You are not logged in.
Pages: 1
Oh, I only have Windows in the internal drive and haven't booted it in months. Only grub's minimal shell, efibootmgr, and windows can find it. I have the files backed up, so losing it wouldn't be any trouble, unless it interferes with booting anything I use. It's been like this the entire time I've had linux on the external drive.
Hello, I have had this trouble ever since allowing ubuntu-derived and debian/devuan to coexist on the same drive. I often lose the uefi boot entries except for one that is labeled 'neon' or 'ubuntu' or just the hard drive's name and windows (internal disk) in the uefi boot menu, and must use uefi boot to add it back, or, until recently, use grub from whichever neon entry I used (anywhere from one to three show). Until last night (when I got stuck on a loop of the the acer logo flashing before rebooting automatically again), I have been able to just enter from the neon grub (regardless of which uefi entry lead to it). While I have shim installed, now trying to boot from neon's grub (so I can use efibootmgr in devuan proper) shows that the shim signature is bad and that the kernel needs to load first. I will try to reboot into devuan as soon as the kde partition manager finishes canceling (It froze at 25%, so after an hour hit cancel) or I surrender my ebooks which I have copies of.
More details:
From within the grub menu, I can run ls, and everything shows up. My typical hard drive is hd0 if no live media is used, and otherwise it is hd1 (with the live boot media being hd0), and the internal drive (which I cannot access or see once booted) is one index higher, and still hd (although it is an ssd). Prior to the latest trouble, I went 85 days without rebooting, then rebooted twice in one day (I doubt this is the issue). I am using an Acer Aspire 515, and my shim versions are below:
shim-helpers-amd64-signed is version 1+15.8+1~deb12u1
shim-signed:amd64 is version 1.44~1+deb12u1+15.8-1~deb12u1
shim-signed-common is version 1.44~1+deb12u1+15.8-1~deb12u1
shim-unsigned:amd64 is version 15.8-1~deb12u1
I copied /etc/os-release to /etc/lsb-release since that was suggested on an arch forum thread, but firmware was a potential issue too, although I haven't messed with firmware lately, and have intel-microcode installed along with firmware-iwlwifi neon only has the additional amd64-microcode, firmware-sof-signed, and iucode-tool, which I am working on installing through chroot now.
Update: installing those did not help, and now I cannot even enter into uefi settings or uefi boot menu; that is just a frozen tty. mokutil shows secure boot is enabled again, as booting into neon sometimes does. But I have the current workaround of mokutil to disable secure boot and efibootmgr to recreate the entry and make it be nextboot.
The release files appear to be expired/invalid in the apt repo http://packages.gnuinos.org since the site went down.
The github link is now broken; is the project cancelled?
I think if you use apt pinning of the current repos to be over 1000, it will downgrade, quickfur.
Oh, sorry, I misunderstood twice that it was required. First in the sticky, then in here.
Yes, that is the one I read; task-mate-desktop kept removing libsystemd0. Now, when I try to install libsystemd0, it wants to remove task-mate-desktop, mate-desktop-environment, mate-desktop-environment-core, caja, gvfs-backends, mate-applets, gvfs, gvfs-daemons, gvfs-fuse, udisks2, network-manager-gnome, network-manager, libpam-elogind, and libelogind-compat
I am currently booted without libsystemd0 because it apparently conflicts, either directly or indirectly, with these packages. I do not think libsystemd0 is harmful, but I seem to be unable to have it, as holding it when installed would not allow installing these packages (well, dummy-elogind allowed some of them, but not all).
This is not important, but is it possible to get this to show up on wifi indicator applets? If not, I may (depending on free time next semester) try to edit the mate applet showing wifi to have a gui for this.
Alright, I've had it fixed for a day, but I just now fixed the DE and halfway fixed the internet. Thank you everyone.
I misunderstood the sticky on systemd mentioning libsystemd0 being used sometimes to determine if systemd was running, and thought that this package was necessary, so I tried to use the dummy-elogind, but this would not allow some packages that were part of task-mate-desktop, so I had a partial de install and dummy-elogind. Fortunately, I figured it out when I looked into what would break when I removed libsystemd0. My wifi applet does not work (that is okay; I don't necessarily need a gui for it as long as it works), but I followed the excellent guide by quickfir and got it working (the only difference was I had to use apt-file to locate iwconfig and ifup... now they work). If I have time next semester, I may try to edit the applet to get a gui method for this workaround.
Thank you! Should I continue where I left off, or do I need to force upgrade the packages that did not upgrade and then reboot again?
Today I started to migrate my Debian bookworm installation (using stable, updates, security, and one package from backports) to daedalus. In preparation, I fully cloned it and booted into both the original and the clone using the grub installed in my old KDE Neon install (which I am now using). After seeing that both worked, I followed the migration guide. Initially, after the apt upgrade, I noticed that a few packages were upgradable but not upgraded (dbus-bin, dbus-daemon, dbus-x11, dbus, libdbs-1-3, and init... also octave*). However, looking ahead I saw -f install, rebooting, purging, and autocleaning, so thought that maybe this was as intended. I suspect this is where I made a mistake. Installing eudev and sysvinit-core did force init to install without me manually doing that. Rebooting after the -f install I had a headless install, as expected (I use mate), but it said at the top
startpar: services returned failure: nscd network-manager . . . failed
* octave is upgradable but not upgraded since I edited the .deb from the debian repo to remove the gui since I use geany for it and have no qt except on the neon install; I highly doubt this has any effect but am telling out of precaution
Now, what would be the safest way to continue? I chrooted into it to get that list of upgradable and not upgraded packages (I'm lazy and didn't write it down), and could install that way, similar to installing enough packages to set up wifi on a debootstrap-install. Alternatively, I could download the remaining packages and install them when truly booted into that install (well, efibootmgr will say I'm booted into neon since I'm using this grub to avoid ambiguities for the time being, but I assume it is okay since grub gets me in the correct one). I do not like the other options of using the clone or cloning the clone over this attempt, because I want to keep at least one working debian/devuan install at all times and cloning takes a day and a half. Is it safe to continue the migration from a chroot (since I already rebooted), at least enough to get wifi working? This is my preferred method as it is easiest.
Pages: 1