I have a VLAN set via MACVLAN that, under v7.0, worked just fine.

 

After upgrading to 7.1.1 and now 7.1.2, the containers on that VLAN and the host are broken in both directions: Host to Container and Container to Host.

 

I made no other changes to the network config before or after the 7.1.1 or .2 updates.

 

I have 'host access to custom networks' enabled.

knoxx-diagnostics-20250517-1748.zip

Edited May 17, 20251 yr by aglyons

I have since found two other cases of this issue popping up.

 

https://forums.unraid.net/topic/190244-712-docker-containers-cant-access-pihole-also-in-docker

 

https://forums.unraid.net/topic/190273-712-docker-no-shim-network/

 

 

And a fourth.

 

https://forums.unraid.net/topic/189927-solved-710-host-access-to-custom-networks-no-longer-works/

 

The posted 'solution' does not read as a real solution but rather a temporary patch. It is a series of steps that must be performed after each system reboot, so the 'fix' is not persistent.

Edited May 18, 20251 yr by aglyons

9 minutes ago, aglyons said:

And a fourth.

 

https://forums.unraid.net/topic/189927-solved-710-host-access-to-custom-networks-no-longer-works/

 

The posted 'solution' does not read as a real solution but rather a temporary patch. It is a series of steps that must be performed after each system reboot, so the 'fix' is not persistent.

Yeah - it's a workaround not a solution but it works for now.

 

Hoping we see a fix in .3!

Yep, can confirm this is just a workaround.  Luckily @bonienl has a script proposed for 7.1.3 that will force re-create the network from my understanding.  The issue for me seems completely related to whether a docker container using br0 is started automatically or not, because it seems to have interfered with the creation of shim-br0. br0 is created fine, but shim-br0 is not.

I'm not sure if this is the same problem, but I'll post this reply just in case it perhaps helps someone discover the solution for this particular issue. I don't use VLANs but it might be similar.

 

My Configuration

In my environment, I typically host my Docker containers on the same network (10.0.0.0/24) that all of my devices are configured on - including the Unraid server at 10.0.0.10. I created a macvlan based custom Docker network on that same network, and each Docker container has an explicitly assigned IP address in the same network.

 

My unRaid has a dual 10GB network card with the interfaces bonded with 802.3ad on two aggregated ports on a Unifi 10GB switch (interface br1)  with the 10.0.0.10 statically assigned.

 

For some reason, my Docker containers on the same network could not communicate with the unRaid host and vice versa. The final solution I found worked was to manually create a a second interface called "br1.macvlan0" on its own IP that explicitly bridged with the br1 interface. After that, the Docker containers and host (Unraid server at 10.0.0.10) could not communicate seamlessly.  

 

I installed User Scripts and added one to run the following explicitly at array start but you can have it start whenever appropriate for your situation.

 

#!/bin/bash
ip link add link br1 name br1.macvlan0 type macvlan mode bridge
ip addr add 10.0.0.2/24 dev br1.macvlan0
ip link set dev br1.macvlan0 up

 

To be fair, I'm not sure what - if any - technical network security issues this could create but I can't easily think of them as if I wanted that segregation I could explicitly create it.

 

Hope this helps those above resolve what appear to be a similar issue. If you want more explanation of those commands, I suggest throwing them into one of the GPT LLMs (ChatGPT, Gemini, Claude, etc) and ask for more detailed explanations.

Edited May 18, 20251 yr by RandoUser

On 5/18/2025 at 2:53 AM, aglyons said:

And a fourth.

 

https://forums.unraid.net/topic/189927-solved-710-host-access-to-custom-networks-no-longer-works/

 

The posted 'solution' does not read as a real solution but rather a temporary patch. It is a series of steps that must be performed after each system reboot, so the 'fix' is not persistent.

I finally decided to update to 7.1.2 from 6.12.5 today and proceeded to spend an entire evening trying fix this issue. Thank you for linking the temporary solution mentioned in that post, it immediately got things working for me again. Really hope to see this fixed asap!

My issue upon updating from 7.0.1 to 7.1.2 was that when the docker service started and the shim-br0 bridge was created I intermittently lost connectivity to most devices on my br0 LAN subnet - noticed when my UPS connection via SNMP repeatedly went offline (lots of email notifications over night).

This is with the following Docker network settings (and a couple of IPv4 custom networks)

Docker custom network type: ipvlan

Host access to custom networks: Enabled

I seem to have been able to get the network stable again by -> Unraid network settings - changing the Bonding mode from "active-backup (1)" to "balance-alb (6)"

NB. you need to stop the docker service to be able to make the change

Remains stable for me after reboots also.

Hopefully useful info for someone....

I've updated from 6. something to 7.1.3 and now have this issue too, all our npm domains located in dockers time out.

ipvlan and enable access custom network is set. I'm yet to find a solution that works!

A few users have connectivity issues with 7.1.3, if that0s your case, try Unraid OS Version 7.1.4-rc.1 available - Prereleases,

it seems to resolve the issue for all affected users.