Damn, I changed it already and restarted 
Anyway, it’s up and running again and I’m not touching it again until tomorrow evening.
BTW, Kodi never crashed while it was up and running. It was always happening after a restart.
Just to sort out any confusion, your system still crashing when under-clocked does NOT mean that the power supply is good. If the PSU had drifted just a little out of spec then under-clocking might have decreased the load just enough to make the problem go away (at least temporarily) and allowed for a quick answer without having to acquire new hardware. Since it crashed both with and without the under-clock then basically that test told you nothing regarding the integrity of the power supply.
It is entirely possible that this is a power supply issue.
One more thing to note is that powering a hard drive from the pi itself is not recommended (even if it had been working for you). Best practice says that it should be run from a powered hub.
Got it, thanks darwindesign.
The idea now is to try to pin point when and what is causing the issue by doing one change at the time.
Kodi is not crahsing since yesterday, so I’ll try with HDD in the evening.
The testing order is the following:
- Fresh install PASSED
- Edit fstab and mount the USB HDD
- apt-get update & apt-get distr-upgrade -y
- Install UFW
- Install couchpotato
- Install sickchill
My bet is on the power supply since SD card turn out to be OK.
After the fresh install was up and running for 18 hours, I’ve connected the HDD and booted up successfully.
I was able to browse through the HDD and watched a movie without problems, so I decided to edit fstab and mount the HDD:
-
blkid to get UUID
-
sudo mkdir /mnt/clouddrive
-
Added the line UUID=E63054C530549DFD /mnt/clouddrive ntfs-3g defaults,permissions,nofail to fstab
-
sudo useradd couchpotato
sudo useradd sickchill
sudo groupadd htmc
sudo usermod -G htmc osmc
sudo usermod -G htmc www-data
sudo usermod -G htmc sickchill
sudo usermod -G htmc couchpotato
sudo chown -R www-data:htmc /mnt/clouddrive -
reboot
The issue appears again - Kodi crashes in a loop.
Here is the latest log.
So why not just use the automount and forget trying to mount in fstab?
Besides having CouchPotato and SickChill, I want to install Nextcloud with the data directory on the USB HDD.
From the guides I’ve read, my understanding is that the drive needs to be mounted in fstab in order to do that.
Also, this set up has been working up until recently…
Another update.
I’ve done the following with success:
-
blkid to get UUID
-
sudo mkdir /mnt/clouddrive
-
Added the line UUID=E63054C530549DFD /mnt/clouddrive ntfs-3g defaults,permissions,nofail to fstab
-
reboot
As soon as I do the following, the issue is back:
sudo useradd couchpotato
sudo useradd sickchill
sudo groupadd htmc
sudo usermod -G htmc osmc
sudo usermod -G htmc www-data
sudo usermod -G htmc sickchill
sudo usermod -G htmc couchpotato
sudo chown -R www-data:htmc /mnt/clouddrive
I’ve started from the beginning again but this time I’ve used Webmin for the above and there was no issue. I’m not sure if Webmin does anything different when creating users and groups, or when you change permission and ownership of a directory.
After this, I’ve managed to install SickChill, CouchPotato, Samba, Transmission and set everything up.
It’s working properly for a day now, so I’ve created two images - one using DD command and the other using Ubuntu’s create image (Disk application function).
I’ve generated SHA256 hash for both images using Win32 Disk Imager and they are different. I’m not sure why there would be difference, but I’m more inclined to trust the image created using DD.
What remains is to install Nextcloud and set it up.
If everything goes well and the pi works without problems in the next few days, I’ll consider this resolved, but I would sure like to understand what happened 
After successfully installing Nextcloud, everything is running fine for a week now.
I’ve even managed to upgrade to the latest OSMC version without problems.
At the end, I don’t understand what caused the issue, but I’ll close the topic.