No "datapath" ovs bridge found after setup a k8s cluster with fastdp enable

What you expected to happen?

What happened?

I have setup up 1 controller and 2 worker nodes cluster. All these 3 nodes are enable fastdp mode.

Worker1 is a stand server. And the Worker 2 is another server(PCIE card server) with a PCIE interface, which plugin into the PCIE slot of the Worker1. So the IP between these servers are over PCIE.

And the network topology are strange. There are datapath bridge and VxLAN port on both controller node and Worker2 nodes(PCIE card server). But the datapath bridge and VxLAN port are missing on Worker1 node(stand server).

But the Worker2 and Worker1 can communicated with each other via weave network.

So I wonder, is it normal that there is no datapath bridge and VxLAN port with fastdp mode. And in fastdp mode without datapath bridge and VxLAN port, what’s the traffic flow via weave network

How to reproduce it?

Anything else we need to know?

weaveworks/weave-npc tag:2.7.0 weaveworks/weave-kube tag:2.7.0

Versions:

$ weave: 2.7.0
$ docker: 19.03.12
$ uname -a
 Work1:  Linux  3.10.0-1.bd7848f.VCA+
 Work2:  Linux 4.19.97-1.decb955.vca+
$ kubectl 1.19.3

Logs:

$ docker logs weave

or, if using Kubernetes:

$ kubectl logs -n kube-system <weave-net-pod> weave

Network:

$ ip route
Work1: 
default via 10.240.224.1 dev enp3s0f0 proto static metric 100
10.32.0.0/12 dev weave proto kernel scope link src 10.44.0.0
10.240.224.0/24 dev enp3s0f0 proto kernel scope link src 10.240.224.149 metric 100
172.18.0.0/16 dev docker0 proto kernel scope link src 172.18.0.1
172.32.1.0/24 dev eth0 proto kernel scope link src 172.32.1.254
172.32.1.1 dev eth0 scope link

Work2: 
default via 172.32.1.254 dev eth0
10.32.0.0/12 dev weave proto kernel scope link src 10.36.0.0
172.18.0.0/16 dev docker0 proto kernel scope link src 172.18.0.1
172.32.1.0/24 dev eth0 proto kernel scope link src 172.32.1.1
172.32.1.254 dev eth0 scope link

$ ip -4 -o addr
$ sudo iptables-save

weaveworks / weave