I have a bunch of docker containers and I'd like to assign each one its own IP. When I assign the containers to the br0 interface using macvlan, I get call traces and after awhile the server hangs. When I use the host or bridge interface and skip br0 the call traces disappear. My motherboard has 2 integrated NICs, eth0 and eth1.

 

I was thinking maybe I could connect eth1 and create a new bridge interface, br1, and assign the containers to the new br1 bridge as a test? Does that make sense, and what is the easiest way to do this?

1 hour ago, ensnare said:

I was thinking maybe I could connect eth1 and create a new bridge interface, br1, and assign the containers to the new br1 bridge as a test?

My solution was to create a VLAN specifically for docker containers.  You might try assigning them an IP address on br1, but, for some, they also got call traces on br1.

 

Interestingly, the VLAN I created (br0.3) is on the same physical NIC as br0 yet br0.3 is call trace free whereas they occur frequently on br0.

 

Here is a thread with the long saga of discovering and resolving call traces on br0.

 

And this is the thread I followed to create the VLAN in unRAID and on my switch (your switch/router must support VLANs).

4 hours ago, Hoopster said:

My solution was to create a VLAN specifically for docker containers.  You might try assigning them an IP address on br1, but, for some, they also got call traces on br1.

 

Interestingly, the VLAN I created (br0.3) is on the same physical NIC as br0 yet br0.3 is call trace free whereas they occur frequently on br0.

 

Here is a thread with the long saga of discovering and resolving call traces on br0.

 

And this is the thread I followed to create the VLAN in unRAID and on my switch (your switch/router must support VLANs).

This was super helpful. Thanks. I encountered the same call traces on br1. I don't really want to vlan my whole network, so I just gave up and went back to host/bridge networking. What a weird bug.