35[00:33:26] <[E]sc> my internet connection periodically
disconnects. i'm not sure why, and i have to manually restart
NetworkingManager.service. Is there a fix for this?
58[00:45:39] *** Quits: ledeni (~ledeni@replaced-ip) (Remote host closed the connection)
59[00:47:14] <[E]sc> sney: dlink systems.
60[00:47:24] <somiaj> sappheiros: Might want to start with what
install medium did you use? How did you copy it to the usb drive?
And what was the error/failure point, and we can help.
67[00:50:43] <sappheiros> somiaj: USB installer, couldn't
get an HDMI signal to my monitor from my graphics card; CPU and DRAM
lights red on the motherboard, wondering if i broke the CPU without
realizing it by pushing hard to attach the heatsink latches
68[00:51:19] <somiaj> sappheiros: okay, back to the questions I
actually asked. Whick install image? How did you copy it to the usb?
69[00:51:40] <somiaj> sappheiros: Also, does your machine even
POST?
72[00:52:06] <[E]sc> tor browser connects though. chrome comes
and goes. youtube on chrome fails sometimes, but connects with tor,
sometimes youtube-dl doesn't connect.
74[00:52:27] <sappheiros> no POST beeps, and i used the version
10 USB image from the debian site
75[00:53:12] <sney> [E]sc: is the wifi access point in your
control? beacon issues can sometimes be solved by disabling 802.11b
at that end (so you only support 802.11g and up)
76[00:54:04] <[E]sc> sney: let me see try that.
77[00:54:22] <sney> it's also *possible* that the newer
kernel in buster-backports might have some iwlwifi improvements, but
I'd look at the router first
79[00:55:28] <somiaj> sappheiros: yea sounds like your problem
is hardware first, until you can get your machine to POST and get
into its firmware/bios, I wouldn't worry about installing
anything.
80[00:55:58] <sappheiros> right, that's why i messaged
inbe4 in #offtopic since it was about hardware
81[00:56:11] <sappheiros> *responded to inbe4 (not /msg)
99[01:13:02] <segamain> Does anybody here know a good hardware
channel (freenode ##hardware doesn't count it's not
helpful) This is not cross posting I've already asked on oftc
and got no helpful answers
117[01:20:28] <segamain> abortman: how can I ask #freenode what
do you mean ?
118[01:20:39] <segamain> ahhh it's channel right ?
119[01:21:17] <abrotman> Is <tab> really that hard?
120[01:22:52] <segamain> soimaj: just curious here. How deep do
these rules go ? if a channel doesn't respond or can't
help me am I not allowed to ask on other channels ?
130[01:26:46] <abrotman> And you were told already it seems?
131[01:27:17] <somiaj> segamain: It is more politness, no strict
rule, but usually give a channell 30mins or so to respond.
132[01:27:43] <somiaj> !cross-post
133[01:27:44] <dpkg> Posting the same question in several places
at the same time (IRC channels, news groups, mailing lists, forums)
is impolite; your time is NOT more valuable than everyone
else's. Your question might be answered elsewhere, meanwhile we
are wasting our time doing research for a problem you've
already solved. Cross-posting can also make you look like a spammer
and get you k:lined. See also <multiple ask> <hurry>.
267[03:26:38] <ajayahmed> hi i added packages.sury.org as an apt
source so i can install multiple versions of php but on a new debian
10 server every time i try to install or update something like
php7.4-opcache, it also tries to install php8.0-common,
php8.0-opcache. `apt-cache depends php7.4-opcache` shows no
dependency on anything php 8 related so i have no idea why this is
happening. can someone help please?
268[03:27:09] <sney> ,v php8.0-common
269[03:27:10] <judd> Package: php8.0-common on amd64 -- sid:
8.0.5-1
270[03:27:40] <sney> ajayahmed: do you have sid sources for some
reason? if not, sounds like ondrej messed something up on his repo
282[03:35:36] <somiaj> the only thing I can think of is maybe
let it install those extra packages, then run 'aptitude why
php8.0-common' that should tell you what pulled it in.
283[03:35:49] <mutante> all have some priority, maybe try to
change priority so that is sury is preferred?
284[03:36:18] <somiaj> ajayahmed: if you add
--no-install-recommends do you also get the same beavhior.
285[03:36:32] <ajayahmed> somiaj yep
286[03:36:34] <somiaj> ajayahmed: apt policy packagename should
tell you what repo a package would come from if mutante suggestion
is correct.
287[03:38:21] *** Quits: Filo (~filohuhum@replaced-ip) (Ping timeout: 240 seconds)
304[04:14:15] <Edu4rdSHL> Hello, I'm trying to setup
networking in Debian, it's a Nspawn container, I have created
the corresponding .network file to work with systemd-networkd but it
isn't working. The service starts correctly but seems that it
isn't reading the .network file because the interface is still
"down" and "unmanaged" by systemd.
305[04:15:08] <sney> is the interface listed in
/etc/network/interfaces? some other networking services will ignore
any interfaces in that file
306[04:16:25] <Edu4rdSHL> sney, I dont have the /etc/network
folder, it's a minimal Debian installation
307[04:16:59] <Edu4rdSHL> I dont know if debian put some
"default networking services" even in minimal
installations
308[04:17:06] <Edu4rdSHL> It's the .network file
replaced-url
310[04:17:36] <sney> /etc/network/interfaces *is* the default
networking service configuration in a minimal installation. so if
you don't have that file, it's not the problem.
312[04:19:02] <somiaj> do you have do anything special to turn
systemd-networkd on (it is disabled by default)
313[04:19:11] <Edu4rdSHL> sney, so, if I create a own container
with `debootstrap --components=main,universe buster container-name
repository-url`, it should create /etc/network ?
314[04:19:23] <Edu4rdSHL> somiaj, just enabling the service.
315[04:20:04] <sney> 'universe' is an ubuntu thing,
but otherwise, normally any debian installation of any size will
have ifupdown installed, and /etc/network/interfaces
316[04:21:27] <sney> the man page for systemd.network(5)
recommends putting Name=* in the [Match] section to avoid a warning,
maybe add that and see if it works (perhaps it's being enforced
more strictly)
320[04:23:20] <PMT> Hi all, I've got a Debian bullseye VM,
just recently I started having a strange behavior where if I run
"sudo su -" and then run "cp /foo /bar" inside
the session, the cp succeeds unremarkably, but when I type exit,
sudo segfaults.
321[04:23:36] <sney> !sudo su
322[04:23:36] <dpkg> Typing "sudo su" is like typing
"cat file | cat". If you want a non-login shell as root,
just use "sudo -s". If you want a login shell, "sudo
-i". There's no need to run two different
superuser-elevation programs chained back to back. See also
<buster su>
323[04:24:17] <PMT> Sure, I'm doing that wrong. My question
stands.
324[04:24:37] <sney> does the behavior persist if you get a
login shell as the factoid suggests, with sudo -i?
325[04:24:59] <PMT> The behavior persists if I just do
"sudo cp /foo /bar", so I'm going to guess
"yes"
331[04:27:49] <sney> can you elaborate on the source and dest
files, do they have to be at the fs root or is that just a generic
example? also, I recognize you from #zfs, is the example fs zfs?
332[04:28:09] <sney> Edu4rdSHL: it's more likely to be
freedesktop enforcing the Name field, and the man page hasn't
caught up yet.
333[04:28:57] <Edu4rdSHL> sney, I see. Thank you very much <3
334[04:29:00] <PMT> The FS is ext4, but I am playing with
ZFS-related files - the destination file is a ZFS library under
/lib/x86_64-linux-gnu/, the source can be literally anything.
339[04:33:30] <sney> indeed, there's nothing zfs-ish in ldd
$(which sudo) that I can see. you said this started recently, can
you match a package in /var/log/apt/history.log to the approximate
date it started happening?
345[04:37:08] <PMT> I don't see anything obvious in
dpkg.log - the package installs in the last hour were me installing
gdb and libc6-dbg to debug this (so, presumably, it started before),
and before that, just "python-is-python3" an hour prior.
346[04:37:31] *** Quits: catman370 (~catman@replaced-ip) (Quit: See you later..)
348[04:37:52] <somiaj> wonder if trying to install the newer
version from experiemntial to see if the behavior is still there
could at least help debug it
352[04:38:28] <PMT> I can try that, this is just a testbed VM,
so I don't really care if the world burns down in it. :)
353[04:38:29] <somiaj> PMT: so this only happens with very
specific destination files and cp?
354[04:40:41] <PMT> somiaj: if I do not run cp [somewhere] [that
library], it does not segfault on exit. cp [somewhere]
/lib/x86_64-linux-gnu/libdumbme.so does not cause it, whether that
so path existed already when I ran sudo or not.
355[04:42:09] <PMT> (It also works for overwriting other ZFS
libraries, even with themselves - I have not yet tried overwriting
other libraries.)
363[04:48:16] <PMT> sudo from experimental also segfaults
364[04:49:12] <somiaj> you could also use snapshot.debian.org to
see if older sudo's had this behavior, but wonder if there is
some underlying feature happening. Are the libaries you overwriting
currently loaded by some running program?
365[04:49:29] <somiaj> though still seems odd, no clue what
oculd be going on
366[04:49:46] <sney> weird segfaults in an oddly specific
location can also be memory problems
367[04:50:03] <sney> just that one part of that one dram
395[05:04:36] <PMT> so, i should perhaps be more precise
396[05:04:41] <PMT> the following statements are true
397[05:05:12] <PMT> * a login shell as my user does not report
crashing on exit, but i'm also not sure i'd see the
message before it displayed the login prompt again
406[05:09:53] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
407[05:11:52] <PMT> i'm reasonably confident this is memory
corruption, since i just tried to save a core dump by running sudo
gdb su and causing su to segfault, but it saved an invalid file and
crashed.
408[05:12:33] <somiaj> could it be just something with the vm,
or maybe the host? Maybe run memtest86 overnight
409[05:13:07] <PMT> not impossible, but none of the other VMs
have had any unusual behavior yet.
411[05:14:58] <somiaj> well running multiple passes of memetest
if you can afford the downtime could at least answer the question,
or maybe it is just something strange about that particular vm.
465[06:15:28] <ASDX> is there an apt/dpkg command to check if
files for a particular package have been changed since being
installed/upgraded ("rpm -qV" equivalent)?
478[06:43:28] <somiaj> ASDX:
/var/lib/dpkg/info/packagename.md5sums will contain the md5sums of
all files installed by that package. So you could use that to
manually verify, debsums is just the tool to automate this.
479[06:44:50] <somiaj> ahh looks like dpkg as a --verify option
that uses those md5sums
481[06:47:04] <somiaj> though seems debsums has some features
not included with --verify
482[06:58:59] <ASDX> somiaj: thanks. shame "dpkg -V"
needs to be run against every package. takes a while on my system to
run a result. i had an 'apt upgrade' complain that one or
more files in /etc/pam.d/common-{auth,account,password,session} have
been locally modified" so was trying to confirm which file(s)
were changed and why. but "dpkg -V | grep libpam-runtime"
doesn't return anything. i see there's a
/var/lib/dpkg/info/libpam-runtime.md5sums
483[06:58:59] <ASDX> file but these common-* files have
different file locations (eg "<md5sum>
usr/share/pam/common-auth" when I was expecting
"<md5sum> /etc/pam.d/common-auth" but looks like
there's a /var/lib/dpkg/info/libpam-runtime.postinst file that
actually move the files into their end location but the md5sums are
not considered in their end location with "dpkg -V" as a
result.
484[07:00:01] *** Quits: riff-IRC (~riff2@replaced-ip) (Remote host closed the connection)
486[07:02:41] <somiaj> ASDX: you can state what package to test
if you don't want to test them all
487[07:03:12] <somiaj> many files in /etc are conf files
controled by debconf, those you will have to deal with seperatly
488[07:04:32] <somiaj> yea, the message you are getting is there
is a change to the configuration file, and it is asking if you want
the new configuration file or your old one. I have had cases where I
get this even on files I haven't edited, but sometimes scripts
will edit them after debconf so they look like they have changed to
debconf.
587[09:40:14] <afidegnum> hi, i have added an additional ip to
the network interface but i can't still ping it. what am i not
doing right ?
replaced-url
591[09:45:01] <Tigryss> hi, I have some problem with ipv6/4. I
try ro use my android with usb tethering and it's working,
debian can only reach network with ipv6. Example "ping
8.8.8.8" response network unreachable but "ping
2001:4860:4860::8888" working fine
592[09:45:56] <Tigryss> is there anyone who had this issue
before?
653[10:53:12] <alkisg> Tigryss: Also check if you do have an
ipv4 with `ip a`, and if your phone does give a DHCPOFFER, with
`sudo /usr/lib/klibc/bin/ipconfig -n usb0`
668[11:23:16] <shtrb> [trying to do live backup of akonadi]
Should mariahdb file storage be able to correct and start from
backed up files in case files had been copied while mysqld was
running (but no write operations had been running at the point in
time) ?
669[11:24:04] <shtrb> I'm even wondering if I could just do
a filesystem snapshot and try to load from it
670[11:24:46] <jelly> mysql server is supposed to survive the
latter, yes, it's equivalent to a recovery after a power loss
671[11:24:58] <shtrb> thanks
672[11:25:31] <jelly> but if you want proper backups, use
mariadb-backup
701[11:45:57] <dpkg> You have to especially tell the packaging
system to reinstall config files because when they are gone, it is
assumed that you want them to stay deleted. "aptitude -o
DPkg::Options::='--force-confmiss' reinstall
$packagename" will restore them (man dpkg for details). If the
package uses <ucf> for config file management, ask me about
<ucf confmiss>.
702[11:47:02] *** Quits: v01d4lph4 (~v01d4lph4@replaced-ip) (Remote host closed the connection)
712[11:55:53] <jelly> hegemoOn, if you can purge the package,
install it again, and then the file appears, that would mean
something like confmiss is happening
713[11:57:07] <hegemoOn> will test
714[11:57:09] *** Quits: Grldfrdom (uid391113@replaced-ip) (Quit: Connection closed for inactivity)
715[11:57:36] <jelly> that is, you or a custom tool is deleting
the file
798[13:06:53] <PamiPetteri> I am recompiling the kernel on
Debian: "make -j`proc` deb-pkg". How come when I compiled
with bindeb-pkg I did not get the PREEMPT string in my uname -a? I
was expecting it. Also, my graphics card configuration was
complaining something about mmio nouveau or nvidia, when I booted
up. Should I expect that I get a better result now that I recompile
with deb-pkg? I had also changed .config "on the fly"
799[13:06:59] <PamiPetteri> while compiling. Now I have a ready
.config from the start. Is this compile going to have better
results, regarding the "PREEMPT" string and the NVidia
graphics configuration?
800[13:08:12] <PamiPetteri> I do not really know what I am
doing. This is my hobby.
801[13:12:56] <petn-randall> PamiPetteri: What is CONFIG_PREEMPT
set to in your .config?
810[13:23:00] <Tigryss> hi, I have some problem with ipv6/4. I
try ro use my android with usb tethering and it's working,
debian can only reach network with ipv6. Example "ping
8.8.8.8" response network unreachable but "ping
2001:4860:4860::8888" working fine
814[13:25:11] <jelly> Tigryss, also [08:53:12] <alkisg>
Tigryss: Also check if you do have an ipv4 with `ip a`, and if your
phone does give a DHCPOFFER, with `sudo /usr/lib/klibc/bin/ipconfig
-n usb0`
815[13:25:34] <jelly> oops, forgot to dehighlight alk'isg
816[13:26:16] <Tigryss> ping 64:ff9b::8.8.8.8" works? No
817[13:27:43] <jelly> Mr. Dagger knows more about ipv6 and NAT64
than me, I'm just pasting the stuff you may have missed last
time you were here
820[13:30:13] <Tigryss> sudo /usr/lib/klibc/bin/ipconfig -n
usb0 no device to configure
821[13:30:40] <Tigryss> jelly thanks I read back :D
822[13:31:29] <petn-randall> PamiPetteri: If you set that, and
then compiled the kernel, it should be preempt. Not sure when that
shows up in `uname -a`.
823[13:32:54] <PamiPetteri> petn-randall: it might be on without
showing it in uname -a. but I am trying to recompile - mostly
because my graphics card configuration was complaining upon boot.
824[13:34:58] *** BrianG61UK__ is now known as BrianG61UK_
826[13:36:13] <alkisg> (02:30:13 PM) Tigryss: sudo
/usr/lib/klibc/bin/ipconfig -n usb0 no device to
configure => what is the device name? `ip a` will tell you. Also,
what's the output of `ip r`, is there the correct route
registered there?
845[13:44:56] <alkisg> Tigryss: that's the usb interface
name? Oh well, ok, check with that one. And also check the route,
`ip r`, and if you can ping the dhcp server == android
854[13:48:46] <alkisg> Tigryss: try again the `ipconfig`
command, but now without the -n
855[13:49:01] *** Quits: ecsim (~Thunderbi@replaced-ip) (Remote host closed the connection)
856[13:49:08] <alkisg> This should register the correct route so
that you can use ping etc. But you need to configure it properly,
with network-manager or dhclient etc
858[13:50:04] <alkisg> This is the output of `ip r` with
bullseye+network manager: 192.168.42.0/24 dev usb0 proto kernel
scope link src 192.168.42.118 metric 101
859[13:50:33] <alkisg> Without an IP and a route for
192.168.42.x, ipv4 won't work
860[13:51:57] <Tigryss> hmm can be that the problem is that the
modem manager is removed?
861[13:52:17] <alkisg> I think you should answer that one first:
which program are you using to get an IP
862[13:52:31] <alkisg> Are you using network-manager, or
/etc/network/interfaces (ifupdown), or wicd, or dhclient...
863[13:52:39] <Tigryss> network manager
864[13:53:07] <alkisg> Can you paste the output of: nmcli
890[14:07:26] <petn-randall> PamiPetteri: If you don't have
any issues I'd say those are purely cosmetic. What graphics
card(s) do you have in this machine? Is it a laptop?
891[14:07:40] *** Quits: ecsim (~Thunderbi@replaced-ip) (Remote host closed the connection)
892[14:08:24] <PamiPetteri> petn-randall: NVidia GeForce. this
is a tower PC from 2013, K5130 ASUS PC. I will ask screenfetch or
neofetch about my graphics card.
899[14:11:43] <PamiPetteri> I am now using deb-pkg instead of
bindeb-pkg in case this works better. where exactly is the source
code and where is the upload and where is it uploaded?
906[14:15:25] <petn-randall> PamiPetteri: "bindeb-pkg"
will create a binary package to install. "deb-pkg" will
also create a source package, which you could use to build the
binary packages from.
908[14:16:43] <petn-randall> None of those will actually upload
anything. "Preparing an upload" is a term used for
preparing the source tarball for upload, so that Debian build
servers can build the binary packages.
910[14:17:15] <euri10[m]> greetings, I upgraded my buster
desktop to bullseye yesterday without any issue except a minor
inconvenince with bluetooth I reported. The only thing that does not
work, but it's not Debian's fault, is the
`nvidia-container-runtime` for docker : some research reveals that
they dont support cgroups v2 yet, which is now the default. There
seem to be a fix that consists in forcing systemd to use cgroup v1
(in the grub cmdline) but I wanted to
911[14:17:15] <euri10[m]> know before I test that option what
are the pros/cons of not using the cgroups v2 default ?
998[16:29:17] <PamiPetteri> my graphics crashed when my computer
was hibernating under GNOME in Xwayland. I did a "killall -1
Xwayland" as root from the console and when I logged back in
GNOME, the computer refused to connect to my IRC shell machine, but
gave proper "ping" replies from the machine. What is this?
When I rebooted the computer there was a 1 min 30 sec stop job from
cron running. Was there some sort of upgrade going
999[16:29:23] <PamiPetteri> on, or what else could it have been?
Am I going to have a better experience using MATE with X than GNOME
with Xwayland?
1002[16:31:36] <PamiPetteri> I suppose the graphics problem had
to do with this new kernel, that is giving graphics errors upon
boot. Cannot read some addresses and cannot write some addresses. It
might have to do with the intel-microcode install I did with apt. Or
just a kernel issue with the graphics driver.
1003[16:31:41] <PamiPetteri> jelly: ok.
1004[16:31:42] <jelly> you can keep gnome just switch back to
gnome on X, I think
1005[16:32:25] <jelly> that ought to be one of the session
options on the login screen (display manager)
1017[16:47:30] <daft_dutch> Hi im bought a vpn account and it
uses openvpn openvpn <script> works but i like to run it at
the background. I tried to set default/openvpn as good as i tried
but im stuck
1022[16:51:54] <mirko> after i realised yesterday that my firefox
v88.0 package fetched from sid and installed on bullseye is
considered a frankedebian, i now built the firefox v88.0 package on
bullseye and can reproduce the crash. now i also have debug symbols
and a meaningful backtrace:
replaced-url
1025[16:57:17] <daft_dutch> openvpn <script> runs. but if i
run it as service it says May 5 16:55:06 mole ovpn-login[11271]:
Options error: Unrecognized option or missing or extra parameter(s)
in /etc/openvpn/login.conf:1:
1029[17:01:27] <somiaj> mirko: this channel is for stable
support, so we assume that users are running stable. Installing sid
packages in testing is a common practice in #debian-next on
irc.oftc.net
1030[17:01:49] <somiaj> mirko: though with bullseye about ready
to be releaed, don't depend on sid packages, espically ones
that will never migrate to stable if you stabalize with bullseye
1031[17:03:03] <somiaj> mirko: if you grab older firefox packages
from snapshot.debian.org, do they also crash or is this only with
firefox v88? You could file a bug report, might get a bit more
useful response.
1032[17:04:56] <mirko> somiaj: the crash was introduced with v88
- downgrading back to v87 everything is fine again
1036[17:06:15] <mirko> i checked already and it doesn't seem
common, hence i suspected my setup and figured maybe the backtrace
might ring a bell what's wrong here
1045[17:09:54] <somiaj> and you ahve tried the binary compiled by
mozilla?
1046[17:10:07] <mirko> somiaj: no
1047[17:10:22] <somiaj> I would try that, see if it is a debian
package issue vs firefox
1048[17:10:39] <somiaj> if the binary from mozilla has the same
problem, I'd report a bug upstream as it may get more response
than reporting it to debian.
1049[17:11:33] <mirko> understood, currently suspecting
fGetString() so, maybe i can get to the bottom myself
1060[17:24:31] <Xalys> Does the debian-security repo use phased
updates / gradual rollout? A security update for bind9-host was
published to the repo on 29 Apr 2021, but unattended upgrades only
updated it on 2 May 2021
1061[17:25:08] <petn-randall> Xalys: No, it doesn't haven
staggered rollouts.
1062[17:25:18] *** Quits: sinaowolabi (~Sina@replaced-ip) (Remote host closed the connection)
1063[17:25:31] <petn-randall> Xalys: Did it install earlier on
another machine?
1064[17:26:47] <Xalys> petn-randall: No, I have three machines
that updated the package all in the same night. They're all
configured to run unattended-upgrades every 4h and install updates
from Debian-Security. But when I look in bind9-host's
changelog, I see a release date of 29 Apr 2021. Perhaps the
changelog was finalised before publishing the package? Is that
common?
1072[17:33:12] <petn-randall> Xalys: If you look at the
"news" section, you'll see that it was uploaded on
may 1, so I'm assuming the changelog entry was prepared
earlier.
1100[18:14:27] <junialter> I'm trying debian Bullseye on my
raspberry pi 2 in order to make it a stratum 1 NTP
1101[18:14:55] <junialter> What I'm missing is how I can
change the gpio pin for the pps signal which I did on raspbian with
dtoverlay in /boot/config.txt
1102[18:15:21] *** Quits: wintersky (uid453465@replaced-ip) (Quit: Connection closed for inactivity)
1138[18:56:05] <PamiPetteri> I should remember not to fix what is
not broken. I will reboot now and see how my current setup reacts to
firmware-misc-nonfree and firmware-realtek. I am not too excited
about using flashrom, because that could end up in unwanted results.
I do not want to break my computer's hardware.
1139[18:56:21] *** Quits: Filohuhum (~filohuhum@replaced-ip) (Remote host closed the connection)
1192[19:46:56] <kevindd992002> I'm running Debian Buster and
while I was upgrading qemu-extra-blocks, it brought up the usual
sshd_config reconfiguration screen (in my SSH terminal) saying that
a new version of the sshd_config file is to be installed and either
I proceed with it or keep the local file that's currently being
used. I accidentally pressed Esc and it just continued
1193[19:46:56] <kevindd992002> with the installation. Immediately
after pressing Esc, here's what I saw:
1194[19:46:57] <kevindd992002> Setting up openssh-server
(1:8.4p1-2~bpo10+1) ...
1195[19:46:57] <kevindd992002> Installing new version of config
file /etc/init.d/ssh ...
1196[19:46:58] <kevindd992002> Installing new version of config
file /etc/ssh/moduli ...
1197[19:46:58] *** kevindd992002 was kicked by debhelper (flood)
1200[19:47:31] <kevindd992002> I'm running Debian Buster and
while I was upgrading qemu-extra-blocks
1201[19:47:44] <kevindd992002> it brought up the usual
sshd_config reconfiguration screen (in my SSH terminal) saying that
a new version of the sshd_config file is to be installed and either
I proceed with it or keep the local file that's currently being
used. I accidentally pressed Esc and it just continued with the
installation.
1202[19:48:00] <kevindd992002> I restarted the server and tried
ssh'ing again. What I noticed is that after I input my user
password, it takes more than a minute for it to get in and when I do
a systemctl status I see a degraded status with this failed units:
1210[19:52:03] <greycat> So the problem is "ssh-ing in takes
over 1 minute, and log shows user@1000.service loaded failed".
I'd check the ownership of /run/user/1000/ because I've
seen that get screwed up before.
1211[19:53:08] <cosimone> hello, this is probably a pedantic
observation, but i noticed that ocaml-nox in the standard buster
repository, *suggests* ocaml-doc, which is in the non-free
repository. is it correct for a package in the standard repo to
suggest packages in the non-free repo?
1226[19:56:26] <greycat> OK, then I don't know. I might just
reboot. systemd and dbus are still kinda mysteries to me.
1227[19:56:30] <cosimone> it seems surprising to me that packages
in the standard repository might suggest packages in the non-free
repo, but if this is allowed, then i have no complaints
1228[19:56:52] <somiaj> kevindd992002: unsure if the
configuration issue during the install is affecting this, but you
should have either a backup of your old config or a copy of the
original dpkg config in /etc/ssh/
1229[19:57:22] <greycat> ssh configuration should not be causing
systemd --user services to fail
1230[19:57:29] <kevindd992002> I already rebooted. I noticed the
issue after rebooting after the upgrade of qemu-extra-blocks
1231[19:57:30] <jhutchins> cosimone: It's OK for a package
in main to recommend or suggest a non-free package, but it would be
wrong to depend on non-free.
1232[19:58:08] <somiaj> jhutchins: are you sure about that? I
don't think main can have any depends in contrib/non-free.
1233[19:58:14] <jhutchins> What's userservices trying to do?
1247[20:00:47] <somiaj> cosimone: often time docs have unmutable
parts which puts them in non-free, though I must have not noticed
that suggests can be outside of main.
1274[20:09:35] <kevindd992002> what do you know, installing
scdaemon fixed it!
1275[20:09:46] <kevindd992002> both ssh and user manager problems
1276[20:10:01] <kevindd992002> I don't understand why though
1277[20:10:37] <kevindd992002> I have another Debian system that
doesn't have scdaemon installed too but I don't see that
no scdaemon message when I do sytemctl status on that box
1278[20:10:43] <jhutchins> kevindd992002: It's likely the
usermanager failure was causing the delay, waiting to complete.
1279[20:11:01] <greycat> does it have gpgconf from backports? or
is that unique to the one that was having issues
1280[20:11:04] <kevindd992002> Right, that's what I thought
1286[20:14:24] <kevindd992002> I'm assuming that if if
qemu-extra-blocks was the one responsible for upgrading other
dependencies (including gpgconf) from the backports, then it should
do the same for both systems, no?
1302[20:30:32] <kevindd992002> Before all this, what happened was
an older backports version of qemu-block-extra was installed and it
was tagged as "not upgraded"
1303[20:30:54] <johnjay> is /etc/sudoers.d/wheel a standard file
in general or just debian?
1304[20:31:02] <kevindd992002> I then did apt list --upgradeable
-a and it showed two possible upgrades for that package which caused
it to not be upgraded
1305[20:31:11] <kevindd992002> one from stable and one from
backports that has a higher version
1306[20:31:39] <kevindd992002> that's why I did apt -t
buster-backports upgrade qemu-block-extra which caused the upgrade
of a lot of packages from backports too
1307[20:31:53] <kevindd992002> Did I do that incorrectly?
1308[20:32:22] <greycat> johnjay: sounds more like a BSD thing. I
have no such file.
1309[20:32:34] <johnjay> ok
1310[20:32:49] <johnjay> i do have the folder itself on debian
though
1311[20:32:59] <greycat> yes, mine just has a README
1312[20:33:08] <topoi_> I want to mount a device that is
inserted, do some tasks and umount it automagically thereafter; how
would I accomplish this? Would it be feasable to parse syslog and
fork the rest or are there simpler ways?
1313[20:33:10] *** Quits: dubbie (~strength@replaced-ip) (Remote host closed the connection)
1314[20:34:37] <jmcnaught> kevindd992002: it would have been
better to do "apt -t buster-backports install
qemu-block-extra" which would have been more targetted. Because
you used the upgrade verb, you probably upgraded every package that
could be upgraded to backports version, which is not recommended.
1319[20:36:10] <kevindd992002> jmcnaught: argh, that's
exactly what I thought. I was going to use the install verb
initially but thought upgrade was more targeted. Is there anyway to
revert those changes? Or would a reinstall of Debian be faster at
this point?
1320[20:37:20] <jmcnaught> kevindd992002: downgrades aren't
really supported by apt/dpkg but you can try the remedy suggested by
"/msg dpkg partial downgrade"
1321[20:37:20] <somiaj> kevindd992002: you could go figure out
which packages were upgraded and manually downgrade them
1322[20:37:35] <somiaj> results may varry depending on number of
pcakges.
1323[20:38:56] <kevindd992002> yeah, too much work, I guess.
judging from the apt history.log, there's a lot that have been
upgraded. i'll probably reinstall debian and charge this to
experience.
1324[20:39:10] <kevindd992002> i was luck to still have ssh
access as I did this remotely, lol
1325[20:39:15] <kevindd992002> lucky*
1326[20:40:49] *** Quits: ecsim (~Thunderbi@replaced-ip) (Remote host closed the connection)
1327[20:41:03] <somiaj> kevindd992002: backports just have less
testing and no directy security support, so though they are often
good packages, there is less testing on how they interact together,
hence best to minimize what backports you use based on need.
1381[21:22:58] <sney> apt is a newer tool that has more functions
in a single binary.
1382[21:22:59] <somiaj> xuxx: apt is meant to be a user front
end, and apt-get is meant more for scripts.
1383[21:23:03] <sney> but they are essentially equivalent for
this
1384[21:23:12] <jhutchins> xuxx: same outcome, different internal
process.
1385[21:23:30] <somiaj> well apt-get was heavily used in scripts,
and due to this its options have to be static, so a new binary apt
was created that could be a bit more modern in terms of output and
behavior
1386[21:23:37] <greycat> dpkg, apt vs. apt-get
1387[21:23:37] <dpkg> apt is promoted over apt-get for
interactive use. It uses fancy colors and has output format
differences. It removes the .deb files that it downloads during an
install or upgrade. It installs new packages during upgrades.
apt-get has a stable command-line interface and is promoted for
scripting.
1388[21:24:21] <somiaj> xuxx: Though almost an XY problem here,
what are you trying to do? Get a specific set of packages on a newly
installed system?
1389[21:24:26] <xuxx> sney: '-y' skip prompts even if
it works ?
1390[21:24:57] <xuxx> somiaj: I installed something and I want to
do a script for my friends so it's easier for them
1405[21:30:06] <sney> you could also make a metapackage .deb that
depends on the packages needed, then they could just "apt
install ./custom-whatever.deb" and apt will handle the rest
automatically
1406[21:30:35] <xuxx> sney: not my level XD
1407[21:30:48] <sney> it's basically a text file, you
didn't even check
1409[21:30:55] <xuxx> also I have config file to edit
1410[21:31:01] <xuxx> I planned to do that in the bash script
1411[21:31:21] <saint_> Anyone using Digital Ocean to host Debian
VM with XFCE by any chance ?
1412[21:32:06] <greycat> I don't understand why you would
put a desktop environment on a remote server.
1413[21:32:25] <saint_> I'm using X2GO to remote to the X11,
and it's pretty sluggish. I was wondering if Adding RAM , since
I can't add ram on the graphical board , would change anything
?
1422[21:35:41] <greycat> Any plain window manager.
1423[21:35:52] <saint_> sney of course. but my users are
worldwide. scratch this option for me.
1424[21:35:52] <greycat> There are literally dozens.
1425[21:36:01] <sney> ^ or lxqt/lxde are less involved than xfce
as well
1426[21:37:15] <jhutchins> saint_: lxde is rumored to be leaner,
but not as advanced.
1427[21:37:18] <sney> last time I needed a gui on a remote server
I used windowmaker. it had the added benefit of nobody knowing wtf
it was, so I didn't really need to lock it either
1428[21:37:19] <saint_> ok. I'll pump this VM up with
another GB Ram, and if not working I'll try lxqt. Thanks
1431[21:40:40] <jelly> saint_, the sluggishness is a function of
latency mostly. The clients need to be close to the remote GUI
server, and they need decent bandwidth.
1432[21:40:55] <jelly> RAM will not help
1433[21:41:47] <jelly> that said, there are better and worse
remote GUI solutions. vnc is pretty crap. X11 over ssh is worse. the
fastest-feeling solutions are proprietary
1434[21:41:51] <saint_> jelly i went cheap for test purpose with
only 1GB Ram. Clients sadly are on remote oil rigs or thousands of
miles away with "ok" connections. I think 100Mbs/download,
5Mbs/upload (if that)
1435[21:41:55] *** Quits: mezzo (~mezzo@replaced-ip) (Quit: leaving)
1436[21:42:04] <jelly> saint_, 1GB is not enough for GUI
1438[21:43:02] <jelly> but if your clients are on oil rigs you
probably have money for AirDesk or something
1439[21:43:10] <saint_> jelly you'd think ...
1440[21:43:33] <sney> xrdp is reasonably fast even over a crappy
connection, again not brilliantly fast or anything but still usable
1441[21:43:39] <saint_> jelly i was thinking aout VNC, but i dont
remember where I read that security wise, it's a no-no, while
x2go runs over ssh tunnel
1442[21:43:59] <saint_> i m resizing the VM with 2GB Ram.
I'll see in a few once it s done.
1443[21:44:34] <jelly> AnyDesk, not AirDesk
1444[21:44:56] <greycat> VNC can also run over an ssh tunnel
1445[21:44:58] <jelly> saint_, you can run vnc over an ssh
tunnel.
1446[21:45:10] <jelly> but that is also not a good idea wrt
latency
1447[21:45:30] <jelly> use a datagram-based vpn tu secure the
transport instead
1448[21:45:34] <jhutchins> I think forwarding a whole desktop
should be avoided if at all possible. Forwarding one or two apps is
not too bad, but general control and management should be done with
a plain console.
1458[21:48:20] <jelly> count the number of remote exploits it
had, on its native platform
1459[21:48:44] <jelly> secure the traffic with vpn.
1460[21:51:56] <saint_> yeah. i'm either doing vpn or ssh.
but we have to connect those guys to multiple machines, on multiple
data centers, which are not connected together. so they'd need
a vpn connection per destination, which is going to be a major pain
in the butt to maintain on both ends. i'd rather use ssh
tunnels. X2GO allows me to set them up with remote login / IP ,
mount the ssh, and off they go. one application for them, multiple
servers / desktops to
1461[21:51:56] <saint_> connect to. easier on their end.
1462[21:52:36] <alkisg> The most important part is the network
connection... I vnc over ssh to my office, but I have a 100 mbps
line on both sides (and low latency); so it's very snappy
1463[21:54:10] <alkisg> saint_: is "remote disk" an
option instead of "remote desktop"? If so, it feels a lot
faster. E.g. VM over SSHFS...
1467[21:57:16] <jelly> I have my own machines in 5-6 datacenters,
all in the same /24
1468[21:58:01] <saint_> jelly I cant connect the endpoints
together. they're all different companies
1469[21:58:38] <saint_> alkisg not an option. the goal is to run
proprietary applications locally on those servers. good point
though.
1470[21:58:46] <jelly> saint_, you can still connect them to the
same vpn server and allow traffic routing accordingly
1471[21:59:41] <saint_> jelly i could. but then i'm adding a
point of failure with this new server in the middle. and it's
going to add latency instead of them ssh'ing directly to the
end servers
1473[22:01:00] <jelly> saint_, ssh is still tcp-inside-tcp and a
decent vpn won't be
1474[22:01:01] <saint_> I added 2GB Ram on my test VM. seems
smoother. I'll keep doing testing and see how it goes. Might
try LXDE tomorrow.
1475[22:01:14] <alkisg> RDP over SSH port forwarding is also
possible, when double encryption/security makes sense... Or just rdp
with some firewall restrictions for some added security...
1476[22:01:42] <jelly> if you want to do GUI with X give the
system at least 4GB RAM (well, depends on the app workload)
1479[22:02:04] <alkisg> saint_: one thing to test is with or
without compositor. Also, MATE has a "detect remote desktop and
stop effects" infrastructure by sunweaver
1480[22:02:18] <jelly> I wouldn't put a Linux GUI on a
machine with < 4GB, and I wouldn't run a browser without 8GB
RAM
1481[22:02:55] <saint_> jelly ok. let me try to increase size of
this test vm. definitely better now with 2gb than 1.
1482[22:03:07] <jelly> saint_, measure first, then increase
1484[22:03:41] <jelly> run "vmstat 1", log in and click
things, see whether there's enough cache and whether
there's active swapping
1485[22:04:33] <jelly> (si, so columns = reading from swap,
writing to swap)
1486[22:04:44] <saint_> you'd be surprise on how those
companies are so cheap. drives me nuts. we have about 2,700 users to
run on this new setup , connecting to 387 servers. say hello to my
summer nightmare .
1487[22:05:16] <jelly> figure out what you need to make things
smooth, then tell them why and how much it's going to cost
1488[22:05:30] <saint_> but that's better than a job where
you have nothing to do and get bored i guess, lol..
1489[22:05:52] <jelly> sure, with 400 VMs adding 1GB to each adds
up fast
1490[22:06:19] <saint_> jelly their mentality is "if you
can't do what we want you to do, at the price we want or
cheaper, then we'll use someone else". I wouldnt care if i
was running under loads of $$$$ but that s not the case...
1491[22:06:58] <jelly> I'd do a PoC where everything works
great, then tell them how much $$$
1492[22:07:01] <saint_> i spoke to another company about this
client. someone they used to deal with. those guys told me
"yeah you can keep them. nobody wants them" ...
1494[22:07:58] <alkisg> Connecting to amazon workspaces via linux
clients also works well with low bandwidth. If they can afford it,
it'll be quite easy to set up
1495[22:08:01] *** debhelper sets mode: +l 1044
1496[22:08:07] <saint_> you know the client who's a pest and
will call you for every little thing that they cant point out? like
"My computer is not working" ... well.. define "not
working"... like.. not turning on? not on the network?
application crashing? "It's not working, check it
out"... that's this type of client. E V E R Y fuck!ng day.
1497[22:08:26] <saint_> alkisg yeah! I was thinking about that.
Used Amazon WorkSpaces in the past, loved it.
1498[22:08:37] <jelly> and then you charge per call and per hour
1499[22:08:40] <saint_> THat might be a good solution actually.
I'll try to bring this up again.
1500[22:08:56] <saint_> jelly if I do that -no disrespect-
they'll send their business in India..
1502[22:09:50] <jelly> 'swhy I don't like working with
end users
1503[22:09:51] <saint_> alkisg the only thing though - i dont
rememebver - does your workspace keep the same IP address when you
connect to it then have to go out on the internet ?
1504[22:10:24] <saint_> jelly what country are you located in ?
1507[22:11:19] <alkisg> saint_: I helped in setting it up but
I'm not using it myself so not sure. I guess it would keep the
ip as long as you close the rdp client without fully logging out
1508[22:11:57] <saint_> alkisg i have an AWS account, i'll
mount one up and try it again. thanks for that. i totally forgot
about this.
1509[22:13:18] <saint_> jelly i work with americanos :) like in
the movies. they want it now at a dirt cheap price. good business at
the end, but someitmes you have to suffer. no good health care here.
and nowadays with this shit virus going on and business closing, i
do my best to achieve expected goals, and even better. i like the
challenge though. keeps your brain going .. :)
1520[22:16:30] <saint_> so basically an internal address, then i
need to setup a VPC, and this will guaranty the same public IP
address. i'll look into it.
1521[22:16:42] <saint_> on this note, back to my tests. thanks
all for the chat and hints.
1529[22:30:05] <saint_> I added an extra 2GB, and it's night
and day with 4GB RAM now. Thanks for whoever recommended this. I was
going cheap :) Lesson learned.
1575[23:19:34] <ld> I want to do the 4.3.3 method so I can
include nonfree firmware during install
1576[23:20:18] <greycat> !firmware images
1577[23:20:18] <dpkg> There are <live> system and
<netinst> and DVD images containing non-free Debian
<firmware> packages available from
replaced-url
1578[23:21:49] <jmcnaught> ld: then either use an unofficial
non-free firmware image, or create a partition in the remaining
space after following the steps in 4.3.1
1579[23:21:50] <ld> It wouldn't allow me to create a
partition in the remaining free space after 4.3.1
1604[23:33:57] *** Quits: Ekis (~Ekis@replaced-ip) (Remote host closed the connection)
1605[23:34:18] <somiaj> ld: both are supported, though installing
bullseye directly would be one less step
1606[23:34:42] <somiaj> ld: just note you won't have full
security support until the actual release, but I hear it is going to
be in the next 1-2 months.