replaced-ip) (Remote host closed the connection)
2153[18:17:04] <dermoth> if that's just it, fine. if oyu
need HDX., there's a bug in the installer where the x86_64
version checks if arch is armhf, and fails because it's not the
right arch - change the check for x86_64
2156[18:18:13] <dermoth> I reported it to Citrix, then they
realized I made it to work so they figured the issue was resolved
and closed it... go figure
2178[18:26:01] <jhutchins> If you run a command with su
you're still the original user.
2179[18:26:08] <Numero-6> jhutchins: yea that back in current
user
2180[18:28:34] <jelly> yanmaani, perhaps it was set somewhere in
systemd settings using hostnamectl ? No idea where those get stored
2181[18:28:40] <Numero-6> that better to use su - root than su -l
?
2182[18:28:46] <Numero-6> just a banal question
2183[18:28:55] <jelly> Numero-6, "su -" is enough
2184[18:29:11] <jelly> and the shortest of all those equivalent
options
2185[18:29:19] <Numero-6> yea because in past i used only
"su"
2186[18:29:30] <jelly> !buster su
2187[18:29:30] <dpkg> In buster, su no longer overrides PATH by
default, requiring that you use "su -" or "su
-l" for login shells (which is not really a new thing at
all...). See
replaced-url
2188[18:29:31] <Numero-6> and now with buster that make probleme
with sbin
2189[18:29:52] <Numero-6> Ok so always su -
2190[18:29:56] <jelly> yes
2191[18:29:59] <Numero-6> i have write in my brain
2192[18:30:04] <Numero-6> :D
2193[18:30:26] <jhutchins> su - <user> gives you a shell as
<user>, and <user>'s environment.
2194[18:30:37] <Numero-6> and that more easy you just do su -
<user> -c "cmd"
2195[18:30:41] <Numero-6> yea
2196[18:31:04] <jhutchins> Numero-6: That may not have the
results you're after.
2198[18:31:49] <jhutchins> Numero-6: Although I use that format
to run commands as root, so it might work for you.
2199[18:32:01] <Numero-6> i just no use sudo, so when you have
lot of cmd requierd root and other no, you do always switch
2200[18:32:22] <Numero-6> and i have always a root terminal open
for check and other
2201[18:32:57] <jhutchins> Yeah, I never got into the habit of
using sudo for single commands. I just become root, do whatever I
need to do, then drop back to my user.
2202[18:33:03] <Numero-6> and when i make a doc about a tutorial
i want write easy just like copy/past
2203[18:33:19] <Numero-6> so that why i ask for run a cmd in
normal user, in session root
2206[18:35:58] <rendar> i have uninstalled a package,
postgresql-11, and i use `dpkg-query -l "$@" >/dev/null
2>&1` this script to check if the package is installed or
not, but it doesn't work because with that command, the package
results installed, since i get this line: `un postgresql-11
<none> <none> (no description available)`, how to fix
this?
2207[18:36:06] <rendar> i think that i could ditch all lines that
starts with 'un' ?
2208[18:36:11] <rendar> but i'm checking the dpkg-query
return value there
2209[18:36:13] <rendar> any clues?
2210[18:37:00] <slop> ty ty on Citrix tips dermoth, jhutchins
will try those
2220[18:42:01] <giaco> hello! When apt upgrade is performed and
there are some dkms modules installed, the system compiles the dkms
modules so that they would be ready to modprobe before next kernel
boot. How does this system works?
2223[18:42:44] <giaco> I'm asking this as I would like to
know how this is working on a real debian, to fix what is
non-working in debian-like raspbian
2224[18:42:59] <jhutchins> giaco: I think what happens is on the
boot after the kernel upgrade, dkms checks the incompatible modules
and re-runs the build.
2225[18:43:21] <jhutchins> !raspian
2226[18:43:21] <dpkg> Raspberry Pi OS (previously called
Raspbian) is a distribution <based on Debian> made
specifically for the <Raspberry Pi>. Raspbian is not Debian
and it is not supported in #debian. Please use #raspbian (or
#raspberrypi) on irc.freenode.net for support.
replaced-url
2227[18:43:24] <giaco> jhutchins: so it happens on next boot, not
before next boot
2228[18:43:54] <jhutchins> As far as I know, and I use it for my
nvidia drivers.
2229[18:44:17] <jhutchins> giaco: There's always the
possibility that raspian does it differently.
2235[18:44:56] <giaco> I'm not asking anything about
raspbian
2236[18:45:07] <rendar> any help?
2237[18:45:14] *** 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!! >>)
2246[18:49:39] <BCMM> giaco: which bit of the system are you
wondering about? if it helps, a "something-dkms" package
usually depends on the dkms framework, which it invokes in postinst
2247[18:49:41] <jhutchins> giaco: I just want to emphasize that
they are different, they change things, and we can't know what
they do and don't change.
2248[18:49:53] <rendar> karlpinc: i know that, hence i added:
18:36 | <rendar> but i'm checking the dpkg-query return
value there
2249[18:50:11] <rendar> how can i check if postgresq-11 package
is installed or not, by checking exit code of some command?!
2250[18:50:25] <BCMM> jhutchins: i'm sorry, but giaco made
it 100% clear that they were asking about how debian works, and are
prepared to work out for themselves how raspbian differs
2251[18:50:27] <greycat> rendar: dpkg -l should suffice
2258[18:51:24] <karlpinc> rendar: You can't use the exit
status of dpkg-query to tell anything other than whether the query
you gave to dpkg-query was valid or otherwise failed.
2259[18:51:34] <greycat> karlpinc: that sucks.
2260[18:52:02] <greycat> karlpinc: ... and appears to be untrue.
2262[18:52:17] <greycat> "dpkg -l sdhkfjsd" gives me
exit status 1 here
2263[18:52:31] <rendar> greycat: the point is dpkg -l returns 0.
replaced-url
2264[18:52:37] <giaco> BCMM: I know what dkms do, as I have
manually created some dkms modules that would "dkms
autoinstall" nicely. I also see that debian has some system to
kick-in dkms recompilation when apt update installs a new binary
kernel and kernel-headers. So what I am asking is how this automatic
"dkms autoinstall" system works when debian is in the
condition of running a kernel but ready to boot in next one
2265[18:52:37] <karlpinc> greycat: The man page says: 1 The
requested query failed either fully or partially, and 2 Fatal or
unrecoverable error due to invalid command-line usage.
2266[18:52:39] <giaco> on next restart.
2267[18:52:41] <greycat> rendar: it does not do that for me.
2268[18:52:53] <rendar> greycat: look at that printing, it says
(un) flag to postgresql-11
2299[18:56:36] <rendar> greycat: yeah i saw that, but do you
refer to Desired=XYZ or Status=ABC? I can't see if (un) refers
to Desired=Unknown or Status=Unpacked
2300[18:56:37] <karlpinc> greycat: It may be that the exit status
depends on whether or not the supplied query identifies a package
that exists....
2314[19:00:11] <BCMM> particularly as it looks like
/etc/kernel/header_postinst.d/dkms would trigger whenever a new
kernel headers package is installed
2315[19:00:15] <giaco> I was expecting something on
kernel-headers upgrade
2316[19:00:18] <BCMM> (hook installed by the dkms package)
2317[19:00:33] <giaco> nice spot BCMM, thanks
2318[19:00:53] *** Quits: fflori (~fflori@replaced-ip) (Remote host closed the connection)
2319[19:02:29] <greycat> dkms doesn't build things at boot
time, no. It builds things at package installation time.
2336[19:07:29] *** Quits: gravitos (uid355353@replaced-ip) (Quit: Connection closed for inactivity)
2337[19:07:34] <BCMM> so looks like anybody could put stuff in
/etc/kernel/header_postinst.d to have it trigger on kernel header
installation, but currently only dkms actually does
2338[19:07:37] <giaco> next unrelated question. Is there a file,
in debian, containing the info about the kernel that will boot next,
while still running old one?
2339[19:07:56] <BCMM> giaco: as in, telling you which kernel
would be booted?
2340[19:08:01] *** debhelper sets mode: +l 1236
2341[19:08:14] <giaco> BCMM: yeah, that short version :D
2342[19:08:33] <BCMM> not sure, but i don't think as it
would presumably depend on the bootloader you are using
2343[19:08:40] <giaco> k
2344[19:08:50] <BCMM> but i don't think dkms tries to detect
that
2345[19:09:25] <quadrathoch2> dkms builds for every kernel
available
2346[19:09:25] <BCMM> i think it just does it for every kernel
that's installed
2381[19:27:04] <sponix2ipfw> I rebuilt wireguard-dkms and now
when I rmmod wireguard and modprobe wireguard only the wireguard
module itself and the ipv6 stuff is loading, NOT ipv4 stuff that I
need for it to function
2382[19:27:19] <sponix2ipfw> Anyone have a quick solution for
this that doesn't involve a reboot ?
2383[19:27:52] <zutat> how is the kernel image package called? i
can find linux-config, linux-doc, linux-headers and so forth, but
can't see an image in the list
2389[19:31:47] <zutat> hmm. for some reason, the search on
package.debian.org wouldn't list any linux-image- packages for
buster-backports, but just browsing the packages lists several
packages named like that
2421[19:48:06] <dpkg> The version number of a package has a
prepended number called the "epoch". It is only added when
the system for upstream version numbers changes. Example: in sarge,
X was version 6.8 but in etch it was 1.1 (xfree86->xorg). But 1
< 6, so we add an epoch "2:" to signify that everything
with 2: is newer (if there is no : the epoch is assumed to be
"0"). See section 5.6.12 of <policy> or ask me about
<compare versions> <debian revision>.
2448[20:07:24] <EdePopede> just ran a test session on the other
PC after the first one seems to run like forever, killed the other
processes (tracker-miner-fs, tracker-miner-apps, applet.py, agent)
manually, waited for checked after one and a half hours now, and the
systemd tree is gone.
2449[20:07:48] <sponix2ipfw> I changed routers and my network
also went from 192.168.1.* to 192.168.50.* -- for some reason my
default gw still comes in as 192.168.1.1 on every boot, even though
it should just be pulling a dhcp lease
2450[20:07:59] <greycat> EdePopede: sounds like a question for
#systemd
2451[20:08:02] <sponix2ipfw> anyone know where I might have
manually configured this or something ?
2452[20:08:24] <EdePopede> greycat: won't they send me back
to #debian because debian?
2455[20:08:48] <queip> (cross git) any editor that shows 2 files
line by line and allows to select easily to take each line from file
A or B? kind of like meld, but not just working on diff and each
line is separate?
2458[20:09:47] <EdePopede> i thik i'll give it a 3rd try
first and not kill anyhing. if the processes end anyway than maybe
this is how it *should* work normally.
2493[20:28:05] <JordiGH> A few years ago I added some images to a
pdf. I can't remember what program I used to do that. I just
need to add my signature to a pdf form. Any clue what I used?
2494[20:28:38] * greycat spins the wheel of guessing ... you used ...
adobe acrobat!
2495[20:29:28] <JordiGH> No, no non-free software.
2496[20:29:48] <JordiGH> I thought I converted to flpsed but that
doesn't allow adding images.
2497[20:29:49] *** Quits: HeXiLeD (~grumpy@replaced-ip) (Remote host closed the connection)
2515[20:44:21] <zutat> i installed a linux-image from backports.
it displays an error about a missing firmware file on boot and
freezes there. does it need some additional packages?
2541[20:59:46] <sponix> slop: that is the thing, netstat -rn
shows 192.168.1.1 as the default route/gw after a restart. This was
the network for my OLD router, the new one is 192.168.50.1 --
Everything was dhcp, and it was still getting the old gateway
2591[21:40:42] <Kocane> Anyone who can help me wih some
xorg/xinit/xrandr related? I got an Intel NUC running Kodi that
starts with /usr/bin/xinit thru a systemd file. But it seems to me
it doesnt output full range RGB and I'm having some trouble
figuring out how I got about configuring it correctly. I've
seen something With Xrandr it seems possible to output it with
"--
2592[21:40:42] <Kocane> set "Broadcast RGB"
"Full"" but I don't get how to configure that if
I'm using xinit.
2593[21:42:57] <zutat> wouldn't a xorg configuration file
help?
2595[21:43:35] <sponix> slop: still working on it, somehow my
eth1 is getting both the new and the old networks (192.168.1.144
192.168.1.1) (192.168.50.144 192.168.50.1) -- and they conflict
2596[21:43:43] <EdePopede> is ps from debian not able to show
processes of a user as tree? or is just its manpage total crap?
2597[21:43:55] <greycat> The "f" option does the tree
thing.
2598[21:43:59] <Kocane> zutat: does that mean in
/usr/share/X11/xorg.conf.d?
2599[21:44:09] <greycat> ps f -u someuser
2600[21:44:15] <EdePopede> i always use faux and i have no idea
how to not show everything
2601[21:44:15] <sponix> I seem to have hardcoded the old network
in the OS somewhere. But it isn't in dhcpclient.conf or
/etc/network/interfaces where I normally look
2602[21:44:16] <greycat> note the lack of a - before f
2607[21:46:03] *** Quits: auk (auk@replaced-ip) (Ping timeout: 256 seconds)
2608[21:46:18] <zutat> Kocane: yes. you should be able to use the
options in a config file there. i would guess that they go to a
"monitor" section of such a file
2609[21:46:22] <EdePopede> great. with faux it displays the
systemd group, then the others, with f -u it reverses the order
2619[21:51:30] *** Quits: mthe878 (~mthe@replaced-ip) (Quit: Lost terminal)
2620[21:52:27] <Kocane> zutat Hmm, thats where I came across
"Intel doesn't have an option to set colorspace in their
xorg.conf yet. If you want to set the colorspace then you will have
to create an autostart.sh file and use the following"
2622[21:55:14] <zutat> Kocane: i don't have access to an X
machine, but there should be a way to put the same options in a
config file as what you use with xrandr
2650[22:11:20] <wr> how would i install a Ralink RT2870 Linksys
WUSB600N v1 Dual-Band Wireless-N Network Adapter on debian 10.7?
2651[22:11:35] <sney> !ralink firmware
2652[22:11:35] <dpkg> Firmware from userspace is required by the
<rt61pci>, <rt73usb>, <rt2860sta>,
<rt2870sta>, <rt2800pci> and <rt2800usb> drivers.
Ask me about <non-free sources>, then install the
firmware-ralink package to provide.
2653[22:11:40] <CommunistWolf> ralink are weird - or were
2682[22:36:47] <sponix2ipfw> slop: when I get home, I am going to
go with a different approach, I'm just going to change the IP
scheme on the router to match what my Linux rig thinks it must be
LOL
2683[22:37:22] <sponix2ipfw> slop: appreciate your help though.
at this point I think doing it on the router is going to require
less effort
2701[22:44:27] <sney> ok? if lsmod shows the rt2800usb driver, it
doesn't matter how many times it's in the list, that just
means it's loaded. and the device is "installed".
2702[22:44:35] <sney> and you can now configure it
2707[22:49:10] <wr> sney, on my NM it doesn't show up, so i
would have to use terminal?
2708[22:49:52] <sney> if it's not visible in n-m, check if
it already has an entry in /etc/network/interfaces, as n-m will
ignore any interface specified in that file
2728[23:02:00] <wr> sney, seems so, i am just checking which is
which cause i have three cards and or the naming confused or it
wasn't on list but now is
2775[23:18:53] <dpkg> We're sorry your distro's channel
isn't being helpful, but that doesn't make it appropriate
to use #debian for non-Debian questions. Please go back to your
channel and wait patiently for better help, or install Debian and
party with us.
2818[23:39:01] <pfred1> jhutchins There have been releases where
they had serious problems too
2819[23:40:35] <jhutchins> pfred1: Not too many with Debian. I
got caught by a few bad ones on Mandrake. (like dovecot
wouldn't start on my mailserver because the config file format
was obsolete, and obviously nobody had tested it).