rrq | though, despite the stupidity of the notion, it really is just a choice (and fundamental change) of how debian developers now a) install files and b) refer to them from scripts and programs. They really *need* to install programs at one path and then use them from another.. otherwise they can't sleep in the night. | 00:06 |
---|---|---|
rrq | obviously they do this back-assed; moving the programs first and then sort out references... (maybe just so it should be some breakage on the way to show how important they are) | 00:12 |
rrq | but really: how invested in particular pathnames are you or Hurd? Does it matter that much? | 00:16 |
rrq | it's just "code" | 00:18 |
fsmithred | As I understand it, all the commands will be found if you have the merged usr symlinks. | 00:40 |
fsmithred | I think they did it just to prove that they could make us do it the new way. | 00:41 |
rrq | yeah; good way to reduce trust in debian as software platform | 00:49 |
* rrq ot | 00:49 | |
fsmithred | ok, back on topic... | 01:11 |
fsmithred | kernel install is still failing for 6.6.9 on one system. Error is wiht initramfs-tools. | 01:13 |
rrq | due to some /usr/... path I guess | 01:16 |
fsmithred | I added set -x to /etc/kernel/postinst.d/initramfs-tools and /usr/share/initramfs-tools/hooks/lvm2 | 01:16 |
fsmithred | want to see the output? (say yes) | 01:16 |
rrq | yes :) | 01:17 |
fsmithred | there's a weird lookiing path in it | 01:17 |
rrq | try https://paste.rrq.au if you like :) | 01:17 |
fsmithred | https://termbin.com/kj2v | 01:18 |
fsmithred | I can't copy/paste | 01:18 |
rrq | right https://paste.rrq.au is my transfer.sh site | 01:18 |
fsmithred | look for lines with mkinitramfs that start with cp | 01:18 |
rrq | you'll need set -x in /usr/share/initramfs-tools/hooks/lvm2 | 01:20 |
fsmithred | I did | 01:20 |
fsmithred | maybe it's the /lib/udev... but I have merged usr in this system. | 01:21 |
rrq | missing 69-lvm.rules ? | 01:22 |
rrq | or hmm maybe it requires them missing and they are not | 01:23 |
rrq | (no; seems like it wants to copy them into that temp dir) | 01:23 |
fsmithred | it's not in /etc | 01:24 |
rrq | OOM ? | 01:24 |
rrq | in /var/tmp | 01:24 |
fsmithred | not in /lib | 01:25 |
fsmithred | only thing in /var/tmp is mkinitramfs_<some-letters> | 01:25 |
rrq | where is 69-lvm.rules ? | 01:26 |
rrq | maybe lvm is half configured? or left residue when removed? | 01:27 |
fsmithred | in the lvm2 package, but not the one I installed, I guess | 01:27 |
fsmithred | I'm trying 'aptitude reinstall lvm2' | 01:27 |
fsmithred | this will take a couple minutes | 01:27 |
fsmithred | now hooks/dmsetup is failing | 01:29 |
rrq | dome /usr/.. moved path I guess :) | 01:30 |
fsmithred | I need to purge one kernel so this goes faster | 01:30 |
fsmithred | shouldn't matter if I have the symlinks, should it? | 01:30 |
rrq | no, it shouldn't. only if you got the stupid-links recently set up and that process lost stuff | 01:31 |
fsmithred | lost stuff? | 01:32 |
fsmithred | They couldn't write a script that copied everything over? | 01:33 |
rrq | well, that would be one way for not finding it now; there is no telling what the kiddies can do | 01:33 |
fsmithred | reinstall gave me 69-lvm.rules | 01:34 |
fsmithred | so I guess I need to reinstall dmsetup | 01:34 |
fsmithred | to find out the next one that's missing | 01:35 |
rrq | yeah; good way to reduce trust in debian as software platform | 01:36 |
fsmithred | I'm not seeing what the error is now. Didn't save it in a file and the console is spammed with Warning: behavior of -n is non portable | 01:40 |
rrq | yes, it's the | 01:40 |
fsmithred | and the first part of the output gets lost in the console. shift-pgup works for a couple time and then I scroll up into previous commands. First output is lost. | 01:41 |
fsmithred | also, when I quit my pager (less) the file I was looking at is still in the console. | 01:41 |
rrq | yes, it's the "cp" maintainer (Michael Stone <mstone@debian.org>) who's got another good idea | 01:41 |
fsmithred | phone | 01:41 |
rrq | (could possibly be upstream cp I guess) | 01:44 |
fsmithred | back. I think the problem is fixed. | 02:15 |
fsmithred | cp is the problem? | 02:16 |
fsmithred | I think I fixed it by reinstalling lvm2 and dmsetup | 02:17 |
rrq | at least we haven't got them to be systemd-sed and systemd-cp yet | 02:18 |
fsmithred | no boot. cryptsetup still can't find any devices | 02:19 |
fsmithred | lol, coming soon, I'm sure. | 02:19 |
rrq | you have an error from cryptsetup? ...and/or can you flash the script that is failing? | 02:21 |
fsmithred | cryptsetup says it can't find device - uuid it shows is the right luks partition. | 02:21 |
fsmithred | you want a picture? | 02:21 |
rrq | :) | 02:21 |
rrq | no | 02:21 |
fsmithred | I can do that. It's in a vm. | 02:21 |
fsmithred | ok, say more | 02:22 |
rrq | add "console=ttyS0" to the boot command line, and then start it on the command line (as you already do I guess) | 02:22 |
rrq | that'll make noise in that terminal | 02:22 |
fsmithred | ok | 02:23 |
fsmithred | I think I have a button for that option | 02:24 |
fsmithred | some key combination to see the console in my terminal? | 02:25 |
fsmithred | well, that screen (host computer) seems to be locked up | 02:26 |
fsmithred | crytpsetup: Waiting for encrypted source device UUID=blah... | 02:27 |
fsmithred | Gave up waiting for root filesystem | 02:27 |
onefang | You can get qemu to dump the console in real time to the terminal you started it from. It abuses virtual serial ports. | 02:29 |
fsmithred | yeah, that's what I was trying to do. | 02:29 |
fsmithred | more message... | 02:29 |
fsmithred | ALERT! UUID=(a different uuid) does not exist. Dropping to a shell. | 02:30 |
fsmithred | initramfs | 02:30 |
fsmithred | oh, maybe a resume file | 02:30 |
fsmithred | no | 02:30 |
fsmithred | there's no swap partition | 02:30 |
rrq | you'll need "-serial mon:stdio" for console to terminal maybe | 02:36 |
fsmithred | yeah, I think I need both | 02:37 |
rrq | so this is encrypting before lvm mastering or lvm first? | 02:38 |
* rrq needs to make lunch here so will be intermittent | 02:39 | |
onefang | I use " -serial pipe:/tmp/guest && cat /tmp/guest.out" Don't think the linux kernel needs to know in this case. | 02:41 |
fsmithred | no lvm | 02:41 |
fsmithred | I'm gonna paste a few lines here | 02:42 |
fsmithred | Begin: Mounting root file system ... Begin: Running /scripts/local-top ... [ 4.572242] device-mapper: core: CONFIG_IMA_DISABLE_HTABLE i> | 02:42 |
fsmithred | [ 4.573865] device-mapper: uevent: version 1.0.3 | 02:42 |
fsmithred | [ 4.574620] device-mapper: ioctl: 4.48.0-ioctl (2023-03-01) initialised: dm-devel@redhat.com | 02:42 |
fsmithred | cryptsetup: Waiting for encrypted source device | 02:42 |
fsmithred | UUID=8f10bf6f-5c5c-440e-96af-8bd90e2f69dc... | 02:42 |
onefang | Oh wait kernel has " console=tty0 console=ttyS0,115200n8" | 02:42 |
onefang | So I get both. B-) | 02:42 |
fsmithred | some of that line got cut off | 02:43 |
fsmithred | CONFIG_IMA_DISABLE_HTABLE is disabled. Duplicate IMA measurements will not be recorded in the IMA log. | 02:44 |
rrq | mmm maybe inspect the initrd for /usr/... anomalies | 02:53 |
rrq | like hardcoded /usr/something in the init scripts | 02:53 |
fsmithred | it's working now | 03:00 |
fsmithred | I reinstalled lvm2 and dmsetup and I guess that supplied the missing files | 03:00 |
fsmithred | wait a minute | 03:01 |
fsmithred | I started to boot and walked away. I might have booted 6.5 | 03:01 |
fsmithred | yeah, I'm in 6.5. So I can check initrd | 03:01 |
fsmithred | not right now | 03:01 |
fsmithred | Is there a reason why /etc/init.d/elogind is not executable in my ceres? | 13:07 |
fsmithred | This is in the one that user can run startx and get to wm | 13:07 |
fsmithred | On another ceres, elogind is executable. That's the system that won't boot 6.6.9 kernel. Boots with 6.5 but I can't start X as user or as root. Permission error. | 13:25 |
rrq | "not executable" as in file mode x missing? | 13:34 |
fsmithred | correct | 13:44 |
fsmithred | and that's the one that works | 13:44 |
fsmithred | I just tried removing the x on the broken system and it didn't help | 13:44 |
fsmithred | actually had to boot live-iso to fix it. After entering login name and password, the cursor moves down a fdew lines and then nothing hapens. | 13:45 |
rrq | might be unrelated? perhaps you trialled starting with disbaled elogind at some point? | 13:45 |
rrq | startx | 13:45 |
rrq | (before installing seatd) | 13:46 |
fsmithred | I've been running starx with elogind disabled just fine on 6.6.9 kernel on the system that was debootstrapped | 13:46 |
fsmithred | what would that do? | 13:46 |
rrq | seatd provides user access to input devices | 13:46 |
rrq | for Xorg | 13:46 |
fsmithred | yeah, I'm not the one who disabled elogind on the working system | 13:47 |
rrq | as alternative to using elogind | 13:47 |
DPA | It also handles the vt switching, I thionk. | 13:47 |
DPA | And access to the /dev/dri/ cards | 13:47 |
fsmithred | mmdebootstrap onto existing encrypted partition, added only a few things, | 13:47 |
fsmithred | oh, maybe I removed elogind on that system and it disabled itself? | 13:47 |
rrq | sounds odd | 13:48 |
fsmithred | the broken system was installed from devuan live ceres that I made and then it broke with the last upgrade | 13:50 |
fsmithred | started that one back in June and it's been working until the new kernel 5,5,9 | 13:51 |
fsmithred | I extracted initramfs from both systems and compared them in meld. I saw no significant differences in config files. | 13:53 |
fsmithred | and both had merged usr | 13:54 |
rrq | there are times when debian updates of forked packages suprecedes the forking, and installing/upgrading (accidentally) at then is usually a problem | 13:57 |
fsmithred | in that case, all up-to-date installs should have the problem, right? | 14:01 |
fsmithred | I got it so that root can startx, but I can't use mouse on desktop. | 14:01 |
rrq | probably yes | 14:01 |
rrq | with seatd? | 14:02 |
fsmithred | yes | 14:03 |
fsmithred | well, full xfce desktop and everything that goes with it. elogind, pkit, and I think seatd is installed, too | 14:03 |
rrq | any Xorg log complaint | 14:04 |
DPA | Do you have xserver-xorg-input-libinput installed? | 14:04 |
fsmithred | I'll check. booting now | 14:05 |
fsmithred | yes | 14:06 |
fsmithred | open /dev/dri/card0 No such device | 14:07 |
fsmithred | sorry No such file or directory | 14:07 |
fsmithred | same for fb0 | 14:08 |
rrq | but those pathnames exist as normal? | 14:08 |
fsmithred | there is no /dev/dri | 14:09 |
rrq | so no graphics ? | 14:10 |
fsmithred | root can see the desktop | 14:10 |
fsmithred | lightdm did not show up even when it was running | 14:10 |
fsmithred | I disabled it | 14:10 |
fsmithred | so more files got lost when I ran usrmerge? | 14:11 |
DPA | That sounds like either udev not working, or the modules for the kernel not being found. | 14:11 |
fsmithred | I had to reinstall lvm2 and dmsetup to recover a couple of lost files | 14:13 |
fsmithred | maybe I need to reinstall xorg? | 14:13 |
fsmithred | or maybe reinstall all the packages? | 14:13 |
rrq | eudev at least | 14:13 |
fsmithred | hm, I did that, then init 1, then log in as user and startx | 14:16 |
fsmithred | cursor hangs, then I hit ctrl-c and then some output came up like it was going to start X. | 14:17 |
fsmithred | It's responding slowly. | 14:17 |
fsmithred | timeout in locking authority file | 14:17 |
fsmithred | unable to connect to xserver. Connection refused. | 14:18 |
DPA | Missing files could also be caused by hard drive problems. | 14:18 |
fsmithred | qemu | 14:18 |
fsmithred | I'm gonna have to stop, and I probably won't be able to work on this at all in the next week. | 14:20 |
fsmithred | I'm reinstall xorg, xinit and every package with 'xserver' in the name. | 14:51 |
fsmithred | also reinstalled elogind libpam-elogind and policykit-1-gnome - no help | 15:10 |
fsmithred | oh, something did help there. I can now boot the 6.6.9 kernel in the broken system. Cryptsetup no longer cries about not finding the source device. | 15:12 |
fsmithred | and user can startx. | 15:13 |
fsmithred | I re-enabled lightdm, rebooted and login screen comes up - again and again every time I log in. | 15:16 |
fsmithred | I'm done for now | 15:17 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!