2[00:03:05] <ryouma> i have heard that apt(1) installs
recommendations and suggests. is this true? does it make sense to
turn one or both off for a dist-upgrade from stretch to buster?
3[00:03:23] <sney> it installs recommends by default but *not*
suggests
4[00:03:47] <sney> for upgrades, it's best to leave it at
defaults, just make sure to disable any extra sources
5[00:03:57] <ryouma> is that the same as apt-get?
6[00:04:06] *** kish`_ is now known as kish`
7[00:04:30] <sney> default behavior is the same regardless of
which frontend is used, at least in this case
10[00:07:58] <somiaj> ryouma: note recommends are supose to be
packages that are not a hard dependency, but you will loose
functionality without, while suggests are sometimes more related
software that is sometimes used in conjunction.
11[00:08:42] <somiaj> Of course there is some grey area,
sometimes I think packages are recommended when they should only be
suggested, but it might dependon the user if the functionality that
is used is considered optional or not.
12[00:09:06] <sney> it's pretty easy to get into a broken
state when disabling recommends by default.
13[00:09:20] <somiaj> so in general not installing recommends
may mean you loose functionality of software that you want, hence
the default is to include them.
14[00:09:36] <sney> I definitely use --no-install-recommends on
the command line in specific situations from time to time, but I
would never do it without seeing what was going to be installed
first.
16[00:11:34] <mutante> ~ recommended to install recommended
software but suggested to not blindly install all suggested packages
17[00:12:46] <ryouma> ok, yeah i was mostly concerned about the
(apparently false) idea that apt installs suggests by default
18[00:13:21] <ryouma> i do know, however, that i have to do
apt-get -o "Apt::AutoRemove::SuggestsImportant=false"
purge ... to get rid of cruft. but idr what that apt variable
actually means and don't know what man page it is in. (i think
it had something to do with an asymmetry where installing a package
installs some stuff and then purging does not purge it.)
19[00:13:43] <somiaj> Also note debian allows non-free software
to be suggested, so be careful if you choose to install suggests by
default but dont' want non-free software.
20[00:14:14] <somiaj> well that is different, autoremove cannot
tell if a package that is suggested by another package is wanted by
you or not
21[00:14:32] <ryouma> i wouldn't install suggests by
default
22[00:15:15] <somiaj> what is happening in that case is a
package is installed as a depends/recommended of another package,
but one on your system also suggests that package, so though apt
wouldn't install it by default, it also won't remove it
since it dosen't really know why that package is there, it only
knows that a pakage you want suggests it.
30[00:17:19] <somiaj> it can, autoremove by default only
removes packages that are auto installed, and not dependend
on/recommneded/or suggeted by another manual package (or its
dependencies)
31[00:17:47] <somiaj> or you installed something that was
suggested then marked it as auto, there are lots of ways a suggested
could have gotten installed
32[00:17:49] *** Quits: short-bike (~short-bik@replaced-ip) (Quit: and on that note...)
40[00:22:09] <somiaj> sometimes due to the complex dependency
tree it is useful to just look though what you have installed and
think about if you need it, aptitude why packagename can often be
useful
69[01:03:41] <ryouma> not really. just: so as not to get
overwhelemed, to make copying faster, to make seraching fastger, to
make integrity checking faster, to work on small devices, to reduce
complexity in dist-upgrades, to find things quicker, to make man -k
smaller, and to know what i have that i need among a bunch of
similar choices because i kept only that one package.
70[01:04:15] *** Quits: Filohuhum (~filohuhum@replaced-ip) (Remote host closed the connection)
71[01:04:31] <ryouma> also to make dependency issues cleaner
upon ordinary upgrades, installs, and purges.
96[01:32:13] *** Quits: Filohuhum (~filohuhum@replaced-ip) (Remote host closed the connection)
97[01:32:32] <maxrazer> When I use slim display manager to
login I notice that the background theme for the login manager
becomes my desktop for i3wm. Is that some side effect? I so no way
to set wallpaper from slim, just login theme itself.
101[01:35:11] <LuKaRo> Yes, i3wm isn't clearing the screen
when launching. So if the display manager does not do that as well,
it's content will become your background.
102[01:35:44] *** Quits: Filohuhum (~filohuhum@replaced-ip) (Remote host closed the connection)
103[01:36:14] <maxrazer> Ok, that is kind of what I was
thinking. I'm not sure if slim has an option to do that, which
would leave me with a black background again. Though, if I use some
program like feh to run wallpaper I think that will overwrite it.
110[01:40:18] <maxrazer> I noticed that the Debian wiki says
that the startx also does surprisingly even though it isn't
even a display manager like XDM.
123[01:43:13] <maxrazer> As far as disabling display manager the
only way I've found is to disable the systemctl service for
whatever display manager is set to default. (I think)
124[01:43:55] <maxrazer> I'm going to try lightdm and
reboot.
164[02:18:11] <Hash> What's the best way to switch from a
diff distro to debian but keep /home and try to prerve /etc/ and
other configs (going from a debian based distro to debian)
165[02:18:20] <Hash> Don't ask me why. Just help me.
Thanks!
166[02:19:33] <Hash> This machine needs conversion!
167[02:19:49] <abrotman> you should ask the distro you're
going to install to ensure their installer doesn't wipe the
partition
168[02:20:39] <Hash> Well, the goal is to take this machine,
repalce ubuntu with debian stable, and try to preserve /home and
perhaps some settings from /etc so I don't have to redo
everthing. Try to save me some work.
173[02:21:35] <abrotman> when you run the installer,
there's an option to not format a filesystem
174[02:21:58] <abrotman> we may need to know more about the
filesystem if there's something special like LVM or luks or
whatever
175[02:22:47] <Hash> No I'll be installing a different
partition altogether, so incase anything gets messed up, I can have
the original ubuntu parition. Oh just direct partition access, on
layers in the middle, GPT disk, 5800x cpu
176[02:24:21] <Hash> I'm not wanting to save the package
list, as that maybe different (diff package names etc.) but common
configs like nginx, and various other stuffs, I think I'll just
install debian to a diff partition, and then mount /etc from
original partition, and copy over whatever I need for whatever
ackage.
178[02:24:47] <Hash> Thing I wanted to worrya bout was, config
file differences perhaps in the same software but diff distro mabye
cause issues? I dunno this.
182[02:25:54] <Hash> The rest of the most of the configs would
be in /home and again the worry is whether programs migh have
changed their config file format or something from ubuntu 20.04 lts
stable to debian stable latest.
183[02:26:23] <somiaj> abrotman: due to differences it might be
easier to just backup /etc and /home, and then copy over the stuff
you want, this also is good practice so you have a backup of your
data.
190[02:27:41] <somiaj> though note, software configurations in
$HOME should carry over just fine provided the versions match, often
times configurations can go from older to newer version, but not
newer to older as nicely (and this varies between software). Hence
if you just backup and copy over stuff, you can test things out and
see what configurations still work and don't.
191[02:27:57] <Hash> HYmm
192[02:28:06] <Hash> I will back up /home/usr/.config an any
.dot files
193[02:28:10] <somiaj> I once tried to share /home across a few
dual boots, and quickly found out it was a pain due to different
versions of software being slightly different.
194[02:28:13] <Hash> That's a sound idea
195[02:28:30] <Hash> yeah, that's I discovered once too
which is why I'm very cautious now
196[02:28:32] <somiaj> I would jsut backup all of /home/usr (or
all of /home), as backups are always good ot have floatinga round.
199[02:28:52] <Hash> well I mean just the dot stuff.
200[02:29:14] <somiaj> Plus I find personally when I do this, I
have changed, and so sometimes it is just easier to reconfigure to
my new desires from defaults vs some old config that I have had
floating around for many years.
208[02:43:24] <abstrn> Maybe this isn't the best place to
ask this, but I haven't found an answer anywhere else: I have
an external hdd that stays plugged in most of the time but sometimes
I "safely remove" it via udisksctl --unmount/--power-off
and then unplug it. Then when I plug it into the same USB port
shortly after, it is not at all recognized by the system.
'lsblk' doesn't show it, it doesn't show up
anywhere in
209[02:43:25] <abstrn> /dev/sd* that I can find. If I plug it
into another USB port, it shows up and I can mount fine. And if
maybe a week goes by, I can plug it into the original port and
it's recognized again. Does anyone know what's going on?
Does the system somehow cache which harddrives are ejected, and
ignore them for a while after they're removed?
210[02:49:06] <somiaj> abstrn: that sounds like hardware issues,
could that usb port just be failing?
211[02:49:27] <abstrn> perhaps, but it
212[02:49:36] <abstrn> 's happened on multiple usb ports
213[02:49:45] <somiaj> maybe an issue with the usb controler?
214[02:49:50] *** Quits: Filohuhum (~filohuhum@replaced-ip) (Remote host closed the connection)
215[02:50:17] <somiaj> have you looked at 'demsg'
right after you plugged the drive back in, you should get a hardware
event even if it doesn't assign it a blk device
220[02:53:53] <abstrn> there's no indication in dmesg that
it's been plugged in... maybe it is a hardware issue. but
another strange thing is that when i plug in a usb hub into the
original port, and plug the hdd into the hub, it is recognized
222[02:58:34] <somiaj> could be the usb drive is having issues,
the usb controler, have you tried this in another computer? Do you
have this issue with other usb devices?
223[02:59:06] <somiaj> but what you are describing sounds like
hardware to me, if the kernel isn't triggering any event when
you plug it in (which should appear in demsg), then some hardware
isn't triggering the event to send to the kernel.
240[03:12:54] <abstrn> makes sense, probably hardware then :(. i
thought i remember it happening on a debian server a while ago but
that was an old machine and could also have had a hardware issue.
253[03:27:36] <ryouma> abstrn: i am curious about udisksctl
--unmount/--power-off as i have not heard of them. have you
considered hdparm -B and umount?
268[03:54:55] <abstrn> huh, looking again at the manpage,
udisksctl power-off says that "the USB device will be
deconfigured followed by disabling the upstream hub port it is
connected to". so maybe that is the issue? ryouma: i
haven't used hdparm to power off a device, i thought -B just
changes power management/preformance?
270[03:55:43] <abstrn> or is unmounting the device enough? i
guess i just got in the habit of doing udisksctl unmount / power-off
before removing a device, maybe the second step isn't
necessary?
274[04:06:53] <ryouma> idk if an unmounted partition's
drive will still spin, i would only guess that it does. but ime the
power management can be used to power off effectively.
275[04:08:01] *** debhelper sets mode: +l 1010
276[04:08:44] <ryouma> abstrn: look at sdparm man page. eject,
stop, and " In the Linux 2.6 series, especially with ATA disks,
using sdparm to stop (spin down) a disk may not be sufficient and
other mechanisms will start the disk again some time later. The user
might additionally mark the disk as "offline" with
'echo offline > /sys/block/sda/device/state' where sda
is the block name of the disk. To restart the disk "offl
277[04:08:44] <ryouma> ine" can be replaced with
"running". "
278[04:09:07] <ryouma> there -y in hdparm but maybe only for an
obsolescent bus
279[04:09:34] <ryouma> although i believe it worked ok for
external sata long ago
280[04:10:12] <ryouma> so it sounds complex, unfortunately
281[04:15:05] <abstrn> ryouma: hm, thanks for the info,
i'll take a look at sdparm
282[04:16:33] <ryouma> i use -B for those bus powered externals
283[04:16:51] <ryouma> idk if i should be using more lately,
haven't checked
287[04:18:22] <ryouma> another question is whether you are
improving anything other than power. for example, does it affect
power surge response. or does arm parking status affect anything.
408[07:16:27] <Lope> In my /etc/network/interfaces I have a
bridge, br0 configured to include eth0 in it's bridge-ports. In
/etc/NetworkManager/NetworkManager.conf I have
unmanaged-devices=interface-name:br0;interface-name:eth0;interface-name:enp34s0
409[07:16:31] *** Quits: darunesh (~darunesh@replaced-ip) (Remote host closed the connection)
410[07:16:43] <Lope> And yet when I boot up, eth0 receives a
DHCP IP address from the LAN
411[07:16:50] <Lope> It's extremely annoying.
412[07:17:00] <Lope> That IP is on the same subnet as my static
IP on br0
413[07:17:14] <Lope> So I have to manually remove the IP from
eth0 on every boot.
415[07:18:11] <somiaj> Any reason you want to use network
manager and myabe just result to only using the interfaces file?
416[07:18:33] <somiaj> and you don't have any dhcp setup
for eth0 or the bridge in the interfaces file?
417[07:18:57] <somiaj> maybe (a) disable network manager, and
make sure that is the culprit that is causing you the problem, if so
one solution is just uninstall network manager.
418[07:19:06] <alkisg> Does `nmcli` show this device as
unmanaged? If yes, the problem would be in /etc/network/interfaces,
you could pastebin its content
419[07:19:33] <somiaj> Or let someone who knows network manager
give a better way to check if it is managed or not.
420[07:19:37] <ryouma> i have network manager purged and my
system seems to work, fwiw. i was nervous about purging it.
422[07:19:52] <ryouma> and didn't change anything to purge
it, just puirged it
423[07:20:00] <Lope> somiaj, I only have NetworkManager because
it makes it easy to connect to a wifi network. But I'm not
currently using NetworkManager at all. I don't even have a
wireless adapter installed/attached at the moment, so the
NetworkManager widget in KDE is empty.
426[07:20:38] <somiaj> I personally use wpa_supplicant in roam
mode and wpa_gui for this from my interfaces file,
/usr/share/doc/wpasupplicant has a good guide how to configure this
427[07:21:06] <somiaj> If you want a non networkmanager way to
manage wifi networks, but this maybe beyond your actual problem, and
more avoids it. I would first folow alkisg advice and double check
that network manager is really causing the problem.
428[07:21:46] <Lope> alkisg, `nmcli` shows that every single
network interface is unmanaged by NetworkManager, currently.
430[07:22:27] <alkisg> Lope: then I would gather that the IP is
set by ifupdown. Can you pastebin the contents of
/etc/network/interfaces? Hide macs/ips if you want
431[07:23:14] <Lope> oh... bastard. I found
/etc/network/interfaces.d/setup
432[07:23:32] <Lope> I've not used files in there before.
433[07:23:46] <Lope> In that file it adds eth0 dhcp
434[07:24:04] <somiaj> yea, more and more things are splitting
up configuration files to make things more modular, but now you no
longer have a single file to look through.
435[07:24:07] <Lope> (Ive not used interfaces.d before) seems
like a relatively new thing.
436[07:24:18] <Lope> yeah. Thanks
437[07:24:41] <Lope> That explains the BS I've been
experiencing on my other PC as well :/
439[07:24:52] <somiaj> Lope: yea, more and more there are foo.d/
directories all through out /etc
440[07:25:05] <somiaj> so just make sure you keep your eyes out
for them.
441[07:25:18] <Lope> yeah
442[07:25:41] <Lope> /etc/daddy.d/ who's your daddy?
443[07:26:40] <alkisg> But note that each .d has its own rules;
/etc/network/interfaces.d requires "no extension", while
e.g. /etc/dnsmasq.d accepts "almost any extension, even
.bak"
445[07:27:16] <Lope> somiaj, thanks for the tip man! I've
only used wpa supplicant a handfull of times, because I found that
configuring wifi interfaces via config files was very arduous.
446[07:27:30] <Lope> somiaj, but wpa_gui sounds promising.
I'll check it out.
448[07:27:51] <somiaj> Lope: It is a very simple tool, and I
rarely use it, but yea I like this as wpa_supplicant is installed
and used anyways, so use its features vs something else
449[07:29:04] <Lope> I actually don't like NetworkManager
at all, I just find that it's helpful for connecting to WiFi.
I've also used the network manager widget previously when
testing VPN connections (only from very new live distros) which was
quite nice, it even handled wireguard.
450[07:30:12] <Lope> But generally I don't like magic,
it's not often the magic does what I want specifically. And
when it breaks, then if I've been reliant on magic, then I
suddenly find myself with broken production systems, at step 1 of
knowledge and negative time available to fix the issue.
451[07:30:16] <alkisg> It's also nice that it displays an
icon with the current status, so in desktop environments you can
easily detect when the network cable has been unplugged :)
453[07:31:33] <Lope> alkisg, yes, although I tell networkmanager
that it's not allowed to touch my wired interfaces, so that
functionality doesn't really work for me.
454[07:32:03] <Lope> I found NetworkManager with bridges to be
very unreliable.
455[07:32:22] <Lope> And I often use bridges when I run
VM's.
457[07:32:26] <alkisg> I'm using network-manager with
bonding in about 100 schools with no issues, it's rather stable
here
458[07:32:36] <alkisg> Not bridging though; just bonding
459[07:33:02] <Lope> is bonding like raid 1 or like failover?
460[07:33:42] <alkisg> Both. It doubles the speed with 2 NICs,
and can also work with a single NIC
461[07:34:03] <Lope> So you need 2 NICs on each end both
configured to bond?
462[07:34:30] <alkisg> No. E.g. a server has 2 NICs, and 10
clients have 1 NIC. This means that the server can send/receive with
2 gbps instead of 1 gbps
464[07:36:04] <alkisg> (all in the same simple switch)
465[07:36:11] <Lope> oh, wild, so can you connect the
server's 2 NICs to an unmanaged switch, and bond the nics to
have a single static IP?
466[07:37:12] <alkisg> Yes, I even set the same MAC for both
NICs from network-manager, so that it's easier to detect the
server with arp from elsewhere
472[07:37:48] <alkisg> And it's just an nmconnection file,
nothing more
473[07:37:49] <Lope> The switch is a dumb unmanaged switch, and
the other nodes on the network don't know?
474[07:37:54] <Lope> Amazing
475[07:38:02] <alkisg> Exactly, cheap switches here, schools are
poor :D
476[07:38:09] <Lope> Wow, that's awesome
477[07:38:25] <Lope> I've got a spare gbps port going to
waste.
478[07:38:54] <Lope> It's a pity that SFP+ and 10GBE is so
expensive.
479[07:39:39] <Lope> I mean, I could afford picking up some
10GBE cards at $100 each but then I ask myself "Do I really
need it?" and the answer is "Not for a few hundred
dollars."
480[07:39:41] <Lord_Devi> I've got something similar setup
with my home. 2 broadband connections at home, each tunneled into
its own remote cloud server. I don't think it was a bonded
setup, but they work in fail over mode and will under most
circumstances boost my bandwidth.
481[07:40:15] <Lord_Devi> I wanted a setup, where I could have 2
ISPSs, and any one of them could go down - but I'd still have
internet. And if both were up, I might be able to download or upload
things a bit faster.
482[07:40:36] <Lope> Lord_Devi, interesting, I planned to do
that previously. Are you using any special sauce to make it work?
483[07:41:04] <alkisg> I'm hoping that we'll soon see
switches that will support 2x 10GBE (bond) for the server, and the
newish 2.5 GHz standard for the clients with the existing cabling.
With the HDD > SSD upgrade, local disks are now much faster than
LAN, and this will lower the gap...
484[07:41:15] <Lope> Lord_Devi, I actually was going to do the
same thing, but then it turned out that my secondary ISP was such a
piece of crap and so unreliable as to be pretty much a waste of
money, so I cancelled it.
485[07:41:18] <Lord_Devi> Well it is OpenBSD based, and I was
using ifstated to monitor the interface state. Then pf to make sure
return traffic was going over the right interface. But honestly it
wasn't working 100% right.
486[07:41:32] <Lord_Devi> I'm still working on fixing the
quirks with a person I found to help.
487[07:42:26] <Lope> alkisg, the thing is though, that 10GB
copper ethernet (whatever it's called) doesn't need to be
expensive. and it even works over Cat5e over short distances (like
one room to another).
488[07:42:41] <Lope> They're just charging early adopter
tax on it.
489[07:43:07] <alkisg> Ah really? My use case is "a single
computer lab of e.g. 25 m2", so it would work there with the
existing cabling, nice!
490[07:43:36] <Lope> alkisg, I can't bring myself to buy
any 2.5Gbps equipment. You can basically get almost the same bonding
2x 1gbe ports together, and I've always got spare 1gbps nics
banging around.
494[07:44:01] <Lord_Devi> There is a 4 port 10Gb on amazon for
like only 140 I think. It supports Copper as well as SFP+. Adapter
needed for copper.
495[07:44:07] <alkisg> Unfortunately we can't use bonding
on the clients, not enough ports for that
496[07:44:26] <Lord_Devi> The copper 10Gb's use way more
power than the fibre optics though, and run really hot.
497[07:44:30] <alkisg> Last time I checked, the 10Gb NICs were
also expensive...
498[07:45:09] <Lope> alkisg, and I almost always find it a waste
of money to spend $30 on 2.5Gbps instead of $100 on 10Gbps, becauee
invariably, I buy it initially, just because I want it. Then later I
find out that I need it, and then the $30 I spent on 2.5Gbps is
basically a mediocre solution that I have to discard, so the $30 was
actually just a waste and I should have just spent $100 from the get
go. Cheap is expensive. So I tend to delay pulling the trigger when
possible,
499[07:45:09] <Lope> but then go big when I do upgrade.
500[07:45:10] <Lord_Devi> I'd generally only really use the
copper if I had to run a longer 10Gb I think. NOT AN EXPERT though..
lol.. I just found that long fibre was really difficult to acquire,
and then work with.
501[07:45:57] <Lord_Devi> Lope: Yeah it really is a good idea to
wait those extra couple of months to be able to just afford the
correct solution the first time. If possible.
502[07:46:21] <alkisg> A quick check here, says 30€ for 2.5
Gb, and 143€ for 10 Gb. Hmm, for e.g. 12 clients, schools
couldn't afford 10 Gb yet :/
503[07:46:42] <Lope> But I even feel like 10Gbps is a bit of a
joke. I've seen 40Gbps QSFP+ cards on ebay for a reasonable
price. Problem is I'm in South Africa and shipping costs with
"the virus" is worse than ever. Also apparently those old
used QSFP+ server cards are very power hungry, which is also
off-putting.
504[07:46:46] <Lord_Devi> Schools don't know how to handle
money.
505[07:46:52] <alkisg> I guess we'll need to use SSDs for
caching for a few years :D
506[07:47:09] <alkisg> (fortunately most of our traffic is
cache-able)
507[07:47:28] <Lord_Devi> Lope: lol, i found out about the QSFP+
just after I had pulled trigger on ordering all my 10gb SFTP equip.
Like.. the next day. I had thought I dug too!!
508[07:47:33] <Lope> So I'm kind of irrationally waiting
for reasonably priced 100GbE or at least like 50Gbe etc to feel like
it's a worthy reason to spend money on a network upgrade.
It's gotta at least reach something that resembles a single SSD
speed.
509[07:48:03] <Lord_Devi> Ideally on gear not produced or
manufactured by skynet.
511[07:49:25] <Lope> Lord_Devi, apparently getting SFP+ switches
are cheaper, but then if you want to get the copper adapters it ends
up being a lot more expensive.
512[07:49:39] <Lope> So apparently there are 2 cost effective
ways to homelab fast ethernet.
513[07:49:55] <Lope> 1. QSFP+ with direct attach copper cables
"DAC cables"
514[07:50:26] <Lope> 2. Proper RJ-45 10GBE copper switch and
NIC's. The switch prices have come down a LOT. The NIC's
are unfortunately still $100.
516[07:51:35] <Lope> Lord_Devi, interesting, yes I've heard
that coppper 10Gbe uses a lot more power but haven't looked
into quantifying it
517[07:52:13] <Lord_Devi> Yeah. That mattered to me too, because
I was trying to build an ultra low power home lab. I even went so
far as to get this special ATOM processor for my openbsd firewall.
518[07:52:55] <Lord_Devi> Although... I think I am kind of not
sure if I want to continue that trend. I really like the idea of my
network being able to resist a ton of power outage. But at the same
time, I am not super pleased with these low power atoms.
519[07:55:50] <Lord_Devi> MikroTik has this really nice 16 port
10Gb I got, CRS317-1G-16S+RM, to tie into a 24 port 1Gb switch
(CSS326-24G-2S+RM) that also has 2x 10Gb ports (to tie to the big
ass 10Gb). Didn't cost much!!
521[07:56:37] <Lord_Devi> The 16 port mikrotik 10Gb was the most
expensive part, but still I was shocked at how affordable it was. I
think it is a bit more now.. costs like $350. I think I got mine for
about $250...
531[08:03:36] <Lord_Devi> Yeah it says it is a 16w TDP. So
I'm not 100% if that's exactly power from wall you asking
for, but it is related. I get a little confused with those specifics
honestly.
532[08:03:58] <Lord_Devi> RUNS great. Takes FOREVER to boot.
533[08:07:03] *** Quits: szorfein (~daggoth@replaced-ip) (Remote host closed the connection)
534[08:07:16] *** Quits: ChubaDuba (~ChubaDuba@replaced-ip) (Remote host closed the connection)
550[08:32:20] <TheBigK> hi, I have a developer here that does
have a root partition encrypted but the password which worked
yesterday is not working anymore.
replaced-url
551[08:32:26] <TheBigK> does anyone have any idea about this?
552[08:32:47] <TheBigK> i saw some errors logged in the smart
memory... but the lukscontainer looks fine to me?!
553[08:32:56] <TheBigK> is there some sort of check mechanism
which i can use?
554[08:33:17] <hwpplayer1> Yes But this is a Security hole
555[08:33:28] <hwpplayer1> As in my understanding
556[08:33:40] <hwpplayer1> You need to prove the ownership
557[08:33:43] <TheBigK> hes 100 % sure that he knows the
password
558[08:34:12] <TheBigK> is there a destroying mechanism for too
many tries? no, right?
578[08:39:55] <hwpplayer1> i saw some errors logged in the smart
memory... but the lukscontainer looks fine to me?! // How did you
realize that ? With a tool I mean or by hand ? I ask because I will
check tools for that Okay ?
579[08:41:41] <hwpplayer1> TheBigK: Can I pm you ?
582[08:54:32] *** Joins: mezzo (~mezzo@replaced-ip)
583[08:56:41] <Lope> Lord_Devi, in my experience it's
pointless worrying about power consumption if you don't have a
device that measures power consumption from the wall, because
sometimes you'll do "optimizations" that you think
save power, but they cause more power to be used, or sometimes you
do "optimizations" that make the system painfully slow,
but only save an irrelevant amount of power. So I now regard having
such a device as a requirement before I do any power
584[08:56:41] <Lope> tweaking. They're not expensive.
585[08:57:27] <Lope> Expecially in the case of a desktop PC. At
least with a laptop, if you run it from battery, powertop can tell
you relatively accurately how much power you're using.
588[09:00:49] <hwpplayer1> TheBigK: Are you there ?
589[09:01:26] <Lord_Devi> Lope: Yeah. I agree with the need for
a meter of some kind. I got a special kind of power supply bar for
my rack that will measure the power drain at each outlet in it. You
can connect to the power remotely to fetch data from it too.
ALTHOUGH that's home lab stuff I haven't gotten to playing
with yet.
590[09:02:06] <Lord_Devi> I don't have anything like that
for my desktop or various vdu's though.
592[09:03:22] <Lord_Devi> I am getting a sense for that problem
you raise issue with though: The tweaking for 'irrelevant
amount of power' differences. I probably could have gotten a an
ATOM with a bit more power draw, like say.. 30 TDP and gotten WAY
more performance out of it.
598[09:04:44] <Lord_Devi> We were discussing network gear, and
their power usage.
599[09:04:50] <Lope> hwpplayer1, Lord_Devi said he setup a low
power atom server for his homelab.
600[09:05:08] <hwpplayer1> When it comes to server we usually
run a server without a GUI
601[09:05:13] <hwpplayer1> funabashi: ^
602[09:05:14] <Lord_Devi> We were talking about 10Gb SFP+
networking gear, and I had mentioned how the copper adapters draw
way more power on them than the fiber does.
624[09:12:20] <Lord_Devi> Yeah I am in the same boat with that.
It's that general feeling that lead me to just digging into
'pf' for openbsd, rather than bothering with pfsense.
625[09:12:28] <hwpplayer1> Lord_Devi: It seems that I talk about
K8s or docker when i say microserver yes you are right
626[09:12:40] <Lord_Devi> Similar scenario... Not cisco related.
I was never interested in cisco for anything.
627[09:13:11] <hwpplayer1> I learned Cisco it is funny but I do
not want any of their device or software
628[09:13:12] <Lord_Devi> I'm JUST learning bout k8's
now. It is actually kind of why I'm looking into Debian more
seriously now.
629[09:13:42] <hwpplayer1> Debian is like Red Hat Ubuntu is like
more like Fedora
630[09:13:51] <Lord_Devi> I am trying to figure out what a good
server farm for a lot of static websites might look like. K8's
are sounding like a good option for me to be playing with. But
TOTALLY new to me.
631[09:14:09] <Lord_Devi> I'm just... like, a linux user
from the 90's who's use virtual machines his whole life.
632[09:14:29] <hwpplayer1> You can handle if you know what is
the kernel side and how UNIX is designed nothing new there
633[09:15:18] <dreamer> hwpplayer1: eh?
634[09:15:27] <Lord_Devi> Oh interesting grouping! I
wouldn't have used that. I would have grouped them the other
way around. "Ubuntu is like a mutant Debian." and,
"Fedora is a mutant Red Hat".
635[09:15:29] <hwpplayer1> dreamer: ?
636[09:15:35] <dreamer> "< hwpplayer1> Debian is like
Red Hat Ubuntu is like more like Fedora"
637[09:15:39] <dreamer> this doesn't make any sense
638[09:15:43] <dreamer> and is not based in any reality I know
of
639[09:16:17] <hwpplayer1> Okay I mean Debian and RHEL is more
focused on stability and the others are focused on fast development
642[09:16:36] <dreamer> ubuntu is not focussed on "fast
development"
643[09:16:44] <Lord_Devi> Ubuntu is based on Debian/sid. Red Hat
kinda stems from Fedora. I say kinda.. because they have an odd
chicken and egg thing going on.
644[09:17:01] <hwpplayer1> From my experience I see that But
Fedora is really focused on fast development
645[09:17:39] <tarzeau_> ubuntu, redhat, fedora is offtopic
here. thanks
646[09:17:40] <Lord_Devi> I was toying with the idea of trying
fedora for my k8 project. I like the idea of CoreOS and of
silverbleu's immutable operating systems.
647[09:17:48] <Lord_Devi> But ultimately.. RH are dicks, and
don't like using their tech. lol
649[09:18:26] <hwpplayer1> tarzeau_: I know I am teaching the
K8s to focus on it's own Debian instances, trying him to
understand Debian
650[09:18:29] <Lord_Devi> I went to the k8 homepage, and their
instructions.. IMMEDIATELY dived into instructions for building k8
clusters with Debian. That was a great sign to me.
651[09:19:05] <Lord_Devi> Just found out today Docker K8's
are a no-go though. I wasn't aware of that.
652[09:19:13] <Lord_Devi> Docker Pro? No thanks. Podman it is I
guess.
653[09:19:21] *** Quits: silverwhitefish (~hidden@replaced-ip) (Quit: One for all, all for One (2 Corinthians 5))
654[09:19:25] <hwpplayer1> Google suggests or offers in their
shell Debian as i know
655[09:20:11] <hwpplayer1> In short you can run Debian for
anything Lord_Devi
656[09:20:47] <hwpplayer1> tarzeau_: What do you think about the
cryptsetup issue above ?
657[09:21:01] <hwpplayer1> Do you know any forensics team for
Debian ?
658[09:21:34] <Hash> I'm working on a debian set of meta
packages
675[09:24:20] <Hash> you just need debian, and some meta
packages to install said categorized tools.
676[09:24:22] <hwpplayer1> Debian is enough for your needs just
apt install some tool
677[09:24:24] <Hash> easy peasy
678[09:24:34] <Lord_Devi> I hear you about Kali though.
679[09:24:40] <Hash> IT's bloated bloatware
680[09:24:42] <Lord_Devi> Ah yeah meta packages. Like
build-essentials?
681[09:24:50] <Hash> I want debian menu to show these programs,
no matter the wm.
682[09:25:00] <Hash> I use Xmonad only.
683[09:25:04] <Lord_Devi> Sorry, it has been a while since I
loaded up Debian. Refreshing myself on it right now.
684[09:25:16] <Hash> meta packages are those that depend on
other packages. Meta means about.
685[09:25:37] <Hash> They are packages about packages, that
install other packages. Or a collection fo packages by installing
just one.
686[09:25:38] <Lord_Devi> Yeah Xmonad is nice. I don't
haskell though. But I used Xmonad for a few weeks on my laptop. It
was actually the very first tiling window manager I ever used, and
fell in love INSTANTLY.
687[09:25:46] <gry> nice
688[09:25:52] <Hash> you amek a meta package (*or jsut a
package) and just put in whatgever it should dependon
689[09:26:06] <Hash> Then you install that pckage, and
everything it depends on wil be auto isntalled.
690[09:26:18] <Hash> You can consider them tasks kind of like,
in tasksel, but not really
693[09:26:26] <Lord_Devi> Yeah. Good to know there are cool
pentest meta packages to look at. Also, sounds incredibly useful to
perhaps just make a few of my own.
694[09:26:33] <Lord_Devi> TASKSEL!! I remember that
697[09:27:40] <Hash> I'm doing my 2nd bachelors in science
these days
698[09:27:48] <Hash> Cybersecurity this time, first was comp.
sci
699[09:28:42] <Lord_Devi> Oh very cool. There is going to be an
insatiable demand for security experts. The ransomware is spreading
fear amoung the corps like crazy.
700[09:29:23] <Lord_Devi> Plus it is fun. I may be wrong about
this, but I think Cyber Security is as yet, the only real IT
profession that is nearly purely gamefied.
701[09:29:56] <Lord_Devi> You literally build your resume by
playing security wargames. Way cooler than building a resume by
having HR count your lines of codes and number of commits.
707[09:32:16] <Lord_Devi> Yes it encompasses all the rest of the
IT professions.
708[09:32:17] *** Quits: HeXiLeD (~grumpy@replaced-ip) (Remote host closed the connection)
709[09:32:54] <Hash> You need to know multiple disciplines in
order to be successful at it. You need to know networking, software
dev/programing for software/buffer/etc exploit, you need to be able
to use various OS comfortably, use various OSes command line tools
and internals in depth, sytem admin work, settign up services and
sercuring/hardening them, penetration testing, social engineering
etc. etc.
711[09:34:04] <Lord_Devi> Yeah, "I don't like Mac, so
i'm not going to touch it." is not something a security
expert gets to say. About any OS really.
712[09:34:47] <Hash> yup
713[09:34:52] <Lord_Devi> Lord, I watch some of these videos of
people walking through some of those pico challenges and wargames.
It is CRAZY the obscure knowledge these guys need to get through the
challenges.
714[09:34:56] <Hash> no matter how annoying you find an os, you
gotta know abou it
715[09:35:10] <Hash> Anyway we're offtopic.
716[09:35:22] <Lord_Devi> hah, yeah. ##security I guess.
717[09:35:33] <Lord_Devi> Didn't know about that. I'll
poke my headi n later. Thanx
729[09:50:54] <TheBigK> i ended up reinstalling the system. the
dev needs to work. we gonna do a header backup now... that we can
recover for next time
730[09:50:58] <TheBigK> but i dont know what happened
739[09:54:18] *** Quits: bigpresh (~bigpresh@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
740[09:54:19] *** Quits: Swant (swant@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
741[09:54:19] *** Quits: nicole (ilbelkyr@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
742[09:54:19] *** Quits: tomaw (tom@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
743[09:54:19] *** Quits: JonathanD (~JonathanD@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
744[09:54:20] *** Quits: kloeri (~kloeri@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
745[09:54:20] *** Quits: mniip (mniip@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
746[09:54:20] *** Quits: WrathOfAchilles (unit193@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
747[09:54:21] *** Quits: kline (~freedom0@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
748[09:54:21] *** Quits: Stx (~stx@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
749[09:54:21] *** Quits: Unit193 (ukikie@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
750[09:54:21] *** Quits: mquin (~mquin@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
751[09:54:21] *** Quits: niko (~niko@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
752[09:54:22] *** Quits: jess (jess@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
753[09:54:23] *** Quits: Sigyn (sigyn@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
754[09:54:23] *** Quits: eir (bot@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
758[09:54:39] *** Quits: ChanServ (ChanServ@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. If you
care about your data, drop your NickServ account NOW before that
happens.)))
762[09:54:48] *** Quits: grumble (~Thunderbi@replaced-ip) (Killed (grumble (My fellow staff so-called 'friends'
are about to hand over account data to a non-staff member. I'm
leaving freenode forever and I recommend you drop your NickServ
account to keep your data safe.)))
774[09:56:27] <Hash> I think he's fine. A few sentences of
venting is fine. Shit happened. It's natural. So I didn't
follow your issue, but what happened? You upgraded something? Or
when did yo notice disk was inaccessible?
777[09:57:02] <TheBigK> from one day to the other... he tried
older kernel as well... we tried live system. his cryptsetup
password did not get accepted
778[09:57:25] <Hash> Something must have changed on the system.
779[09:57:36] <Hash> Any package updatges or forigen repos?
780[09:57:41] <Hash> Audit everything you can
781[09:57:42] <TheBigK> how should i check when i cant access
the device without the password :D
782[09:58:01] *** debhelper sets mode: +l 1010
783[09:58:08] <Hash> The man should know if he did anytning to
the system
826[10:02:59] <TheBigK> we r not in north korea here... he would
have admitted it... trsut me
827[10:03:13] <Hash> you can't say for certain anything
about another's mind.
828[10:03:26] <Hash> Hel
829[10:03:34] <Hash> One can rarely ever even be certain of
one's own mind.
830[10:03:36] <Hash> ffs
831[10:03:38] <Hash> :)
832[10:03:49] <TheBigK> true... i mean... its not the biggest
deal in the end of the day... annoying surely... but not a big
deal.. everything he was working on, was checked in....
833[10:03:55] <Lord_Devi> I've totally nuked data in the
past. Owned up to instantly. Kept that client for years.
834[10:03:57] <another> Hash: Don't read my mind, please.
877[10:21:14] <Hash> And about the recently oper kills, thre was
this gy in ##security claiming he had 0 days aginat this very ircd
and he was going to use them if the ops iddn't take his
findings seriously
878[10:21:24] <Hash> he said he would prove it. I wonder
what's happening.
879[10:21:47] <TheBigK> #freenode is the right place i would say
919[10:58:25] <queip> channels might need to move to other
ircnet, esp if they are oriented around various aspects of free
networks. I guess debian has OFTC #debian, so lucky for us
931[11:07:43] <SymbioticFemale> hey everyone. theres a dumpster
fire going on in freenode right now. if you have debian support
questions or some copper wire to burn the insulation off of, come
join the dumpster fire.
932[11:07:51] <SymbioticFemale> sorry. in #freenode
933[11:08:29] *** EnchanterTim is now known as Hash
964[11:24:18] <Hash> OFTC debian is official for irc.debian.org
and will remain
965[11:24:32] <Hash> Makes no difference, as all official debian
channels and project is there.
966[11:25:27] <shtrb> Odin (OFTC) would make the benches ready
for a glorious feast
967[11:25:28] <Hash> Should something happen to freenode, there
is always irc.debian.org #debian and all official channels. People
will simply start joining there.
968[11:25:53] <Hash> They should certainly get ready to expect
more users perhaps. Who can say
1040[12:45:15] <funabashi> Hi is there any lightwight debian
version? i dont need any GUI.
1041[12:45:30] *** Quits: Iamthehuman (~noname@replaced-ip) (Quit: reboot for upgrades)
1042[12:46:41] <ratrace> funabashi: yes. just don't install
one during.... installation.
1043[12:46:50] <shtrb> or uninstall
1044[12:47:28] <ratrace> more precisely, don't _select_ the
(Default (GNOME)) GUI during installation .. or uninstall which may
be a bit more finnicky to do
1113[13:46:27] <Lope> any recommendations for chunk-size for raid
0 of NVME?
1114[13:46:31] <Lope> 3x nvme drives
1115[13:48:57] <ratrace> Lope: depends on your file sizes
histogram
1116[13:49:16] <jelly> about yay big
1117[13:49:20] <ratrace> Basically you want your most sized files
to fit in 3 chunks
1118[13:50:01] <ratrace> 3 chunks *or more
1119[13:50:37] <ratrace> you don't want less. so for example
if you use chunk size of 64k, any file smaller than that won't
be spread in 3 stripes. only files 3x64k or bigger would.
1120[13:51:33] <ratrace> I use 8k for my steam games 2-way raid0,
as an example.
1121[13:51:38] <Lope> ratrace, I'm confused. I thought mdadm
--level 3 will spread data across the 3 devices so that any block
stored on the md, will consume space across all 3 devices?
1122[13:52:13] <ratrace> Lope: yes but how do you think
it'll make 3 stripes for chunk size N if the file is smaller
than 3xN ?
1123[13:52:13] <Lope> if I specify --chunk=12K does that mean
it'll put 4K on each device, resulting in a 12k chunk that I
should format my filesystem with?
1138[13:55:02] <Lope> How about 8K, that way it should work well
on my NVME SSD's which might have 8K blocks internally.
1139[13:55:06] <shtrb> ratrace, that the problem I don't get
an error (in any log I could find) I do ssh -X , do "flatpak
run com.skype.Client" no error presnted but no drawing of X
locally. I had even minimized it to local machine only (to be sure
it's not a firewall or something like that issue ) so locally
ssh -X differentuser@localhost would still fail
1140[13:55:18] <Lope> ratrace, I'm going to be storing large
files, videos.
1149[13:56:11] <ratrace> Lope: I did benchmarks. the biggest
difference was for file size fitting in the full stripe width. for
8k chunks that was 16k or larger
1150[13:56:40] <ratrace> but between 8k or 64k or 512k, the
difference was negligible in my tests in fact, coutner intuitively,
it seemed as if smaller chunkss were faster
1151[13:57:44] <ratrace> (16k or larger, for a 2-wide raid0, thus
2x8k)
1152[13:57:47] <jelly> also you probably want to avoid giving
SSDs too many tiny writes, even if their firmware and caching and
write blocks are supposed to Deal With It
1153[13:58:40] <ratrace> that too. eraseblock sizes matter
1154[13:58:51] <jelly> SSDs have like 1M-4M blocks
"internally"
1155[13:59:16] <Lope> jelly, interesting
1156[13:59:59] <ratrace> Lope: one thing though .... raid0 can be
_slower_ than other levels, depending on the use case
1160[14:01:42] <ratrace> so for a video hosting use case, it may
happen that you're better off with a mirror, if performance is
what you seek, assuming you can spare the space, if you have
multiple processes contending for reads. it's not as clear cut,
you should benchmark as close to your use case as possible
1161[14:03:38] <ratrace> the reason is relatively simple. A
single read transaction will engange the full stripe width of disks,
whereas for mirrors, a single reader process might use one disk,
another will use anotehr disk and yet another will use the third.
1165[14:04:24] <ratrace> so 3-way raid1 can satisfy 3 readers
independently, wheree a 3-wide raid0 can satisfy 1 reader for same
drive engagement, bandwidth and iops per drive.
1166[14:05:05] *** Quits: Haudegen (~quassel@replaced-ip) (Quit: Bin weg.)
1167[14:05:28] <ratrace> (writes are the opposite: single writer
will concurrently hit 3 disks for 1/3rd of bandwidth of single,
compared to mirror)
1174[14:06:18] <Lope> ratrace, I'm so lost reading about
chunk size
1175[14:06:29] <Lope> I find it very weird that the default is
512K
1176[14:06:30] <ratrace> I was too, so I ran tests :)
1177[14:06:31] <Lope> Seems so insane?
1178[14:06:34] <ratrace> it is
1179[14:06:37] <Lope> hahaha good idea.
1180[14:06:41] <Walex> ratrace: that comparison between 3-wide
RAID0 and RAID1 is way, way too simplistic, it depends a lot on
access patterns, stripe size, rotational latency.
1181[14:06:44] <Lope> I suppose tests are the only way.
1182[14:07:05] <ratrace> Walex: correct. my point was that
raid0==faster is not necessarily true, depending on use case
1183[14:07:07] <Walex> Lope: even tests can be misleading in the
general case
1184[14:07:25] <Lope> Walex, right, so I'll test with my
specific workload
1185[14:07:36] <Lope> Okay, will start with 8K for shits.
1186[14:08:07] <Walex> ratrace: thart is well expressed.
1187[14:08:26] <ratrace> however, the way mdadm works, it is my
experience, maybe biased, that it's PID related. btrfs
definitely is PID related (single PID will NOT balance and engage
multiple disks on >1 async depth)
1191[14:10:29] <Walex> ratrace: IIRC MDRAID is not PID related
but the block subsystem and elevators may introduce similar
behaviour.
1192[14:11:27] <kline> libera.chat is open btw
1193[14:11:45] <ratrace> yeah I did not dig deeper, just
observing differences between mdadm+ext4, btrfs and zfs on same
2-way raid0 and 2-wide raid1 examples
1194[14:11:51] <qrpnxz> did you guys see the global ann from
jess?
1196[14:11:59] <ratrace> (the other way around, 2-way raid1,
2-wide raid0)
1197[14:12:48] <ratrace> qrpnxz: so it seems the Fat Lady is
gurgling holy water and doing final vocal exercises before climbing
the stage....
1198[14:13:04] <Walex> Lope: this is a discussion of chunk size
and latency:
replaced-url
1199[14:13:14] <qrpnxz> lmao sorry geez
1200[14:13:18] <ratrace> whoops that ..... came out wrong. that
was in relation to previous "ain't over until the fat lady
sings" commentary about Freenode death
1201[14:13:28] <qrpnxz> kline, thx
1202[14:13:50] <qrpnxz> ah ok :)
1203[14:14:46] <ratrace> time to wipe out the nickserv profile
.... see y'all on OFTC when this one implodes
1262[15:06:41] <sappheiros> goodbye >_> since official
debian is on OFTC
1263[15:06:51] <sappheiros> thanks for your help
1264[15:06:52] <ratrace> OFTC is OFTC, Debian has official
channels on OFTC; this on FReenode is unofficially official;
libera.chat is supposedly a fork of Freenode by ops who are
resigning now en masse
1346[15:44:24] <shtrb|work> ratrace : -ChanServ- [#debian]
Welcome to #Debian. This is a discussion channel; if you have a
question about Debian GNU/Linux, ask and we will try our best to
answer it. Newcomers should read the channel's guidelines by
typing "/msg dpkg guidelines". Please do not paste in the
channel; use #flood instead. Thank you.
1370[15:52:12] <imMute> shtrb|work: irc.libera.chat is the IRC
network. libera.chat seems to be using GitHub Pages' CDN, which
honestly isn't a terrible idea until they get server sponsors
sorted out
1371[15:52:12] *** Quits: kopiyka (~user@replaced-ip) (Remote host closed the connection)
1372[15:52:35] <shtrb|work> Oh that explain why I can't
connect :D
1373[15:52:43] <ratrace> imMute: no the actual IP address of the
IRC server is in Github's ASN
1465[16:45:50] <Strum> one webpage says they couldn't get it
to run either and I also found a forum post saying they had the same
error with no replies
1466[16:46:16] <Strum> othanks anyway ratrace
1467[16:46:39] <ratrace> Strum: seen the linuxbabe article?
1468[16:46:46] <Strum> yes
1469[16:46:55] <ratrace> then I have nothing more to add, sorry
:)
1480[16:48:27] <jelly> it's a debug tool showing system
calls. Need to install it first. if you're doing it for the
first time ever you're probably not going to figure out much
1509[17:03:38] <LunarAgent> LunarIRC (irc.lunarirc.net) - the
friendliest IRC network around + free BNCs! We welcome retrogamers,
coders and wizards. Drop by, feel at home, create and re-create
channels and enjoy your stay! LunarAgent, over and out.
1523[17:10:12] <aecalaz> Hi! I have a problem with the video
drive. Using i915 firmware on Debian Buster and kernel 4.19.0-16.
Using chrome mainly, but also using gnome-shell. The error message
on dmesg is "[drm] GPU HANG: ecode 7:0:0x...."; i saw this
error on other distros (Ubuntu and Fedora). I tried upgrading the
kernel; but it failed. Some idea about it?
1527[17:11:37] <sney> aecalaz: there are some parameters you can
pass to i915 to change behavior, some are for working around
specific gpu quirks and may help you. see the list of parms at the
bottom of 'modinfo i915'
1528[17:11:53] <sney> however, if it happens with any kernel in
any distro it might also be a hardware defect.
1532[17:13:13] <sney> there are some more details here as well.
replaced-url
1533[17:13:37] <ratrace> Strum: snap is a containerized software
package manager like flatpak. snapstore is the (only) central hub
for it.
1534[17:14:18] <ratrace> Strum: for software that's not
regularly packaged as .deb in the repos, installing via snap or
flatpak is a very good and functional alternative to installing
random tarballs manually
1535[17:14:43] <aecalaz> @sney thanks; I had this problem before,
on a different computer... and works correctly under Windows and
FreeBSD.
1536[17:14:43] <aecalaz> This computer was working perfectly with
Ubuntu a week ago; I switched it to Debian and the problem persists.
I'll check that info!
1537[17:14:46] <aecalaz> Thanks again!
1538[17:15:10] <sney> np
1539[17:15:22] <Strum> i'll have a look at it i guess,
thanks ratrace
1577[17:34:12] <Strum> was expecting to get flamed and called a
n00b for asking
1578[17:35:14] <ratrace> nah. we only flame questions that are
hyper-lazy-i-couldn't-even-bother-to-ask-google :)
understanding snaps is a bit more involved ;))
1579[17:35:19] <Strum> was using arch before debian
1580[17:35:37] <Strum> and got a lot of that sort of reaction in
their channel
1660[18:34:52] <GenTooMan> chat not check sigh my hands are doing
other things :D
1661[18:35:33] <ddsys> or libera?
1662[18:36:04] *** Quits: dselect (~dselect@replaced-ip) (Quit: ouch... that hurt)
1663[18:37:22] <somiaj> !oftc move
1664[18:37:22] <dpkg> irc.debian.org moved to OFTC on June 4th
2006, see
replaced-url
1665[18:38:01] *** debhelper sets mode: +l 1031
1666[18:38:17] <somiaj> Debian's offical irc network is
oftc, and has been for some time. This channel is the only one left
here to provide users support.
1668[18:38:32] <cybrNaut> how might i list all debian packages
that I installed, but filter out all the dependencies that installed
automatically as a result of my command? "dpkg -l" lists
/all/ pkgs and the manpage only shows how to filter by name
1669[18:38:33] <ratrace> just reg'd and enabled cert SASL
for OFTC.
1681[18:43:57] <somiaj> aptitude search '~i !~M' --
also works, and once you are on bullseye (apt 2.0), you can use
search flags with apt list '~i !~M'
1693[18:47:18] <teratorn> somiaj: I'm not sure what version
that is, but it's probably newer, and probably yes it would
suite my needs. I just like to be on latest release of plasma, which
isn't always easy unless you're on ubuntu PPAs
1694[18:47:22] <somiaj> unixbsd: please keep your off topic noise
to yourself, that isn't debian support.
1695[18:47:39] <teratorn> what is in debian 10.9 would probably
suite my needs too...
1696[18:47:41] <somiaj> teratorn: bullseye is almost ready to be
released (next month or two), for a desktop system it is reasonable
to upgrade to.
1697[18:47:47] <unixbsd> stats showd that tthe linux poppulation
uses at 70-80 pct Ubuntu, especially developers.
1700[18:48:15] <ratrace> 98.67% of statistics is made up on the
spot
1701[18:48:22] <somiaj> teratorn: You can upgarde now, though
there are a few rc-bugs left and slow security support until the
release actually happens. For a desktop system it is reaonsable and
many are already using it.
1702[18:49:03] <unixbsd> ratrace: the rule of stats, higher n,
the most accurate.
1703[18:49:04] *** wyatt8750 is now known as wyatt8740
1704[18:49:04] <somiaj> teratorn: so if you want a newer kde now,
upgrading isn't unreasonable (though it isn't offically
released yet, but it has been frozen for a while). Or you could just
make do with 10.9 until the release then upgrade then.
1713[18:53:06] <teratorn> well I probably won't miss
anything major
1714[18:54:03] <somiaj> teratorn: Yea, debian stable is a frozen
release, and part of its stability is not tracking upstream. I know
some who just run testing/sid to try to keep up to date (though
there is still a lag to create the debian packages).
1717[18:55:10] <somiaj> teratorn: Note that debian's
philosophy isn't to provide newest software in its stable
releases, but to provide frozen well tested software. Something to
consider when choosing a distro. You could also choose to just build
it yourself from upstream, though here again, desktops are big
collections of code and this takes a lot of work (hence why there is
a lag even in unstable to get the newer kde)
1718[18:55:57] <somiaj> teratorn: note running testing/unstable
too has its issues, I use to run sid, but now only stable, as I find
I don't really need shiny new stuff
1719[18:56:01] <teratorn> yeah, that is a TON of C++ etc to
compile :/ and I have no idea just how much tweaking of upstream is
done in debian package patches
1721[18:57:21] <unixbsd> tthe kde libs are quite dirty, you
cannot compile modern QT like that., apt-get install * and run ;
automake ;) joke
1722[18:58:00] <somiaj> teratorn: well my suggestion is just use
the version in debian stable, and about every 2 years get to
upgrade. Or maybe run stable for most of its release then near the
end upgrade to testing as it starts to freeze and just deal with
limited security support.
1723[18:58:06] <teratorn> yeah, generally I prefer the older,
stable, more-well-tested releases of most software... but I /like/
Plasma, and so prefer to be on latest release :/ if only to make bug
reporting more relevant
1724[18:58:37] <somiaj> teratorn: yea, desktops are just very big
and have a messy dependency tree. I haven't seen them
backported to stable.
1725[18:58:38] <unixbsd> QT was commercial and is still
commercial. it is not made to be nice code, but to give a nice
desktop
1726[18:59:09] <unixbsd> QT 1 was easy to compile in comparison.
See artic
1796[19:56:33] <queip> cp: oops IO error! what if I try again?
oops IO error! what if I try again? oops IO error! what if I try
again? oops IO error! what if I try again? oops IO error! what if I
try again? oops IO error! what if I try again? oops IO error! what
if I try again? oops IO error! what if I try again?
1797[19:56:41] <queip> mc copy: oops IO error! ok moving on, lmao
1846[20:28:27] <cybrNaut> I cloned my drive. The UUIDs are
different (which is expected when using rsync to mirror). I ran
"chroot /stagingmirror grub-mkconfig" then "chroot
/stagingmirror update-grub"
1847[20:28:40] <cybrNaut> when it boots, GRUB still looks for the
old UUID
1848[20:29:02] <ratrace> cybrNaut: did you prepare the chroot
with appropriate mounts?
1849[20:29:35] <cybrNaut> but when I look in
/stagingmirror/boot/grub/grub.cfg, the UUIDs are correct (no mention
of the old UUID)
1850[20:29:53] <cybrNaut> ratrace: yeah.. did the rbind mounts
1851[20:30:09] <ratrace> sounds like you missed some. do you have
a separate /boot?
1852[20:30:31] <cybrNaut> e.g. for i in /run /tmp; do mount
--bind $i /stagingmirror$i; done && for i in /dev /proc
/sys; do mount --rbind $i /stagingmirror$i; done
1873[20:36:50] <ratrace> cybrNaut: I'm not 100% sure if
install-grub doesn't also prime the stage1 for location of
/boot/grub/grub.cfg to use; so you might want to do the grub-install
from the chroot as well
1874[20:36:54] <cybrNaut> i remove the original drive, so the
target drive is the sole drive in the machine.
1879[20:38:52] <Oberon> okay this isn't explicitly Debian
related, but... does anyone know what key combinations you use on
Putty to send a literal 0x0D, 0x0A, and 0x1A characters? Those are
carriage return, new line, and (sub), whatever that is
1880[20:38:53] <cybrNaut> this process actually worked for me,
but then I later had to re-rsync. I repeated the process in my
notes. Not sure why i'm getting this problem now, but according
to my notes i never ran grub-install within chroot
1881[20:38:54] *** Quits: Night-Shade (~TimF@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
1882[20:40:32] <ratrace> cybrNaut: okay so to be clear, what
exactly you mean by "grub still looks for the old UUID"
... stage1 loader is looking for config, OR rootfs's UUID?
1883[20:41:04] <jhutchins> Oberon: What platform are you running
putty on?
1901[20:50:01] <Oberon> Okay, I think I need to look at the
actual packets here, what's the packet capture tool whose name
I should be able to remember but can't because it's been
like 15 years?
1905[20:51:32] <ratrace> and you can save to a file with -w and
analyze later with wireshark (or you can use wireshark on teh same
machine)
1906[20:51:44] *** Quits: prahou (shvehlav@replaced-ip) (Remote host closed the connection)
1907[20:52:07] <Oberon> thanks
1908[20:52:24] <ratrace> Oberon: but uh... if that's over
ssh, you won't see anything in the packets. ssh is encrypted
unless you use the null cipher thingy
1910[20:52:49] <Oberon> it's not, it's plain
unencrypted HTTP on port 80
1911[20:52:59] <cybrNaut> ratrace: when grub boots, it gives an
error like drive <old UUID> not found
1912[20:53:05] <Oberon> I can see the error log but that's
not showing me exactly how the request is malformed, just that
it's bad
1913[20:53:26] <Oberon> I'm writing some code that's
going to run on a cellular modem, so it's like...
1914[20:53:35] <Oberon> ...yeah
1915[20:53:51] <Oberon> gotta look at the actual packets to
figure out what's wrong, apparently
1916[20:54:04] <ratrace> what kind of low level stuff are you
even doing there for http?
1917[20:54:18] <cybrNaut> ratrace: and i find that bizarre,
because when i examine /boot/../grub.cfg, there is no mention of the
old UUID
1918[20:54:21] <Oberon> making a.... thing... that will...
1919[20:54:27] <Oberon> it's complicated
1920[20:54:49] <Oberon> in the end there will be a piece of
hardware that's small and has an embedded cellular modem which
will interact with a server via HTTP
1921[20:55:03] *** Parts: ricorambo (~rico@replaced-ip) ("Using Circe, the loveliest of all IRC clients")
1922[20:55:28] <ratrace> Oberon: there's small http
libraries even in C
1923[20:56:03] <ratrace> cybrNaut: can you post the _exact_
verbiage or imgur a screenshot?
1924[20:56:14] <Oberon> I'm sure, but I don't know how
to get one of those libraries onto this piece of hardware
1925[20:56:24] <Oberon> or if it's even possible
1926[20:56:30] <cybrNaut> ratrace: i'll work on that.. will
take some time
1951[21:08:07] <cybrNaut> ratrace: yes. It's short, so
i'll just type it. Exact output is: "error: no such
device: 12345-..." \n "error: unknown filesystem." \n
"Entering rescue mode..." \n "grub rescue> _"
1956[21:09:46] <cybrNaut> i have no idea what could cause it to
look for the old UUID because that UUID does not appear in
/boot/grub/grub.cfg of the new target
1957[21:10:23] <alkisg> cybrNaut: I just glanced at the last
lines so I might be out of topic, but grub also checks the UUIDs of
/etc/fstab
1984[21:16:36] <ratrace> cybrNaut: so that looks like inability
to find /boot .. the UUID in question is of the /boot filesystem?
1985[21:17:20] <ratrace> cybrNaut: is there EFI involved or is
this purely a legacy boot (gpt or mbr?)?
1986[21:18:47] <cybrNaut> ratrace: the UUID it looks for is /boot
of the previous filesystem. But /boot/grub/grub.cfg correctly
references the UUID of the new /boot
2000[21:22:43] <somiaj> Achylles: I'm unsure on the current
state of what is happeneing (and maybe nothing will become of this),
but it could be that debian just no longer has channels on any
network except oftc (not clear idea yet, and it really depends on
how this all pans out)
2001[21:23:18] <cybrNaut> the partition table is a GPT partition
table, which /normally/ only works on UEFI, but I created a bios
partition which makes it possible to use GPT on a BIOS machine
2002[21:23:44] <cybrNaut> actually it's a partition of type
"bios boot"
2007[21:24:38] <somiaj> but you can always find debian channels
(and many more than just #debian), on the offical network
irc.oftc.net (which is what irc.debian.net points at)
2008[21:24:50] <cybrNaut> run this to see what i mean =>
sfdisk --label gpt -T | grep -i bios.boot
2009[21:24:52] <ratrace> cybrNaut: so you mean you have GPT
_layout_ and among GPT partitions there's the bios_grub one
(required for grub on GPT), the /boot one and the root fs?
2010[21:24:59] <BlueMatt> as someone with personal experience
with the new freenode owners outside of irc,.....yea run away
2015[21:25:51] <cybrNaut> ratrace: that's right, plus a 4th
for swap
2016[21:25:58] <ratrace> cybrNaut: and the old disk is nowhere
physically present on the system?
2017[21:26:09] <Achylles> somiaj, ok. But, I like this one.
That's why I asked...
2018[21:26:09] <somiaj> BlueMatt: freenode survived the crisis
that almost disolved it in 2006, so I wouldn't jump the gun too
quickly, I'm of the opinion to just wait and see what happens.
2019[21:26:14] *** Quits: Jerrynicki (~niklas@replaced-ip) (Remote host closed the connection)
2020[21:26:19] <cybrNaut> that's right.. i removed the old
drive
2021[21:26:22] <somiaj> Achylles: a lot of the users (espically
the ops) are in both.
2022[21:26:48] <somiaj> Achylles: though oftc isn't as
populated, it has a lot of good users and many of the main helpers
you'll see on oftc as well as here.
2023[21:27:11] <ratrace> cybrNaut: well that doesn't make
sense... unless grub-install bakes in the location of /boot
2024[21:27:20] *** Quits: Oberon (~Oberon@replaced-ip) (Remote host closed the connection)
2031[21:28:38] <ratrace> cybrNaut: yes, but like I said I
don't know if it also bakes in it the location of /boot, or
taht's 100% autodetection by stage1 . in which case, there HAS
to be traces of the old /boot somehwere on the drive physically
which the stage1 finds first
2043[21:31:13] <lenswipe> got a broken debian system here im
trying to reapir.
2044[21:31:32] <lenswipe> before i do that, i want to backup my
cron jobs so I want to try and view (not edit, just view) where they
live at rest on disk.
2045[21:31:42] <lenswipe> I've looked in /var/spool/ and
there's no cron folder
2046[21:31:46] <lenswipe> open to any other suggestions
2047[21:31:49] *** adminn is now known as zeroed
2048[21:31:55] <ratrace> cybrNaut: you can use `strings` on it,
yes
2049[21:32:13] <ratrace> assuming it's stored as ascii
string
2050[21:32:14] <mentor> Does the BIOS actually understand GPT?
2051[21:32:18] <cybrNaut> well, i grep'd the bios boot
partition for the old UUID and no hit.. but that doesn't mean
it's not there.. it could be encoded/compressed
2052[21:33:06] <ratrace> cybrNaut: but just to be 100% sure, did
you re-try grub-install from _within_ the chroot?
2053[21:33:11] *** Quits: Christian75 (~Christian@replaced-ip) (Remote host closed the connection)
2054[21:33:29] <cybrNaut> ratrace: i'll try that now
2056[21:34:28] <cybrNaut> it stands to reason that that could be
the issue (that i ran it outside of chroot). My notes don't say
to run it in chroot, but maybe my notes are wrong
2057[21:34:28] <ratrace> mentor: what do you mean? like, if
it's too old for GPT?
2058[21:35:33] <ratrace> cybrNaut: I've made hundreds of
installations on servers, I use a custom ansible and saltstack
deployment scripts with deboostrap, and always used grub-install
from within the chroot, so I don't know what would happen if
you tried from outside
2059[21:35:46] <mentor> A BIOS boot GPT partition was mentioned,
but not if a compatibility MBR was written for it or something
2061[21:36:37] *** Quits: mezzo (~mezzo@replaced-ip) (Quit: leaving)
2062[21:37:18] <ratrace> protective mbr should be done by
grub-install afaik
2063[21:37:34] *** Quits: Noisytoot (~noisytoot@replaced-ip) (Remote host closed the connection)
2064[21:37:43] <cybrNaut> ratrace: what i suspect happened is
perhaps i ran grub-install from within the chroot the 1st time when
shit worked, for good measure in addition to running it outside of
chroot.. and I may not have noted that I did that possibly critical
step. I'll know in a few min.. i'm off to reboot now
2077[21:47:16] <cybrNaut> I had to run grub-install from within
chroot to ensure the bios boot partition had a pointer to the UUID
of the correct /boot
2261[22:45:35] <ax562> ratrace reason I was asking. was thinking
most would jump ship.
2262[22:46:03] <ratrace> ax562: I'm not sticking around on
the Andrewnet on principle
2263[22:46:31] <cybrNaut> suppose i partition a drive manually
with LUKS partitions, some using keys and one using a password. Will
the Debian installer give me a chance to enter the password to mount
the encrypted partition, and then skip the partitioning step?
2264[22:46:36] <EdePopede> a mass leave would be nice while the
new dude is online
2269[22:47:29] <ratrace> cybrNaut: the partitioning step is still
used, you select existing ones (you manually managed) to point at
rootfs, boot, etc....
2270[22:47:44] <ratrace> cybrNaut: that'd be manual
partitioning step tho
2271[22:47:48] <imMute> EdePopede: the (now former) staff
regaining control of the domain names would be even nicer.
2272[22:48:01] *** debhelper sets mode: +l 1016
2273[22:48:03] <ratrace> imMute: hear hear. but eh...
2274[22:48:08] <EdePopede> that's why i'll stay here
for the moment...
2275[22:49:02] <cybrNaut> EdePopede: i haven't read about
the drama.. why is the new ownership getting hate?
2276[22:49:24] <EdePopede> this one has a lot of extra links
replaced-url
2277[22:49:52] <cybrNaut> ratrace: will it ask for password or
keys for existing encrypted partitions?
2278[22:50:10] <ratrace> cybrNaut: no, you must unlock them
first, if you manually prepared the LUKS containers
2279[22:50:23] <ratrace> it _will_ detect the mapped devices and
offer them to point at rootfs
2375[23:58:18] <sney> lenswipe: are you sure you are using the
right /dev node for your root filesystem? on my laptop it is not
/dev/mapper/crypt but a different name under /dev/mapper that
includes the hostname