111[00:23:42] *** Quits: Tobbi (~Tobbi@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
112[00:24:16] <jelly> adding repos for an older release does not
have huge stability issues, as versions are always monotonically
rising and it's hard to accidentally downgrade things
183[01:03:11] <Caesar_NayKid> I can have commands run.
184[01:03:23] <ratrace> nano /etc/fstab
185[01:03:24] <sney> there is no indication that you need to
edit fstab.
186[01:03:29] <Caesar_NayKid> My mom literally is the one at the
console.
187[01:03:46] <ratrace> and... what sney said. should pastebin
it to see why it's failing
188[01:03:49] <Caesar_NayKid> I tried to start ssh so i could
connect.
189[01:04:05] <Caesar_NayKid> Kicks it back to emergency mode
prompt.
190[01:04:14] <ratrace> dunno if journald has more clues if you
query the .mount unit
191[01:04:42] <Caesar_NayKid> That particular mount is a backup
hard drive i dont care about.
192[01:05:14] <ratrace> for non-critical mountpoints that may be
missing or broken on boot, but are in fstab, you should use the
"nofail" option in the fstab options columns
193[01:05:16] <Caesar_NayKid> If i could get booted in and
access it via ssh or NoMachine on the LAN i could probably take it
from there
206[01:08:32] <Caesar_NayKid> Not necessarily but it appears to
be just that particular drive. I maybe need the guid?
207[01:09:18] <Caesar_NayKid> I tried "mount -a"
earlier
208[01:09:20] <ratrace> Caesar_NayKid: you can post the fstab
entry for that mountpoint for us to see if anything jumps out
209[01:09:29] <Caesar_NayKid> Parse error in line 1
210[01:09:49] <ratrace> also see if journalctl -u
home-paul-BackupINT.mount has more clues (I don't know if does
for .mount units like it does for service units)
211[01:09:52] <Caesar_NayKid> Do i need to edit the fstab then
to look at it?
212[01:10:02] <ratrace> Caesar_NayKid: you can `cat` it
213[01:10:25] <ratrace> you can pipe it to termbin.com at port
9999 via netcat if you want a quick pastebin
214[01:10:26] <Caesar_NayKid> Cat /etc/fstab
215[01:10:41] <ratrace> eg. cat /etc/fstab | nc termbin.com 9999
and then post here the URL you get. requires netcat installed
216[01:11:17] *** nuala2 is now known as nuala
217[01:11:28] <Caesar_NayKid> Im not familiar with the pastebin
thing
222[01:12:36] <ratrace> Caesar_NayKid: a pastebin or a paste
site is a site where you post larger amounts of text or images and
get an URL for that, to share on IRC
223[01:13:03] <Caesar_NayKid> I think that's too complex
for my mom via internet skype session
224[01:13:06] <ratrace> common netiquette states you
shouldn't post more than 2-3 lines at once, thus teh paste
sites where you post configs, outputs, etc...
225[01:13:16] <Caesar_NayKid> I see.
226[01:13:27] <ratrace> well... I don't know how else we
can help. we need to see the contents of those files and you
can't post them here
254[01:22:41] *** Quits: Mister00X (~quassel@replaced-ip) (Quit: I'll be back! — Arnold Schwarzenegger)
255[01:22:43] <sney> Caesar_NayKid: comment out the bottom 3
lines, save and reboot, the system should come up. then you can
optionally mount those other volumes from the running system.
265[01:23:45] <Caesar_NayKid> Oh i see your comment now. This
irc app is terrible on iOS
266[01:23:50] <ratrace> but then ... I wonder what happened
there since the fstab lines _already_ have the nofail option
267[01:23:53] <johnjay> if i need a static ip for my wlan0 i may
need to put it somewhere else
268[01:24:07] <Caesar_NayKid> Comment out is # yes?
269[01:24:12] <sney> johnjay: if you configure your wlan0 in
/etc/network/interfaces, network-manager will no longer handle it.
270[01:24:30] <johnjay> ok. so if that's true then it
won't connect at all.
271[01:25:19] *** Quits: werneta (~werneta@replaced-ip) (Remote host closed the connection)
272[01:25:20] <sney> you will also need to put your wifi ssid
and psk in /etc/network/interfaces for it to work. the debian wiki
has instructions, and there is a lot of information in 'man 5
interfaces' as well.
273[01:25:21] <ratrace> johnjay: unless you set up
wpa_supplicant directly
323[01:40:52] <ratrace> Caesar_NayKid: speaking of, you have two
fstab entries for the same mount point. is that deliberate? they
can't both be correct at the same time
324[01:41:06] <ratrace> but then, they all have nofail so I
don't understand what kind of problem you even had there?
334[01:42:50] <ratrace> I suppose systemd choked on having two
fstab entries for the same mountpoint
335[01:43:14] <Caesar_NayKid> Probably
336[01:43:21] <Caesar_NayKid> Got another problem
337[01:43:27] <Caesar_NayKid> I updated NoMachine
338[01:43:32] <ratrace> if you intend to enable the mountpoints,
decide which one to remove, for /home/paul/BackupINT mountpoint
339[01:43:46] <Caesar_NayKid> Then it told me something about
the system id changed.
340[01:43:59] <Caesar_NayKid> So it wanted to relog me. Now it
failing
341[01:44:06] <Caesar_NayKid> I connected via ssh in Putty
342[01:44:13] <ratrace> Caesar_NayKid: Bullseye is still in
testing. Even though it's soft frozen, there's no
guarantees it won't break with an upgrade
343[01:44:23] *** Quits: gry (~kvirc@replaced-ip) (Client Quit)
344[01:44:37] <Caesar_NayKid> Any idea how to restart NoMachine
from a ssh command prompt?
345[01:44:50] <ratrace> so you (have) made two mistakes:
upgrading remotely, letting a non-tech user use debian testing with
you not nearby to fix
374[01:56:56] <dpkg> In buster, su no longer overrides PATH by
default, requiring that you use "su -" or "su
-l" for login shells (which is not really a new thing at
all...). See
replaced-url
405[02:16:35] <Caesar_NayKid> The nvme0 is my debian drive
406[02:17:06] <Caesar_NayKid> With separate partitions for /var
/tmp /home
407[02:17:12] <factor> I found the error that I get with Steam
./MarsSteam: /lib/x86_64-linux-gnu/libm.so.6: version
`GLIBC_2.29' not found. Can make a sym link from debian glibc
2.28
408[02:17:34] <factor> being this close I would imagine I can
fake a lib version
434[02:23:49] <ratrace> Caesar_NayKid: you can do that; that
will just try to mount _all_ entries defined in fstab
435[02:23:58] <ratrace> Caesar_NayKid: I'd suggest you
ratehr tried to mount just that one
436[02:23:59] *** Throwawayname is now known as AF04FB9290474265
437[02:24:18] <Caesar_NayKid> Can you walk me through that?
438[02:24:29] <Caesar_NayKid> In terminal?
439[02:24:29] <ratrace> Caesar_NayKid: BackupINT? just run
`mount /home/paul/BackupINT` or what was teh mountpoint. with a
fstab entry in place, you don't need full mount invocation
440[02:24:54] <factor> Now steam does ot work at all
441[02:25:01] <ratrace> factor: what did you do to it?
442[02:25:19] <factor> sudo apt-get install steam
443[02:25:29] <ratrace> factor: and what did you have _before_
that?
444[02:25:37] <factor> Downloaded steam
445[02:25:47] <ratrace> if you already had steam installed
somehow, chances are your ~/.steam is polluted
450[02:26:53] <ratrace> Caesar_NayKid: then if it's
mounted, and `ls -la` /home/paul/BackupINT shows empty dir,
that's it then
451[02:26:57] <factor> okay moved it and its loading the lcient
now
452[02:27:24] <Caesar_NayKid> That's what? The data's
gone you mean?
453[02:27:48] <ratrace> Caesar_NayKid: yes.
454[02:27:52] <Caesar_NayKid> Gotcha
455[02:28:01] <Caesar_NayKid> Wow.
456[02:28:23] <ratrace> Caesar_NayKid: or, to be precise, if
/dev/sda1 is mounted on /home/paul/BackupINT, and you ls -l that
directory, and it shows nothing, it shows the contents of /dev/sda1
filesystem
458[02:28:46] <ratrace> Caesar_NayKid: now, try to umount
/home/paul/BackupINT ... and run ls -l /home/paul/BackupINT
459[02:29:24] <ratrace> it's possible, and happens
sometimes, that a mount was borked, and user copies files thinking
it's copying to the mounted drive
460[02:29:27] <Caesar_NayKid> Well,.. crontab runs in like 30
seconds
461[02:29:42] <ratrace> while in reality it's copying just
into the directory that's supposed to be the mount-POINT, but
on the local fs
462[02:30:13] <Caesar_NayKid> The user meaning my linux
username?
463[02:30:24] <ratrace> I mean a user in general
464[02:30:31] <Caesar_NayKid> Because only rsync copies files
there
465[02:30:34] <ratrace> a person
466[02:30:40] <ratrace> or, okay, an automated script
467[02:31:09] <ratrace> again, point is, broken mount and the
copy is into the directory on the local system instead of a drive
attached to that particular directory, aka "mounted"
468[02:31:17] <Caesar_NayKid> So, what likely happened then, was
two UUIDs attempted to mount in the same location and broke the
mount.
490[02:37:53] <Caesar_NayKid> Where are those saved?
491[02:38:28] <ratrace> on whatever filesystem that directory is
part of, when nothing is mounted on it. rootfs I suppose unless you
have /home/ on another partition
492[02:38:50] <ratrace> oh wait, you do, btrfs right?
493[02:39:01] <Caesar_NayKid> Yes just for that
494[02:39:14] <mutante> I think this is part of the reason why
there is /mnt for mount points
501[02:42:00] <Caesar_NayKid> So, you're saying the files
got moved off of this drive and moved into my other drive? Or they
never were on the seccondary drive in the first place?
502[02:42:13] <Caesar_NayKid> Linux drives and partitions
confuse me.
503[02:42:31] <Caesar_NayKid> Too many years being a DOS/WinN00b
504[02:42:59] <Caesar_NayKid> My disk will fill up fast this way
if it's wrong
529[02:52:51] <ratrace> Caesar_NayKid: then mount that to, say,
/mnt/backups (and stay within /mnt/ dir for future mounts, it's
cleanest)
530[02:52:53] <factor> redownloading everything
531[02:53:14] <factor> I only buy native games.
532[02:53:39] <ratrace> factor: chances are you don't have
to do that. The reason Steam is __still__ using ubuntu 12.04 runtime
is that it's so damn API stable by now, all the linux native
games are almost guaranteed to run
533[02:53:41] <Caesar_NayKid> So sudo mount /dev/sda1
/mnt/backups
534[02:53:57] <ratrace> factor: however ... it's possible
your gpu drivers aren't fully compliant
535[02:54:06] <factor> ahh okay
536[02:54:08] <ratrace> Caesar_NayKid: yes, assuming you want
sda1 mounted
538[02:54:18] <somiaj> factor: what gpu/drivers are you using?
539[02:54:28] <Caesar_NayKid> Oh wow. It doesn't really
need to be mounted?
540[02:54:44] <ratrace> factor: if that's nvidia, make sure
you're using the latest driver from buster-backports. the
buster native 418 is having issues
541[02:54:49] <Caesar_NayKid> What's the point of mounting
if it doesn't need to be?
542[02:54:59] <ratrace> Caesar_NayKid: what do you mean?
543[02:55:09] <Caesar_NayKid> Oh you said if i wanted it mounted
544[02:55:56] <factor> I download and compile my own. to lng
term kernel currentl 5.10.32
545[02:55:57] <ratrace> Caesar_NayKid: yes, I don't know
WHICH drive you want moutned. if that's sda1, then that's
teh command you run, yes.
546[02:55:57] <ratrace> it's better to use UUIDs so
you're sure. EVEN BETTER.... use LABELS
547[02:55:57] <somiaj> factor: you manually install nvidia
drivers from nvidia.com?
548[02:55:57] <factor> somiaj, yes
549[02:55:59] <ratrace> ugh
550[02:56:23] <Caesar_NayKid> I may have to stop for the night.
551[02:56:24] <factor> Yes Surviving mars is not working either.
553[02:57:01] <ratrace> factor: tehre's stuff in
~/.steam/debian-installation/logs/ see if there's hints as to
what's not working
554[02:57:04] <Caesar_NayKid> The drives/mounting/directory
concept on this seems amazingly flexible but it mindblowing for me
at this point still.
555[02:57:06] <somiaj> factor: that could be the issue, debian
will conflict with those. Why not use the 5.10 kernel from backports
and the nvidia drivers from backports?
557[02:57:49] <Caesar_NayKid> I guess i thought i was mounting
folders to drives but maybe im mounting drives to directories
558[02:57:49] <ratrace> Caesar_NayKid: you can also do that on
windows these days, use arbitrary directories as moount points,
instead of drive letters
559[02:57:51] <somiaj> factor: debian can't really support
that setup, since dpkg packages will conflict with what
nvidia.com's self installer does, and due to that you could
have strange conflicts (if this is actually your issue is hard to
say)
560[02:58:25] <factor> somiaj, oh okay will see will remove this
version for now and put my version back in
561[02:58:27] <ratrace> Caesar_NayKid: on unix/linux you
_always_ mount block devices (Drives, partitions, mapped devices,
....) to _directories_
562[02:59:19] <ratrace> factor: the conflict is possibly in
xorg's opengl modules. nvidia installs its own liblg files, but
the packages manage all that transparently
563[02:59:35] <ratrace> I'd recommend purging your manual
driver, and apt install it from backports. it's relatively
latest
564[02:59:36] <somiaj> factor: debian does include a
nvidia-installer-cleanup package to help clean up what the nvidia
package does. My experience is you get intersting issues if you try
to use nvidia.com's installer and not debian's package
with nvidia's non-free binaries. And this is due to
debian's package do a bunch of other thigns to make the driver
play nicely with other gl systems (such as masa)
567[02:59:52] <Caesar_NayKid> Ok. Thank you for your time. I
will read some about that. Help that guy with his game stuff. I may
return later this week after i do a little research on the
drives/filesystem stuff. Certainly there is endless articles and
videos about that. You guys saved me to get booted again. i really
thank you
568[03:00:25] <dvs> Caesar_NayKid: not necessarily filesystems
as it is mounting.
569[03:00:50] <Caesar_NayKid> Well. All of the above i need to
do a refresher course.
570[03:01:21] *** Quits: OPK (~OPK@replaced-ip) (Remote host closed the connection)
576[03:06:54] *** Quits: Numero-6 (~Numero-6@replaced-ip) (Quit: << - Qui etes vous ? - Je suis le nouveau numero 2 -
Qui est le numero 1 ? - Vous etes le numero 6 - Je ne suis pas un
numero ! Je suis un homme libre!! >>)
593[03:20:17] <somiaj> uxfi: are you running buster?
594[03:20:24] <uxfi> yeah
595[03:20:26] <uxfi> the latest debian
596[03:20:26] <dD__> I am trying to set an environment variable
on a server I have. The server has a log in banner and I want to
remove it. It's bitnami so I can't just remover it the
normal way, but I need to set the environment variable before the
log in prompt. How could I do this?
597[03:20:28] <somiaj> uxfi: and this is just a stndard
secuirty/regular upgrade.
605[03:22:17] <somiaj> uxfi: unsure here, I'm not quite
sure where the problem would be. Wonder if something in
/var/lib/dpkg/info/openjdk-11-jre-headless\:amd64* is corrupt
causing this usse.
606[03:22:18] <mutante> uxfi: I wonder what you get when you run
'apt-cache policy openjdk-11-jre-headless'
607[03:23:12] <somiaj> uxfi: arg my system is amd64, but
hopefully you see those files, maybe you can just add the final new
line to any of those files if one is missing
609[03:23:56] <uxfi> N: Unable to locate package
openjdk-11-jre-headle
610[03:24:15] <uxfi> ah mine is a arm system
611[03:24:21] <uxfi> well the one im testing debian on
612[03:25:33] <mutante> uxfi: do you know how it got installed
in the first place?
613[03:25:35] <somiaj> uxfi: you are missing two ss, it should
be openjdk-11-jre-headless
614[03:26:12] <uxfi> I assume with a update
615[03:26:25] <uxfi> so
616[03:26:28] <uxfi> should I pruge openjdk-11-jre-headles?
617[03:27:12] <somiaj> uxfi: in buster openjdk-11-jre-headless
is in the armhf repository, so the fact it isn't seeing it says
something
618[03:27:30] <mutante> an update should not pull new packages
619[03:27:39] <somiaj> uxfi: according to your error you
didn't state the name of the package correctly, if it
can't see the package, maybe just share the output of 'apt
cache policy' with us. But it should see that package.
620[03:27:54] <mutante> was it installed with a manual dpkg -i
or something?
621[03:28:41] <uxfi> according to your error you didn't
state the name of the package correctly, if it can't see the
package, maybe just share the output of 'apt cache policy'
with us. But it should see that package.
634[03:29:33] <dpkg> It's considered impolite to paste many
lines of text on IRC. Please don't do it. Pasting one line is
fine. Pasting two lines you can usually get away with. Pasting three
lines will get you insulted. Pasting four or more lines will get you
kicked. If you want to paste, ask me about <paste>
635[03:29:43] <uxfi> Sorry
636[03:29:45] *** Quits: qrpnxz (abc4f95c31@replaced-ip) (Remote host closed the connection)
637[03:29:47] <uxfi> I will use a paste site next time
644[03:30:51] <somiaj> uxfi: You could try to purge that package
then reinstall it, but if dpkg is breaking and not allowing that,
you may have to go look at the files I pasted.
645[03:31:10] <somiaj> !raspbian
646[03:31:10] <dpkg> Raspberry Pi OS (previously called
Raspbian) is a distribution <based on Debian> made
specifically for the <Raspberry Pi>. Raspbian is not Debian
and it is not supported in #debian. Please use #raspbian (or
#raspberrypi) on irc.freenode.net for support.
replaced-url
647[03:31:10] <uxfi> My apache is also otu of whack
648[03:31:25] <somiaj> uxfi: also we don't really support
raspbian here, it isn't debian, and uses their own
packages/repos.
651[03:32:37] <uxfi> files list file for package
'openjdk-11-jre-headless:armhf' is missing final newline
652[03:32:37] <somiaj> I hope you are able to get your system
back into a usable state, but since you are choosing raspbian to
run, you should use their support community.
653[03:32:40] <uxfi> weird
654[03:33:00] <uxfi> I know
655[03:33:06] <uxfi> I didnt do anything wrong
656[03:33:12] <somiaj> yea, sounds like something in your dpkg
database is missing a new line, might go manually look at the files
I pointed you at.
657[03:33:16] <uxfi> Im jsut askign from a debian... and in
genenral why woudl this happen?
658[03:33:20] <somiaj> uxfi: it could be a file got corrupted.
659[03:33:27] <uxfi> well
660[03:33:30] <uxfi> also anotehr thing
661[03:33:38] <uxfi> my apache 2 is screwed up and web server
wont spin up
662[03:33:50] <uxfi> I tried cehckign the access.log
in/var/logs... Nothing in the lgo at alll
663[03:33:54] <somiaj> It is hard to say, but you aren't
running debian, maybe there was a bug in one of raspbian's
packages, maybe some corruption on the disk, but here we don't
like to guess what debian derivaives have done.
664[03:34:42] <uxfi> ah
665[03:34:46] <uxfi> ok
666[03:34:51] <uxfi> fair enuff
667[03:34:55] <mutante> uxfi: #raspbian is quite an active
channel though
683[03:53:37] <factor> ratrace, I pinned and installed libc from
testing and everything is working fine. As my stuff is compiled from
my own I guess that is where I needed to be.
684[03:53:42] <factor> thanks for the help
685[03:53:57] *** Quits: kek_ (~kek_@replaced-ip) (Remote host closed the connection)
686[03:55:10] *** Quits: catman370 (~catman@replaced-ip) (Quit: See you later..)
687[03:55:23] *** Quits: wintersky (uid453465@replaced-ip) (Quit: Connection closed for inactivity)
688[03:56:38] <somiaj> factor: are you running debian stable?
689[03:57:03] <factor> mostly still but yes
690[03:57:04] <somiaj> mixing stable/testing is strongly not
supported and is most likely not the best solution to a problem.
691[03:57:10] <somiaj> !don't break debian
692[03:57:10] <dpkg> it has been said that dont break debian is
replaced-url
693[03:57:20] <factor> :)
694[03:57:24] <somiaj> factor: don't do that, at this point
you might as well just finish the upgrade to bullseye and run that
instead (it is going to be released soon)
695[03:57:38] <somiaj> !buster->bullseye
696[03:57:56] <factor> somiaj, I may , seems where I need to be,
but stable has some good points
698[03:58:20] <somiaj> yea, bullseye is really close to being
ready for the next stable release (could be a month or so until
release).
699[03:58:31] <factor> Oh nice
700[03:58:37] <somiaj> Right now its biggest draw back is slow
security support and a few rc-bugs, but not many
701[03:59:03] <somiaj> I think it is fine for a desktop. But
running bullseye is way better than what you have done. Don't
mix debian releases, this is just asking for trouble.
702[04:00:26] <factor> Compiling my own kenrel and nvidia driver
is the same warning.
703[04:01:00] <somiaj> I think it is worth learning how to deal
with things 'the debian way', overall what you are doing
is not supported by #debian.
704[04:01:15] <somiaj> I am using the debian nvidia drivers just
fine and haven't had any issues with steam for multiple
releases
708[04:02:01] <factor> Im also using compiled thrustmaster
driver. Which works , but have to hack some stuff.
709[04:02:19] * dvs shudders at the number of ops
710[04:02:23] <factor> Driver was just official 5.13
711[04:03:19] <somiaj> just sounds like #debian cannot
reasonablly support your setup since you are mix/matching things, my
suggestion is try to stick to debian packages, things actually are
well tested and work nice together.
712[04:06:40] <uxfi> still having issues
713[04:06:55] <somiaj> uxfi: Head over to #raspbian
714[04:07:06] <uxfi> I did
715[04:07:07] <uxfi> waiting
716[04:07:27] <somiaj> You might just need to be paitent, but we
cannot resonablly support debian derivatives here.
717[04:07:44] <uxfi> somiaj; can i jsut dod a system image
restore?
718[04:08:06] <somiaj> That is up to you, I was going to
suggest, that maybe backup/restore might be useful if you have
corruptedy our dpkg database to the point dpkg is broken.
723[04:09:39] <somiaj> Anyways, best of luck, but you should try
to stick to raspbian support if that is the distro of choice. Note
on some pi's you can run pure debian,
replaced-url
727[04:11:01] <somiaj> One advantage of the pi is you can just
get another sd card, and start fo scratch adn have all your data
backuped on the previous one
803[05:30:07] <outoftime> I'm writing sh script that needs
to ask for sudo password but I have no /etc/sudo.conf file. May I
mess up with some configs if I'll just make it exactly as
example in `man sudo.conf` shows?
805[05:33:35] <outoftime> Example from manpage enables some
plugin, that is why I'm a bit confused
806[05:36:22] <somiaj> what are you wanting to put in the
sudo.conf file?
807[05:36:58] <outoftime> askpass path
808[05:37:53] <somiaj> Yes, if a configuration file is not
present, you can create it. Would setting up NOPASS in sudoers for
the command needed to run also be an option?
809[05:38:32] *** Quits: ax5623 (~NickServ@replaced-ip) (Remote host closed the connection)
811[05:44:18] <outoftime> I have bunch of regular command in a
script and want to prompt a password for mounting partition, that is
all. sudo have `--askpass` option but it needs to know what program
should handle it. I wanted to add path to /etc/sudo.conf but there
is not such file at all and I'm wondering is it safe to make it
empty with only one line specifying path for askpass program.
814[05:49:35] <outoftime> somiaj: I suppose NOPASS will give you
rights to execute potentially dangerous command without asking for
permisions. That is not exactly what I want to achieve.
815[05:49:39] <somiaj> outoftime: is all you need sudo for is
mounting a partition? Is it the same partition?
816[05:49:54] <somiaj> i.e. is what they are mounting have a
UUID on the filesystem?
817[05:50:41] <somiaj> outoftime: you can be very spicific in
the sudoers file and only allow exact commands to be run via sudo
without a password.
818[05:50:47] <outoftime> somiaj: I know about fstab
820[05:51:14] <somiaj> To me if they are mounting a file system
that already has permissions on it, such as ext4, adding this to
/etc/fstab maybe a nice solution
829[05:55:35] <somiaj> noauto makes it not mount during boot,
and user allows user to mount
830[05:55:57] <somiaj> anyways, I'm not exactly sure the
goal of your script, just suggesting things that maybe more
apporpriate than asking for a password.
831[05:58:21] *** Quits: edlou (uid413273@replaced-ip) (Quit: Connection closed for inactivity)
833[06:00:18] <outoftime> somiaj: no, this is actually a good
suggestion. I have tried "default,auto" but it ends with
readonly partition. The only way I managed to mount it is via
`mount` command. That is why I do stick with writing a script. If
something will faild I'll know exactly what and will be able to
fix it.
867[06:33:17] <Razva> Hi. Can you please let me know what's
the syntax for forwarding all traffic on port `80` to port `80` on
IP `192.168.1.100` when using `ufw`? All I'm getting are old
tutorials that instruct me to modify `before.rules`.
891[07:07:55] <outoftime> somiaj: actually there is a
difference. If I'm addding line to fstab and mounting `mount
/mnt/Games`, executable sh says "line 38: ... Permission
denied" but when I do `sudo mount /dev/sda5 /mnt/Games` it
works fine
923[08:06:04] <outoftime> I'd like to find out why it is
not working with just fstab, but I'd rather invest my time in
studing shell scripting than figuring out what went wrong with
fstab. Which sounds strange when I "spell" it.
946[08:29:22] <terr> This is going to appear to be a Krazy
question. And I am not necessarily planning to even try this. 1) can
Linux boot and run in an NTFS partition (I need to be able to run
damn windows and even windows NT. 2) when I set up multiboot before
I recall using dd to lift the mbr. But its been years. I should be
able to install more than one operating system in a single bootable
partition and select which O/S simply by coping the correct MBR form
a group st
962[08:34:28] <somiaj> terr: with UEFI, it is really easy to put
multiple OSes to boot in the same EFI partition. So dual boot is
possible to setup (Though I still use grub to handel the dual
booting)
964[08:35:11] <somiaj> terr: it sounds like you didn't
setup multiboot correctly in the past, dual booting with windows is
a common thing (Though I would also consier nkuttler suggestion, and
run the os you use less often in a VM)
972[08:37:42] <somiaj> Yes since you are basically running two
oses at once, it will be more cpu intensive than just running one,
but my windows VMs work just fine and don't take that much cpu
ussage overall
983[08:45:04] <outoftime> terr: in the past I had several
windows versions on ntfs partition and linux on separate ext. grub
allowed to switch between windows\linux and windows' loader
allowed to select which version you want.
985[08:46:00] <somiaj> biggest problem use to be install linux
second because windows destroys the MBR, though with UEFI, this is
far less of an issue, just make sure your EFI partition is big
enough.
1012[09:07:30] <backupluis> Hi ppl, I have a new internet
provider with no access to admin page of the router, also I have all
the inbound ports closed and no public IP4, its something called
cgnat, but I have a public ip6 address, my problem is with my linux
box, on windows I can download torrents, maybe using the ip6 because
I see some peers with ip6 but on linux nothing happens, on both
system got ip4&6 from dhcp, tested many torrent clients and
1013[09:07:31] <backupluis> all the same, not even connect with
trackers, any clue?
1025[09:16:13] <terr> The configuration you set up sounds perfect
for me. On an internal drive I need nt, 7 and 10 64 bit. Possibly a
32 bit version but I am scratching my head on that one because these
a very low power machines with 2 GB.
1026[09:17:22] <terr> On an external which I will use as a backup
probably a copy of every os I use which includes for a Raspberry Pi
1094[10:01:48] <wrksx> I'm authed, and typing /join #httpd
sends me to ##checkyourconnection, wth is happening to me?
1095[10:01:52] <wrksx> so sad
1096[10:02:14] <ratrace> wrksx: well, it's offtopic here in
#debian. if you have a question about apache on Debian, ask :)
1097[10:02:29] <wrksx> haha
1098[10:02:48] <ratrace> also, everyone is leaving freenode for
libera.chat, so there could be closed channels, channel modes
changes, and no ops around to help.
1150[10:15:51] <alkisg> To be honest, the ability to start with
text chat, and then selectively use voice or screen sharing, is
great for support. But I don't like the fragmentation of
slack/discord/matrix/whatever, it was nice that all developers were
on IRC for a decade and more
1165[10:19:54] <jelly> I'd accept whatever protocol as long
as the UI looked like irc, dense text for easy reading. All the chat
apps waste space and are making it hard to follow conversation.
Slack and Discord are somewhat close but still horrible
1166[10:20:16] <terr> Hey if you new boots want something better
than IRC, try ZOOM
1167[10:20:45] <jelly> !slap terr
1168[10:20:45] * dpkg strikes a resounding *THWAP* across terr's
face
1169[10:20:49] <wrksx> Well said jelly. And images and vieos
don't help
1195[10:28:39] <jelly> personally I plan to stick around until
most of freenode #debian users, so about 500 at least, appear on
either irc.debian.org or irc.libera.chat channel by the same name
1196[10:29:45] <jelly> this channel has been technically
"unofficial" since
1197[10:29:47] <jelly> !irc move
1198[10:29:47] <dpkg> irc.debian.org moved to OFTC on June 4th
2006, see
replaced-url
1199[10:30:03] *** Quits: uvolmer (~uvolmer@replaced-ip) (Remote host closed the connection)
1200[10:30:06] <ratrace> welp, I finally planted my derriere into
OFTC/#debian
1201[10:30:11] <TheBigK> will do same then... i guess
1269[11:25:39] <PMT> Hi all, I've got a slightly odd issue -
I'm playing with an HPPA VM in qemu, and in qemu-system-hppa,
networking works fine, but if I shut down that VM, mount its root
filesystem, and chroot in using qemu-hppa-static, networking works
fine...for everything except apt, which hangs forever on "0%
[Working]" if I try apt update or install. wget -O /dev/null
replaced-url
1270[11:25:45] <PMT> DNS, and
-o'Debug::Acquire::http=true' doesn't output
anything. Anyone have any thoughts on what to try next?
1272[11:27:52] <petn-randall> PMT: Broken IPv6 in the machine is
my first guess.
1273[11:28:03] <Razva> Hi. Is this correct for forwarding port
`65432` on eth0 to port `65432` on eth1 to 192.168.1.2 : `ufw route
allow in on eth0 out on eth1 to 192.168.1.2 port 65432 from any port
65432 comment "SSH Sea"` ?
1274[11:28:04] <PMT> petn-randall: I tried -o
[...]forceipv4=true, no change.
1294[11:37:26] <petn-randall> PMT: There are also other debug
options mentioned in `man apt.conf`. I'd try
Debug::Acquire::http. What is the output of `host
ftp.kr.debian.org`, and can you connect to it via curl/wget?
1295[11:37:44] <PMT> The latter, yes; the former, as I said,
never outputs anything.
1296[11:38:17] <petn-randall> Oh right, we tried that.
1297[11:40:00] <petn-randall> PMT: just to clarify, wget -O
/dev/null
replaced-url
1298[11:40:49] <PMT> Fascinatingly, though that wget does work
and report 200 OK, "host" and "nslookup" both
hang requiring a kill -9 to stop. Finding out on what may be
slightly annoying, sec.
1299[11:41:21] <petn-randall> a simple `kill` should be enough,
or not?
1362[12:08:39] <PMT> Rats, even setting up a local forwarding DNS
server on the qemu host and pointing the qemu-static chroot to
127.0.0.1 doesn't help.
1397[12:53:22] <jelly> PMT, at this point you could try and
upgrade qemu blindly and hope for the best, or ask in #qemu over on
irc.oftc.net (= irc.debian.org), that's the official channel
1608[15:17:21] <cybrNaut> sfdisk refuses to allow a BIOS BOOT
partition to be placed in the 1st 2048 blocks of the table. It says
"/dev/sdb1: Sector 34 already used. Failed to add #1 partition:
Numerical result out of range. Leaving."
1609[15:18:02] *** Quits: bitblit (~bitblit@replaced-ip) (Remote host closed the connection)
1610[15:18:29] <cybrNaut> but it's actually a feasible place
for BIOS BOOT, according to
replaced-url
1631[15:31:18] <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>.
1632[15:31:34] <queip> lol last time it was said to mention where
it's crossed
1641[15:33:50] <petn-randall> queip: It's not default
because I'd *want* cp to fail when it can't copy the
symlinks as is. Because there's a huge difference between a
symlink and the real file.
1642[15:33:53] <deadrom> hi. is there an package that gives me a
NAS web gui for debian? I know there are NAS distros, but can I have
this with "regular" debian?
1645[15:34:47] <petn-randall> queip: Imagine 1000 symlinks to a 1
GB file. If you copy that to FAT32 that would all of a sudden use
1001 GB of disk space instead of 1 GB.
1646[15:35:07] <petn-randall> deadrom: What do you want the NAS
GUI to do?
1647[15:35:17] <qman__> deadrom: not specifically as far as I
know, what whould this NAS GUI do, exactly?
1648[15:35:37] <qman__> there are lots of web packages that do
various things that you might install
1663[15:40:27] <deadrom> qman__, well, actually just NAS but NAS
distros have not the community support like large dists. probably
will boil down to that after all, then
1666[15:42:25] <qman__> various hardware NASes and NAS distros do
all sorts of different things in their web GUIs, so you need to
define what functionality you're looking for
1673[15:43:51] <qman__> ok, but what would the web GUI do in that
scenario? that's easily accomplished by installing samba and
filling out the smb.conf
1677[15:50:38] <deadrom> qman__, you know what, you're
right. I was thinking people need to be able to check online status
and usage on a web gui and possibly drw logs, but then again - they
never do
1678[15:51:19] <qman__> if you want usage stats and monitoring,
I'd use prometheus and grafana
1679[15:51:30] <qman__> and there are lots of other options
1741[16:26:05] <jamea77> somiaj: changed (only for test purposes
in 1 test server: 192.168.104.87 sysprodtest1 sysprodtest11,
restarted the server, after, form windows pc: (first ping
192.168.104.87) answered correctly, but not when do ping
sysprodtest1 The ping request could not find the host sysprodtest1.
Check the name and
1861[17:31:04] <somiaj> monsterco: though often times a
filesystem is mounted read-only if it has errors, you may want to
get onto a live system, fsck the file system and make sure there
aren't any errors before doing this, otherwise there is a
chance you destory the file system.
1869[17:32:59] <jelly> monsterco, kernel remounts a filesystem in
read-only mode if a filesystem error occurs, to prevent further
corruption
1870[17:33:34] <jelly> monsterco, inspect "dmesg"
output to confirm that happened. You will have to reboot and let the
boot-time filesystem check fix things
1876[17:36:53] <monsterco> somiaj, this is a container in proxmox
1877[17:36:59] <monsterco> this has happened before
1878[17:37:06] <monsterco> when storage fills up it umounts
1879[17:37:15] <monsterco> is there a way I can empty some files
first?
1880[17:37:28] <jelly> monsterco, no, not in a safe way
1881[17:37:35] *** Throwawayname is now known as AF04FB9290474265
1882[17:37:38] <somiaj> well you may want to ask proxmox support
for how to deal with containers when this happens
1883[17:37:38] <jelly> first reboot and fsck, then delete files
1884[17:38:05] *** Quits: scabootssca (~scabootss@replaced-ip) (Disconnected by services)
1885[17:38:08] *** scabootssca_ is now known as scabootssca
1886[17:38:12] <somiaj> but yes, be careful, mounting a corrupted
filesysstem read/write and writing to it can destroy data
1887[17:38:18] <jelly> long term, set up monitoring for your host
and don't let the _host_ side of the storage fill up when
you're using thin provisioned storage
1888[17:38:30] <monsterco> why deleting files before fsck is not
a good practice/
1908[17:42:10] <dpkg> Proxmox Virtual Environment (Proxmox VE) is
a GNU/Linux distribution <based on Debian>, providing a
virtualization platform with <LXC> and <KVM>. It is not
supported in #debian. There's an unofficial
proxmox channel on Freenode. For official venues, see
##replaced-url
1909[17:42:27] <somiaj> monsterco: ^^ you might try their
support, might find users more familar with proxmox tools that could
help out.
1910[17:42:38] <jelly> monsterco, did you at least look at dmesg
before you turned it off?
1938[17:53:14] <jelly> monsterco, that output implies the host
backend blocked the writes, not the guest, but you'll want to
ask someone who know more about lxc and proxmox to confirm
1943[17:54:23] <jelly> monsterco, it's possible you have set
limits from the host side that are lower than what the guest needs.
It is NOT normal for IO errors to happen when the filesystem is full
or close to full.
1944[17:54:33] <monsterco> jelly, I see
1945[17:54:41] <monsterco> I did pct mount ID there which is from
proxmox
1955[17:55:54] <karlpinc> I'm trying to self-backport
guacamole. Judd says no problem, but I get an error with "apt
-b source guacamole". It does not like -Werror. There's a
patch that turns -Werror off in debian/, but apparently it does not
do enough. What does "apt -b source guacamole" actually
do? (How do I add another patch to the package so that I can make it
compile?)
1964[17:57:15] <jelly> karlpinc, separate apt source ... from
building; omit -b, the go to the source direcitory and do the build
bit there, something like "dpkg-buildpackage -b -uc -us"
1965[17:57:44] <jelly> then when it fails, you can inspect the
remains and try to fix things
2009[18:09:32] <jsync> All of the lvm instruction pages act
2010[18:10:00] <somiaj> karlpinc: seems upstream source might be
a better starting point that the what appears to be broken debian
package.
2011[18:10:53] <jhutchins> It seems like it would be a good idea
to check out other remote desktop options that are better known and
supported.
2012[18:11:12] <karlpinc> somiaj: I'm starting to think so.
I'll try following the apache.org instructions.
2013[18:11:26] <jsync> All of the lvm instruction pages seem
poorly written. Misspelled words in various site pages that are
typically reliable. It's difficult to find instructions that
seem to make sense within an applied context.
2015[18:11:48] <karlpinc> jhutchins: Mebbie. But I like the idea
of granting access via a browser.
2016[18:12:21] <somiaj> karlpinc: I would start with the wiki
jelly linked
2017[18:14:02] <karlpinc> Humm. Apache says: "Normally, you
don't need to build guacamole-client, as it is written in Java
and is cross-platform.". Maybe I'll try backporting the
server and just "use" the client.
2051[18:42:18] <jsync> I'm really uncomfortable with these
instructions. It's not clear in any instructions I've
found as to what drive has how much of what logical volume, not
actually. It's like, various drives comprise one volume group,
& I'm resizing a logical volume for the var folder?
2052[18:42:25] <cybrNaut> i plan to partition a target drive
manually, since the Bullseye installer apparently can't handle
BIOS BOOT partitions
2053[18:43:05] <cybrNaut> question is, when i'm in expert
mode and I ask for a shell, how do I handle dm-crypt stuff?
2054[18:43:10] <PMT> jsync: /win 18
2055[18:43:15] <PMT> sorry, m/t
2056[18:43:18] <cybrNaut> there is no "cryptsetup"
command
2062[18:45:05] <somiaj> cybrNaut: Did you boot in UEFI or Legacy
mode with the bullseye installer?
2063[18:45:18] <jhutchins> cybrNaut: Remember, the testing
installer is for the purpose of testing/developing the installer,
not for installing testing. It will be one of the last components
updated before the release, to accomadate any last-minute changes.
2065[18:45:58] <cybrNaut> somiaj: my hardware is BIOS-only
2066[18:46:01] <somiaj> cybrNaut: afiak if you use UEFI, it
assumes gpt, and if you use legacy is assumes MBR, since those are
the prefered situation for each method.
2067[18:46:10] <karlpinc> jsync: Read the first paragraph of
"man lvm".
2068[18:46:11] <somiaj> cybrNaut: If you think you have found a
bug/issue, you should report it.
2069[18:46:15] <jhutchins> jsync: It might be time to back up the
content of the volumes and re-map the system.
2070[18:46:35] <somiaj> I haven't tested the installer on
legacy hardware, so have no experience if this has changed with
bullseye
2071[18:47:23] <cybrNaut> somiaj: yeah, those assumptions may
simplify things for many users, but I need to partition manually as
I will have a GPT-BIOS (thus bios boot)
2072[18:48:04] <jsync> jhutchins, I'm just starting to setup
lvm, & I haven't found any reasonable instructions, not
actually.
2073[18:48:26] <somiaj> cybrNaut: ahh so you are trying to use
GPT with MBR compadability? I missunderstood, I thought you were
using leagcy boot and not giving the mbr option. Yea, you may have
to load some of the aditional installer componenets and do this
manually.
2074[18:48:39] <somiaj> I'm unsure how sophoiscated the
partition manager in the installer is for these edge casese.
2075[18:48:40] <jhutchins> !lvm
2076[18:48:40] <dpkg> [lvm] the Linux Logical Volume Manager (replaced-url
2077[18:49:32] <cybrNaut> jhutchins: I'm booting the
Bullseye blu-ray release disc of april 14th (from USB). Does that
come with the correct installer for installing bullseye?
2078[18:49:39] <jsync> Some guy writes that I run commands &
format hdds, & he can't spell the difference between where
& were, or to & too, not actually. Instructions aren't
reliable & there are things not explained.
2079[18:49:56] *** Quits: Strum (~Strum@replaced-ip) (Remote host closed the connection)
2082[18:51:15] <somiaj> cybrNaut: the testing installer is in the
rc1 canidate, but it still may have some issues to work out,
providing bug reports and feedback will be useful.
2084[18:51:54] <karlpinc> Well, looks like guacamole-server
builds, using the instructions at apache.org. But for some reason
the backported package of the same version does not. (Something to
do with permissions for '_apt'.)
2086[18:53:06] <cybrNaut> somiaj: the installer's partition
tool is indeed too limiting for my case, so i must partition
manually. My LUKS is also going to be more sophisticated than that
installer can handle too. So I will setup the crypto manually in
advance too. My problem: when i jump to a shell in the installer, it
gives a busybox environment that lacks cryptsetup. So how can I get
my LUKS volumes mounted before the
2092[18:55:51] <cybrNaut> hmm.. i loaded the extra components (it
did not give a choice of what components), and then went to a shell
and "cryptsetup" is still unrecognized
2097[18:57:56] <cybrNaut> the thing is labeled: "load
installer components from installation media", and when I
trigger that it goes off and loads a bunch of stuff
non-interactively
2098[18:58:08] <somiaj> no not that
2099[18:58:23] <somiaj> that loads the default componenets, there
is an option to load additional compoents.
2100[18:59:52] <cybrNaut> i don't see that option. this is
the non-graphical expert install. Do I need to do the graphical
expert install?
2101[19:00:59] <somiaj> no it is there, maybe I'm not
recalling its name correctly, but I can't find a screenshot of
the main expert menu
2102[19:01:02] <somiaj> I know I saw it recentally
2108[19:05:17] <somiaj> cybrNaut: actually that is it, after you
select "Load isnatller components from installation media"
it detects the media then gives me a list of additional things I can
load. I see crypto-dm-modules-5.10.0-6-amd64-di as an option, but I
don't see one for cryptsetup, maybe you'll be better doing
a manual install from a live system using debootstrap
2110[19:07:13] <cybrNaut> crypto-dm-modules-5.10.0-6-amd64-di
probably includes cryptsetup. but i'm not given any options
after triggering "load installer components from installation
media"
2111[19:07:36] <cybrNaut> did bullseye change that?
2112[19:07:39] <somiaj> hmm, it was clearly available to me from
the bullseye installer -rc1
2113[19:07:56] <somiaj> no that has been available for a long
time, I just tested my bullseye netinstall, it was there
2114[19:08:24] <cybrNaut> i'm using RC1 as well. it's
"debian-edu" (what's edu?)
2115[19:08:33] <somiaj> oh I'm using the offical debian one
2116[19:08:37] <somiaj> debian-edu is a pure blend
2117[19:08:39] <somiaj> !debian pure blend
2118[19:08:39] <dpkg> A Debian Pure Blend (formerly Custom Debian
Distribution) is a subset of Debian configured to support a
particular target group out-of-the-box. Not to be confused with
distributions <based on Debian>.
replaced-url
2119[19:09:26] <somiaj> here edu = education, so it is a specific
set of packages designed for education/school labs, etc, I
don't know eactly what is on it, but I wouldn't use the
-edu pure blend unless you know you need it
2120[19:09:54] <somiaj> I don't know if they change the
installer (that would seem weird though), so I wouldn't expect
that is the issue of why your isntaller is behaving differently than
mine
2122[19:10:37] <cybrNaut> shit.. it was hard enough just to find
that. the bullseye blu-ray install ISOs are scarce
2123[19:10:57] <somiaj> why do you want the blu-ray and not just
a netinstall?
2124[19:11:22] <somiaj> again I doubt that is the issue, but it
could be your setup would be best done from a full live system (like
maybe grml) and do a manual install with debootstrap
2125[19:11:59] <jsync> jhutchins, thanks for the urls.
2126[19:12:03] *** Quits: m4rley (~m4rley@replaced-ip) (Remote host closed the connection)
2131[19:12:28] <jsync> That instructional makes sense to me.
2132[19:12:50] <cybrNaut> maybe i'll have to see how
debootstrap works then
2133[19:13:18] <somiaj> cybrNaut: debootstrap requires a network,
but with apt-cacher-ng you can create a caching proxy easy enough
and thus only download from the network once
2134[19:13:29] <cybrNaut> i'm migrating from Stretch to
Bullseye. I'm not interested in dist-upgrades and looking to
skip Buster
2135[19:13:56] <somiaj> if you are doing a lot of local installs
apt-cacher-ng gives you a nice way to have access to the newest
packages and download things only once.
2136[19:14:07] <jsync> So, after I extend my var folder to 4tb
with the additional drive, & I want to copy data out of a
separate drive, I just copy all that data to the var folder?
2147[19:16:45] <somiaj> Unsure, and I highly doubt the
pure-blends are changing the installer outside of providing a
different set of packages by default.
2148[19:17:46] <somiaj> but downloading a full blue-ray before a
release (since there is still going to be various updates) may not
be the best method. If you want to go with a full
blue-ray/dvd/whatever installer, you may want to wait until the
release.
2149[19:18:00] *** Quits: Numero-6 (~Numero-6@replaced-ip) (Quit: << - Qui etes vous ? - Je suis le nouveau numero 2 -
Qui est le numero 1 ? - Vous etes le numero 6 - Je ne suis pas un
numero ! Je suis un homme libre!! >>)
2150[19:18:02] <somiaj> this way you get all the fixes that will
be done before the release in the next month or two
2151[19:18:34] <cybrNaut> somiaj: won't i get all the fixes
by doing "aptitude update"?
2153[19:19:09] <somiaj> Sure, but then you have to do that on
each and every machine, and are now in the downloading stuff which I
thought you wanted to avoid.
2155[19:20:01] <somiaj> hence why maybe apt-cacher-ng might suit
your needs a bit better. Also note you don't need the internet
to use the netinstall. The netinstall contains a fully functional
base system. Once you finish the install, you can then set up your
sources however you want/configure tor/etc, then install the rest of
your packages.
2156[19:20:26] <somiaj> I think many get confused by that name,
and don't realize you can do minimual base system offline
installs with the netinstaller.
2157[19:20:30] <karlpinc> Ok... debian/rules needs
--enable-allow-freerdp-snapshots added to the end of the
dh_auto_configure lines. Then guacamole-server backports.
2158[19:21:03] <karlpinc> But I need UTF-8 in my debootstrap
chroot. What package do I install for that?
2173[19:29:09] <cybrNaut> i've been all over that file tree.
regardless of whether the high level branch is
"unofficial" or "bullseye_di_rc1", either way it
leads to *edu* builds
2174[19:29:39] <somiaj> it could be the installer/dvd team
hasn't started to making images until the actual release
2175[19:29:51] <cybrNaut> this should not have the edu version
=>
replaced-url
2176[19:29:56] <somiaj> usually once the release hits, they go
start creating all those images
2191[19:33:15] <somiaj> I think what you see isn't an issue,
but all the edu is is a different set of packages by default, they
all use the same main repo
2195[19:36:40] <karlpinc> What's a good vnc server? (x11vnc,
tigervnc-standalone-server, something else?)
2196[19:37:11] <jsync> somiaj, I'm not certain how to mount
the new drive, not actually.
2197[19:37:13] <oxek> none of them are good, but I use tiger
2198[19:37:28] <sney> x11vnc is specifically for sharing an
existing X session
2199[19:37:48] <jsync> The Disk utility says that the dev/sde1 is
mounted.
2200[19:37:50] <karlpinc> oxek: That's what the
"vnc-server" virtual package uses.
2201[19:38:01] *** debhelper sets mode: +l 993
2202[19:38:08] <sney> and yeah, it's the vnc protocol that
is kind of crappy so the software doesn't make much of a
difference. try xrdp for a slightly more modern approach
2203[19:38:09] <karlpinc> sney: Humm. I _do_ want to share an
existing X session.
2212[19:43:33] <jsync> Ohh. I found it. I had to run resize2fs
/dev/named-vg/var
2213[19:45:16] <somiaj> jsync: where is it mounted?
2214[19:46:19] <jsync> Uhh, I'm thinking it's mounted
at var. This actually is my first lvm setup.
2215[19:47:52] <cybrNaut> somiaj: glad you mentioned debootstrap.
After looking more into that, it seems like the ultimate expert of
all expert install options. The big blu-ray ISO i already downloaded
can be mounted and debootstrap can grab pkgs from it locally.
It's seems to give the power & flexibility of a scratch
install, but still Debian in the end
2217[19:47:58] <somiaj> here lvm isn't anything special
here, the same issue would appear if you didn't use lvm and
traditional paritions, you need things to be setup where they are.
2218[19:48:00] <jsync> I followed these instructions:
replaced-url
2219[19:48:39] <jhutchins> karlpinc: The correct answer is ssh
2228[19:53:18] <karlpinc> jhutchins: Unfortunately I'm using
rdp to frob a MS Windows box, and somebody needs to be able to watch
what I'm doing -- they want to watch my desktop.
2229[19:54:05] <sussudio> buu: boot rescue disk. there are a few
varities out there.
2241[19:56:28] <buu> Mounting any disk says "no such file or
directory"
2242[19:56:55] <buu> "mounting /dev/sda1 on /mnt failed: no
suchfile or directory"
2243[19:56:59] <buu> And yes /mnt exists, I just made it
2244[19:57:30] <PMT> I wonder if it's spitting that up
because it needs a module for the filesystem and it's not
autoloading? Random-ass guess, I haven't been in that situation
in an age.
2245[19:57:43] <karlpinc> buu: And does /dev/sda1 exist?
2248[19:59:33] <buu> It turns out it needed -t ext4
2249[19:59:43] <buu> A+ error message
2250[20:00:26] <buu> Do I need to like
2251[20:00:34] <karlpinc> buu: The busybox mount command is,
well, minimal.
2252[20:00:36] <buu> chroot and run grub update or something or
can I just update the file
2253[20:01:20] <PMT> I would edit the file.
2254[20:01:51] <karlpinc> buu: chroot, edit /etc/default/grub,
and run grub-update to generate the actual grub config file. (Or,
you could go into the interactive grub prompt and dynamically modify
the kernel command line for one-time use. And then fix it once
booted.)
2259[20:02:49] <dpkg> s390 is a port of Debian, which uses the
Linux kernel on IBM System/390 (aka S/390, zSeries) mainframe
hardware. It was previously a supported architecture from Debian 3.0
"Woody" until Debian 7 "Wheezy". It has been
replaced by the <s390x> port.
replaced-url
2260[20:03:13] <buu> karlpinc: I couldn't convince it to
give me a grub prompt =[
2261[20:03:31] <buu> PMT: I just edited grub/grub.cfg
2262[20:04:00] <karlpinc> buu: You can edit that, but it will go
away when you upgrade the kernel etc unless you edit
/etc/default/grub.
2263[20:04:02] <cybrNaut> thanks.
2264[20:04:22] <PMT> Yeah, the above is the important caveat to
hand-editing grub config files.
2265[20:04:58] <buu> karlpinc: Well, the offending argument
isn't in etc/default/grub so I'm slightly confused
2266[20:05:06] <buu> But I just rebooted, lets see what happens
2267[20:05:23] <buu> I assume I got myself into this mess in the
first place by specifying a /dev/sdX as my root= device
2274[20:09:48] <cybrNaut> the "debootstrap" install
method is documented in "Appendix D. Random Bits", which
gives a feeling of incompleteness. Shouldn't it get a chapter
of its own?
2298[20:35:35] <cybrNaut> after cloning a debian installation to
another drive, the clone cannot install stuff. "aptitude
install <anything>" gives "[ ERR] Writing extended
state information" \n "[ ERR] Building tag database"
\n "E: Failed to execute process to save dpkg selections, dpkg
or trying to execute it exited with status/errno: 2" \n
"E: failed to save selections to dpkg database"
2299[20:36:36] <cybrNaut> is root supposed to own
/var/lib/apt/extended_states?
2300[20:37:01] <cybrNaut> i wonder if rsync didn't preserve
ownership, which may be causing this
2315[20:43:58] <cybrNaut> aptitude leaves no further clue about
the problem, so then i tried apt-get instead, which said "dpkg:
error: cannot scan updates directory
'/var/lib/dpkg/updates/': No such file or directory"
2316[20:44:22] <PMT> was it the --prune-empty-dirs
2317[20:44:24] *** Quits: catman370 (~catman@replaced-ip) (Quit: See you later..)
2318[20:44:27] <cybrNaut> mkdir /var/lib/dpkg/updates/ <=
fixes it for both aptitude and apt-get
2349[21:00:56] <cybrNaut> i just realized that if i compensate
for the issue by doing "mkdir /var/lib/dpkg/updates", i
risk creating it with the wrong perms. So it's perhaps smarter
to "touch /var/lib/dpkg/updates/dummyfile" before doing an
rsync
2394[21:29:21] <cybrNaut> PMT: i used to use reportbug to form
the bug report msg, which mixmaster could then send. I think the
mixmaster network still runs, and hopefully one day i'll scrap
together a client
2395[21:29:32] *** Quits: catman370 (~catman@replaced-ip) (Quit: See you later..)
2407[21:39:44] <cybrNaut> i should say the bugs don't get
lost.. i just report them against ubuntu instead so at least they
have a chance of making it upstream
2408[21:40:19] *** Quits: catman370 (~catman@replaced-ip) (Quit: See you later..)
2417[21:48:41] <jhutchins> PMT: Well, if it's an upstream
bug, then yeah, it needs to get reported upstream, but that's
something the Debian people shold do after triaging them.
2418[21:48:45] <sney> package maintainers are typically in touch
with their software's upstream, so as long as the maintainer is
active, relevant non-pebcak issues usually get passed along. but a
cc never hurts.
2419[21:49:07] <cybrNaut> jhutchins: i suppose it's down to
people's motivation. I post the bug report on Ubuntu. From
there, a developer might mirror partially upstream to debian, or
fully upstream to the top level.
2420[21:49:10] <sney> having a public record of the communication
is arguably better than the maintainer just pinging someone on irc,
e.g.
2421[21:49:18] *** Quits: jxel (~Thunderbi@replaced-ip) (Remote host closed the connection)
2422[21:49:38] *** Quits: czesmir (~stefan@replaced-ip) (Quit: Lost terminal)
2424[21:50:13] <jhutchins> When I was working with Mandrake, the
upstream folks took reports from the Mandrake people much more
seriously than users. I was often told to install from source and
check for the bug.
2425[21:50:44] <cybrNaut> it's a sad state of affairs
because the top of the stream is often MS Github and many users
won't go there
2426[21:51:13] <jhutchins> On the other hand, I just got blown
off for reporting a bug from buster backports because the version
was too old to be relevant. So much for supporting stable.
2427[21:52:18] <PMT> On a Debian bug?
2428[21:52:48] <sney> bpo is more of a convenience than anything,
it can't be prioritized higher than the testing cycle. were you
able to reproduce that issue in bullseye?
2443[22:03:35] <cybrNaut> both sfdisk and gdisk are incapable of
writing to sector 34. What other tools can I try on a GPT?
2444[22:04:09] <ratrace> cybrNaut: writing what?
2445[22:04:23] *** Quits: Vizva (~Vizva@replaced-ip) (Remote host closed the connection)
2446[22:04:24] <PMT> I do not know of anything that can reproduce
a Windows partition layout like that (IIRC the last time I saw this
was Windows doing it, and being frustrated nothing would do it)
2447[22:04:43] <cybrNaut> a bios boot partition, following this
guide:
replaced-url
2448[22:05:01] <cybrNaut> it says " In fdisk or gdisk create
a new partition starting at sector 34 and spanning to 2047 and set
the type"
2449[22:06:28] <ratrace> you don't have to do exactly that.
I run bios_grub partitions at 1 MiB offset, 1MiB in size
2450[22:07:01] <ratrace> the _first_ 34 sectors have meaning for
MBR, but there's no rule taht you have to place the bios_grub
one right after taht
2451[22:07:24] <cybrNaut> right, i know i have a choice
2452[22:07:33] <ratrace> BUT ... if you insist... parted can
start partitions at sector granularity
2453[22:07:58] <PMT> IIRC parted complains if you try this, but
I'd love to be wrong!
2454[22:08:01] *** debhelper sets mode: +l 980
2455[22:08:15] <ratrace> it warns about misalignments
2456[22:08:25] <cybrNaut> thanks i'll try it
2457[22:08:28] *** Quits: kline (~freedom0@replaced-ip) (Disconnected by services)
2458[22:08:31] *** Quits: nicole (ilbelkyr@replaced-ip) (Disconnected by services)
2459[22:08:33] <ratrace> but you can switch to "s"
units (Sectors) and it won't matter, it's always aligned
2472[22:15:14] *** Quits: epony (epony@replaced-ip) (Remote host closed the connection)
2473[22:15:41] <cybrNaut> so i issued these commands: "unit
s" then "mkpart primary 34 2047" and it came back
with "Error: Unable to satisfy all constraints on the
partition."
2496[22:21:44] *** Quits: digitalD (~dp@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2497[22:21:56] <cybrNaut> apparently parted cannot place a
partition above others. even though it already had a GPT label, i
did a "mklabel gpt" and that wiped out the existing table
2498[22:22:17] <cybrNaut> after that, it accepted my "mkpart
primary 34 2047" command
2499[22:22:18] <ratrace> after it warned you it'll do so
2501[22:23:26] <cybrNaut> well the warning is kind of vague. This
is what it said: "Warning: The existing disk label on /dev/sdb
will be destroyed and all data on this disk will be lost. Do you
want to continue?"
2503[22:24:01] <PMT> how is "all data lost" vague?
2504[22:24:02] <ratrace> well, the label will be destroyed, but
not data. it's possible to recreate it manually as those are
just offsets in the GPT table sectors
2505[22:24:21] <cybrNaut> well, perhaps not vague.. i just would
have thought that the label change would not be destructive if going
from gpt to gpt
2506[22:24:29] <ratrace> the _tables_ will be destroyed. GPT
keeps them at the beginning and end of the physical device
2508[22:24:51] <PMT> I thought GPT kept them solely at the end
with a fake one at the front?
2509[22:24:58] <ratrace> if you remember, or have in the screen
scroll buffer, the offsets before you mklabel'ed, you can
recreate the layout
2510[22:25:16] <ratrace> PMT: no it has two copies, so if
one's corrupted, it can restore from the backup copy
2511[22:25:17] *** Quits: szorfein (~daggoth@replaced-ip) (Remote host closed the connection)
2512[22:25:34] <PMT> Huh. That makes sense, I just didn't
know it did it.
2513[22:25:59] <cybrNaut> okay.. bit annoying though. I'd
like to use sfdisk to write all the partitions except BIOS BOOT,
then use parted to just add BIOS BOOT. that seems impossible
2514[22:26:25] <ratrace> why don't you use parted for
everything
2515[22:26:53] <ratrace> also, sfdisk is pretty versatile, it CAN
do what you want. lemme re-check the manpage
2530[22:32:39] <ratrace> what I like about sgdisk is that you can
set relative sizes/offsets. with parted you have to calculate the
start of the next partition based on size of previous
2531[22:33:03] <ratrace> with sgdisk you can just 0:+<size>
and it adds them at the end of last one
2539[22:38:03] <mspe> you don't have to calculate the
starting offset, cause the final offsets of existing partitions are
displayed; but you have indeed to calculate ending offsets of new
partitions you create
2543[22:40:25] <cybrNaut> i found that for my drive, if i let the
last partition I create use the rest of the disk space, it's
not aligned on a 4k boundary which supposedly hinders performance.
2544[22:40:31] <ratrace> mspe: you have to calculate the starting
offset in scripted, noninteractive situations
2545[22:40:54] <cybrNaut> so i sacrifice a couple k to ensure
alignment
2546[22:41:08] <ratrace> cybrNaut: the _beginning_ misalignment
hinders performance, not the ending
2547[22:41:31] <ratrace> it barks about that because it's
dumb. misaligned ending affects the _next_ partition but if
that's the end of the drive it doesn' matter
2548[22:41:59] <ratrace> that said .... I always partition
shorter. I've been bitten by disk replacements that claimed
same byte size but weren't really in sector count
2549[22:42:02] <cybrNaut> interesting.. that will make things
easier then
2553[22:45:19] *** Joins: another (~another@replaced-ip)
2554[22:45:54] <cybrNaut> parted makes no mention of stdin..
guess the commands must be all on the CLI? does parted need to be
invoked for each command then?
2565[22:50:41] <ratrace> well I could be wrong, I seem to recall
you could chain them; but it's been a while since I used parted
that way. sgdisk is what I use these days
2584[22:59:54] <ratrace> the first param for mkpart is the
(PARTLABEL) name
2585[23:00:20] <ratrace> ALSO set flag. set 1 bios_grub on
2586[23:00:39] <cybrNaut> ratrace: that's not what the man
page or the interactive docs say.. i think we have a version
difference
2587[23:01:07] <ratrace> I know what the manpage says and
it's wrong. I've been doing this for many years with
parted :) also TIAS. check /dev/disk/by-partlabel/ after you've
created it that way
2593[23:03:08] <cybrNaut> "PART-TYPE is one of: primary,
logical, extended" <= and part-type is a mandatory 1st arg.
So if that's not the case, the docs must be wrong
2594[23:03:46] <ratrace> TRY IT and see for yourself :) mkpart
bios_boot 34 2047 and then see that /dev/disk/by-partlabel/bios_boot
has been created
2606[23:12:06] <ratrace> but if that's scripting, perhaps
you should use -Z and compeltely wipe out the tables before new
layout. check the manpage for differences
2607[23:13:02] <cybrNaut> -Z is apparently for a damage drive,
and it also quits after executing
2608[23:13:28] <ratrace> -Z does a wipe and doesn't bother
to check if existing tables are okay or not
2636[23:31:23] *** Quits: yans (~yans@replaced-ip) (Quit: chaos is the only true answer)
2637[23:31:37] <cybrNaut> sgdisk docs suck too
2638[23:32:06] <ratrace> now what :)
2639[23:32:38] <cybrNaut> "For example, type sgdisk -A
4:set:2 /dev/sdc to set the bit 2 attribute (legacy IOS bootable) on
partition 4", but if i do sgdisk --list, BIOS boot partition is
"ef02"
2640[23:33:47] <cybrNaut> so the "2" in
"4:set:2" is undocumented
2641[23:34:13] <ratrace> 0xef02 is BIOS_GRUB type for Grub's
GPT stage 2 loader . it's not the same thing as
"bootable" flag for MBR
2642[23:34:31] *** Quits: Grldfrdom (uid391113@replaced-ip) (Quit: Connection closed for inactivity)
2643[23:34:33] *** Quits: NeoCron (~neocron_@replaced-ip) (Remote host closed the connection)
2653[23:38:32] <ratrace> so if you're looking to define the
Grub's GPT "bios_grub" type 0xEF02 partition,
that's not the "bit 2; legacy BIOS bootable" flag