12[00:15:43] <jhutchins> I've disabled a couple of sites I used to host, and now certbot renewal fails becuse it tries to validate the sites. How can I get around this?
13[00:16:31] <jhutchins> If I need to just remove the existing certificates and files from letsencrypt and generate a new cert, which files do I remove and which do I retain? Can I just purge certbot and reinstall it?
53[00:57:14] <lembron> hi, grub-update today/yesterday has me confused... i just updated a kernel like a week ago if even - everything was alright, but now with "only grub in upgrade" it aborts and yells about beeing unable to write boot - to any of my disks (nvme0, nvme1, (and obvs.) md2) - so.... whats up there? "grub-install: warning: this GPT partition label contains no BIOS Boot Partition; embedding won't be
65[01:07:56] <petn-randall> lembron: Oh, you have a EFI system partition. This might mean that you're not booted in EFI mode, which might confuse grub.
66[01:08:08] <lembron> also mounts has: "/dev/nvme0n1p1 /boot/efi vfat"
68[01:09:02] <petn-randall> lembron: To clarify, you *should* be booted in EFI mode, as you apparently installed the system in EFI mode. Which I assume from the fact that you have an EFI system partition.
75[01:22:41] <jhutchins> petn-randall: I don't have that folder, so I guess I'm booted in bios mode, but when I start, I get a message "invalid partition table". If I hit any key, the boot continues normally.
76[01:23:12] <jhutchins> petn-randall: Any idea what I need to do to fix that? (This is on stretch).
77[01:24:31] <petn-randall> jhutchins: Invalid partition table sounds like an error from the BIOS? Not sure what that means.
78[01:25:52] <petn-randall> lembron: If you run `grub-install /dev/nvme0n1`, what does that give you? You might also want to use -v and/or --recheck.
79[01:26:50] <jhutchins> fdisk doesn't show an efi partition. I think I might have started an efi build, but it failed for some reason and I went back to bios.
80[01:27:14] <jhutchins> I should look in the bios when I'm ready to upgrade to buster, see what it thinks it's doing.
82[01:27:33] <petn-randall> jhutchins: If you install and boot via legacy mode, that's completely fine. In that case you don't have a EFI system partition on your GPT, but rather a BIOS boot partition.
83[01:28:28] <jhutchins> D'Oh, I do have an efi partition.
84[01:29:15] <petn-randall> lembron: Now I'd try running the same apt command as before to see if the issue is gone. Might be that you had an outdated device map, or had grub configured to install to some other device.
86[01:29:40] <petn-randall> I've had grub issues so seldomly that I've never bothered to dig to the root cause.
87[01:31:17] <lembron> petn-randall well i think ive basically skipped past apt by telling it "ignore" so i dont know what exact command was run there.. a) should i run that for nvme1n1 too then? b) i should have another identical server with the grub update pending, i may try to run it there before apting? (both disk again id presume)
110[01:39:45] <sney> in fail2ban, the filter scans the logs and reports matches. the action is what happens as a result. it's possible for one to work without the other. if you haven't looked at your config, you should do that now, because there may be something missing or needing you to make a choice
111[01:39:57] <lembron> hop depending on what it works on there might be lag, or the rule it sets is ineffective (wrong ports / broken order), see iptables-save
112[01:40:28] <hop> most ips are banned, just some are not. and as i said, this is the debian standard config
113[01:41:53] <hop> besides, if i were able to identify a problem with the config files, i wouldn't have come here. i've tried.
126[01:47:45] <aaro> hop: how frequent are the logs for those addresses? doesn't fail2ban has a time frame to apply a ban?, like ie 3 attempts in less than five minutes
129[01:48:28] <hop> aaro: we are talking dozens a minute, hundreds in the 600s configured by default
130[01:49:01] *** debhelper sets mode: +l 1127
131[01:49:08] <hop> hm. i'm running out of test cases. it seems the storm is passing :-/
132[01:50:52] <hop> it also seems the baddies got wise to the default configuration. most ips keep well away from the 5 in 600s of the default installation
133[01:52:17] <hop> is it conceivable that f2b just can't cope above a certain level? because now that the attempts have gone down, i can't see the problem anymore
170[02:25:16] *** Quits: bertbob (~bertbob@replaced-ip) (Remote host closed the connection)
171[02:25:21] <lembron> petn-randall f2b? well he said "all default" and i belive default is just new
172[02:25:23] *** Quits: voidSurfr (~todd_dsm@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
173[02:25:59] <lembron> also, made reboot on the node, seems all well... - gonna try with another one now and run the grubcmd there before the install to see what happens :P
192[02:40:58] <timothywcrane> trying to compile a list of languages that are used or included in the packages of a min netinstall server. So far I have come up with C, Python (2&3), Perl and AWK, and bash if you include it as a language... anybody care to drop the names of others?
193[02:41:48] <timothywcrane> only utilities and ssh teasksel included... nothing web or DE included
220[02:58:35] <sney> CyberManifest: the images are semi-unofficial but maintained by a debian developer, as you can see in the page footer. this is because installing to a rpi involves debootstrap, which is more difficult. you're still getting a pure debian OS out of the deal.
221[02:58:57] <quadrathoch2> yeah wanted to say that
222[02:59:13] <quadrathoch2> if you want to do it yourself, maybe also head over to #debian-raspberrypi on OFTC
224[02:59:51] <CyberManifest> sney: if it's semi-unofficial the how is it any different than say Raspbian or Raspberry Pi OS ?
225[03:00:04] <CyberManifest> then*
226[03:01:01] <sney> CyberManifest: because the *image* may be unofficial but the *contents* are not. raspbian is a separate project with separate developers and repos. this is more like the nonfree firmware installer, it's "unofficial" but you're still installing debian
227[03:01:01] <CyberManifest> also are these 32 bit images or 64 bit ?
240[03:06:24] <quadrathoch2> Eh on the graphical stuff i can’t really talk about. As I am using it as a small server. But vlc shouldn’t be an issue imho.
241[03:07:19] <quadrathoch2> And 4) shouldn’t even be at all an issue as Debian doesn’t use the raspi kernel
243[03:09:10] <CyberManifest> sney: I have a mSATA over USB HAT attached to my Raspberry Pi 3 B+ and I'd like to be able to "Install" debian to that mSATA drive; after reading at raspi.debian.net, it mentions no installer; is there a way I can get a qualified installer to use on that mSATA drive?
244[03:09:46] *** Quits: Tom01 (~tom@replaced-ip) (Remote host closed the connection)
245[03:10:36] <sney> the regular debian armhf/arm64 installer might work. I don't have any personal experience but I imagine they were tested on the rpi due to its popularity
246[03:10:59] <sney> I would expect some possible issues with configuring the bootloader but I'm sure that information is available
247[03:11:11] <quadrathoch2> You would still need an sd card cybermanifest :/
251[03:12:22] <CyberManifest> quadrathoch2: temporarily, the HAT comes with software that negates the need for an SD card, you can read about it: it's an element14 Pi Desktop
253[03:12:52] <quadrathoch2> How would that work, as the firmware doesn’t support it (except you got the beta)
254[03:13:54] <CyberManifest> quadrathoch2: yes the firmware does support it on the 3 B+ as I've booted from the mSATA on numerous occassion without so much as an SD card in the slot
255[03:14:07] <CyberManifest> occasion*
256[03:14:19] <quadrathoch2> Ahh kk, so at some point I really need to get myself a new pi
257[03:15:05] <CyberManifest> quadrathoch2: the way it works, is it still checks the SD card if nothing is there then it takes a while and eventually boots over USB
309[03:46:09] <CyberManifest> quadrathoch2: already having hardware issues... the key repeat is sensitive, it automatically repeats and once after doing an install package it kept doing a carriage return
310[03:50:32] <CyberManifest> like I couldn't even open a .sh file in nano, it was scrolling the screen and doing carriage returns, and when I go to exit it wants to know if I wanna save the buffer, because of all those carriage returns
332[04:10:28] <CyberManifest> I want to pull in some raspberry pi tools like rpi-config from Raspbian's sources, but they are 32 bit, is there a way to enable 32 bit support here in 64 bit debian?
339[04:13:31] <lembron> petn-randall some more follow up on the grub thing - just noticed munin (or rather icinga rrd check) complains df-_dev_nvme0n1p1-g.rrd : data are too old, 30-Jul-2020 02:55:07 -- nvme0n1p1 itself is and was totaly "there" in fdisk at least - and "looks the same" as on a not-yet-upgraded machine --- and all boot fine so... whatever i guess...
344[04:16:44] <nvz> CyberManifest: its pretty simple.. single command to enable i386, and you can also specify in the sources.list [arch=amd64] or such for specific lines
345[04:17:15] <CyberManifest> nvz: yeah I'm seeing that
346[04:17:38] <nvz> the specification isnt required or anything, its just an option
348[04:18:11] <nvz> CyberManifest: things like wine wouldnt really work without multiarch
349[04:19:36] <nvz> CyberManifest: its just a bit of a pain cause you need seperate copies of libs and such and when installing you just put :i386 or such after the package name
351[04:19:59] *** xparanoik9 is now known as xparanoik
352[04:20:19] <CyberManifest> nvz: thanks for the tips ;)
353[04:20:37] <sney> mixing arm64 with armhf will probably be a slightly different experience than the 32/64 x86 situation, but still good stuff to know
356[04:23:23] <CyberManifest> so on that page I'm seeing an example of: deb [arch=armel,armhf] replaced-url
357[04:23:46] <nvz> yes everything is more annoying on arm.. that time I fubarred my debian install and was trying to kexec the debian kernel from android I ran into problems and after much digging found some undocumented kexec option that made it work
359[04:24:45] <nvz> CyberManifest: you can get by not specifying anything.. the sources list specifications are only if you want to limit what it'll get from that repo.. once you add the architecture, apt will get all archs you added from any repo that doesnt specify
360[04:24:47] <sney> CyberManifest: I know it's tempting to skip the instructions and go right to the examples, but "It is not normally necessary on Debian" is true. the [arch=] fields are only necessary in some situations
362[04:25:35] <sney> add the source as normal and only specify the arch if apt yells about it. and yes, you can put just 1 arch in that field
363[04:25:45] <nvz> CyberManifest: one case you may specify in sources.list is if the repo in question doesnt have all archs your package manager has.. and you want to avoid errors/warnings on update
364[04:25:54] <CyberManifest> sney: it did yell at me saying it was 32 bit
381[04:35:27] <CyberManifest> I don't understand I can't find rpi-config in the repos, I found it on github but I don't know what it's dependency looks like
382[04:35:55] <sney> ,i bin/rpi-config
383[04:35:56] <judd> No package named 'bin/rpi-config' was found in buster/amd64.
392[04:38:22] <sney> yeah, so you can apt install devscripts and build-essential, download that source tree, do a 'fakeroot debian/rules binary' and see what happens
393[04:38:59] *** Quits: fnaticrisk (~fnaticris@replaced-ip) (Remote host closed the connection)
394[04:39:16] <sney> only debhelper in build-depends so that will Just Work probably. won't know until you try though
401[04:44:26] <CyberManifest> it's not totally working cause all the little programs it depends on aren't installed. just got an error for wpa-supplicant
402[04:44:39] <CyberManifest> what package is that in ?
424[04:54:36] <judd> No packages in buster/armhf were found with that file.
425[04:54:53] <CyberManifest> sney, even with the dependancies it gives me errors saying it can't communicate with the wpa-supplicant
426[04:54:55] <sney> there, now you know as much as I do. (you can also /msg judd, so you don't clog the channel)
427[04:55:10] <sney> did you install wpasupplicant?
428[04:55:25] <CyberManifest> sney: said it was already installed and the newest
429[04:55:49] <sney> it may be looking for some raspbian-specific thing that debian does differently.
430[04:56:18] <CyberManifest> sney: and I didn't figure rpi-update would be in debian's repos, I meant raspbian's repos; not sure how to check or what the proper name of the package is to check for
431[04:56:33] <CyberManifest> sney: yeah probably
432[04:57:13] <sney> if you just want to connect to wifi using wpa, replaced-url
434[04:58:25] <sney> if you have raspbian's repo added and you can't see packages that you know should be present (e.g. apt policy rpi-update) then you may need to run apt update again.
435[04:59:04] *** Quits: k4nz (~Thunderbi@replaced-ip) (Remote host closed the connection)
437[04:59:20] <sney> doing something like 'apt list rpi*' would show you available packages that begin with that string, without spamming your screen with 1000 short descriptions
438[04:59:25] <CyberManifest> sney I did, I'll check that wiki for debians method, honestly I wanted this thing to set my locales like keyboard, language, time-zone etc.
441[05:01:03] <CyberManifest> sney, I figure if I change my keboard and locals my keysensativity will stop, as of right now local says POSIX instead of en-UTF8
443[05:02:48] <sney> I'm not sure what you mean by key sensitivity, but anyway: dpkg-reconfigure is your friend. call it with 'locales' and 'keyboard-configuration' for those
444[05:03:13] <sney> !refcard
445[05:03:14] <dpkg> A short reference of Debian commands - quite useful for someone new to Debian - can be found at replaced-url
446[05:03:31] <CyberManifest> sney: like dpkg-reconfigure locales ?
447[05:03:35] <sney> yes
448[05:03:47] <CyberManifest> sney: awesome, thank you :)
449[05:06:09] *** Quits: JustTheDoctor (~iam@replaced-ip) (Remote host closed the connection)
450[05:07:57] <CyberManifest> sney: that keyboard-configuration didn't let me pick my keyboard type: 103 key vs 104 key etc.
451[05:08:47] <sney> what happens instead?
452[05:08:51] <CyberManifest> sney: and dpkg-reconfigure locales let me pick my locale but didn't change anything... when I run locale it still shows POSIX
453[05:09:09] <sney> locales is session-dependent so log out and log back in to be sure
454[05:09:18] <CyberManifest> sney, for the previous question, It only let me pick my keyboard language, that was it
455[05:10:06] <sney> if you 'apt policy keyboard-configuration' is your installed version from debian?
468[05:17:59] *** Quits: yonder (~yonder@replaced-ip) (Remote host closed the connection)
469[05:18:12] <CyberManifest> sney ^
470[05:18:16] <sney> that's less than illuminating.
471[05:18:58] <CyberManifest> sney: well you asked about: apt policy keyboard-configuration
472[05:19:16] <sney> you're kind of getting into frankendebian territory now, just fyi. you may want to look at the docs for apt pinning (don't ask me, I don't use it, look it up on the debian wiki and/or the apt manual) and directly set the raspbian repos to a lower priority
473[05:20:00] <sney> but anyway, according to this, the keyboard-configuration package *should be* identical between debian and raspbian since they didn't change the version string.
474[05:20:37] <CyberManifest> sney: actually, I think I'm going to remove the raspbian repos and 32 bit arch, it's not serving me how I thought / intended anyways so I'll just have to adjust to debian ways
476[05:21:28] <sney> I think your best bet for informed support is to go to #debian-arm on OFTC. it's a smaller channel and you probably won't get responses as quickly, but they will do less guessing than I'm doing now.
477[05:22:10] <sney> but yes removing the raspbian repos is a good idea
478[05:22:13] <CyberManifest> sney: alright, thank you for all your help :)
483[05:31:06] <CyberManifest> sney: guess what... after removing the repos and removing 32 bit support, I ran the keyboard-configuration thing again, and now it's working and showing me all my options, lol
484[05:32:09] <sney> I'd take this opportunity to yell at plugwash for modifying the package without changing the version string, but he isn't in this channel.
486[05:33:10] <sney> there's nothing *really* special about the pi outside its weird bootloader so you should be set with pure debian stuff from now on. probably.
489[05:36:42] <CyberManifest> sney real quick with that wifi wiki link you sent me, it shows a command and I can't tell if it's literal instruction or an example it says: # ip link set wlp2s0 up #do I do the same or will mine look like: ip link set wlan0 up #?
490[05:37:33] <sney> again, you have to read the instructions part, not just the example. the 'ip a' and 'iw dev' are there so you can find out what your interface is called before setting it to up
501[05:41:19] <sney> (full disclosure: I wrote this section of the wiki, and maybe I should split the 'find' and 'bring up' steps, but it seems to have done the job just fine for some time now)
502[05:41:53] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
505[05:42:23] <CyberManifest> sney: also do I need to do these steps if I installed wicd ? and I know my SSID and now my interface name ?
506[05:42:44] <sney> no, if you're using wicd, you can just use wicd.
507[05:42:48] <nvz> yes well I didnt know of this iw command.. I used iwconfig, iwlist, but if the output is that clean you could make it a oneliner with a subshell even
508[05:43:01] <nvz> especially if this command has options to even just output only the device like some do..
509[05:43:05] <sney> since you were messing with rpi-config and wpasupplicant it sounded like you weren't using a gui.
510[05:43:07] <nvz> then you dont need to parse the output
512[05:44:02] <nvz> ip link set $(iw dev | cut -b ..) up
513[05:44:06] <nvz> or something like that
514[05:44:12] <nvz> sed, perl, awk, whatever
515[05:44:14] <CyberManifest> Ok, nevermind, I just read: "Your wireless interface should not be referenced within Debian's /etc/network/interfaces file. "
516[05:44:17] <sney> nvz: I try to avoid putting oneliner kludges in documentation, it can be... off-putting to people who don't know what they're looking at. better to give people the tools.
517[05:44:34] <nvz> sney: yeah I know the feeling.
518[05:45:11] <sney> I've ripped a few nasty one-liners out of dpkg factoids too. space conservation is one thing, but if you have a sed mess full of punctuation, nobody's learning anything anyway
521[05:45:40] <CyberManifest> sney: well I'm not presently using a GUI but I intend to, do I need to do something else to get wifi to work on the terminal when I'm not using the GUI ? like when I log out of my X Win session?
524[05:46:14] <sney> CyberManifest: wicd will only maintain your wifi connection when you're logged in. if you want it up all the time, use /etc/network/interfaces
525[05:46:34] <sney> (or just stay logged in)
526[05:46:49] <nvz> or just use network-manager
527[05:46:55] <sney> I wouldn't on arm
528[05:46:59] <CyberManifest> sney: if I use /etc/network/interfaces will it interfere with wicd ?
529[05:47:10] <nvz> oh.. forgot that bit of context
530[05:47:25] <nvz> wicd is just shit.. I dont want to say that, but its true
531[05:47:34] <sney> wicd can co-exist with /etc/network/interfaces to some degree. I think you'll still get the applet showing signal strength, but if you try to override settings with wicd things will get weird
532[05:47:37] <nvz> there is a wicd curses interface though
533[05:47:45] <nvz> one of the many things I like about it
534[05:47:53] <nvz> I just wish wicd was reliable codebase
535[05:47:58] <CyberManifest> sney: alright, thanks for the heads up
554[05:50:32] <nvz> its user friendly, the problem is just sadly that in my experience wicd itself is just buggy
555[05:50:35] <nvz> very buggy
556[05:51:13] <nvz> last I used it was on my orangepi lite.. and I compared n-m, wicd, and /e/n/i and I got fewest disconnects and connection failures with /e/n/i and the most with wicd
557[05:51:13] <sney> I've never run into an experience-breaking bug with it and the first time I used it was on squeeze. it's definitely a bit clunky though
558[05:51:22] <nvz> in addition to wicd just crashing of course
559[05:51:24] *** pingflood is now known as pingfloyd
560[05:51:41] <nvz> na the user interface is superb..
562[05:51:59] <nvz> its just the stability of the code and ability to deal with any quriks in connectivity that is shit
563[05:53:27] <nvz> where the stability of n-m has gotten great over the years
564[05:53:36] <nvz> I hated it early on.. and was using wicd back then
565[05:53:48] <CyberManifest> sney: not sure I understand this wicd-curses I selected my ESSID and hit enter and it keeps showing Wireless: Obtaining IP address... \(spinning) and at the top ti shows IPv6: ADDRCONF(NETDEV_UP): wlan0: link is not ready
571[05:54:43] <sney> CyberManifest: try getting the connection up with the /etc/network/interfaces example, it's easier to debug. if it works fine then we can proceed to troubleshooting wicd.
572[05:54:46] <nvz> yeah back when n-m was new to debian wicd was fairly new to it as well.. and I loved the look of wicd.. I wanted it to succeed..
573[05:54:52] <nvz> but then it got worse and n-m got better
574[05:55:01] <CyberManifest> sney: alright, thank you
575[05:55:15] <nvz> the kinda crap he just experienced is EXACTLY what I mean with wicd
576[05:55:25] <nvz> it /does/ that.. and crashes with python tracebacks and stuff
577[05:55:33] <nvz> but its so damn elegant looking..
578[05:55:40] <nvz> just doesnt work right all the time :P
587[05:57:11] <nvz> opposite for user interface :P
588[05:58:12] <nvz> n-m worked more than wicd did but on low ram like the opi lite had (512mb) it was painful
589[05:58:31] <sney> I rarely use anything but e/n/i or ifupdown unless the machine is going to be portable
590[05:58:41] <sney> I guess I have n-m on this desktop but that's just... complacency
591[05:58:55] <sney> (it was already there, you know)
592[05:59:32] <nvz> my only gripe with n-m these days is if you ever looked into hooking events.. certain up events are in one place, then down events dont follow same convention.. its just weird like two different people worked on sections of code and implemented things differently
593[05:59:43] <nvz> its not immediately apparent where to put scripts for certain events
594[06:00:19] <nvz> but other than that, the cli is easy to use, the applet is great.. and its documentation and overall stability and such have became noteworthy
602[06:03:03] <CyberManifest> sney: using the wifi chip on the raspberry pi 3 B+ and it's being shown as wlan0
603[06:03:12] <nvz> like I was fooling with aircrack to demonstrate something the other day.. and I didnt even bother doing a check kill, and n-m just played it cool as can be while I manipulated the crap out of a device it was controlling
604[06:03:20] <sney> I've seen pulse pull some dumb stunts, and the fact that "output to both of these devices, please" requires jumping through hoops is still dumb but it normally just works and I guess that's the standard
605[06:03:45] <sney> CyberManifest: if you do 'dmesg|grep -i firmware' what do you see? you don't need to pastebin it, just look for warnings/errors
606[06:03:48] <nvz> as soon as I did airmon-ng stop wlp3s0, n-m went right back to working that instant like it never happened
607[06:03:52] <nvz> not even a hiccup
608[06:04:40] <nvz> well wlp3s0mon :D
609[06:05:02] <nvz> when you do it like I said before, your wifi will still be absent from both the left and right click menus of the n-m applet :D
610[06:05:39] <nvz> point I was making though is n-m seems real solid anymore.. though its still not as solid as /e/n/i
611[06:05:48] <CyberManifest> sney: no warnings/errors... brcmfmac: ...Firmware BCM4345/6 wl0
612[06:06:02] <sney> b43, gross
613[06:06:11] * nvz shivers and utters "brrrroadcom"
614[06:06:12] <sney> oh yeah but it's a rpi I guess that makes sense
617[06:08:27] <CyberManifest> nvz: so in wicd-curses it shows my ESSID in green and it shows Connected to Wireless at 100% with an IP so I guess it's working ?
650[06:17:06] <sney> routing with split domain can be a little weird, if you know that the wifi network is set up properly then you can mostly not worry about this and just reboot
651[06:17:14] <nvz> CyberManifest: if you changed state of ethernet while on wifi, you may have screwed up your routing table
652[06:17:32] <CyberManifest> nvz so reboot and try again ?
653[06:17:50] <nvz> or just clear out the routing table or disconnect everything including wifi and reconned
674[06:29:12] <CyberManifest> wait, how do kernels work on raspi ? I noticed there is a firmware partition and root patition, is the kernel in /boot in the root partition ?
675[06:29:46] <CyberManifest> or is it somehow in the firmware partition ?
676[06:30:33] <CyberManifest> would it be more advisable to boot raspi from sd and somehow do a proper debian install onto the mSATA?
677[06:30:55] <sney> on any system, there's a kernel (and initrd usually) in /boot, and then depending on the hardware there are bootloaders and/or firmware that point the cpu to the right spot on disk to boot the system
678[06:31:34] <sney> the rpi's bootloader has always been a weird animal. you're probably going to have a bit of a tangle no matter how you do it. if it's working now, keep it
679[06:32:03] <CyberManifest> sney: well I wanted a separate /boot / <swap> /home
681[06:32:24] <sney> if all of that is on 1 volume: why?
682[06:33:23] <CyberManifest> sney, to retain /home during updates, /boot for kernel testing and Swap cause I do some fancy VSWAP on the PI to help performance
683[06:33:28] *** Quits: torbo (~user@replaced-ip) (Remote host closed the connection)
690[06:35:05] <sney> there are a lot of logs from this channel of people with separate /boot that was too small and ran out of room generating the initrd, for example
699[06:36:54] <sney> if it's the same fstype as / then why not just leave it? we're not in the 90s anymore, your hardware doesn't care how big the first partition is. the only reason separate /boot makes sense anymore is for edge cases
700[06:36:55] <CyberManifest> sney: these used to be typical user conventions
701[06:37:29] <sney> emphasis on the "used to"
702[06:37:43] <sney> they were already kind of obsolete when I really started being a linux nerd, in
703[06:37:44] <sney> hmm
704[06:37:46] <sney> 2003
705[06:37:55] <CyberManifest> sney: they're still suggested in redhat docs and arch docs and such
706[06:38:26] <sney> docs maintain old stuff to establish better understanding, not so new users can cargo cult old ideas for no rational reason
707[06:38:49] <CyberManifest> sney: I like how you put that :)
709[06:39:47] <sney> I should find a better term than "cargo cult" but thanks
710[06:40:37] <CyberManifest> sney it wouldn't be hard for me to seperate the partitions, aside from /boot... but this still side steps my question of wether or not to put on SD and do a proper install to mSATA (if can) or just keep things as they are on the mSATA
711[06:41:06] <sney> if it's working and booting predictably now, what would you gain from starting over?
715[06:41:58] <sney> once it's booted, is it comparable?
716[06:42:32] <CyberManifest> sney: and if I did a proper install I may have more control over the partition scheme and easier method of obtaining the packages I want rather than one by one
717[06:43:51] <CyberManifest> sney: once it's booted Debian (Raspi) is faster and more attentive by far than Raspbian but I'm still occasionally having issues with the keyboard input being too sensative.
718[06:44:13] *** Quits: swift110-phone__ (uid50036@replaced-ip) (Quit: Connection closed for inactivity)
719[06:44:56] <CyberManifest> it repeats characters at random
720[06:45:16] <sney> the obsession with boot times is another 90s/00s relic. we don't turn off our computers every day anymore, so who cares? if it's 30s or a minute+ it's downtime that we might worry about twice a month at the *most*
727[06:47:27] <CyberManifest> sney: theoretically if I were to place raspi on an SD card, how would I go about an installer to another location ? I mean is there an installer package I can download?
728[06:47:33] <sney> it's summer in north america right now, there's some heat, but every qualified opinion I've ever heard about rpi overheating has been "don't worry about it, it's within spec" - a lot of embedded stuff is happy to run hot because that's how it was designed.
729[06:48:19] <sney> the arm64 installer can be found at replaced-url
730[06:48:22] <CyberManifest> sney: but I enable overvolting and my VRAM is taxing on the processor
731[06:48:46] <sney> note that qualified opinion doesn't include randos on reddit
732[06:49:00] <sney> you shouldn't trust me either, tbh. if it's from rpi.org it's probably legit
733[06:50:10] <CyberManifest> sney: so I burn that debian-10.4.0-arm64-netinst.iso to SD card and boot and follow allong like typical ?
734[06:50:45] <sney> that's how it usually works. I haven't tried it on a pi3 so I can't tell you what to expect, but might as well give it a shot
735[06:50:57] <CyberManifest> doesn't include randos on reddit what's that ?
736[06:51:56] <CyberManifest> sney: well I tried the aarch64 mini.iso and it didn't work and neither did armhf mini.iso
737[06:52:23] <CyberManifest> so I'll give this a try and see, if not I know how to get back to where I am, thanks to all your help :)
738[06:52:31] <sney> the key word is "qualified" - a lot of people on the internet, forums, reddit, have opinions and personal experiences that may or may not be useful. qualified opinions come from actual experts. in the case of the pi, that's raspbian developers and people from the rpi foundation. everyone else is guessing
739[06:52:53] <sney> if you're running arm64 on your pi3 now, then arm64 is the correct arch, not any of the other ones. (armel might work but you would hate it)
746[06:55:47] <sney> CyberManifest: aarch64 is a different name for arm64. if you haven't tried the debian arm64 iso, then try it. debian's armhf doesn't support the original rpi cpu, which is the whole reason raspbian exists. maybe it would work on the pi3?
749[06:57:54] <CyberManifest> sney: I'm on the pi3, I'll give that ISO from the link you provided a try, thank you again for all the help and support, you're a standup person :) Kudos
768[07:08:21] <Aurora_v_kosmose> Half of my VMs on one host got grub_calloc error.
769[07:10:19] <diogenes_> Aurora_v_kosmose, it showed an error executing some pearl script /tpm/xyz but it's because my /tmp is mounted with noexec, other than that all seem to work.
770[07:10:43] <Aurora_v_kosmose> I got "error: symbol `grub_calloc' not found."
771[07:13:02] <CyberManifest> sney: the XFCE image isn't booting either :( I don't understand... If I was running ARM64 before why wouldn't these images boot too ?
772[07:13:35] <CyberManifest> sney: I guess it's cause of the non-free boot blobs or something ?
778[07:18:57] <Aurora_v_kosmose> It's almost as if grub stopped to ask for interactive dialogue, failed because I was updating via a script, but exited with 0
824[07:59:03] <Ashleee> cyberbanjo, with a bit of luck you can grab kernel and dtb and then debian aarch64 installer netinst tarball and put it on sd card and make it boot and then "reinstall" debian onto the sd card you just booted from (or anywhere really), but I don't know how the boot process on RPi works because it boots via the GPU that loads the IPL etc
999[10:13:51] <themill> timothywcrane: there's a difference between a source package and a binary package. A source package can make a lot of different binary packages.
1032[10:31:42] *** jjasddsfs is now known as b30wulf
1033[10:32:43] <mav> hello :) anyone know if font rendering is broken in bullseye/sid? only getting boxes with hex values in them for non-ttf fonts (pcf/bcf etc), ttf/otf seems to work fine
1093[11:16:35] <Psy-Q> i've inherited an snmp configuration on jessie (!) that since this night's libnsnmp upgrade seems to reject the extensions configured in snmpd.conf: /etc/snmp/snmpd.conf: line 111: Warning: Unknown token: extend.
1094[11:16:40] <tarzeau> Haohmaru: here's my favourite fonts for (setfont in console/fb): replaced-url
1128[11:48:09] <dpkg> The reasons we hate polls are because: 1) they are a massive waste of IRC and bandwidth and screen real estate when asked in a 700 person channel. Did you want 700 answers? 2) They're often a prelude to isolating one helper and barraging him or her in /msg. BAD! 3) We don't care about Ubuntu, this is #debian.
1129[11:48:14] *** Quits: mthe878 (~mthe@replaced-ip) (Quit: Lost terminal)
1131[11:48:37] <tarzeau> from the binary, unpacked, uncompressed :)
1132[11:48:41] <ratrace> sfx2496: you can create a systemd unit in any way you want. systemctl edit is a convenience method when using an editor. don't forget to run systemctl daemon-reload after you change a unit file
1133[11:49:04] <tarzeau> Haohmaru: the ocr one is manually drawn ocr-b font for 8x16 vga done by me
1136[11:50:05] <Haohmaru> i don't like this one.. i do have a 8x16 font which is hardcoded into a program: replaced-url
1137[11:50:14] <Haohmaru> but it's too big for code
1138[11:51:09] <Haohmaru> i also like fixedsys
1139[11:51:22] <Haohmaru> it's big, but not too big
1140[11:51:23] <WormholeTraveler> BTW I have a switch that does VLAN. but my router does not, am I okay?
1141[11:52:02] <sfx2496> ratrace: would /etc/systemd/system/apt-daily.timer.d/override.conf be the file to edit, or will that be overwritten on reload?
1142[11:53:22] <ratrace> sfx2496: with that form (aka "drop-in units") you can have units that only contain entries that override default. a /etc/systemd/system/apt-daily.timer unit would completely replace the default unit
1143[11:54:00] <ratrace> sfx2496: also, #systemd would be more helpful with systemd specific help, this is not debian specific
1144[11:54:16] <tarzeau> Haohmaru: should be easy to patch?
1145[11:54:26] <sfx2496> k, ty
1146[11:54:31] <Haohmaru> patch what?
1147[11:54:39] <tarzeau> Haohmaru: i'm waiting for flashrom to support nvidia bios (the fonts are embedded and bios roms) i'd like to patch my video cards
1148[11:54:57] <tarzeau> Haohmaru: the program that hardcodes that bitmap font
1163[12:05:01] *** Quits: in1t3r (~LordShiva@replaced-ip) (Remote host closed the connection)
1164[12:05:51] <sarthe> ugh. i am drunk and have forgotten... how to check the installation status of a name package? I thought it was `apt-get status <package>`
1167[12:06:09] <Haohmaru> i come from crapdows, and as crapdows becomes more and more crap, i figured i need to get more serious into using linux.. i asked for a recommendation on what to use, given that i am interested in video/audio/music/photo/3D/whatnot, someone recommended ubuntustudio .. some time later i figured that this is just a crapbuntu pre-configured to install a certain bunch of programs, plus customized
1168[12:06:09] <Haohmaru> fancy UI theme and wallpapers.. stuff that you can do by installing a normal crapbuntu and then installing the programs you need
1169[12:06:33] <WormholeTraveler> you didnt like ubuntu? why
1170[12:06:35] <Haohmaru> WormholeTraveler then i heard bad things about crapbuntu, so i went to debian
1171[12:07:01] <ratrace> Haohmaru: very much offtopic here tho
1172[12:07:08] <Haohmaru> i know :/
1173[12:07:10] <WormholeTraveler> Hows that offtopic :|
1193[12:12:28] <ratrace> WormholeTraveler: wayland is a protocol, so the question then becomes about specific compositors supporting the protocol. gnome's mutter is one. kwin another (afaik). sway is i3-wm implementation for wayland. and then some
1194[12:13:10] <WormholeTraveler> does gnome + wayland use less resources?
1195[12:13:56] <Haohmaru> i kinda doubt.. for less resources look into LXDE or xfce4
1196[12:14:14] <ratrace> if anything it uses more due to CSD and moving the renderer into client's domain. but that's just pure speculation and conjecture, I never ran benchmarks :)
1197[12:14:33] <Haohmaru> those two don't use wayland tho
1198[12:16:17] <WormholeTraveler> thanks. Ill just stick with cinnamon.
1199[12:16:23] <ratrace> I think phoronix ran some benchmarks, but take those numbers with a trainful of salt; as one should always with moronix^W phoronix
1234[13:11:14] <ElectroXexual> How do I hide (or enable auto hide) the kde panel (kicker) from linux shell? I want to do it so I can Hide/Show it through a global shortcut.
1235[13:12:09] *** Quits: Conradish006 (~conradish@replaced-ip) (Remote host closed the connection)
1248[13:36:36] <password2> how do i mount an existing raid array?
1249[13:38:44] <ratrace> password2: assuming linux software raid (md), and mdadm package installed, it should be automatically detected and assembled, so you mount the corresponding /dev/md* device.
1277[13:49:19] <ratrace> I don't know debian's live image too well, but if it's like some other distros' live images, maybe you can apt install mdadm ?
1278[13:49:22] <password2> i run 'ls /sbin/ | grep md' to check
1279[13:49:47] <password2> where would i get package?
1280[13:50:06] <password2> i tried apt0get isntall mdadm
1287[13:54:50] <dpkg> A suitable /etc/apt/sources.list for "Buster" has the lines: "deb replaced-url
1288[13:55:26] <ratrace> password2: see if these entries at least are present in /etc/apt/sources.list ; my take is it _should_ be able to use apt, even if it maybe isn't set up to do so by default
1329[14:09:20] <Enissay> Is there a simple trick to replace each line from /etc/hosts to transform for example: 127.0.0.1 localhost loopback to: 127.0.0.x prefix_localhost prefix_loopback x is a digit & prefix is the same for all hosts names (can be 1 to 4 hostnames per line)
1330[14:09:20] <ratrace> check if it assembled, cat /proc/mdstat . see if lists something like md127: active raid... sda...
1362[14:19:13] <ratrace> according to this, mdadm _should_ work tho: replaced-url
1363[14:19:17] <ratrace> !pastebin
1364[14:19:17] <dpkg> Do not paste more than 2 lines to this channel. Instead, use for text: replaced-url
1365[14:20:16] <jrtc27> Enissay: gawk '{ n = split($0, a, " ", b); if (a[1] == "127.0.0.1") { a[1] = "127.0.0.x"; for (i = 2; i <= n; i++) { a[i] = "prefix_" a[i] } } line = b[0]; for (i = 1; i <= n; i++) { line = line a[i] b[i]; } print line}'
1367[14:20:29] <ratrace> password2: so you could try mdadm --assemble --scan --force and see if that creates any /dev/md* devices, visible under /proc/mdstat also
1368[14:20:45] <jrtc27> Enissay: that'll do exactly what you asked for, though we're all a bit confused what you _actually_ are trying to do
1376[14:23:49] <password2> unkown filesystem type LVM2_member
1377[14:23:53] <ratrace> password2: but according to teh synology docs I posted above, there's maybe LVM atop of that so you need to activate it with vgchange -ay , then you mount /dev/vgX/lvY /mnt where vgX is the actual volume group and lvY is the actual LV
1378[14:24:13] <password2> its unde /dev/md/
1379[14:24:14] <ratrace> password2: you might need to apt install lvm2 first
1453[15:23:45] <jstolarek> so, yesterday's updates to grub broke my dual-boot with Windows :-/ Linux boots fine, but selecting Windows from grub results in blank screen and nothing happening. I can boot Windows though if I select windows bootloader from bios
1479[15:51:13] <g0zzy> I've messed up the perms on /etc/sudoers and sudo refuses to perform. Unfortunately that was before i changed the root password so i can't su. Any ideas without unmounting the root partition and fiddling with the files?
1481[15:52:09] <dpkg> For GRUB: 1) press 'e' to edit the kernel setting in the grub command line (add 'init=/bin/sh' to the end of it) 2) 'fsck' your root file system, 3) 'mount -o remount,rw /', 4) 'passwd root' 5) 'mount -o remount,ro /' 6) 'reboot -d -f' (exec /sbin/init should work); For LILO: 1) 'Linux init=/bin/sh' at the LILO boot prompt (hold Shift while booting), steps 2-6 are the same; For yaboot: 1) 'Linux init=/bin/sh' at yaboot prompt.
1505[16:16:39] <kittonian> hi all. is anyone here familiar with running davical?
1506[16:16:47] <dvs> !anyone
1507[16:16:47] <dpkg> Please do not ask if anyone can help you, knows 'something' or uses 'some_program'. Instead, ask your real question. (If the real question _was_ "does anyone use 'some_program'?" ask me about <popcon> instead.) See <ask> <ask to ask> <polls> <search> <sicco> <smart questions>.
1524[16:25:46] <kittonian> i am looking at alternative to radicale and one that looks promising is davical. All clients are Mac (desktops and mobile devices) and currently with Radicale we are seeing very slow update times between devices. Is this a better solution or is there something else I should be looking at?
1575[16:59:34] <neilthereildeil> i was looking at the new boot hole vulnerability discovered in grub (CVE-2020-10713). the original blog post said i dont need grub2 in order to be vulnerable. how is that possible, if its a memory corruption in grub2?
1576[16:59:54] <greycat> CrystalMath: too vague. *how* early? what has to run before it?
1577[17:00:28] <CrystalMath> the script creates a static /dev file that is too unpredictable to have statically on the disk
1578[17:00:34] <CrystalMath> the dev file in question is /dev/rtcX
1579[17:00:46] <CrystalMath> it requires only sysfs
1580[17:00:54] <CrystalMath> it is needed to have RTC access
1582[17:01:07] <CrystalMath> by rtcX i mean it will create rtc0, rtc1, etc...
1583[17:01:15] <freenoodle> can someone give me a headsup with apt pinning? I am using Mobian, which is debian for the pinephone, and while some packages exist in both repos, the one from mobian should be preferred, which is what /etc/apt/preferences.d/mobian (or whatever the exact filename is) should ensure. apt policy libhandy-1-0 says:
1584[17:01:17] <freenoodle> libhandy-1-0:
1585[17:01:17] <freenoodle> Installed: 0.84.0-1
1586[17:01:17] <freenoodle> Candidate: 0.84.0-1
1587[17:01:17] <freenoodle> Version table:
1588[17:01:18] *** freenoodle was kicked by debhelper (flood)
1592[17:01:58] <greycat> CrystalMath: so, you either insert it (the S* symlink to it) into the rc2.d or rcS.d directory, whichever is appropriate, or you wimp out and use rc.local
1593[17:02:06] <greycat> in either case, you stop using sysvinit ASAP
1594[17:02:13] <greycat> it is a horror show
1595[17:02:26] <CrystalMath> i know, but debian doesn't package BSD init
1596[17:02:33] <CrystalMath> i wish debian had slackware's init system
1597[17:02:45] <CrystalMath> i want things to be even more retro, not more modern :
1598[17:02:49] <CrystalMath> :P
1599[17:02:58] <Nothing4You> is there an easy way to get a summary of how much storage all the volumes in an lvm thin pool can take up?
1600[17:03:20] <CrystalMath> greycat: i got rid of udev, as a part of the retroization process
1601[17:03:24] <CrystalMath> but now, no rtc :(
1602[17:03:52] <CrystalMath> i'll email the kernel folks about bringing back device 10,135
1603[17:04:02] <neilthereildeil> heres the blog post: replaced-url
1604[17:04:05] <CrystalMath> this was present before kernel 2.6
1605[17:04:12] <neilthereildeil> they say theres a vulnerablility in grub that can bypass secure boot
1606[17:04:30] <CrystalMath> i don't have EFI, so i don't have to worry :P
1607[17:04:31] <neilthereildeil> but this is the line i didnt understand "The vulnerability affects systems using Secure Boot, even if they are not using GRUB2."
1608[17:04:58] <neilthereildeil> how can a machine not using grub2 be affected, if the memory corruption is inside grub2?
1609[17:05:12] <neilthereildeil> if grub2 isnt installed, what parser will get corrupted?
1611[17:07:16] <oxek> neilthereildeil: it's a problem because even if you're not using grub2 right now, an attacker with physical access to the machine (yeah I know) can replace your current bootloader with the previous vulnerable version of grub2 in order to bypass secure boot
1612[17:07:26] <oxek> hence all the necessary key revocations happening
1613[17:07:33] <neilthereildeil> ahh lol
1614[17:07:42] <JyZyXEL> neilthereildeil: the dbx revocation list is updated for all systems uefi
1616[17:08:44] <neilthereildeil> JyZyXEL: yea, but whether or not the dbx revocation list is updated in my system, if i am not running grub2, i dont have to worry (as long as someone doesnt install vulnerable grub2)
1617[17:09:09] <JyZyXEL> someone else could run it for you
1618[17:09:25] <oxek> you don't have to worry about this vulnerability if no one malicious has physical access to your machine or root privileges
1619[17:09:56] <CrystalMath> BIOS is safe from the vulnerability :D
1620[17:10:03] <JyZyXEL> protected by not affording a house maid
1621[17:10:13] <CrystalMath> on my machine, you can boot whatever you want off a floppy
1623[17:11:12] <greycat> after spending about 5 seconds trying to understand what that page is talking about, my understanding is that this "hole" only matters if you were actually TRYING to use Secure Boot as some sort of security mechanism against... something... instead of the Microsoft inflicted impediment to installing Linux that most of us perceive it as
1624[17:11:36] <CrystalMath> greycat: i suppose most people who see it that way disable it
1625[17:12:27] <oxek> at least they fixed a few bugs in grub2, which is nice... but it seems to be breaking systems, which is not so nice
1626[17:12:44] <CrystalMath> i upgraded grub recently
1627[17:12:48] <CrystalMath> hopefully it broke nothing
1628[17:12:59] <CrystalMath> meh, i'll go back to lilo
1629[17:12:59] <oxek> it seems to only break non-EFI systems
1638[17:14:02] <CrystalMath> if i see GRUB fail to boot, that's it, i'm switching to LILO forever
1639[17:14:13] <CrystalMath> maybe i should install grub-legacy?
1640[17:14:37] <oxek> have a recovery plan ready then. Seems to involve downgrading to a vulnerable version as the only workaround now.
1641[17:14:43] <CrystalMath> i don't need EFI, it literally does not exist on my machines (either because they're old enough, or becuase i use legacy boot and don't acknowledge its existence)
1663[17:19:38] <greycat> What do people actually try to *use* Secure Boot for, that they would care if it had a "hole"?
1664[17:19:57] *** Quits: jstolarek (~jstolarek@replaced-ip) (Remote host closed the connection)
1665[17:20:24] <oxek> greycat: I guess it's a hobby, might be fun to play with. I tried importing my own signing keys, and signing bootloader & kernel.
1666[17:20:28] *** Quits: ich (~ich@replaced-ip) (Quit: Ex-Chat)
1667[17:20:31] <klys> greycat, probably for embedded efi devices
1668[17:20:33] <oxek> was a fun exercise, but I gave it up
1678[17:24:37] <klys> a while back (long time), grub2 had a bug where it couldn't be installed to a partition, so folks started using it in the mbr. install-mbr is in teh mbr package. sort of like fdisk /mbr
1679[17:24:50] <CrystalMath> klys: i installed grub 1
1699[17:32:45] <CrystalMath> it's using root=UUID=...
1700[17:32:50] <CrystalMath> i'm a little worried about that
1701[17:32:53] <CrystalMath> i have no udev
1702[17:32:53] <greycat> "Secondly, in grub-legacy, the partition number starts at 0, not 1 like in Linux, so subtract 1 from the linux partition number. And in grub2, it starts at 1."
1703[17:32:56] <greycat> Yikes.
1704[17:33:00] <CrystalMath> hopefully that's not a problem
1741[17:38:53] <shibboleth> CrystalMath, windows up until win8 had terrible track record when it came to it booting under diff configurations
1742[17:39:28] <shibboleth> this wasn't due to the device "name" but due to the mass storage driver windows was set up to use
1743[17:40:04] <shibboleth> for linux/grub, basically, the same block device might be named sda or hda based on whether the bios is set to ide/legacy or ahci/sata
1744[17:40:23] <CrystalMath> okay, i won't use IDE legacy
1745[17:40:23] <shibboleth> and the defaults are there to prevent that being a dealbreaker in terms of booting
1877[18:21:13] <CrystalMath> the bootsector seems fine
1878[18:21:19] <CrystalMath> i looked at it manually
1879[18:21:32] <klys> -snapshot : Write to temporary files instead of disk image files. In this case, the raw disk image you use is not written back. You can however force the write back by pressing C-a s (see disk_005fimages).
1889[18:44:17] <CrystalMath> klys: i'm gonna write and submit a kernel patch which restores 10,135 if CONFIG_RTC_LEGACY_DEVICE is set (new config option)
1890[18:44:50] <klys> crystalmath, and this is for compatibility with v2.6 ?
1891[18:44:57] <CrystalMath> with pre-2.6
1892[18:45:06] <CrystalMath> 10,135 was removed in 2.6
1893[18:47:12] *** Quits: timur_davletshin (~timur_dav@replaced-ip) (Remote host closed the connection)
1894[18:47:48] <klys> well I've used v1, v2.0.3x, v2.2, and v2.4 before. I remember framebuffers came out with v2.4, the early debians I used (such as slink) were 2.2 and 2.0, and trying out the original kernel with ldlinux.exe
1900[18:53:28] <Regor> why midnight commandar doesnt open more than one file at a time in any tiling window wm... ? any option to fixit ? i checked mc in awesome,bspwm,,, but always the same story. i tried it also in floating mode ..but the same result ..
1962[19:36:09] <dpkg> Ubuntu is based on Debian, but it is not Debian. Only Debian is supported on #debian. Use #ubuntu on chat.freenode.net instead. Even if the channel happens to be less helpful, support for distributions other than Debian is offtopic on #debian. See also <based on debian> and <ubuntuirc>.
1974[19:40:55] <cybrNaut> i don't care which channel has someone who can answer the question
1975[19:41:05] <greycat> !chroot
1976[19:41:06] <dpkg> To chroot into your Debian system boot to your Debian install disk/live CD, switch to the other console (Alt-F2). Mount your root filesystem with "mount -t ext2 /dev/whatever /target" and make /dev, /proc and /sys usable with "mount --rbind --make-rslave /dev /target/dev ; mount -t proc none /target/proc ; mount -t sysfs none /target/sys". You can then chroot into the system with "chroot /target".
1977[19:42:33] <oxek> well that's pretty different from what I have, need to study up on it
2020[20:28:13] <kittonian> hi all. yesterday I was trying out software that required systemd, and used apt to install. however, as this is old system i got a failure message during install regarding overwrite a man page (known error). As I have no intention of using systemd I would very much like to revert back to where it was previously but apt is throwing errors because it hasn't completed the install.
2021[20:28:17] <kittonian> how do I clear it out
2022[20:29:14] <kittonian> the exact error is trying to overwrite '/usr/share/man/man8/halt.8.gz', which is also in package sysvinit 2.88dsf-41+deb7u1
2023[20:29:38] <kittonian> which you can see from this stackexchange post replaced-url
2024[20:29:47] <kittonian> and I don't care about systemd
2025[20:30:05] <kittonian> I just want to get the server back to where it was before it tried to install systemd and its dependencies
2026[20:30:11] <CrystalMath> maybe download the source instead and modify it to not require systemd
2027[20:30:14] <kittonian> or at the very least, resolve this apt error
2029[20:30:31] <kittonian> CrystalMath, no, i don't want the software or systemd. just trying to resolve the apt error
2030[20:30:32] *** Quits: Error451 (~R@replaced-ip) (Disconnected by services)
2031[20:30:47] *** Quits: chris__ (~chris@replaced-ip) (Remote host closed the connection)
2032[20:30:56] <kittonian> can't use apt because it thinks it's in the middle of an installation and I should use --fix-broken install, which of course fails
2071[20:37:02] <CrystalMath> kittonian: but don't worry, i personally *never* restored a backup
2072[20:37:07] <kittonian> me either
2073[20:37:08] <CrystalMath> that is because... there are no backups
2074[20:37:13] <nvz> you dont go /testing/ stuff on a production system you are not fully backed up and prepared to restore a backup of
2075[20:37:17] <nvz> you use a f'n VM
2076[20:37:29] <nvz> /thats/ how you administer a server
2077[20:37:37] <nkuttler> doing it on prod is more fun!
2078[20:37:42] *** Quits: incentive (~rbatty@replaced-ip) (Remote host closed the connection)
2079[20:38:24] <nvz> not really.. I'm still annoyed with having sid on my T440 cause my dumb ass wanted to see the latest MATE and XFCE features I read about upstream
2080[20:38:47] <nvz> now I dont wanna go back to buster cause my SSD is old and I'm afraid to put it through the paces until I can get another one
2081[20:39:15] <greycat> If my Debian 6 system can do "mount -t nfs -o nfsvers=3 a:/foo /bar" just fine, but "mount -t nfs -o nfsvers=3 b:/foo /bar" (different server) always comes back with "mount.nfs: mount system call failed", is this *DEFINIETLY* a misconfiguration on the second server? Or do I need to flip some arcane lever on the client?
2082[20:40:23] <nvz> greycat: what is this server thats failing running? if the one that works is on Debian 6
2083[20:40:53] <greycat> The server is not Debian, and is not mine. I don't know what it is. It's controlled by a different department.
2084[20:41:23] <nvz> and I find it odd you'd even ask such a question.. cause you're the kinda guy that usually goes to lengths to look into things and doesnt sound like you ever did any debug, logging, or anything yet
2089[20:42:27] <nvz> it does sound like it could be anything from incompatible nfs tools, to default mount options.. you name it.. thats not enough information to speculate on really.. other than yeah.. sounds like the 2nd box is the issue
2090[20:43:07] *** Quits: bewees (~quassel@replaced-ip) (Remote host closed the connection)
2091[20:43:08] <greycat> I have another 3 open for "linux nfs4" because apparently I can mount it with NFS version 4, but not with version 3, and I do not know nfs 4, and it looks like it will not be a viable option because it requires coordination with the server end, which I DO NOT CONTRGOL.
2092[20:44:05] <nvz> greycat: presumably if you are doing this you have administrative access to the 2nd machine that doesnt work, no?
2095[20:45:04] <greycat> What part of I DO NOT CONTROL (other than the typo) was unclear...
2096[20:45:27] <greycat> I have no access to the NFS server. It's a share that I'm supposed to be able to mount. I have to exchange emails with the person who admins it.
2097[20:45:56] <greycat> If I'm desperate enough to ask this here, you know it's pretty bad.
2111[20:51:55] <cybrNaut> BCMM: thanks for the tip, but no worries. I have 3 debian-based systems (stretch, mint, ubuntu) and i asked the same question in #ubuntu and #linuxmint. In my client, i can crosspost just by hitting up arrow, and i didn't bother to edit the hostname. But anyway, i got my answer (and that answer applies to all 3 platforms)
2113[20:52:51] <xshockfish> in debian 10 I upgraded today, grub asked me where to install - replied /dev/sdc1 - after reboot system doesn't boot with: error: symbol grub_calloc not found - entering rescue mode - grub rescue
2114[20:53:32] <xshockfish> that /dev/sdc1 was where normally /boot was mounted (and was mounted at time of this aptitude upgrade). how to enter the system and install grub into maybe /dev/sdc this time? or ?
2115[20:53:46] <sney> /dev/sdc would be more likely, yes
2116[20:53:48] <sney> !fixmbr
2117[20:53:49] <dpkg> To reinstall <GRUB> boot to your Debian install disk/live CD, switch to the other console (Alt-F2), mount your root filesystem (mount -t ext4 /dev/whatever /target ; mount --bind /dev /target/dev ; mount -t proc none /target/proc ; mount -t sysfs none /target/sys), chroot into it (chroot /target), run "mount /boot/efi" on EFI and "update-grub && grub-install /dev/whatever". See also <rescue mode>, <dual boot guide>, <supergrub>.
2118[20:54:06] <xshockfish> how to fix that from within grub rescue?
2123[20:55:16] <nvz> greycat: again I/O error is vague but more points toward what may gather information.. and if its an IO error the kernel should certainly know more
2124[20:56:03] <greycat> There does not appear to be anything current in dmesg on the client. Yes, I checked. No, I do not know whether the timestamp in dmesg is current. I am assuming it's not. It's not talking about the correct network interface anyway.
2125[20:56:05] <nvz> greycat: my biggest issues with nfs were always permission type things.. and since you seem to know nothing about this sytem for all you know you could be locked out somehow
2128[20:57:08] <WormholeTraveler> just installed debian, is there a guide I should be looking at for post install?
2129[20:57:16] <WormholeTraveler> 10.4
2130[20:58:51] <sney> WormholeTraveler: debian tries to have sane defaults, so no, there's nothing you need to "fix". if you're looking for something specific it might be in the install guide or release notes
2131[20:59:03] <n_1-c_k> WormholeTraveler, there are probably dozens but here's one, replaced-url
2132[20:59:29] <sney> the handbook is useful too
2133[21:00:00] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
2134[21:00:05] <WormholeTraveler> im mainly looking for xorg tweaks
2138[21:01:15] <sney> also, post-install blog articles about changing stuff in apt or installing random selections of packages for *reasons* are by amateurs and can often make things worse. skip 'em.
2140[21:02:42] <greycat> if you're having actual issues with X, it's usually missing firmware, or you need to switch to a different driver... can you describe the issue you're seeing?
2143[21:03:12] <nvz> greycat: sources as obscure as mcaffee commenting on windows nfs servers (didnt know that was a thing) says that I/O errors /can/ result from incompatible nfs versions replaced-url
2144[21:03:31] <jhutchins> greycat: You're specifying the nfs version, so it's not trying an incompatible version. It sounds like there might be a problem with portmap on server B.
2145[21:04:53] <nvz> or your UID/GID doesnt line up.. more a million other things
2146[21:04:55] <jhutchins> There was a problem with the NFS client on RHEL6(?) back around 2013 that caused a similar error.
2147[21:05:05] <nvz> cause you dont know what the client or the server has doing on
2148[21:05:52] <jhutchins> greycat: Very difficult to diagnose when you can't see the server logs.
2149[21:05:52] <greycat> I mounted a share from a Debian 10 box just to prove that I could. To rule out something like "the client is missing kernel support for NFS v3 entirely".
2154[21:08:04] <nvz> or having a firewall issue is also ruled out
2155[21:08:19] <jhutchins> greycat: Do any utils like showmount work?
2156[21:08:19] <greycat> I know there isn't a firewall on the *client*.
2157[21:08:23] <greycat> jhutchins: yes.
2158[21:08:46] <greycat> The server person gave me the wrong directory name at first, and I had to use showmount -e on the server to look for the right name.
2159[21:09:02] <jhutchins> greycat: If server B didn't have permissions you should get a more meaningful error.
2160[21:09:10] <greycat> So, you understand that I do not have a lot of faith in the admins of this server.
2163[21:10:56] <greycat> The last thing they told me was "can you try mounting /BK_MEG_Lab_Study/MEG_Lab_Study" and that one gave an access denied error, as one would expect.
2167[21:14:23] <nvz> so that kinda rules out UID/GID issues as well.. which I was still unclear about given that you say you can mount this on YOUR machine
2175[21:23:10] <greycat> Mounting it with nfs4 works, but then the directory is owned by "nobody" because crazy bullshit kerberos wtf nfs4 idmap I have no idea.
2180[21:29:28] <nvz> greycat: there is also a lot of bug reports out there or rhel, centos, etc about "port exhaustion" which results in an I/O Error on mount.. options rsvport and norsvport may or may not influence that from what I've seen
2197[21:37:43] <greycat> In a new twist, the server admin opened up the share to all addresses, so I tried mounting with NFS v3 from my Debian 10 workstation. That worked.
2204[21:43:24] <greycat> the box is running 2.6.26 because hardware compatibility reasons with a device that I suspect is no longer connected to it, and therefore that may no longer be a roadblock
2205[21:43:43] <greycat> there is a 2.6.32 kernel in /boot, so I can try that one first, or I can try upgrading it to wheezy....
2206[21:44:35] *** ddsys_ is now known as ddsys
2207[21:44:44] <sney> *indiana jones voice* IT BELONGS IN A MUSEUM
2208[21:45:28] <greycat> and *any* kernel work means I have to be there, which I can do tomorrow, but god damn it covid-19 STOP MAKING LIFE SUCK
2209[21:47:02] <greycat> Is there any "Lazy Man's nfs4 Guide" or something that'll just tell me how to make nfs4 mounts work like nfs3 mounts, with UIDs preserved instead of being force-mapped to "nobody"?
2222[21:51:40] <jhutchins> greycat: I dunno, v4 is pretty rare in actual implementation. It's something that only large institutions seem to use because ordinary admins don't understand it.
2226[21:53:05] <nvz> yes well uid issues are solvable but they require some sort of server side cooperation :P
2227[21:53:33] <nvz> with v3 anon- options no_root_squash or with v4, requires coordinating idmapd settings
2228[21:54:23] <nvz> when I used nfs last, at home, I just used no_root_squash to avoid headaches
2229[21:54:25] <greycat> "the NFS v4 server must have knowledge of the same user and group accounts as the NFS v4 client is using" ... what the FUCK!
2230[21:54:38] <greycat> See, this is why nobody uses it!
2231[21:54:44] <nvz> yes well v3 has to have the same UID
2232[21:54:50] <greycat> INCORRECT
2233[21:54:58] <greycat> In v3 the server doesn't have to have any user accounts AT ALL.
2234[21:55:26] <greycat> The server does not give a flip what the UID of any file is, unless it's 0 and root_squash is in use, in which case, someone's "nobody" gets used.
2235[21:55:41] <greycat> I'm not even sure whose.
2236[21:56:15] <greycat> If two clients mount the same share in NFS v3 and they have disagreement about the username/UID mapping, then the file will show as the wrong user on the client side, but the server doesn't care.
2237[21:56:30] <nvz> I'm talking about for auth purposes to connect.. it doesnt use usernames it uses UIDs
2238[21:56:39] <greycat> There is no auth.
2239[21:56:45] <greycat> Just IP addresses.
2240[21:57:17] *** Quits: nehemiah (~nehemiah@replaced-ip) (Remote host closed the connection)
2242[21:57:58] <greycat> Also, I found the word "NetApp" in the emails, so I believe they're using a NetApp. I don't know how those things handle unix account names for NFS v4.
2243[21:58:55] <sney> like a legacy netapp? those were definitely some flavor of non-linux unix.
2251[22:03:25] <sney> nfs is great if the link between machines is *very* reliable. um, but that's less common in the current era
2252[22:03:30] <nvz> its sole utility seems to be for native permanent filesystem sharing between local *nix hosts
2253[22:04:04] <nvz> once you start trying to use it like other filesharing for over the internet, untrusted networks, casually mounting shares and whatnot.. it goes to shit fast
2254[22:04:19] <sney> and a friend of mine who does ops and automation for a huge datacenter company said they had 2 weeks full of drama hunting down an i/o problem that turned out to be caused by nfs timeouts and a bum switch :D
2255[22:04:24] <nvz> for nfs to be usable you gotta be on your own private network, and have control of clients and servers
2256[22:05:04] <nvz> trying to do more than that with it can end in nothing but tears
2257[22:05:12] <nvz> usually after you've yanked yourself bald
2258[22:05:16] <sney> yep.
2259[22:05:23] * greycat wonders if the wheezy archive servers even offer https
2260[22:05:58] <sney> don't forget on wheezy apt-transport-https is still separate
2261[22:06:15] <greycat> I'm not even *on* wheezy yet, on that box
2283[22:18:39] <greycat> Is there a trick required to make apt-transport-https work in squeeze, beyond changing the lines in sources.list and removing the apt.conf file?
2284[22:20:54] <sney> not that I remember. readme.debian might shed some light if it's not working out of the box?
2285[22:21:28] <Aavar_> WHen upgrading from debian 8 to 10. Is it ok to just edit the sources.list and run an upgrade? Or is there a more correct way?
2286[22:22:22] <greycat> You have to go 8 > 9 > 10. You can't skip.
2288[22:23:28] <Aavar_> greycat: ok, but is editin the sources.list the correct way?
2289[22:23:29] <greycat> but apt-get update just does *nothing*. hangs there saying 0% [Working]
2290[22:23:44] <greycat> dpkg, tell Aavar_ about jessie->stretch
2291[22:24:17] <mutante> could it be missing "https_proxy=" env variable?
2292[22:24:19] <BCMM> Aavar_: there's a whole thing in the release notes about upgrading, but it's pretty much that, yeah.
2293[22:24:49] <BCMM> Aavar_: oh, and of course it's "full-upgrade"/"dist-upgrade", not plain "upgrade"
2294[22:24:51] <greycat> oh wait, there's some progress. now it said "Ign replaced-url
2295[22:26:13] <BCMM> Aavar_: this page (or the equivalent for your version and architecture) has all the things you should watch out for. but it's basically "make backups, don't have stupid repos enabled, edit sources.list, do full-upgrade" replaced-url
2296[22:27:24] <Aavar_> BCMM, greycat: I'll try. It's not a important machine if it breaks anyway :) Tnx
2297[22:27:44] <Aavar_> btw. Is debian 9 still supported?
2298[22:28:08] <greycat> It's under LTS support at this time, not full support, but close enough.
2299[22:28:12] <Aavar_> wait. I see it in the topic :)
2301[22:29:37] <BCMM> Aavar_: and note that LTS is kind of special in Debian - not all packages in Debian are supported by LTS, so anyone intending to use LTS should make sure that the packages they use are supported
2302[22:29:54] <BCMM> (i.e. still get security updates!)
2303[22:29:59] *** Quits: MatrixBottolonie (~matrixirc@replaced-ip) (Remote host closed the connection)
2321[22:42:08] *** Quits: jariyah (uid364305@replaced-ip) (Quit: Connection closed for inactivity)
2322[22:43:39] <greycat> jhutchins: so after a few hours of these Ign lines coming one by one, it might actually work? And I might in theory be able to install packages, however many weeks that takes at this speed?
2351[22:54:25] <wwilliam> Thank you checking tarzeau
2352[22:54:28] <tarzeau> wwilliam: i've seen computers with not enough powerful PSUs, PSU change helped, even power rigs that were a problem...
2353[22:54:44] <tarzeau> wwilliam: the "sensors" output can be useful
2354[22:54:45] *** Quits: mezzo (~mezzo@replaced-ip) (Quit: zouzou)
2355[22:54:54] <jhutchins> greycat: I dunno, v4 is pretty rare in actual implementation. It's something that only large institutions seem to use because ordinary admins don't understand it.ut they postponed it.
2357[22:55:16] <tarzeau> wwilliam: then there's memtester if you got faulty memory
2358[22:55:29] <wwilliam> !sensors
2359[22:55:29] <dpkg> lm_sensors (Linux-monitoring sensors) is a hardware health monitoring package for CPU temperature, fan speed, etc. Packaged for Debian as lm-sensors. replaced-url
2360[22:55:37] <jhutchins> Man, touchpads suck.
2361[22:55:48] <tarzeau> wwilliam: oh wait, you said chrome (the memory hungry thing), out of memory? swapping? check htop of course, activate zram
2362[22:56:02] <jhutchins> They were going to do a point release on the 18th but they postponed it.
2363[22:56:05] <tarzeau> jhutchins: only with wet hands, or gloves
2364[22:56:07] <wwilliam> Yes
2365[22:56:21] <tarzeau> jhutchins: i've found the apple touchpads to not suck, but yeah anything else sucks!
2366[22:57:05] <tarzeau> wwilliam: i also found nohang to do a pretty good job, getting the right process annihilated, unlike the kernel oom killer
2416[23:03:51] <tarzeau> ax562: we got linux desktops of age 0 day - 10 years, from two different manufacturers, some heros with custom/consumer stuff, and there's seldom more than 3 identical machines
2417[23:03:55] <tarzeau> about 200 of 'em
2418[23:04:01] <ax562> i'm running the nvidia-drivers driver
2419[23:04:14] <ax562> haha
2420[23:04:28] <tarzeau> ax562: yeah should be fine with the 450 driver, i'm just not familiar with the *m stuff (mobile)
2421[23:04:42] <ax562> tarzeau what would you recommend for my hardware?
2422[23:04:48] <tarzeau> computers with linux are best if they run 24/7, and you got many of them, to fix one another from another one
2423[23:04:52] <ax562> oh ok but it doesn't work
2424[23:05:08] <wwilliam> OK it just froze again after opening chromium did not have the chance to install nohang.....
2425[23:05:10] <tarzeau> ax562: the cuda.list with that url i gave before, then just apt install cuda-10-0
2461[23:18:01] <tarzeau> CyberManifest: i pity you :) first make it all lowercase, then cat that file sed away all : and ,, and run all words through apt-cache show && echo $? and you've got half of the work done already
2465[23:18:33] <CyberManifest> tarzeau: I find packages.debian.org challenging to navigate, as for one I can't tell software titles from libraries and versions or releases from other versions and releases.... and I'm currently on an Apple Mac and no where near a linux terminal to do an apt-cache check.
2469[23:19:11] <CyberManifest> tarzeau: that's a lot of work, that's why I was asking if anyone knew of an "easy" way.
2470[23:19:20] <tarzeau> CyberManifest: pretty easy: libraries start with lib* (with very little exceptions)
2471[23:19:37] <CyberManifest> tarzeau: besides I don't know sed well enough
2472[23:19:52] <tarzeau> CyberManifest: any macOS has a terminal cmd+space + termin+enter gives you a terminal to ssh into any debian computer with apt
2473[23:20:10] <CyberManifest> tarzeau: that hasn't been my case navigating packages.debian.org
2474[23:20:37] <tarzeau> CyberManifest: that stupid page blocks GET and curl
2475[23:20:59] <CyberManifest> tarzeau: I don't have ssh setup, you're making a lot of assumptions and inferences
2476[23:21:01] <tarzeau> CyberManifest: also links2 -dump and wget -O-
2479[23:22:07] <tarzeau> CyberManifest: basically the content in a file here: echo `cat here` | sed "s/,//g;s/://g/" | while read a; do apt-cache show $a; done
2480[23:22:14] *** Quits: platvoeten (~platvoete@replaced-ip) (Remote host closed the connection)
2481[23:22:18] <tarzeau> CyberManifest: something like that, oh i forgot to make it all lowercase
2482[23:22:31] <tarzeau> CyberManifest: but what you call Task Manager: HTOP -> htop
2483[23:22:40] <tarzeau> CyberManifest: who made that list? and WHY?
2484[23:22:57] <CyberManifest> tarzeau: I did and for my own planning
2490[23:25:22] <CyberManifest> tarzeau: I'm making a setup script for a bare minimum net install and I'm migrating from another distro so I'm not familiar with the package names, I've already ran into one like wpa-supplicant as wpasupplicant or something like that.
2504[23:28:05] <tarzeau> i also remember, we used to have a 20k lines /etc/hosts file at my work place (10+ years ago), i probably also was the only one using it (and still missing it)
2505[23:28:34] <tarzeau> greycat: you know you can attach strace to running processes? (it's been so helpful to me)
2506[23:28:58] <tarzeau> and i've also used it to get an overview of syscall distribution (by amount)
2507[23:29:00] *** Quits: Jade_NL (~JadeNL@replaced-ip) (Remote host closed the connection)
2508[23:29:01] *** debhelper sets mode: +l 1156
2509[23:30:16] <greycat> well, looks like /usr/lib/apt/methods/https is doing *something* because strace -p blah spews stuff faster than I can read it
2510[23:30:18] <CyberManifest> tarzeau: so there isn't a way at packages.debian.org to filter the results to just the primary software and exclude all libraries and help docs and src files ?
2517[23:31:56] <tarzeau> greycat: not useful, if it doesn't hang anywhere. ltrace? ngrep (oh wait only for http useful)
2518[23:32:12] <greycat> I fucking wish I could use http.
2519[23:32:13] <CyberManifest> and that list is filenames from another distro's repo so they're already in package name but the package names are different and sometimes appear completely foreign in a debian search
2520[23:32:14] * tarzeau is probably also the only one without login on his webpage, and all be replaced-url
2521[23:32:20] <tarzeau> greycat: i just do it!
2522[23:32:43] <greycat> Not on my web site. In apt-get.
2523[23:33:06] <greycat> Corporate firewall intercepts it and demands a login or some such bullshit.
2524[23:33:18] <tarzeau> on any port?
2525[23:33:27] <greycat> Using https gets around that, but squeeze apparently doesn't have a GOOD apt-transport-https.
2526[23:33:43] <tarzeau> ahh that's broken, but you know that
2527[23:34:18] <tarzeau> our network guys get pissed off when people plug loops in the network
2532[23:35:17] <greycat> gnutls_handshake() failed: A TLS fatal alert has been received.
2533[23:35:26] <greycat> that's... bad... right?
2534[23:35:31] <CyberManifest> I guess since there isn't an easy solution, I might have to make one later... a website that returns the results of an apt-cache search from a standard input... user types FireFox, they get a return of firefox
2535[23:35:47] <tarzeau> yes that's really bad
2536[23:35:55] <mutante> yea, i think the ciphers used by squeeze will not be accepted anymore
2539[23:36:42] <AndreasLutro> must be some debian mirror that supports https
2540[23:36:47] <ax562> is there a certain order to install?
2541[23:37:14] <mutante> so the TLS handshake does not succeed because client and server can't agree on acceptable ciphers. it's a guess but there have been many security issues with older TLS versions and ciphers
2542[23:37:15] <ax562> I already had nvida-drivers installed. conflict?
2543[23:37:19] <tarzeau> ax562: you have to install their pubkey, and it'll just work
2651[23:54:47] <ax562> well, ubuntu-driver-440 works out the box
2652[23:54:54] <tarzeau> ax562: it's old! i use 450
2653[23:55:19] <tarzeau> ax562: and if you want to do gpu computing, you want nvidia.com drivers, since libcudnn (you need to make an account and download)
2655[23:55:26] <ax562> lol, Im taking this in as fast as I can, remember I'm still noobish
2656[23:55:56] <ax562> yeah, same thing still
2657[23:56:05] <ax562> The following packages have unmet dependencies: cuda-10-0 : Depends: cuda-runtime-10-0 (>= 10.0.130) but it is not going to be installed Depends: cuda-demo-suite-10-0 (>= 10.0.130) but it is not going to be installed
2658[23:56:22] <tarzeau> ax562: if you say: aptitude install cuda-10-0 ?
2670[23:59:15] <CyberManifest> tarzeau: see like... --no-install-recommends xserver-xorg what's the equivalent in debian cause I'm not seeing that package at: replaced-url