Eugen, Thank you very much for your response. What network type do you typically use? flat? vxlan? I don't believe that there is routing between the controller/compute node and the gateway. . .and from my testing that is where the pings are getting lost. Here is an example from the cirros instance pinging the controller/compute node (successfully), and the external gateway(not successfully): $ ping 10.61.157.59 PING 10.61.157.59 (10.61.157.59): 56 data bytes 64 bytes from 10.61.157.59: seq=0 ttl=63 time=1.378 ms 64 bytes from 10.61.157.59: seq=1 ttl=63 time=1.317 ms 64 bytes from 10.61.157.59: seq=2 ttl=63 time=0.901 ms ^C --- 10.61.157.59 ping statistics --- 3 packets transmitted, 3 packets received, 0% packet loss round-trip min/avg/max = 0.901/1.198/1.378 ms $ ping 10.61.157.1 PING 10.61.157.1 (10.61.157.1): 56 data bytes ^C --- 10.61.157.1 ping statistics --- 9 packets transmitted, 0 packets received, 100% packet loss How would I go about determining what is causing the packets to get lost at that point? From the host 10.61.157.59 itself, it CAN ping the 10.61.157.1 gateway just fine. . .and all sites externally (i.e. 8.8.8.8) here is the traceroute output from the controller/compute node, showing a successful output, and with only one hop. . .so I don't think there is anything between the host and the gateway network wise: # traceroute 10.61.157.1 traceroute to 10.61.157.1 (10.61.157.1), 30 hops max, 60 byte packets 1 _gateway (10.61.157.1) 0.374 ms 0.322 ms 0.297 ms