decided to refresh my poor server that had failing disks and bring it to the next level.
– Ubuntu 18.04 LTS on an SSD as a base. (steam and plex still seem to love ubuntu)
– install SSHD and start from scratch.
– differnet harddisks instead of LVM
– replace my AMD FX with a Ryzen 5
– add a GTX1060 for video transcoding and steam
– enable steam link
– throw nextcloud on it
– different mountpoints/drives for nextcloud and plex
– an SSD for system/root
– two factor authentication
– and…of course.. it has to run minecraft server ^^

computing power is x4 now while power consumption has halved, I will probably throw some hypervisor and another SSD for that on it, too. I want to play with pfsense and SDN some more and always need a public host (next to my VPS)

It seems some kind soul on the internet found a way to remove the 2 NVENC stream limit from the non-tesla (quadro) cards

tested and approved ^^

https://github.com/keylase/nvidia-patch

before:

+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.44 Driver Version: 440.44 CUDA Version: 10.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 106... Off | 00000000:1F:00.0 Off | N/A |
| 41% 46C P2 35W / 120W | 921MiB / 3016MiB | 0% Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1442 C /usr/lib/plexmediaserver/Plex Transcoder 331MiB |
| 0 1644 C /usr/lib/plexmediaserver/Plex Transcoder 517MiB |
| 0 1964 G /usr/lib/xorg/Xorg 59MiB |
+-----------------------------------------------------------------------------+

after:

Wed Jan 15 21:16:14 2020
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 440.44 Driver Version: 440.44 CUDA Version: 10.2 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 106... Off | 00000000:1F:00.0 Off | N/A |
| 41% 48C P2 41W / 120W | 1429MiB / 3016MiB | 19% Default |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 1442 C /usr/lib/plexmediaserver/Plex Transcoder 331MiB |
| 0 1644 C /usr/lib/plexmediaserver/Plex Transcoder 517MiB |
| 0 1964 G /usr/lib/xorg/Xorg 59MiB |
| 0 2617 C /usr/lib/plexmediaserver/Plex Transcoder 357MiB |
| 0 2684 C /usr/lib/plexmediaserver/Plex Transcoder 149MiB +-----------------------------------------------------------------------------+

running 5 simultaneous 1080p transcodes on a GTX1060 now (if I want it to…total overkill as usual but hey… ;)

details:

find out what directory nvidia is installed (CACHEDEV2 for me) and add the lib directory to the dynamic linker ld.so.conf

[admin@xxxxlib]# export TERM=xterm-256color
[admin@xxxxlib]# vim /etc/ld.so.conf

add at the bottom:
/share/CACHEDEV2_DATA/.qpkg/NVIDIA_GPU_DRV/usr/lib

then run ldconfig
[admin@xxxxlib]# ldconfig

thanks, https://blog.slowb.ro/fix-terminal-capability-cm-required/

now nvidia-smi should work:

check the driver version, download and install the patch:

https://github.com/keylase/nvidia-patch

for some reason the automatic patching does not work anymore so its manual now.
https://emby.media/community/index.php?/topic/102522-tips-to-get-the-most-out-of-your-nvidia-gpu/

not in the mood right now. fuck nvidia