Creating a bug report/issue I have searched the existing open and closed issues Required Information cat /boot/dietpi/.version G_DIETPI_VERSION_CORE=10 G_DIETPI_VERSION_SUB=5 G_DIETPI_VERSION_RC=2 G_GITBRANCH=‘master’ G_GITOWNER=‘MichaIng’ echo $G_DISTRO_NAME $G_RASPBIAN trixie uname --all …

Hi Michalng, thanks for the reply. The behaviour is always seen in a root user session. I don’t see that DBUS message. It doesn’t seem to matter whether root is logged in on debug console ttyS0 or SSH. I could run a diagnostic command from one when it’s hung on the other if that would be any help? J…

Wait, you mean the reboot/poweroff/systemctl, commands are broken even after a power cycle, so it is not just a transient state right after dietpi-installer ran through in the same session? Then probably something is broken on the base image already. Which one did you use?

Yes systemctl works OK after a cold boot, but once i issue a “reboot” or “shutdown” command the prompt “freezes” and I must use CTRL+C to break out of the hang. After this happens systemctl now does nothing and hangs in the same way - and here CTRL+C will also break back to the prompt. The image I u…

Hmm, that is surprising. The Armbian images should be clean. The reboot and poweroff commands should just call systemctl start reboot.target and poweroff.target respectively, which do not invoke D-Bus. But somehow they seem to hang systemd. Maybe it partially shuts down already, but the shutdown se…

I have been connected to the UART the whole time, because I needed that to run the dietpi migration script which drops the SSH connection. I include the entries from journalctl -bbelow that are between the login and CTRL+C from the hung reboot. Note that echo b | sudo tee /proc/sysrq-trigger will ca…

Right, makes sense, since NetworkManager drops any active connection when it is uninstalled. So it really starts a shutdown sequence, but it does not reach far. What I do not see is cron and dietpi-kill_ssh finishing. From the latter, the dropbear early exit is coming, otherwise the SSH session rem…

OK I disabled cron and dietpi-kill_ssh but the journalctl output is the same: Aug 08 18:53:40 DietPi login[463]: ROOT LOGIN ON ttyS0 Aug 08 18:54:30 DietPi systemd[1]: Removed slice system-modprobe.slice - Slice /system/modprobe. Aug 08 18:54:30 DietPi systemd[1]: Stopped target graphical.target - …

I think I might have found the solution. I noticed in the dmesg output the following: Aug 08 19:04:56 DietPi systemd[1]: dev-ttyS7.device: Job dev-ttyS7.device/start timed out. Aug 08 19:04:56 DietPi systemd[1]: Timed out waiting for device dev-ttyS7.device - /dev/ttyS7. Aug 08 19:04:56 DietPi syst…

I notice that disabling the serial ports in dietpi-config does the same thing as systemctl mask, but it would be nice to prevent the kernel from creating them in the first place. Does that have something to do with an incorrect DTB? I’m not sure about that - and I also see mention of dtoverlay=disab…

Oh yeah that makes totally sense. On unknown SBCs, where we don’t know for sure which UART device is the primary/debug port, we just enable all of them, which is why you see ttyS1 til ttyS7 all enabled, if those nodes /dev/ttyS* do all exist. But usually only 1-3 are actually usable as serial consol…

DietPi Community Forum

OrangePi Zero conversion

Troubleshooting

show post in topic

Home
Categories
Guidelines
Terms of Service
Privacy Policy

Powered by Discourse, best viewed with JavaScript enabled