[Richie]

Power button on the brand spanking new replacement nuc just stopped working. This thing is really starting to piss me off. Hit the button, it hung on load, held the button and it started blinking, wouldn't stop. Unplugged from power, replugged. Button is fried. Now I can only wake on lan again. ARRRRRGHGHGHHGHGHHGHH!!!!!!!!!!!!!!!!

I don't think it's the power button, hardware at least. My scenarios are:

1. Running 3.2.X with colour fixes, using suspend, the NUC just starts itself up again, ie. it won't stay sleeping
2. Running 3.2.X with colour fixes, using shutdown, the NUC shuts down perfectly.
3. Running Gotham nightlies, using Suspend, the NUC gets a 'working' dialogue box and then crashes
4. Running Gotham nightlies, using shutdown, the NUC shuts down perfectly

So my first issue is Suspend. I believe someone from here posted on the NUC forums about the issue, yet to get an answer. Seems to relate to the USB3 host waking up the system, which may also relate to the Suspend crash from Gotham nightlies, I have heard USB3 is still a bit finicky in Linux.

The second issue is, with scenario (3) above, I have to do a hard reset with the power button. The system powers down, but then the button does not work. Even unplugging the power cable doesn't resolve the issue. But!, using the power button on the remote works perfectly. The only other way to get the hardware power button working again is to open the NUC up and take out the ram, S3 suspend keeps it's footprint in ram.

This now leads me to believe that there is an issue with the low level code back to the bios, somehow, OE issues a halt S3 command to the bios, the bios can't process the command, but OE has already halted. The bios then sits there waiting for an un-suspend command (sorry for the lack of technical terms), ignoring any input from the hardware power button (we get similar issues on some HP PC's at my work when they crash), but interpreting the power button on the remote as a wake-from-suspend command (a simple soft power on command). The problem is, hard powering down the NUC kills the XBMC process, so you're back to rebooting from scratch when this happens.

So my conclusion is, this is a bios level issue that Intel need to fix. Intel are normally pretty quick with these, so hopefully we get a fix in a week or so. It's annoying having to use shutdown rather than suspend, as it causes issues with my central SQL database and also takes longer too boot, but it's better than not suspending, at least for me, which leaves my NAS running all the time.

Sorry for the long post, just hoping I can give some insight into what I believe the problem to be.