Hi all,

Using the OpenStack dashboard, we have launched two containers, Container-1 and Container-2, on two different hosts, Host-1 and Host-2. We attached both containers to a virtual network called VirtualNet (10.10.20.0/24). During container creation, both of them get an IP address via DHCP, as provided in the figure attached. Container-1 can successfully ping to 10.10.20.176 (Container-2 eth0 address) and Container-2 can successfully ping to 10.10.20.113 (Container-1 eth0 address).

After this, we created a dummy interface (dummy0) on Container-1 and assigned an IP address (192.168.20.113). We want to ping 192.168.20.113 from Container-2. Therefore, we added a route in Container-2 using the command "ip add route 192.168.20.113 via 10.10.20.113 dev eth0."

The ping is not successful. We captured the packets in

Container-1 using tcpdump command on eth0, and we were able to observe that the icmp echo requests are received and the icmp echo replies are generated. However, the responses are not reaching back to Container-2. The screenshot of the tcpdump on eth0 for Container-1 is attached herewith. In the security groups, we added a rule allowing all ICMP traffic (ingress and egress) to 0.0.0.0/0. Any recommendations or suggestions are greatly appreciated.

Best,
James