On Windows platform, we run containers/VMs on host, and use OVS tunnel to support the container/VMs connections inter two Nodes. We use "ping" command to test the connectivity between the containers/VMs accross the Windows hosts. Then we find several packets are lost. After some capturing the packets, we find that the first packet which expected to be transferred over the tunnel is actually dropped, if the tunnel peer' MAC is not existing on the Windows host's ARP cache.
With tests, we got some output like this,
Pinging 192.168.2.5 with 32 bytes of data:
Request timed out.
Reply from 192.168.2.5: bytes=32 time<1ms TTL=128
Ping statistics for 192.168.2.5:
Packets: Sent = 2, Received = 1, Lost = 1 (50% loss),
To reproduce it, the steps include,
create tunnel port on OVS on each Windows host (there should be 2 Windows hosts at least in the setup)
run one container per Windows host
clear ARP cache on each Windows host
login to the one container, and ping peer container
On Windows platform, we run containers/VMs on host, and use OVS tunnel to support the container/VMs connections inter two Nodes. We use "ping" command to test the connectivity between the containers/VMs accross the Windows hosts. Then we find several packets are lost. After some capturing the packets, we find that the first packet which expected to be transferred over the tunnel is actually dropped, if the tunnel peer' MAC is not existing on the Windows host's ARP cache.
With tests, we got some output like this,
To reproduce it, the steps include,