I have a very specific setup on Hive OS that requires me to use 2 NIC cards.
eth0 is configured for jumbo frames (MTU 9000) and does not require an IP address, which i’ve configured on boot with a service on systemd thats started after hive.service and needed by network-online.target which calls a script with the mtu configuration (ip link set eth0 mtu 9000), deletion of ip address (ip addr flush dev eth0) and additionally deletes the default route on it.
eth1 is the main management and internet connection which is what is set on the hiveos panel
Now the issue i’m facing is that HiveOS likes to rewrite its configuration to all NICs and its constantly overwriting my custom settings (all except for the MTU). Which is an issue since eth0 has no internet connection and is instead a dedicated link for file transfers over a bridged virtualized adapter.
This also brings routing issues since its creating a default route for both adapters for the same network that is my main (eth1) without caring that eth0 is not connected to the internet.
Now if it was connected to the same network switch this would cause an IP conflict which is not good. Hence the title.
Imagine a scenario where you have multiple NICs in case one fails and you want to keep mining, with the current behavior im experiencing this would bring both interfaces online and create that IP conflict.
Does anyone know how to stop this behavior?
Any and all help is greatly appreciated.