Quote from: leenoox on January 21, 2018, 05:43:54 PM

Quote from: NameTaken on January 21, 2018, 05:22:34 PM

Quote from: papampi on January 21, 2018, 03:21:57 PM

Quote from: NameTaken on January 21, 2018, 02:16:09 PM

Quote from: CryptAtomeTrader44 on January 21, 2018, 01:47:55 PM

Quote from: papampi on January 21, 2018, 06:53:39 AM

root partition wont allow to go lower, So I think its the smallest we can get.
But will try to cleanup the OS again, see if I can reduse more.

Why do not you use a lighter linux system like Lubuntu or even, a linux system even more minimalist as there are at least a dozen ...

I know, I say in a vacuum and it's a lot of work but I have often thought about it. I did not have the courage or the motivation to try.

These days, I'm getting ready to make my RIG fatter because 2 GPU is good for testing, but it does not produce enough and not fast enough for my taste.

I don't notice any Sol/s difference as long as the desktop is not being actively used (all my rigs are headless anyway).

Quote from: CryptAtomeTrader44 on January 21, 2018, 01:53:55 PM

Quote from: wi$em@n on January 21, 2018, 06:59:48 AM

Quote from: papampi on January 21, 2018, 06:53:39 AM

root partition wont allow to go lower, So I think its the smallest we can get.
But will try to cleanup the OS again, see if I can reduse more.

I think the only thing left to reduce size and improve stability (maybe?) is abandon Desktop

I thought the same thing but the problem is that nvOC is an easy to use mining system. I think it won't be easy to use for lots of people if you remove gnome desktop or another desktop interface. A lighter interface desktop would be a better solution.

An Xorg instance needs to be running on each GPU (which is what nvidia-xconfig -a or --enable-all-gpus does) for nvidia-settings to change clocks and fan.

Can you share how you run nvOC in full headless mode without losing Xorg.
Any steps necessary for first run ...

May be we can add an option in 1bash for that mode.

I'm actually using a different Ubuntu based OS: https://ba.net/zcash-eth-miner-os/. I like it because it's almost stock Ubuntu with Nvidia driver preinstalled and I use Ansible-playbook for configuration management. An xorg.conf is generated every boot before X starts with this systemd service file:

Code:

[Unit]
Description=Generate xorg.conf
Before=lightdm.service

[Service]
Type=oneshot
RemainAfterExit=true
ExecStart=/usr/bin/nvidia-xconfig -a --cool-bits=28 --allow-empty-initial-configuration

[Install]
WantedBy=multi-user.target

This is not the best solution as the nvidia-xconfig doesn't fully optimize the xorg.conf. I am working on something similar to this, a new script to generate xorg.conf on every boot before starting X.
The main problem, especially with with those 19 GPU motherboards is that BusId is not correctly assigned with the static xorg.conf that we are using. On any mobo when adding/removing cards the Bus is changes... plus there are some busid numbers skipped (used for other PCIe devices) and it is different on all mobos.
The script will use lspci to find all nvidia GPU's and create new xorg.conf with proper busid values assigned and all other tweaks that nvoc uses. Hopefully this will fix the problems like booting into black screen with cursor, booting into pink screen, etc

I have 6 GPU rigs so that hasn't caused any problem although I do plan on building 8 and 12 GPU rigs in the future. That 19 GPU boards seems like it has more problems than it's worth.

https://forum.hiveos.farm/discussion/36/asus-b250-mining-expert