7[00:01:53] <Tenacious-Techhu> jelly, which doesn't
matter; it's a question about openvt and chvt.
8[00:01:57] <jelly> Tenacious-Techhu, and if it's Debian
it's not hard to just nuke all the users with uid 1000-29999,
create a new one and reboot to get rid of all the processes
11[00:03:01] <Tenacious-Techhu> Unfortunately, I have to go AFK
for a bit; I'll have to get back to you about those
complications; but, naturally, that is an entirely separate issue
than getting back to the Virtual Console you opened with openvt with
chvt.
12[00:03:29] <Tenacious-Techhu> I always view helping me out
with the general problem as a secondary issue to answering specific
questions.
13[00:03:36] <jelly> Tenacious-Techhu, as a matter of topic, we
provide help to Debian users here. Ask in ##linux if you don't
actually run Debian, please.
14[00:04:01] *** debhelper sets mode: +l 1272
15[00:04:05] <Tenacious-Techhu> jelly, it's Debian. Please
stop looking for excuses not to answer my question.
16[00:04:12] <Tenacious-Techhu> It'
17[00:04:20] <Tenacious-Techhu> It's a simple question
about openvt and chvt.
24[00:13:01] <jelly> Tenacious-Techhu, sorry. If there's
an easier way to solve user's actual issue as stated I usually
won't bother helping the way they envision a issue should be
dealt with. This veers close to "xyproblem" teritory. Good
luck.
36[00:23:12] <TTT> Hi, pulse-audio with default settings fails
to see my ALSA soundcard (snd-intel) and only creates dummy output.
I can play sound fine using ALSA directly...
37[00:23:55] <TTT> AFAIK pulse uses module-udev-detect, and
logs show "module-udev-detect.c: Found 1 cards.", but does
not way which or what card it detected...
238[04:26:16] <Tenacious-Techhu> jelly, this is nowhere near
"xy problem" territory; it's a legitimate question
with a legitimate answer. You either know it, or you don't. Do
you know whether or not there is a way to chvt back into a Virtual
Console you openvted into, when it doesn't have a number to
pass as an argument to chvt? You either know, or don't know.
Don't beat around the bush implying you are able to solve the
problem when the problem has been prec
239[04:26:36] <Tenacious-Techhu> Either answer it, or say you
don't know. Then you can offer to help with other problems.
240[04:27:04] <Tenacious-Techhu> Implying that more detail will
help you provide an answer to the stated question is misleading
when, in fact, you can't.
245[04:33:33] <Javabean> Tenacious-Techhu: <jelly>
Tenacious-Techhu, sorry. If there's an easier way to solve
user's actual issue as stated I usually won't bother
helping the way they envision a issue should be dealt with. This
veers close to "xyproblem" teritory. Good luck.
246[04:33:57] <Tenacious-Techhu> Javabean, yes, in response to
that.
249[04:34:17] <Tenacious-Techhu> If a user has a question,
answering questions they didn't ask doesn't help anything.
250[04:34:55] <Tenacious-Techhu> If you don't know the
answer, but want to help anyway, make that clear; that way, they can
either turn you down for being off-topic, or consider what else they
might need help with.
251[04:36:40] <Tenacious-Techhu> Generally, I would have turned
down changing the subject; I don't ask questions I don't
care about the answers to; however, I would have been willing to
explore the issue of what else needs to be considered when cleansing
a Debian system of its default users, if he had made clear that he
wasn't capable of answering the original question.
252[04:37:15] <Tenacious-Techhu> Instead, by beating around the
bush, he made me have to drag an answer out of him, even if it was
an answer he wasn't willing to admit.
255[04:38:23] <Tenacious-Techhu> I swear, it pisses me off to no
end when people chalk a simple answerable question up to an "xy
problem", when the question genuinely has a correct answer,
even if it isn't widely known.
256[04:38:34] <Tenacious-Techhu> Some people just ask esoteric
questions! Learn to deal!
257[04:38:34] <_alx_> Tenacious-Techhu: Dude, chill out. It
sounds like you're on a witch hunt
258[04:38:37] <bites> can you rant somewhere else? nobody wants
to read that wall of text.
260[04:39:47] <Tenacious-Techhu> _alx_, when it comes to people
assuming I have an "xy problem", those people should stop
wearing pointy hats, cackling, riding broom sticks, and see a damn
beautician about their sickly palor.
261[04:40:42] <Tenacious-Techhu> I have a "yx
problem"; O
262[04:40:50] <Javabean> Tenacious-Techhu: watch you language
please, this is a publicly logged channel.
263[04:41:23] <Tenacious-Techhu> I want to know y, and people
cannot generally comprehend why I want to know y, so they assume I
really want to know x, when I have no interest in x at all.
264[04:41:57] *** Quits: HugsWontletgo (~HugsWontl@replaced-ip) (Quit: Lost terminal)
265[04:42:04] <Tenacious-Techhu> Nevertheless, so long as y is a
correctly formed question with a factual answer, it's perfectly
reasonable for me to ask, so long as I genuinely want to know.
266[04:42:12] <Tenacious-Techhu> I don't see what people
don't get about that.
267[04:42:44] <_alx_> But you got your answer, why keep going
on?
268[04:42:55] <Tenacious-Techhu> No, I didn't.
269[04:43:05] <Tenacious-Techhu> I was asking about openvt and
chvt.
271[04:43:34] <Tenacious-Techhu> And whether or not there is a
way to pass a non-numeral to chvt when openvt gives you a Virtual
Console without a number.
272[04:44:40] <Javabean> Tenacious-Techhu: do not pm other users
in this channel
273[04:45:34] <Tenacious-Techhu> I deemed the matter of what
exactly about my language needed watching as off-topic for the
channel. How is it not appropriate to then ask exactly what needs
correcting in a pm?
275[04:46:14] <_alx_> I personally know very little about tty,
but it seems logical that you should be able to do it? I don't
know much about openvt and chvt though…
276[04:46:25] *** Parts: Javabean (~Javabean@replaced-ip) ("Once you know what it is you want to be true, instinct is
a very useful device for enabling you to know that it is")
280[04:48:31] <Tenacious-Techhu> _alx_, here's what I find
is happening; very frequently, when I use openvt, rather than
getting "tty1", or "tty" and some number, I get
*just* "tty"; chvt appears to require a number as an
argument, which implies that, when openvt gives you a tty without a
number, you have no way to chvt back to it. I'm here examining
the possibility that chvt's documentation is merely poor, and
that there is in fact some workaround... but the
281[04:48:35] <Tenacious-Techhu> And I
282[04:48:37] <Tenacious-Techhu> Bah.
283[04:48:41] <Javabean> Tenacious-Techhu: see "/msg dpkg
guidelines"
284[04:48:45] <Tenacious-Techhu> And I'd prefer to know one
way or another.
288[04:50:14] <Tenacious-Techhu> _alx_, yes, you can, but then
you risk that vt already being busy; by calling openvt without a
number, you guarantee one that isn't busy.
289[04:50:47] <_alx_> … create a zombie one to use tty ?
lol or use -f
290[04:50:58] <Tenacious-Techhu> Testing whether a vt is busy
would add all sorts of error checking to a script that otherwise
wouldn't need to be there.
291[04:51:10] <Tenacious-Techhu> _alx_, -f?
292[04:51:27] <Tenacious-Techhu> Oh, force it.
293[04:51:34] <_alx_> -f forces open one, regarless of it being
in use
294[04:51:34] <_alx_> yeah
295[04:51:41] <Tenacious-Techhu> Yeah, that might ruin whatever
was keeping that Virtual Console busy.
296[04:51:45] <Tenacious-Techhu> I'm inclined not to do
that.
297[04:52:11] <Tenacious-Techhu> There are definitely
circumstances in which that would be acceptable, but I'm not
inclined to assume this is one of those circumstances.
298[04:52:14] <_alx_> yeah, that part was a joke. But what about
making 2 one after the other instead. one to take care of tty?
299[04:53:15] <Tenacious-Techhu> I'd have to test which one
was a numbered tty, and that would mean passing something from a
child shell to a parent shell, wouldn't it?
300[04:54:26] <_alx_> hmm not sure, i'm really at my limit
of knowledge in this subject man. I never went to school for this
sorta thing lol
301[04:54:37] <Tenacious-Techhu> Yeah, me neither.
302[04:54:44] <Tenacious-Techhu> I tend to exhaust google before
coming here.
303[04:55:48] <Tenacious-Techhu> I tend to be pretty damn good
at researching things, so I only come in places like this if
there's something I legitimately do not understand (rare), or
something I cannot find useful information on (more frequent). And
if the internet doesn't have it, odds are bad that people here
know anything about it.
304[04:56:10] <Tenacious-Techhu> Which makes people being
dishonest about whether or not they have an answer to my *stated*
question really annoying.
305[04:56:46] <Tenacious-Techhu> If I had to guess, it's a
legitimate flaw with chvt; but I don't like guessing, and it
seems like a really weird flaw for chvt to have.
306[04:56:47] <_alx_> not true, i've had good results with
irc in the past. It really depends if the right person is online
307[04:57:07] <Tenacious-Techhu> _alx_, getting the right person
is certainly a challenge.
309[04:58:00] <Tenacious-Techhu> But getting a bunch of people
dancing around the fact that they don't know the answer, and
assume you have some other unrelated problem that you need to solve
doesn't help.
310[04:58:12] <Tenacious-Techhu> I seriously want to strangle
the guy who came up with that "xy problem" crap.
312[04:59:08] *** Quits: cryptodan (~cryptodan@replaced-ip) (Remote host closed the connection)
313[04:59:42] <Tenacious-Techhu> Too often, it's used as an
excuse for people to avoid admitting that they just don't know
the answer to your question, and allow them to insist they have an
answer for you, if only you provided yet another completely
unrelated factoid behind why you want to know the answer in the
first place.
314[05:00:11] <Tenacious-Techhu> Information has value on its
own; it doesn't have to solve a problem today to be useful.
315[05:00:37] <Tenacious-Techhu> And so long as it's
something factual about Debian, I shouldn't be having any
problems getting honesty in here.
317[05:02:28] <_alx_> I get it man, but you're spending a
lot of time on being angry about it. I don't have the time or
the heart to be angry at anyone. Too many problems to solve and too
little time i'm afraid.
318[05:02:39] <Tenacious-Techhu> Javabean, I may be mistaken,
but I think that rule is specifically about package notifications,
and not about in general communication.
319[05:03:25] <Tenacious-Techhu> _alx_, I've said my piece.
Still need an answer to my question. Obviously, I'm going to
have to wait until there are more people online.
324[05:06:20] <Tenacious-Techhu> _alx_, in this case, yes, if
only because you can't specify a command to chvt, like:
"openvt -s -- bash -c " and whatever.
331[05:12:23] <Tenacious-Techhu> _alx_, also, openvt by default
opens an unused Virtual Console; chvt doesn't care whether that
console is busy, it just shows it to you. It doesn't offer any
opportunity for scripted interaction.
332[05:12:29] <Tenacious-Techhu> At least, not on its own.
361[05:47:03] <likcoras> I'm using logrotate to rotate some
logs stored in Logs/*/*/*.log, and I'd like them to be renamed
as Logs/*/*/original-yyyy-mm-dd.log. Will this make logrotate
re-notice those files and try to rotate the old files?
362[05:47:45] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
363[05:48:16] <Tenacious-Techhu> I know nothing of rotating
logs, but it strikes me that a piece of software called
"lumberjack" would be ideal, on account of their
log-rolling contests.
370[05:53:00] <likcoras> Upon re-reading the logrotate manpage,
it seems I need to use olddir. I guess I'll use this for now,
but I wish there is a better way to do this...
392[06:25:34] <likcoras> pantato: stable usually doesn't
get updates unless it's for important bugfixes or security
updates. Have you reported the bug?
402[06:30:12] <likcoras> Hm, probably from downloading from the
archives and installing the .deb manually? Probably not recommended,
though. Also, haven't ever done it, so can't really say
much about it.
440[07:25:20] <Psi-Jack> With an export of `dpkg
--get-selection` to a file, what would be the way to reinstall from
that list of packages?
441[07:26:01] <Psi-Jack> dpkg --set-selections sounds kind of
like it would replace the existing packages (if they differ from
what's installed), but I could be wrong.
442[07:26:01] <dpkg> Psi-Jack: You are person #1 to send an
unparseable request
443[07:30:05] <_alx_> is there a bash way that would make my
life easier? I need all folders (recursively) to change there
permissions to 755, and all files in those same directories to have
644.
444[07:30:12] *** Quits: Uberius (~Uberius@replaced-ip) (Remote host closed the connection)
447[07:33:16] <cfoch> I have created a bootable USB stick with
GNOME Disks but it won't boot if I have UEFI enabled... Ir only
boots using legacy support... Does GNOME Disks actually use dd
internally? How can I boot without disabling UEFI?
448[07:34:01] *** debhelper sets mode: +l 1243
449[07:34:33] *** Quits: kbeflo (~kbeflo@replaced-ip) (Remote host closed the connection)
450[07:37:33] *** himcesjf_ is now known as him-cesjf
508[08:51:42] <Necrony> Hello, i'm using Debian testing
buster 4.16.0-2-amd64 XFCE, trying to install some software from
PlayOnLinux, but they always crash during installation with
"Error in POL_Wine". Any suggestion what I need to do?
564[09:48:12] <TaylorHere> Hi,Guys, I got a issues here, my
/etc/suders broken, and there is no password for root, is there any
solution?
565[09:49:38] <klys> taylorhere, is your bootloader grub?
566[09:49:42] <fpat> TaylorHere: i would boot a livecd/usbdrive
and correct the sudoer file and/or set a password for root (of
course, after chrooting to your system)
688[11:29:08] <Li> maybe I'm crazy so please don't ask
me why > I want to fully install dvd2.iso dvd3.iso files as soon
as finishing installation of debiandvd1.iso
689[11:29:54] <Li> I want to complete/full packages installed at
once .. can anyone give hint on how to do it? I found a link on
apt-cdrom but it's not cleare
690[11:29:57] <Li> clear*
691[11:30:37] <jelly> what is stopping you?
692[11:30:57] *** Quits: graphene (~graphene@replaced-ip) (Remote host closed the connection)
705[11:40:45] *** arahael1 is now known as ArahaelPi
706[11:42:06] <jelly> Li: okay, do you know how to mount .iso
images? How to make /etc/fstab entries for those?
707[11:43:06] <jelly> debian does not guarantee all the packages
are coinstallable. Some packages conflict with others, so you
can't just "install everything" in general
708[11:43:30] <jelly> eg. you can only have one smtp server at a
time
759[12:25:25] <Li> jelly: Yes, I know how to mount them..
it's ok with only one package from the conflicting ones. the
question is still what do next aftering mounting isos?
760[12:25:47] <Li> how to install all (possible) packages
868[14:03:57] <graingert> jelly: it's like a bunch of handy
shell bins
869[14:04:01] *** debhelper sets mode: +l 1280
870[14:04:06] <graingert> like less and gpg
871[14:04:27] <jelly> about half of those are mandatory. Perhaps
you can take a look at standard task
872[14:04:34] <jelly> !standard
873[14:04:35] <dpkg> extra, extra, read all about it, standard
task is a "task" that should be installed on all machines
during the installer stage. It installs packages that are important,
required and standard (i.e. "aptitude install ~pstandard
~prequired ~pimportant"). Packages in the <essential> set
are always installed. Also ask me about <tasksel>.
888[14:18:49] <foxtrotas004> hello, i have a strange problem on
one of my servers , i am trying to configure squid proxy server, but
the strange thing is i can ping my server, i can ssh into it, but
when squid is not working though port 3128 is OPEN , but when i try
to telnet <ipaddress> 3128 i get no route to host found which
is very strange considering, also when i try to telnet
<ipaddres> 22 i can establish a connection no problem
889[14:18:54] <foxtrotas004> has anyone had something like that?
890[14:19:30] <foxtrotas004> at first it looks like a firewall
problem to me
891[14:19:40] <foxtrotas004> so i check my iptables, i have my
port open
892[14:19:50] <foxtrotas004> i think i am tapped out on this
893[14:20:03] <foxtrotas004> service is running i can telnet
localhost 3128 successfully though
894[14:21:42] <fpat> in the first block of text you said telnet
to 3128 leads to an error
895[14:22:45] <p0lyph3m> fpat: <ipaddress> vs. localhost
920[14:35:27] <foxtrotas004> i tried connecting to port 3128
through telnet using ip address i tried remotely
921[14:35:37] <foxtrotas004> and i worked
922[14:35:41] <brokencycle> jelly: thank you!
923[14:35:50] <foxtrotas004> i started to think it might be
firewall problem
924[14:35:56] <foxtrotas004> :(
925[14:35:57] <jelly> ,i less
926[14:35:58] <judd> Package less (text, standard) in
stretch/amd64: pager program similar to more. Version: 481-2.1;
Size: 123.5k; Installed: 277k; Homepage:
replaced-url
927[14:37:08] <p0lyph3m> so you can connect to squid remotly and
localy ? what is it that doesnt work ?
1008[14:58:59] <jelly> foxtrotas004, you're asking in the
wrong place then, but rules in lines 141 and 142 will need to switch
places at the very least
1009[14:59:39] <jelly> so yeah, local firewall
1010[14:59:40] <jelly> !next
1011[14:59:40] <dpkg> Another happy customer leaves the building.
1037[15:14:08] <jelly> well you're on testing, it's bug
reporting time. Perhaps someone in testing/unstable channel has seen
the same issue.
1038[15:14:11] <jelly> !debian-next
1039[15:14:11] <dpkg> #debian-next is the channel for
testing/unstable support on the OFTC network (irc.oftc.net), *not*
on Freenode. If you get "Cannot join #debian-next (Channel is
invite only)." it means you did not read it's on
irc.oftc.net.
1135[16:28:52] *** Quits: RedSoxFan07 (~Thunderbi@replaced-ip) (Quit: Goodbye and [depending on the time] good night.)
1136[16:28:56] <RzR> hi, is there a way to force installation of
package which depends on older package of a version already
installed, I know apt install package subcomponent=$version , but
without telling the version
1289[18:18:24] <devios> "dd if=/dev/md1 | gzip -c | ssh
user@dest-server dd of=/home/user/ddimagemd1.gz" <- I did
this to try to preserve a partition on remote server. Then on the
destination server, i did "gunzip ddimagemd1.gz", then
"mount -t ext4 -o loop,ro ./ddimagemd1 ./ddimagemount",
and i'm getting "mount: wrong fs type, bad option, bad
superblock on /dev/loop0, (Etc....)". Any idea why I can't
mount the image, or anything else i might be doing wrong
1290[18:18:24] <devios> that's preventing this from working?
1310[18:26:39] <greycat> it definitely wasn't a good idea to
copy raw bytes from a mounted partition
1311[18:26:41] <uoods> anyone know if there's any issue
holding up gimp 2.10 from moving from unstable to testing and
stable? I'm not sure there's anything related in the bug
reports
1312[18:26:53] <greycat> (mounted read-write, that is --
read-only would have been OK)
1317[18:28:38] <jhutchins_wk> devios: It looks like you imaged a
device rather than a partition. If that's the case, you'll
have to find the offset to mount the partition within the drive
image.
1324[18:30:48] <devios> sorry tradar and greycat i got
disconnected right after telling tradar the partition was mounted at
time of capture - was a remote server and having a hard time getting
colo to reboot it into recovery mode in a way that I can remotely
access it for the capture.
1325[18:31:39] <jhutchins_wk> Ah, yes, the fstab line.
1326[18:31:49] <greycat> mount output
1327[18:32:07] <devios> ya i pasted a line from output of the
mount cmd.
1352[18:47:11] <devios> md1 is a raid1 of sda5 and sdb5. I wonder
if I could (on the live system) somehow temporarily remove either
sda5 or sdb5 from the md1 raid1, then dd image the removed one the
same way I was trying to image md1. "dd if=/dev/sdb5 | gzip -c
| ssh user@dest-server dd of=/home/user/ddimagesdb5.gz" - is
this possible? do you think it would work?
1354[18:47:55] <GNU\colossus> devios, that would work. but why do
you want to split/destroy your array (for a while)?
1355[18:48:08] <greycat> you'd still be getting corrupted
file system bits
1356[18:48:09] <GNU\colossus> (sorry, I don't have your
prior messages in my scrollback)
1357[18:48:24] <GNU\colossus> ah, so the data on the mirrors is
no good?
1358[18:48:38] <greycat> He copied /dev/md1 (mounted as /
presumably rw) to another box and loopback mount failed and we
don't know if he tried fsck on it yet.
1360[18:49:06] <choice> Hello! I'm trying to backup my
Android phone to my Debian laptop. When I plug it in, I can access
the files in a folder mtp://[usb:...] via the desktop. But how can I
reach the files from the command line?
1361[18:51:03] <Brigo> choice, you need mtp command line tool
1367[18:54:27] <Brigo> choice, i think mtp protocol does not
mount filesystems. I think the file manager is using some tool too.
Maybe you can mount it with fuse ...
1368[18:54:45] <devios> greycat: "losetup /dev/loop0
./TEST" and "fsck /dev/loop0"?
1369[18:54:46] <Brigo> choice, you can check the mounted
filesystem with the mount command
1391[19:00:27] <devios> greycat tradar oops i had already started
"losetup /dev/loop0 ./TEST" and "fsck
/dev/loop0" and got a bunch of "group descriptor #
checksum is invalid. FIXED" and so far have done several
"deleted inode # has zero dtime. Fix<y>? yes". When
this finishes (could be a while) i can re-gunzip and do the
"fsck.ext4 ./ddimagemd1". might take a good long while
before I get there. thx for the help both/all.
1392[19:00:37] *** Quits: faw (~faw@replaced-ip) (Remote host closed the connection)
1487[20:10:10] <SerajewelKS> when a new kernel is installed, dpkg
runs update-grub. fine, great. the config looks fine.
1488[20:10:18] <SerajewelKS> but grub drops to a rescue shell
without giving me any error
1489[20:10:32] <SerajewelKS> if i then boot into the install CD
in rescue mode and run update-grub and reboot, grub gives me a menu
1490[20:10:48] <SerajewelKS> i captured the two grub.cfgs and
compared them and they are identical
1491[20:10:54] <SerajewelKS> so that's weird issue #1 that i
can't explain
1492[20:11:14] <SerajewelKS> weird issue #2 is that every kernel
update, i get "invalid magic" from grub until i reinstall
the updated kernel package
1493[20:11:19] <Brigo> SerajewelKS, the grub versions are the
same?
1494[20:11:28] <SerajewelKS> the only thing that changed is the
kernel
1495[20:11:37] <SerajewelKS> and, i can't stress this
enough, the two grub configurations are identical
1496[20:11:48] <SerajewelKS> for some context, / is in an LVM in
a RAID1
1497[20:11:54] <Brigo> SerajewelKS, not the config, the program
version.
1498[20:12:07] <SerajewelKS> 18:11 < SerajewelKS> the only
thing that changed is the kernel
1499[20:12:18] <SerajewelKS> this happens every kernel update
1500[20:12:51] <SerajewelKS> i have a "bios boot"
partition on all drives
1501[20:13:23] <SerajewelKS> i just can't figure out how
regenerating an identical config causes the system to be able to
boot, unless it's something aside from the file content. such
as which disk block the config winds up stored in.
1502[20:14:26] <Brigo> have you checked the config files with
diff ?
1503[20:14:34] <SerajewelKS> yes, that's how i know
they're identical
1504[20:14:44] <SerajewelKS> they sha1sum the same, too
1517[20:21:23] <SerajewelKS> and on top of it all, the error
messages from grub flash on the screen for a fraction of a second,
so i don't even know what the problem is.
1518[20:23:05] <SerajewelKS> it *seems* like grub is able to load
stage2 but for some reason can't load the config
1519[20:24:23] <jhutchins_wk> SerajewelKS: Is grub maybe writing
something to the logs?
1520[20:24:59] <greycat> there are no writable log files at the
time grub is being executed as a boot loader
1521[20:25:45] <jhutchins_wk> SerajewelKS: Is grub using device
names or UUIDs?
1522[20:25:48] <SerajewelKS> the puzzle i can't figure out
is why regenerating the config file when the produced file is
identical to the "bad" one somehow allows the box to boot
1523[20:25:54] <SerajewelKS> UUIDs
1524[20:26:32] <SerajewelKS> ok i just reinstalled the kernel,
ran update-grub, and then grub-install on all boot devices. now grub
gives me "invalid arch-dependent ELF magic" and failed to
load stage2.
1526[20:27:09] *** Quits: [PinkSpark] (~while@replaced-ip) (Quit: There are paranoid crossroads and shipwreck alleys.)
1527[20:27:13] <SerajewelKS> ten bucks says if i do that all
again (sans reinstall kernel) from rescue mode of the installer, i
will see the grub menu on the next boot
1528[20:27:44] *** Quits: sauvin (sauvin@replaced-ip) (Remote host closed the connection)
1530[20:28:03] <SerajewelKS> it's like my system kernel is
randomly writing bad data to the disk, while the installer CD kernel
doesn't
1531[20:28:35] <greycat> Not the kernel, but the update-grub
stuff.
1532[20:28:52] <SerajewelKS> i would agree except that when i run
that from the installer, i'm chrooted into the system's /
1533[20:28:54] *** Quits: ToBeCloud (uid51591@replaced-ip) (Quit: Connection closed for inactivity)
1534[20:28:58] <SerajewelKS> so it should all be the same stuff
1535[20:29:09] <SerajewelKS> the only piece that's different
is the kernel
1536[20:29:43] <SerajewelKS> BUT i still don't get why
running update-grub fixes the problem, even though the generated
config is identical. they have matching sha1sums.
1537[20:29:44] <xaeB5> do i need to install anything in order to
run a c++ file in debian? looks like i have a "cpp"
program installed already and am trying to file a file.cpp program
1538[20:29:50] <SerajewelKS> one works, one doesn't. i
don't get it.
1541[20:30:36] <SerajewelKS> xaeB5: you don't run C++ files,
you compile them. "cpp -o program file.cpp" will create a
file called "program" that you should then be able to run,
assuming there were no compilation errors.
1543[20:31:12] <greycat> cpp isn't the C++ compiler;
it's the C preprocessor
1544[20:31:21] <greycat> the C++ compiler is named g++ or c++
1545[20:31:37] <SerajewelKS> greycat: i'm trying to come up
with an explanation for this that doesn't involve disk
corruption, and i can't. but it's weird that the rescue
environment doesn't appear to have this problem.
1547[20:31:50] <greycat> it does not sound like file system
corruption
1548[20:32:04] <xaeB5> greycat: should i use g++ or c++ to
compile?
1549[20:32:26] <greycat> xaeB5: if someone else wrote this
program, you should follow their instructions to build it
1550[20:32:31] <SerajewelKS> ok, then, if i have two bit-for-bit
identical grub.cfg's, why would one bring me to the grub menu,
and the other drop me to a grub shell?
1551[20:32:38] <xaeB5> it's just a hello world i wrote
greycat
1558[20:34:14] <SerajewelKS> here's what i observe: running
update-grub from the booted system produces a config that for some
reason can't be read, because i get dropped to a grub shell.
running update-grub while chrooted into the same system from rescue
mode produces a config that has the same sha1sum but works.
1559[20:34:49] <SerajewelKS> running grub-install from the booted
system causes grub to fail to load stage2 with "invalid
arch-dependent ELF magic." doing the same while chrooted into
the same system from rescue mode produces a working grub install.
1560[20:34:53] <SerajewelKS> i don't get it.
1561[20:35:01] <greycat> you keep focusing on "the
config" instead of the other pieces
1562[20:35:19] <greycat> it very clearly talks about executable
code ("ELF magic") which is not config
1563[20:35:34] <SerajewelKS> yes, that's the grub-install
oddity, not the update-grub oddity
1566[20:35:52] <SerajewelKS> if i don't mess with
grub-install, then i get a grub shell with NO error message
1567[20:36:26] <SerajewelKS> so let's ignore grub-install,
since that's not required to resolve the first problem.
that's just me trying to see what could fix it without using
rescue mode.
1570[20:37:11] <SerajewelKS> if i run update-grub in the system
itself, the next boot fails. if i run update-grub in the system
chrooted from the install environment, it works. despite producing
an identical config. so should i copy all of /boot in both cases and
see if anything else changed?
1571[20:37:42] <SerajewelKS> i didn't think update-grub
touched anything except grub.cfg
1572[20:38:06] *** Quits: Darcidride (~Darcidrid@replaced-ip) (Remote host closed the connection)
1585[20:42:18] <greycat> No, it's Debian's asinine and
idiotic "all shell scripts must have set -e" policy.
1586[20:42:32] <dutchfish> greycat, can it be, by accident that
your shell is configured to continue on non zero?
1587[20:42:33] <SerajewelKS> to reiterate from a clean slate: i
upgrade kernel, dpkg kicks off update-grub as it should, i reboot,
get a grub shell with no error. i boot installer in rescue mode,
chroot into my system, copy bad grub.cfg, run update-grub, compare
to bad grub.cfg, they are identical. i reboot, system boots fine.
1588[20:42:34] <greycat> They even put it in /etc/rc.local back
when they still shipped an /etc/rc.local.
1591[20:42:59] <dutchfish> greycat, tes, seen that several times
1592[20:43:51] <dutchfish> greycat, the whole point of set (and
the miriad of changes in latest shells) is to make sure to break at
non zero. at least on -e?
1597[20:45:39] <dutchfish> greycat, bash is not th eonly kid on
the block
1598[20:45:59] <greycat> If you can't see that a set -e is
pointless and stupid in a three-line script where the first line is
the shebang, the second line is set -e, and the third line is exec
something, then this conversation is pointless.
1599[20:46:02] <dutchfish> (just thinking out loud)
1600[20:46:38] <dutchfish> fair enough, i dont use bash btw
1610[20:48:43] <cyclic3> When clicking on the recent posts bar,
one post listed a set of hidden services, which it claaimed linked
to child pornography
1611[20:49:14] <dutchfish> SerajewelKS, try this on a recent
Debian system: ls -la $(which sh)
1613[20:49:46] <dutchfish> and i admit its confusing sometimes
and bash has many faults
1614[20:50:06] <SerajewelKS> greycat: so here's another
headscratcher. i can boot 4.9.0-6 but not 4.9.0-7. with -7 i get
"error: invalid magic number". but running
"file" on /boot/vmlinuz-4.9.0-7-amd64 identifies it as a
kernel.
1615[20:50:45] <SerajewelKS> these are separate issues but
they're painting a scary picture in my head
1616[20:50:48] <cyclic3> any idea on how to report this? I hate
to be one of those people, but this is something that people
generally frown upon
1617[20:50:50] <jhutchins_wk> SerajewelKS: That suggests it might
be a module error.
1633[20:55:58] <SerajewelKS> (IIRC, DOS disklabel has some 512
bytes that were reserved and never used, where grub stores stuff.
GPT doesn't have that wasted space, so an additional partition
is required to store that data.
1634[20:56:32] <SerajewelKS> but only when GPT is used with BIOS
1638[20:57:12] <greycat> So instead of having 1 MB of empty space
at the start of your disk, you have an actual PARTITION there?! Is
this partition ever mounted?
1639[20:57:27] <greycat> If this partition ever gets mounted, I
think we found your problem.
1640[20:57:46] <SerajewelKS> yes, there is a partition with the
"BIOS boot" type. it is not mounted. grub-install writes
data there, similar to how it writes data to the MBR for DOS
disklabels.
1641[20:58:16] <SerajewelKS> remember that / is in LVM in RAID1.
so the grub.cfg does not mention partition numbers, it's
looking for a RAID superblock UUID and then an LVM UUID.
1642[20:58:40] <SerajewelKS> but regardless, the config that
works and the config that doesn't work are identical.
that's what i can't figure out.
1643[20:59:10] <greycat> I'm still thinking that something
is mounting this partition, or fsck-ing the file system if it has
one...
1644[20:59:21] <SerajewelKS> it has no filesystem
1645[20:59:25] <greycat> Which would cause data corruption of
your Master Boot Record (analogue).
1646[20:59:49] <SerajewelKS> IIRC it's the same blob that
grub stores in the MBR of a DOS disklabel
1651[21:04:03] <jim> SerajewelKS, hmm, you must have a GPT
partition table
1652[21:04:12] <SerajewelKS> yep, i said that above
1653[21:04:32] <jim> just sat down, didn't see that :)\
1654[21:05:00] <SerajewelKS> i'm considering switching the
system to a DOS disklabel just to see if that fixes it, but that
means failing one of the RAID1's, repartitioning it, rebuilding
the array, etc etc
1665[21:07:18] <greycat> I didn't source-dive into
grub-mkconfig or even go from the man page to the *shudder* info
page.
1666[21:07:50] <SerajewelKS> right now i'm in the processing
of fscking / to make sure that there's not some lingering
problem
1667[21:08:02] *** Quits: kirk781 (~kirk781@replaced-ip) (Remote host closed the connection)
1668[21:08:11] <SerajewelKS> FWIW i copied /boot and then re-ran
update-grub from the install CD (in chroot) and did a diff -r, there
were no differences
1669[21:08:17] <SerajewelKS> but the box booted after doing it
from the install CD
1670[21:08:34] <SerajewelKS> so i'm even more perplexed
1671[21:09:15] <SerajewelKS> i'm not exactly 100% sure how
stage2 loads grub.cfg or even knows where to find it
1672[21:10:02] <SerajewelKS> i suspect that's the part
that's breaking down but i don't know enough to figure out
why
1673[21:10:32] <greycat> Try #grub maybe.
1674[21:11:56] <SerajewelKS> wish this server didn't take 5
minutes to POST, makes debugging this crap really annoying
1677[21:16:18] <SerajewelKS> googling around all of this stuff
has led me to a thread that suggests that broken UDMA in the BIOS
can cause it, and that turning off UDMA will make grub work
correctly. then the kernel ignores the BIOS UDMA anyway. so
that's what i'm going to try next.
1686[21:22:12] <jak2000> i type these command: ln -s /backups/
/home/jak/ftp/backups but hen the server restart loose the link, how
to do perrmanently? thanks
1687[21:22:18] *** Quits: Iridos (~iridos@replaced-ip) (Remote host closed the connection)
1688[21:22:22] <SerajewelKS> i guess i can't disable UDMA
from the BIOS :/
1689[21:22:52] <greycat> jak2000: figure out why your
/home/jak/ftp/ directory is being altered without your knowledge or
consent. Then stop it from doing that.
1716[21:44:42] <SerajewelKS> greycat: 100% reproducible, if i
grub-install from inside the system i get "invalid
arch-independent ELF magic" but if i grub-install chrooted into
the system from the install CD, it will boot
1720[21:50:35] <plantroon> can you guys give me reasons why to
install Debian over Parabola Linux (Arch Linux with free software
only)? Is this question ok to ask here? :D
1721[21:50:37] <SerajewelKS> looking around with ls, i can see
the stuff in /boot/grub just fine
1722[21:50:47] <SerajewelKS> but "insmod normal" fails
with the same error, like it's reading the wrong content
1723[21:51:06] *** Quits: symtex (~symtex@replaced-ip) (Remote host closed the connection)
1724[21:51:19] <SerajewelKS> i can insmod test, hfs, lvm, other
random crap... but not normal
1726[21:51:50] <SerajewelKS> plantroon: because you want to use
Debian instead of Parabola Linux
1727[21:53:17] <plantroon> SerajewelKS: I am trying to justify
it. But even regarding stability, I've never had problems with
Arch/Parabola for years. Basically I work at a company that uses
Debian and I would like to switch all of my own servers/VPSes to it.
1728[21:53:20] <annadane> arch is rolling, debian stable's
software is older, that's the most important difference
1729[21:53:41] <annadane> there's sid if you want something
newer or backports/compiling your own software
1730[21:54:16] <annadane> well, for servers, i'd use stable
1731[21:54:55] <plantroon> definitely, sid seems very unstable
and in the past I had package breakage. Like there were some weird
conflicts that I didn't bother solving :D
1735[21:56:44] <SerajewelKS> you don't need sid unless you
know a specific reason why you need sid. and even then you probably
don't need sid.
1736[21:57:14] <plantroon> on servers I would only go stable to
live a carefree life :D but I have that now with Parabola so really,
if I switch it would just be "cosmetic" and not worth the
time maybe.
1737[21:57:18] <SerajewelKS> i'm not going to try to justify
switching to debian for you. that's up to you.
1744[22:01:26] <annadane> (i'd also recommend stable for
general desktop use, i only mentioned sid because they mentioned
their arch background so more used to solving problems, but stable
is fine, one only needs newer software in special circumstances)
1745[22:01:44] <annadane> and stable is... stable, in every sense
of the word
1746[22:02:58] <plantroon> yea, I have a lot of experience with
Debian stable, none of which is bad. Basically the thing I slightly
dislike about Debian is that it modifies upstream packages more than
Arch does.
1756[22:06:35] <plantroon> somiaj: the mariadb user being named
debian-sys-maint does not fit that description. That is an
unnecessary upstream modification imo.
1771[22:11:09] <annadane> i'd definitely recommend stable
over arch for stable use, rolling distros aren't good for
servers
1772[22:11:12] <annadane> otherwise, your choice
1773[22:11:24] <annadane> debian has a large repository, there
are pros and cons in every area otherwise
1774[22:11:41] <annadane> !why debian
1775[22:11:41] <dpkg> Debian strives to maintain your freedom
whilst also paying close attention to the technical aspects of
making a great OS. Debian is stable, upgradable and well tested. See
also
replaced-url
1776[22:11:54] <plantroon> I've run Arch on servers for
years, for some reason nothing ever broke. At least not any more
frequent than Debian.
1777[22:12:24] <annadane> i guess that's fair
1778[22:12:26] <SerajewelKS> i'm still getting "invalid
magic number" trying to boot 4.9.0-7, while 4.9.0-6 boots fine
:/
1779[22:12:30] <annadane> i've heard of some people running
arch in their servers
1782[22:15:00] *** Quits: madage (~madage@replaced-ip) (Remote host closed the connection)
1783[22:15:41] <symtex> I'm using openvswitch and
configuring bridges/ip addresses via the interfaces file. for some
reason during boot two services openvswitch and networking hang to
raise network interfaces until it times out 5 minutes later causing
a slow boot. any ideas?
1803[22:25:46] <somiaj> SerajewelKS: correct, grub-install
installs grub into the mbr or the efi partition of the disk
(depending on your setup) grub-install will configure grub to be
able to find /boot/grub.
1804[22:25:55] <somiaj> update-grub only modifies the files in
/boot/grub afiak
1805[22:26:58] <SerajewelKS> right. sigh. no idea why this
won't work.
1806[22:27:01] *** Quits: _0bitcount (~Big_Byte@replaced-ip) (Remote host closed the connection)
1807[22:27:15] <SerajewelKS> i'm starting to suspect the
BIOS and/or system firmware
1826[22:32:22] <SerajewelKS> considering i'm chrooted, the
only part that's different is the kernel. as well as whatever
ephemeral state is created by the services running on the system
when it's booted normally
1827[22:34:16] <SerajewelKS> this on the other hand somewhat
refutes faulty hardware, unless the "bad" environment is
completely reliably triggering whatever the hardware problem is, and
the installer CD environment reliably does not
1855[22:52:33] <somiaj> It is often suggested to keep /boot
outside of arrays, but unsure on the details here
1856[22:52:53] <SerajewelKS> i have a non-zero mismatch_cnt, but
i have swap in the raid1 and my understanding is that pages being
swapped but not yet written to both disks that get unmapped
prematurely can result in an abort of the write, leaving a mismatch
in the swap space, which is not an issue
1857[22:53:15] *** Quits: platvoeten (~platvoete@replaced-ip) (Quit: Lost terminal)
1858[22:53:52] <SerajewelKS> and the layer that verifies mirrors
doesn't know or care where the swap areas are
1859[22:54:26] <somiaj> this is just a bit outside my experience,
only going off of what litle I know about arrays.
1860[22:54:30] <SerajewelKS> in hindsight i should have made the
swap areas raw partitions outside of the raid
1861[22:54:46] <somiaj> do you have a second disk you could put
/boot on (just for testing) to see if the problem is with the array
or not?
1862[22:54:47] <SerajewelKS> grub2 does understand linear lvm as
well as dmraid
1863[22:55:04] <SerajewelKS> i have a second disk but it has no
unallocated space
1864[22:55:13] <somiaj> grub2 can even do encrypted /boot too,
but I don't think debian supports that by default.
1865[22:55:34] <somiaj> it could not be the problem, but my
limited experience is to keep /boot outside of the array
1866[22:55:51] <SerajewelKS> yeah i just didn't want to futz
around with having two /boots that i have to keep in sync
1867[22:56:28] <SerajewelKS> the point of a mirror is to mirror
:P
1868[22:56:58] <SerajewelKS> plus the installer set all of this
up. so i assume it is a supported configuration.
1869[22:58:11] <somiaj> my info could be outdated, as it may not
be that big of a problem anymore, as I said I hear grub can even
support encyrpted /boot
1870[23:00:03] <SerajewelKS> yeah i think the overall issue is
one of three things: (1) bad hardware doing weird things, but
reliably, or (2) grub not properly understanding either lvm or
dmraid, or (3) mirror failure of some kind
1871[23:00:12] *** Quits: CyberHacker (~CyberHack@replaced-ip) (Remote host closed the connection)
1872[23:00:27] <SerajewelKS> i'm going to try to rule out
(3) by forcing a rebuild
1874[23:01:02] *** Quits: jhutchins_wk (~jonathan@replaced-ip) (Quit: Lost terminal)
1875[23:01:15] <SerajewelKS> it's just weird that the issue
reliably presents when booted into the system, and reliably does not
when booted into the installer
1876[23:01:48] <SerajewelKS> and /boot seems to be the only thing
that is affected, and on a regular basis
1881[23:06:17] <SerajewelKS> this is not the first weird boot
issue i've had. i had a VM where update-initramfs would
generate presumably-bad images that grub would complain about.
1882[23:06:27] <SerajewelKS> but again, booting the CD in rescue
mode and chrooting in would generate a good one
1883[23:06:35] <SerajewelKS> i compared the two and the raw files
were different
1884[23:06:37] <SerajewelKS> but the contents were not