** No longer affects: kolla-ansible -- You received this bug notification because you are a member of Yahoo! Engineering Team, which is subscribed to neutron. https://bugs.launchpad.net/bugs/1922923
Title: [L3] arp issue in router namespace in compute node Status in neutron: New Bug description: Hello, I have an issue with Openstack Victoria. Since I moved to Openstack Victoria I still have an issue. Sometimes the VM cannot be accessed on the FIP and it gets fixed only after I clear the arp table for the private IP of the VM from the network namespace. I did some troubleshooting and I found out that the port seems to down in the OVS. 7(qr-4affa6db-67): addr:00:00:00:00:00:00 config: PORT_DOWN state: LINK_DOWN speed: 0 Mbps now, 0 Mbps max The MAC address of the VM port is this one: [root@compute-38 ~]# arp -a | grep 87 ? (10.10.13.87) at fa:16:3e:ee:d1:57 [ether] PERM on qr-4affa6db-67 And the ping it's now working. [root@compute-38 ~]# ip r 10.10.13.0/24 dev qr-4affa6db-67 proto kernel scope link src 10.10.13.1 169.254.107.94/31 dev rfp-9b2225f1-b proto kernel scope link src 169.254.107.94 169.254.110.46/31 dev rfp-9b2225f1-b proto kernel scope link src 169.254.110.46 [root@compute-38 ~]# ping 10.10.13.1 PING 10.10.13.1 (10.10.13.1) 56(84) bytes of data. 64 bytes from 10.10.13.1: icmp_seq=1 ttl=64 time=0.050 ms ^C --- 10.10.13.1 ping statistics --- 1 packets transmitted, 1 received, 0% packet loss, time 0ms rtt min/avg/max/mdev = 0.050/0.050/0.050/0.000 ms [root@compute-38 ~]# ping 10.10.13.87 PING 10.10.13.87 (10.10.13.87) 56(84) bytes of data. ^C --- 10.10.13.87 ping statistics --- 3 packets transmitted, 0 received, 100% packet loss, time 87ms [root@compute-38 ~]# The solution is to clear the ARP for 10.10.13.87 from the namespace. [root@compute-38 ~]# ping 10.10.13.87 PING 10.10.13.87 (10.10.13.87) 56(84) bytes of data. ^C --- 10.10.13.87 ping statistics --- 3 packets transmitted, 0 received, 100% packet loss, time 87ms [root@compute-38 ~]# arp -d 10.10.13.87 [root@compute-38 ~]# arp -a | grep 87 ? (10.10.13.87) at fa:16:3e:99:08:a5 [ether] on qr-4affa6db-67 [root@compute-38 ~]# ping 10.10.13.87 PING 10.10.13.87 (10.10.13.87) 56(84) bytes of data. 64 bytes from 10.10.13.87: icmp_seq=1 ttl=64 time=0.322 ms 64 bytes from 10.10.13.87: icmp_seq=2 ttl=64 time=0.239 ms ^C --- 10.10.13.87 ping statistics --- 2 packets transmitted, 2 received, 0% packet loss, time 12ms rtt min/avg/max/mdev = 0.239/0.280/0.322/0.044 ms And now the FIP started to answer: ~]# ping 10.40.131.220 PING 10.40.131.220 (10.40.131.220) 56(84) bytes of data. 64 bytes from 10.40.131.220: icmp_seq=2500 ttl=61 time=1.60 ms 64 bytes from 10.40.131.220: icmp_seq=2501 ttl=61 time=0.462 ms 64 bytes from 10.40.131.220: icmp_seq=2502 ttl=61 time=0.536 ms ^C --- 10.40.131.220 ping statistics --- 2545 packets transmitted, 46 received, 98% packet loss, time 2544013ms rtt min/avg/max/mdev = 0.305/0.571/1.606/0.272 ms And it's weird that after it's working the ports still looks down. 7(qr-4affa6db-67): addr:00:00:00:00:00:00 config: PORT_DOWN state: LINK_DOWN speed: 0 Mbps now, 0 Mbps max From what I see it seems like a bug as it gets fixed after I do this workaround. And this is happening to the brand new VMs that are being deployed, but not all of them. Do you have any idea how can I fix this issue? I updated the containers last week to the latest stable release of Victoria. Thanks! Andrei To manage notifications about this bug go to: https://bugs.launchpad.net/neutron/+bug/1922923/+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