| hagbard | Currently trying do install devuan on yet another machine, using the excalibur netinst iso image. The image boots, but fails right at the beginning with "failed to mount the cdrom". Any ideas? | 19:35 |
|---|---|---|
| freem | hagbard: so you are installing on a virtual machine? | 19:37 |
| hagbard | physical | 19:37 |
| freem | so, you put the iso on an usb key? How did you prepared it? | 19:37 |
| hagbard | Just dd , as always. | 19:38 |
| freem | is there is a way to run dmesg in a different PTY? | 19:38 |
| hagbard | yes. The usb key is recognized. But the following lines say "Can't lookup blockdev" | 19:40 |
| freem | no idea what blockdev is, so I'm going to rtfm for a bit. Not sure if I'll be able to help after that though | 19:41 |
| freem | BLOCKDEV(8) blockdev - call block device ioctls from the command line | 19:42 |
| freem | well... that won't be much helpful | 19:42 |
| hagbard | Works fine with the deadalus image, though. | 19:43 |
| freem | https://bbs.archlinux.org/viewtopic.php?id=298097 this is an archlinux post, which mentions the error: | 19:43 |
| freem | ERROR: device '/dev/mapper/root' not found. Skipping fsck. | 19:43 |
| freem | mount: /new_root: fsconfig system call failed: /dev/mapper/root: Can't lookup blockdev. | 19:43 |
| freem | do you have the same message, or does it looks a bit different? | 19:44 |
| freem | dmesg(1) may have more information after failed mount system call. | 19:44 |
| freem | hm... /dev/mapper/foobar is about using LLVM and RAID-like stuff, right? | 19:44 |
| freem | huh... LVM, not LLVM. Coder bias. | 19:44 |
| hagbard | That's about something entirely different. | 19:45 |
| freem | yeah, I can guess, I'm mostly asking about the error message, to confirm it's alike | 19:45 |
| freem | I am almost certain you have more useful context, but I understand that on a physical machine you try to install, it's not necessarily easy to give accurate reports (need a camera to take a pic, or to write the stuff down if you have only one computer... went there) | 19:47 |
| hagbard | I can get it to work for me just fine, by using the deadalus image with excalibur package sources, no problem. But that means the excalibur installer needs some troubleshooting. | 19:47 |
| freem | indeed | 19:47 |
| freem | I have to admit I have been installing my debians with debootstrap and own scripts since years, probably more than a decade, because debian installer sucks | 19:48 |
| freem | I don't remember about devuan's, I may very well never have tried it :/ | 19:48 |
| n4dir | pretty much the same | 19:49 |
| n4dir | at a certain point you can choose the init system, but the rest seems just the same. | 19:49 |
| freem | still, a bug report with this little information will most likely be barely useful... does dmesg really gives anything? | 19:49 |
| n4dir | perhaps i forgot something | 19:49 |
| freem | heh, then it still sucks terribly | 19:49 |
| freem | last time I tried debian's installer, you even *had* to use grub because the lilo option was broken | 19:50 |
| freem | and grub2... thanks, but no thanks. | 19:50 |
| hagbard | It's the same, just with some differing configuration. But I had nor issues with it, even with more complex installs, neither in debian nor devuan. | 19:50 |
| n4dir | running in a broken lilo won't happen to you anymore | 19:50 |
| freem | lilo worked better than grub | 19:51 |
| freem | I use syslinux those days, though | 19:51 |
| freem | the problem I always have had with debian's installer... the several ones, actually: questions are spread, you *must* stay near computer. Better have a beer storage at disposal to kill time... hardly any useful choice, too. And automation is more painful than just scripting your way with debootstrap | 19:52 |
| freem | that's just my opinion, though. | 19:52 |
| freem | and it does not helps with the situation at hands | 19:53 |
| n4dir | if you did it often enough you know pretty well when you have a bit of time, and when to stay | 19:53 |
| freem | I did it too often, so I got tired of that and wrote myself a better script | 19:53 |
| freem | but ofc it's rather barebone and only adapted to my needs and mindset, it must be lacking way too much :) | 19:54 |
| freem | I can only find archlinux related stuff for https://duckduckgo.com/?q=%22Can%27t+lookup+blockdev%22&kae=t&ia=web | 19:55 |
| freem | and only 3 entries, too | 19:55 |
| freem | all the stuff is from 2024 | 19:56 |
| freem | maybe the computer is very new and fresh? | 19:56 |
| freem | so might be some firmware missing... no, that'd be illogical | 19:56 |
| golinux | hagbard: Perhaps be the one who creates the isos and ensures that they work. The job is open: | 20:02 |
| golinux | https://files.devuan.org/devuan_excalibur/installer-iso/WhereAreTheISOs_current_status.txt | 20:02 |
| hagbard | I can put effort into that, if someone can get me started. Unfortunately, I don't have any experience with building installer images. | 20:05 |
| hagbard | I know how to read/modify jenkins groovy, if that helps. | 20:07 |
| golinux | It requires more than just building the isos. Bugs need to be fixed first. ;) | 20:10 |
| hagbard | Well, if someone shows me what to do, I can keep doing that. | 20:18 |
| golinux | Sadly, the bugs are constantly changing. :( | 20:35 |
| freem | lol | 20:36 |
| freem | golinux: that's actually a good thing, it means the same errors are not repeated all the time | 20:37 |
| golinux | freem: Have a look https://git.devuan.org/devuan/installer-iso | 21:04 |
| freem | I'd have a hard time considering #18 as a bug tbh :D | 21:05 |
| freem | --merged-usr *is* the bug | 21:05 |
| golinux | Debian is the bug these days | 21:05 |
| freem | (but I understand it can't be fixed without tremendous amount of work) | 21:06 |
| freem | I agree. | 21:06 |
| freem | debian is as universal as a hammer | 21:07 |
| freem | it *can* handle screws, but the result won't be neat | 21:08 |
| armin | freem: If you're determined, you can absolutely use a hammer to drive a screw. | 21:30 |
| freem | yep | 21:30 |
| freem | that's useful even, when screw's is damaged and you *still* want to put it deeper | 21:31 |
| freem | but that's as far as debian goes, when it's about being universal | 21:31 |
| freem | it's just yet another systemd+gnome distro | 21:31 |
| armin | The weird thing is that on a non-systemd distribution, the times I was wishing to be on a systemd one were pretty damn close to 0. | 21:32 |
| freem | those projects represent most of all I dislike in softwares. But that is offtopic. | 21:32 |
| freem | to me it's strictly 0. | 21:32 |
| armin | reasonable. | 21:33 |
| freem | I tried systemd, long ago. It's not worth it. | 21:33 |
| armin | I totally agree. | 21:33 |
| freem | The idea of declaractive style for configuration is *good* | 21:33 |
| armin | That would lead to NixOS, not systemd. | 21:33 |
| freem | but the project went way too far and lost the most important thing for a project to be good: it's scope. | 21:33 |
| armin | correct. | 21:33 |
| armin | scope creep. | 21:33 |
| freem | so in the end I'm happier with runit, by far. | 21:34 |
| freem | service dependences? Sure. Tell me when exactly you need that. | 21:34 |
| armin | yea or OpenRC or whatever | 21:34 |
| freem | I have had _one_ need for it, and the archaic way of daemontools actually was still quite reliable anyway, while being much simpler. | 21:34 |
| armin | freem: yea, and you can perfectly well just drop daemontools/runit *ON TOP* of systemd if you need it. | 21:35 |
| armin | try that the other way around. | 21:35 |
| freem | complex shell scripts? The only one which requires more than 5 lines is actually the one for systemd-dbus, ain't it ironical? | 21:35 |
| freem | I see no reason to do that | 21:35 |
| armin | I have absolutely 0 problems dealing with hundreds of lines of BASH. | 21:35 |
| freem | I do | 21:36 |
| armin | I have *major* problems going through literally *hundreds* of man-pages that have a name starting with systemd* though. | 21:36 |
| freem | shell script are tricky to write correctly and to read | 21:36 |
| armin | I have no problems writing robust shell scripts. | 21:36 |
| freem | ah, yeah, I have a bigger problem with that | 21:36 |
| freem | reading thousands of lines just for one tool, knowledge I can't reuse elsewhere | 21:36 |
| freem | armin: well, I never claimed to be smart :) | 21:37 |
| armin | me neither :) | 21:37 |
| armin | :D | 21:37 |
| freem | I'm limited to C++, and I configure compiler nazi-style because I'm too dumb. Turns out, that plus some discipline in software architecture prevents all those memory problems they all say C++ got troubles with. But I don't do the work: the compiler does it. | 21:38 |
| freem | it constantly insults me. | 21:38 |
| freem | which is fine. | 21:38 |
| golinux | This should probably go to #devuan-offtopic. ;) | 21:39 |
| freem | anyway, we're off-topic and will be slapped if we continue :) | 21:39 |
| freem | see | 21:39 |
| golinux | Beat you to it! | 21:39 |
| freem | :) | 21:39 |
| golinux | I go to wash dishes and look what happens! | 21:39 |
| armin | golinux: thx for pointing us there, will join. :) | 21:40 |
| golinux | :D | 21:40 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!