Public bug reported: I have a new Icehouse deployment up and running with a controller, 2 network nodes, and 4 compute nodes. I have found that I am unable to boot certain VMs (specifically Ubuntu 14.04) because the instance hangs when making a _specific_ call to the metadata service.
For testing, I brought a CirrOS instance. From there I can curl any metadata URI I want, _except_ for this: curl http://169.254.169.254/openstack/2013-10-17/meta_data.json Looking closer, I ran a tcpdump on the network node (in the neutron router's namespace): cdeng@network1:~$ sudo ip netns exec qrouter-df3f082c-b746-4fa4-9734-f4f976be3645 tcpdump -i any -nn -s0 port 80 tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on any, link-type LINUX_SLL (Linux cooked), capture size 65535 bytes 13:37:14.678641 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [S], seq 573754999, win 14600, options [mss 1460,sackOK,TS val 19504399 ecr 0,nop,wscale 3], length 0 13:37:14.678712 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [S.], seq 384294989, ack 573755000, win 28960, options [mss 1460,sackOK,TS val 126506530 ecr 19504399,nop,wscale 7], length 0 13:37:14.679191 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [.], ack 1, win 1825, options [nop,nop,TS val 19504399 ecr 126506530], length 0 13:37:14.679210 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [P.], seq 1:178, ack 1, win 1825, options [nop,nop,TS val 19504399 ecr 126506530], length 177 13:37:14.679279 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], ack 178, win 235, options [nop,nop,TS val 126506530 ecr 19504399], length 0 13:37:14.746920 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506547 ecr 19504399], length 1448 13:37:14.746981 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [P.], seq 1449:1609, ack 178, win 235, options [nop,nop,TS val 126506547 ecr 19504399], length 160 13:37:14.747439 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [.], ack 1, win 1825, options [nop,nop,TS val 19504416 ecr 126506530,nop,nop,sack 1 {1449:1609}], length 0 13:37:14.749387 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506548 ecr 19504416], length 1448 13:37:14.953392 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506599 ecr 19504416], length 1448 13:37:15.361396 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506701 ecr 19504416], length 1448 13:37:16.177396 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506905 ecr 19504416], length 1448 13:37:17.813398 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126507314 ecr 19504416], length 1448 13:37:21.085398 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126508132 ecr 19504416], length 1448 13:37:27.629418 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126509768 ecr 19504416], length 1448 13:37:40.717398 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126513040 ecr 19504416], length 1448 13:38:06.893427 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126519584 ecr 19504416], length 1448 I see the response, from the metadata proxy, being sent back to the instance (192.168.1.30). However, the instance never receives it. The result is that the requestor (in this case, curl) hangs forever. ** Affects: neutron Importance: Undecided Status: New -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to neutron. https://bugs.launchpad.net/bugs/1328623 Title: Certain requests not being routed back to an instance. Status in OpenStack Neutron (virtual network service): New Bug description: I have a new Icehouse deployment up and running with a controller, 2 network nodes, and 4 compute nodes. I have found that I am unable to boot certain VMs (specifically Ubuntu 14.04) because the instance hangs when making a _specific_ call to the metadata service. For testing, I brought a CirrOS instance. From there I can curl any metadata URI I want, _except_ for this: curl http://169.254.169.254/openstack/2013-10-17/meta_data.json Looking closer, I ran a tcpdump on the network node (in the neutron router's namespace): cdeng@network1:~$ sudo ip netns exec qrouter-df3f082c-b746-4fa4-9734-f4f976be3645 tcpdump -i any -nn -s0 port 80 tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on any, link-type LINUX_SLL (Linux cooked), capture size 65535 bytes 13:37:14.678641 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [S], seq 573754999, win 14600, options [mss 1460,sackOK,TS val 19504399 ecr 0,nop,wscale 3], length 0 13:37:14.678712 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [S.], seq 384294989, ack 573755000, win 28960, options [mss 1460,sackOK,TS val 126506530 ecr 19504399,nop,wscale 7], length 0 13:37:14.679191 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [.], ack 1, win 1825, options [nop,nop,TS val 19504399 ecr 126506530], length 0 13:37:14.679210 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [P.], seq 1:178, ack 1, win 1825, options [nop,nop,TS val 19504399 ecr 126506530], length 177 13:37:14.679279 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], ack 178, win 235, options [nop,nop,TS val 126506530 ecr 19504399], length 0 13:37:14.746920 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506547 ecr 19504399], length 1448 13:37:14.746981 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [P.], seq 1449:1609, ack 178, win 235, options [nop,nop,TS val 126506547 ecr 19504399], length 160 13:37:14.747439 IP 192.168.1.30.53737 > 169.254.169.254.80: Flags [.], ack 1, win 1825, options [nop,nop,TS val 19504416 ecr 126506530,nop,nop,sack 1 {1449:1609}], length 0 13:37:14.749387 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506548 ecr 19504416], length 1448 13:37:14.953392 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506599 ecr 19504416], length 1448 13:37:15.361396 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506701 ecr 19504416], length 1448 13:37:16.177396 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126506905 ecr 19504416], length 1448 13:37:17.813398 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126507314 ecr 19504416], length 1448 13:37:21.085398 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126508132 ecr 19504416], length 1448 13:37:27.629418 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126509768 ecr 19504416], length 1448 13:37:40.717398 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126513040 ecr 19504416], length 1448 13:38:06.893427 IP 169.254.169.254.80 > 192.168.1.30.53737: Flags [.], seq 1:1449, ack 178, win 235, options [nop,nop,TS val 126519584 ecr 19504416], length 1448 I see the response, from the metadata proxy, being sent back to the instance (192.168.1.30). However, the instance never receives it. The result is that the requestor (in this case, curl) hangs forever. To manage notifications about this bug go to: https://bugs.launchpad.net/neutron/+bug/1328623/+subscriptions -- Mailing list: https://launchpad.net/~yahoo-eng-team Post to : yahoo-eng-team@lists.launchpad.net Unsubscribe : https://launchpad.net/~yahoo-eng-team More help : https://help.launchpad.net/ListHelp