)]}'
{"neutron/tests/fullstack/test_l3_agent.py":[{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"a61c0c902721a67cbe1353697f800477de574172","unresolved":false,"context_lines":[{"line_number":298,"context_line":""},{"line_number":299,"context_line":"        # wait until all floatingips are up and working"},{"line_number":300,"context_line":"        for vm, fip in vm_fips.items():"},{"line_number":301,"context_line":"            vm.block_until_ping(fip[\u0027floating_ip_address\u0027])"},{"line_number":302,"context_line":""},{"line_number":303,"context_line":"        # wait a bit more for the cluster to settle"},{"line_number":304,"context_line":"        time.sleep("}],"source_content_type":"text/x-python","patch_set":1,"id":"ffe62b97_e1272e12","line":301,"updated":"2017-03-24 21:18:17.000000000","message":"for some reason, sometimes all VMs are rendered not pingable via the floating IP from inside their own namespace. In tcpdump, we see that ICMP gets into router namespace, gets replied, but reply is missing on VM side.\n\nI tried to trace it, but tcpdump doesn\u0027t work for ovs bridges and veth pairs. Maybe dumping flows could help to understand what happens.\n\nTo debug, I put the line 301 into try: except: and put breakpoint there.","commit_id":"f6ed0d2a8f4b1b459ec54fb98836e47aadb65064"},{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"b08d67a480ed349242b3724e1e9060af0a0fbfef","unresolved":false,"context_lines":[{"line_number":298,"context_line":""},{"line_number":299,"context_line":"        # wait until all floatingips are up and working"},{"line_number":300,"context_line":"        for vm, fip in vm_fips.items():"},{"line_number":301,"context_line":"            vm.block_until_ping(fip[\u0027floating_ip_address\u0027])"},{"line_number":302,"context_line":""},{"line_number":303,"context_line":"        # wait a bit more for the cluster to settle"},{"line_number":304,"context_line":"        time.sleep("}],"source_content_type":"text/x-python","patch_set":1,"id":"ffe62b97_bd76e3a3","line":301,"in_reply_to":"ffe62b97_9a17a5a4","updated":"2017-03-25 07:34:11.000000000","message":"OK I think I found a way to trigger the connectivity issue when br-tun is used consistently, see next PS.","commit_id":"f6ed0d2a8f4b1b459ec54fb98836e47aadb65064"},{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"9610f60035530306bf5db38b116a1bc6947b6c88","unresolved":false,"context_lines":[{"line_number":298,"context_line":""},{"line_number":299,"context_line":"        # wait until all floatingips are up and working"},{"line_number":300,"context_line":"        for vm, fip in vm_fips.items():"},{"line_number":301,"context_line":"            vm.block_until_ping(fip[\u0027floating_ip_address\u0027])"},{"line_number":302,"context_line":""},{"line_number":303,"context_line":"        # wait a bit more for the cluster to settle"},{"line_number":304,"context_line":"        time.sleep("}],"source_content_type":"text/x-python","patch_set":1,"id":"ffe62b97_9a17a5a4","line":301,"in_reply_to":"ffe62b97_babca9dc","updated":"2017-03-25 06:57:19.000000000","message":"And I can reproduce it for non-HA routers. Probably an issue with how we set up connectivity between fake machines and routers in fullstack environment.","commit_id":"f6ed0d2a8f4b1b459ec54fb98836e47aadb65064"},{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"b2ffe46411a287cbce04e6b49115bbf18974e454","unresolved":false,"context_lines":[{"line_number":298,"context_line":""},{"line_number":299,"context_line":"        # wait until all floatingips are up and working"},{"line_number":300,"context_line":"        for vm, fip in vm_fips.items():"},{"line_number":301,"context_line":"            vm.block_until_ping(fip[\u0027floating_ip_address\u0027])"},{"line_number":302,"context_line":""},{"line_number":303,"context_line":"        # wait a bit more for the cluster to settle"},{"line_number":304,"context_line":"        time.sleep("}],"source_content_type":"text/x-python","patch_set":1,"id":"ffe62b97_babca9dc","line":301,"in_reply_to":"ffe62b97_e1272e12","updated":"2017-03-25 06:42:03.000000000","message":"I think I see the pattern: if the master agent qr is plugged into br-int for the VM, then it works; if it has br-tun in between, then it fails to return ICMP reply back.","commit_id":"f6ed0d2a8f4b1b459ec54fb98836e47aadb65064"},{"author":{"_account_id":9656,"name":"Ihar Hrachyshka","email":"ihrachys@redhat.com","username":"ihrachys","status":"Red Hat Networking Systems Engineer"},"change_message_id":"7d3e3f4985b68d3d339b2b4cf8aab50bff9b65e0","unresolved":false,"context_lines":[{"line_number":297,"context_line":"            vm_fips[vm] \u003d fip"},{"line_number":298,"context_line":""},{"line_number":299,"context_line":"        # wait until all floatingips are up and working"},{"line_number":300,"context_line":"        for vm, fip in vm_fips.items():"},{"line_number":301,"context_line":"            vm.block_until_ping(fip[\u0027floating_ip_address\u0027])"},{"line_number":302,"context_line":""},{"line_number":303,"context_line":"        # wait a bit more for the cluster to settle"},{"line_number":304,"context_line":"        time.sleep("}],"source_content_type":"text/x-python","patch_set":2,"id":"ffe62b97_f1ed8b8b","line":301,"range":{"start_line":300,"start_character":8,"end_line":301,"end_character":59},"updated":"2017-03-27 13:33:44.000000000","message":"I was hoping that we can just replace these lines with ACTIVE status check for FIPs and adopt it into the base patch.","commit_id":"5b4c7bb27a29c63f99a792d7b5dc53a4a0e8271a"}]}
