chpasha

that would be great. just do dpkg -i *.deb right? And I need only firmware-aml, linux-dtb, linux-image from the archive or anything else? what about linux-u-boot-amlogic-s905x and linux-stretch-root-amlogic-s905x - shoud I leave them as-is at version 5.37?

no, I didn't modify anything. It doesn't look like overheating issue to me - it happens instantly and after reboot cpu temperature is very low (37-40 C). Can I ask you to upload a couple of older images of debian server (with armbian below 5.37)? I would like to try them - it really looks like software issue, especially after reading the topic with similar symptoms

I've found a topic which seems to deal with similar issue but I'm not sure I can follow the workaround steps described in that one like downgrading the kernel and/or upgrading to some nightlies because we use some custom builds, don't we?

Hello everybody, I have an Amlogic s905 Box with 1gb ram (Model MX Plus - they say it is a full clone of Tronsmart Vega95 Pro). I tried two images (installed to nand) - ubuntu server 3.14 and now debian server with same kernel (ARMBIAN 5.37 3.14.29) - unfortunately I experience the same issue with both installations, at some point of time (maybe after a couple of days) - one of processes starts to use all cpu time so the box becomes unresponsive (I can ping it, but cannot ssh-in anymore) - only reboot helps. Yesterday I left it in the such state for a whole night, in the morning it still was pingable but I could not ssh-in, even stopping some services via remote ssh command didn't work - so the system cannot recover. Both systems were relatively clean - openvpn, transmission-daemon, docker with couple of images (tvheadend, acestream) - nothing that would cause problems on libreelec (I must confess, that I run libreelec on different box, so there is still chance that it would also make problems on this one). I've got a couple of "hang-ups" even while compiling some docker image. I don't exactly know what process causes the problems since I cannot log-in anymore when it happens, so will have to setup some kind of monitoring (but like I said - there are only few of them). Did anybody experience similar problems? Any ideas how to fix or at least find the root cause of this?