dolphin_oracle's respins (updated June 23, 2024)

[dolphin_oracle]

updated my systemd_only and sysVinit_only spins to conform with mx23.3. link in first post.

Mostly conform anyway. there are some testing items in there. there is a slightly different mechanism starting and restarting pipewire on the sysvinit only spin. (technically its present on the systemd-only spin, but not used).

there is also a development live initrd on both spins. this is something we've worked on for a while, but I thought it might be worth including for development purposes. that would only affect live users, as installed systems do not use it, and the goal is that the end user doesn't notice a change.

[a_freed_man]

I have installed the "bare systemd only" respin in a VM and am trying to get it to boot only to a text console and not having any luck.

Started by using the boot options GUI tool to disable the graphical theme, and that worked.

I have changed the kernel parameters from the default "quiet slash slasht" to only "quiet", used "systemctl disable lightdm".

Also tried changing kernel param from "quiet" to "quiet 3" and to "text".

Nothing I've tried works so far. I want the default to be just a text console, but have the option to start a normal X GUI session with startx or equiv cmd line.

[dolphin_oracle]

Those old run level boot codes don’t work on systemd. You need to disable lightdm service.

[a_freed_man]

Those old run level boot codes don’t work on systemd. You need to disable lightdm service.

I have changed the kernel parameters from the default "quiet slash slasht" to only "quiet", used "systemctl disable lightdm".

I did, as I posted above. Apparently you missed that.

All I get is a black screen, not even a flashing cursor.

[dolphin_oracle]

Those old run level boot codes don’t work on systemd. You need to disable lightdm service.

I have changed the kernel parameters from the default "quiet slash slasht" to only "quiet", used "systemctl disable lightdm".

I did, as I posted above. Apparently you missed that.

All I get is a black screen, not even a flashing cursor.

I did.

possibly this then.

https://unix.stackexchange.com/question ... th-systemd

[a_freed_man]

I did try the systemctl set-default to multi-user.target, but not with the -f.
There were a few other suggestions in that link, I'll give them a try.
Thanks!

[a_freed_man]

I doubt the -f flag on the disable makes a difference in this case.
I verified the systemd symlinks are valid units for graphical.target & multi-user.target.
I get a warning about runlevel defaults when I use systemctl [status|enable|disable] lightdm.
setting the default.target has no obvious affect on lightdm's unit status. I know thereis an interplay between init.d scripts & systemd, so this issue isn't as pure as it would be under only systemd or sysV.

With X I have the network configured and can reach the Internet. When I disable lightdm (+ chg default, + diff kernel opts: text systemd.unit=multi-user.target) I get a login console prompt after several minutes delay.

Apparently the network doesn't start. See image

[dolphin_oracle]

being targeted at the desktop, the spins are set up to be used with network-manager. there is a cli interface for it. systemd-network likely isn't needed, but if you want to use manually networking, you'll need to set it up.

[a_freed_man]

The NM created the /etc/network/interfaces and added nameservers to /etc/resolv.conf.

I did several things, not sure what was essential. I disabled the network-manager with systemd, but interfaces still wouldn't come up. I may have forgotten to stop it tho.

I created /etc/init/network-manager.override file containing "manual". Restart was fast and I had networking afterwards.

I can even switch back to the desktop with startx.

[a_freed_man]

I just discovered the MX Workbench respin and it might be just what I need to resolve a bootup issue I have on an old Lenovo i5 system I'm trying to install TrueNAS on.

I added a dual SATA M.2 PCIe adaptor to use as a RAID system drive to install it on, and 2 SATA 2TB HDDs for the main storage. I booted the TueNAS installer from a USB flash drive and installed it onto the M.2 SATA drive array without a problem, TrueNAS detected the drives just fine. However the machine's BIOS did not, and TrueNAS wouldn't boot.

I upgraded the BIOS from 2014 version to the latest one released in 2022, still didn't detect the M.2 SATA drives.

I ran into the exact same problem on an old Dell system a coupla years ago running MX21, which I solved by booting it off a USB drive and essentially chrooting to MX installed on the SSD the BIOS refused to boot from. Fro what I see online this is a fairly common problem with older hardware.

Now TrueNAS is based on Debian but uses the ZFS filesystem. I see MX. Workbench provides tools to support ZFS but I'm not sure how to use them. I see the ZFS daemon for FUSE access is started at boot, but I don't see how to mount the TrueNAS drives, mount fails. I'm hoping perhaps I might be able to boot MX Workbench from USB and resolve the TrueNAS boot problem in a similar manor as I did for the Dell system.

I can imagine 2 possibilities:
1) change the MX EFI on the USB to point to the M.2 SATA drives to boot TrueNAS (I believe EFI has support for PCIe devices)
2) MX mounts the M.2 drive (possibly using chroot) and starts its' systemd / init start process.

I'm open to other suggestions for this boot problem.