November 8, 2025Nov 8 Hello,a few weeks ago I deactivated the LAN-Port (1GBit/s) of my server within the BIOS.On the same day I switched to an external USB-Ethernet-Adapter (2.5G). I assigned the old IP-address to the new adapter after.Now, some weeks later, my router told me there are 2 clients using the same IP-address, both of them online - the new and the previous devices. But, there is no cable connected to the old (and deactivated) LAN-port.Is there an explanation & a workaround for this ?
November 8, 2025Nov 8 Community Expert You need to identify the nic using each of those Mac address. Then assign them different IP addresses. The image has the word proxmox in it which strikes me as a little unusual for the scenario you described.
November 9, 2025Nov 9 Author My network system called Unifi identifies clients automatically. Sometimes it is non-sense.And, is has nothing to do with bridging/bonding for docker containers?And, mcvlan or ipvlan ?Looks like this … Edited November 9, 2025Nov 9 by Blebbens
November 9, 2025Nov 9 Community Expert If you can post your diagnostics, perhaps someone like jorgeb will show up and see something useful to help.
November 10, 2025Nov 10 Community Expert IIRC that's normal with a Unifi router and ipvlan, try macvlan.
November 10, 2025Nov 10 Author I tried to find that option… please tell me how and where to switch from ipvlan to macvlan.And, are there any concerns switching from one to the other ?
November 10, 2025Nov 10 I see both the vhost and hardware mac addresses on the same IP using macvlan in unifi. I don't recall having that on ipvlan but its been ages since I tried using it. It doesn't actually cause issues as far as I can tell but it is annoying.
November 11, 2025Nov 11 Community Expert Settings - Docker Settings (advanced view must be enabled)8 hours ago, _cjd_ said:I don't recall having that on ipvlan but its been ages since I tried using it.Maybe I have it the other way around.If you are using macvlan, try ipvlan.
November 11, 2025Nov 11 I live with the alert because it's much easier to manage network policy when I can see each docker at the network layer. I've just tagged the vhost as such so I know what's what and can verify the alert is for an expected issue and move on. It's been years and no observed issues.I might try bouncing to ipvlan and back on the chance it fixes eth0 being unavailable to docker after 7.2 though so may have a recent experience to report on shortly.
November 11, 2025Nov 11 Community Expert On 11/8/2025 at 3:36 PM, Blebbens said:Hello,a few weeks ago I deactivated the LAN-Port (1GBit/s) of my server within the BIOS.On the same day I switched to an external USB-Ethernet-Adapter (2.5G). I assigned the old IP-address to the new adapter after.Now, some weeks later, my router told me there are 2 clients using the same IP-address, both of them online - the new and the previous devices. But, there is no cable connected to the old (and deactivated) LAN-port.Is there an explanation & a workaround for this ?this is a known issue with how unraid maks a shim bridge and what unifi detects.I treat teh shim brdige like a servce that needs its own static ip. I made a user script to solve this issues as it unfi network spefic detecting unraids shim bridge to conect teh macvlan docker networsk on the hst and on the lan it self...see script:https://forums.unraid.net/topic/178033-bmartino1-user-scripts/#findComment-1492333
November 11, 2025Nov 11 Community Expert 1 minute ago, bmartino1 said:this is a known issue with how unraid maks a shim bridge and what unifi detects.I treat teh shim brdige like a servce that needs its own static ip. I made a user script to solve this issues as it unfi network spefic detecting unraids shim bridge to conect teh macvlan docker networsk on the hst and on the lan it self...see script:https://forums.unraid.net/topic/178033-bmartino1-user-scripts/#findComment-1492333you must use ip a and get the interface name and replace with what your unraid sees as its shim bridge in v6 i saw vhost. in v7 its shim-br0 etc etc...#!/bin/bash # Reset and configure vhost0 interface sleep 80 ip link set shim-br0 down # Bring vhost0 down ip addr flush dev shim-br0 # Remove any existing IPs #use static IP?: ip addr add 192.168.201.101/24 dev shim-br0 # Assign the desired IP #use dhcp?: #ifconfig vhost0 0.0.0.0 #dhcpcd vhost0 ip link set shim-br0 up # Bring vhost0 up
November 15, 2025Nov 15 Author Is there an advantage or disadvantage switching from ipvlan to macvlan or the other way around ?Will there be anything else I have ro change after switching ?As far as I understand, I have to switch. Using a script would be my last option.Ubiquiti is investigating this issue.
November 16, 2025Nov 16 Community Expert 13 hours ago, Blebbens said:Is there an advantage or disadvantage switching from ipvlan to macvlan or the other way around ?Will there be anything else I have ro change after switching ?As far as I understand, I have to switch. Using a script would be my last option.Ubiquiti is investigating this issue.YES! Many!!!Please watch:https://www.youtube.com/watch?v=bKFMS5C4CG0https://medium.com/@dyavanapellisujal7/docker-macvlan-and-ipvlan-explained-advanced-networking-guide-b3ba20bc22e4https://forums.unraid.net/topic/132949-macvlan-vs-ipvlan-a-discussion/ unifi and their network traffic doesn't always read the other end of the bridge and its on unraids first array start that unifi over detects 2 macs, the shim bridge and the unraid interface. This is entirely over scan on unifi networking. The big differ is in how networking is routed into the docker container. Macvlan randomly generates a mac address (unless the docker ID is destroyed that random mac stays and is viable via docker inspect commands) Ipvlan reuses the host interface's mac address that means dockers get the same mac as the host. And depending on configurations, Layer2/layer 3 routeing over the host using the same mac address.Originally unraids default shipped with mac vlan and due to changes with how linux networks and how docker made edits and changes to the docker mutated and evolved today...(there was a bug and issues during this which forced the devs to make a report and they fixed most of it and worked around the other...) Macvlan trace bugs(many still on the form...)... and how macvlan isolated from the host... where the move to iproute2 application even changed old linux networking setup due to how layer 2 and layer 3 network traffic was sent...Reviewhttps://docs.docker.com/engine/network/https://docs.docker.com/engine/network/drivers/ipvlan/https://docs.docker.com/engine/network/drivers/macvlan/There is still use case scenarios where I would want to use one over the other... Unless you need the additional lan identity such as running pihole, adguards, vpn, etc you would want that additional mac and network isolation... where if using vlans, you would want to use the ipvlan for layer 2 layer 3 routes...https://bmartino1.weebly.com/guide-dockernetworks.htmldue to ip route and same subnet It is hard to generate a dhcp pool to use both ipvaln and macvlan on the same interface... Due to how unraids generate custom br0 its a pain to configure and make stable upon reboot.teh deletion of a docker network removes the network from the dockers templates and other due to docker network id changes... pros cons... the point...Unifi network scan is overzealous when it detects unraids making the shim bridge on unraid. And depending on how unraid auto starts and spins up dockers on the custom br0 unifi false positive picks up the interface due to how the shim bridge is generated on unraid.as the nature of the shim bridge not using or generating a random mac or unraid not programing in a random mac... the command at boot, network creation duplicates the mac address and dhcp due to ipv4 forwarding sends and gets a dhcp request. unfi detects rouge dhcp server flags this as 2 of the same macs requesting a ip and say no one is a rouge system and trying to hack in...Thus the unfi warning and overzealous hit due to unraid bridge creation.as what happens is eth0 mac gets cloned and makes br0 then br0 mac gets cloned to make shim-br0 unfine doesn't know what unraid did nor see the separation with the interfaces... unifi detects and see 2 of the same mac and logs a error....Further issues with dockers on ipvlan as they can trigger and do the same...This is why I recommend using macvlan first and ipvaln if you have errors... pfsense and opensense if the settings are high enouth can detect and see this. but I've only encountered this issues and errors on unify due to how they handle clients and record previous client connections...unfi is overzealous in its rouge dhcp scan. unraid doesn't make the shim-bridge interface very well. This is why my user script to flush the shim bridge and static assign a different ip which will make the shim bridge interface generate a random mac... I've seen unifi client spam due to unraid reboots and the use of my workaround script... The workaround is more for the shutup unifi false positive dhcp rouge detection... I've talked to both unraid devs, and unifi on this before in the past... Edited November 16, 2025Nov 16 by bmartino1 spelling
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.