TallMan206 Posted September 8, 2024 Posted September 8, 2024 (edited) Hi all, I use 4 NIC's for my Unraid server. 2x 10G and 2x 1G eth0 =10G (static IP, Main/GUI) eth1 =10G (Docker) eth2 =1G (nothing) eth3 =1G (nothing) Only eth0 (main/GUI) has an IP address (192.168.100.30) the rest of the NIC's are not set. I have Pi with Pihole (192.168.100.2) which is my DNS and Dhcp. My ISP router is just the gateway 192.168.100.1. Today I had issues that I never had before, with my Plex docker. I have this since a power fail 2 days ago. It could not claim the server (it dissappeared from my Plex account) , and the server was saying there was an update. It did not work. My docker is set to only to be using eth1 (the 2nd 10G NIC) . My docker advanced setting is: Subnet: 192.168.100.0/24, gateway 192.168.100.1, docker dhcp: not set. This clearly does not hand out DNS to the dockers with an own IP (like plex on 192.168.100.35) How can I make sure my Plex (which I set to use custom: eth1 (docker) to use the DNS? I'll add screenshots.of my settings. Edited September 8, 2024 by TallMan206 Quote
MAM59 Posted September 9, 2024 Posted September 9, 2024 (edited) How do you expect devices on ETH1 to send/receive data if ETH1 has not a valid address??? Why dont you put the Dockers bridged on ETH0 ? (and yes, DNS settings are on ETH0 only because they run above routing and therefore are global to all NICs. This is just a GUI thing, normally they would be in a seperate section for themselfs) Btw: you cannot set DNS on ETH0 to your pihole! This results in a chicken&egg problem at boot time. DNS is needed already, but Dockers are started last. So there is a lot of waiting, timeout, retries... and the outcome is not defined. Same happens if you stop the array. You can use Pihole for all devices on the LAN, but NOT FOR UNRAID! (set to 8.8.8.8 or something instead!) Edited September 9, 2024 by MAM59 Quote
TallMan206 Posted September 9, 2024 Author Posted September 9, 2024 59 minutes ago, MAM59 said: How do you expect devices on ETH1 to send/receive data if ETH1 has not a valid address??? Why dont you put the Dockers bridged on ETH0 ? (and yes, DNS settings are on ETH0 only because they run above routing and therefore are global to all NICs. This is just a GUI thing, normally they would be in a seperate section for themselfs) Btw: you cannot set DNS on ETH0 to your pihole! This results in a chicken&egg problem at boot time. DNS is needed already, but Dockers are started last. So there is a lot of waiting, timeout, retries... and the outcome is not defined. Same happens if you stop the array. You can use Pihole for all devices on the LAN, but NOT FOR UNRAID! (set to 8.8.8.8 or something instead!) Hi MAM59, PiHole is NOT running on the UnRAID box (PiHole is running on a Raspberry Pi) so UnRAID should not have a problem while booting to find the DNS. Quote
TallMan206 Posted September 9, 2024 Author Posted September 9, 2024 I've done it completely different. I made 802.3AD bonding for the 10G NIC's and the 1G NIC's. I anyway, need to set --dns 192.168.100.2 in docker containers that use bond0 with an own IP (Plex and Roonserver have an own IP) so in extra parameters I just put --dns 192.168.100.2 and it works perfectly fine. I hoped to solve this on NIC level, but that is not working. Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.