hrip6

| __ )  __ _ _ __   __ _ _ __   __ _  |  _ \(_) |  _ \ / _` | '_ \ / _` | '_ \ / _` | | |_) | | | |_) | (_| | | | | (_| | | | | (_| | |  __/| | |____/ \__,_|_| |_|\__,_|_| |_|\__,_| |_|   |_| Welcome to ARMBIAN 5.31 stable Debian GNU/Linux 8 (jessie) 3.4.113-sun7i System load:   0.62 0.50 0.23   Up time:       3 min Memory usage:  12 % of 984MB Up and running within about 7 minutes.  On-screen when process is running: http://sprunge.us/ELKE.
Before and after, all function working without problem.
Have a look here: http://wiki.lemaker.org/LeMaker_Guitar:Debug_COM

FYI headless legacy kernel Orange Pi Plus 2E xenial and jessie updated fine.
Thank you, developers! 

Only minor problems with cubox-i 4pro.
Because there is now a hdmi soundcard MPD soundcard has to be set new.
GOVERNOR was reset from "performance" to "interactive".
That's all!! Thanks!!

Why not, they make up for an ultra low performing NAS? Just take care that you don't get 1st gen Raspberries (so start at least with B+) since those first models did not already contain RPi under-voltage detection. Unfortunately the mythical RPi foundation started with the shitty idea to use Micro USB for DC-IN many other board makers copied later (see your Banana Pi) and since so many people failed with this, ran in instabilities and even data corruption due to the LDO regulators on the first boards they changed two things with next generation Raspberries:
 
using linear regulators so SD cards couldn't be undervolted any more which helped a lot with SD card corruption issues instead of replacing the shitty Micro USB jack with a sane barrel plug they implemented under-voltage detection circuitry which toggles a GPIO as soon as undervoltage will be detected so you get some sort of a feedback when using underpowered Raspberries (no board supported by Armbian is capable of doing so but at least with Bananas it's possible to read out under-voltage, see below)  
Those two well known problems you claim to not exist (at least not for you) are discussed in depth with focus on Raspberries for example here: http://tech.scargill.net/a-question-of-lifespan/
 
Good luck with different distros and Raspberries -- at least underpowering will be somehow compensated there.  
 
In case you want to know whether you suffer from under-voltage on a Banana Pi running mainline kernel it's as easy as: 
 

Exactly. Feel free to use another distribution (focus on software if you experience the strong feeling that it must be software related) if it's that important to reject reality.
 
I already tried to gave you some hints how to check in which specific problems you might run, I tried to explained why your methodology updating individual packages might introduce another bricked situation (dtb and kernel need to be updated at the same time), I tried to explain that all of this is useless if your problem is not software but hardware related (high IO activity triggering filesystem corruption), I even wasted my time again (since Igor already did it) to show you that 1) the network specific update problem has been fixed with 5.31 and that 2) there is NO GENERAL UPDATE problem at all.
 
You don't want to accept this and refuse to follow any of the suggestions so why wasting more time? BTW: Appreciate that you remain nice and polite and follow most basic forum rules (using english as language and trying to describe what's happening).

I just test the upgrade once again. Starting with this image: https://dl.armbian.com/bananapi/archive/Armbian_5.25_Bananapi_Debian_jessie_next_4.9.7.7z
 
apt update
apt upgrade
 
Zero problems. 
 
Logs: http://sprunge.us/QjUT

Hi,


my "Mt5572 wifi-modul wl- um01ebs- 5572- v 1. 0 27*17.7mm usb 2. 0" arrived from china.
The ra5572-chip is working on R1-lamobo with 3.4.108-kernel and backports-3.12.8-1-drivers.
Mainline kernel works out-of-the-box , but I need CSI-cam-support in 3.4.
(see http://forum.armbian.com/index.php/topic/157-add-wifi-driver-ralink-rt5572-rt2800usb/?p=3456)

During my x-mas vacation I tested the driver with an "CSL 300Mbit USB WLAN"-USB-ra5572 from Amazon
and my wifi worked very stable for 14 days till I ended the duration test to solder the new onboard-module.
So anyone who wants to test his software/drivers for the ra5572-mod could test the ra5572 with this Amazon-usb-wlan-Stick
:-)
I think the The orig/old 8192cu-Module on the R1 lamobo  will never be stable using it with a client loading big datastreams if a wifi-client has some distance to the R1-8192cu-chip.
 

 
Thanks again to badrianiulian for this great idea!
New durationtest with the soldered-onboard-ra5572 is running STABLE for 8 days now, never did with 8192cu.
regards