It’s nothing consistent but once or twice I rebooted and nothing happened and once or twice I rebooted and it shut down and didn’t restart, so I had to power cycle. But mostly it’s working as it should.
Does sound PSU related. I believe during boot, the XU4 power usage peaks dramatically (eg: all usb devices are powered at once), CPU is probably running full clock. So any issues with “unstable power” would be most apparent then.
Although if i remember correctly, you upgraded the PSU on your XU4?
copying exynos5422-odroidxu3.dtb as root from a fresh 127 image that has not been updated to 132 to my 132 emmc solved the boot problem for me, but I’m still at kernel 3.10.96+ and not at 3.10.103+ while the kernel 103 is available on the boot partition.
I noticed the problem way back at v122 (?) and stopped using Dietpi because of it. I have used my cloudshell with OMV, Ubuntu 16.04, and android with no power off problems. If the problem was with the power source you would think the issue would show up with other distros - but it seems to only be a problem with Dietpi. I am back to Dietpi because I love it, but it is not usable without a fix to the power down problem.
I have a cloushell with a 320G HD - the HD was used with the other distros also …
I tested the power off issue with a clean Dietpi - immediately after file system expansion and no other software installed and experienced the same problem. Hope this helps.
So just to clarify, the issues you and other users are experiencing are: 1. When you run the command poweroff, the XU4 will not power back on again? Either using the power on button, or, physically reinserting the PSU? 2. When you physically remove the PSU without shutting down, the XU4 fails to power on again?
Can all users who have experienced these issue please list:
huch - The Installation takes a while.
Tried on MicroSD 16GB
i’ve done this and the xu4 reboots. I only can read 6 Lines on the Display the rest seems to be black text-foreground … but the xu4 react on entered commands (reboot, poweroff).
If you could run the same tests on the image above, i’d be extremely greatful. One of our other users has confirmed the issue is now resolved with that image, but, we’d like to get a few more results before closing this one.