)]}'
{"neutron/agent/rpc.py":[{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"faf7d11fbdd38d103f21733fc7a98a9fea94c41b","unresolved":false,"context_lines":[{"line_number":112,"context_line":"        1.4 - tunnel_sync rpc signature upgrade to obtain \u0027host\u0027"},{"line_number":113,"context_line":"        1.5 - Support update_device_list and"},{"line_number":114,"context_line":"              get_devices_details_list_and_failed_devices"},{"line_number":115,"context_line":"        1.6 - Add param agent_restarted to update_device_list"},{"line_number":116,"context_line":"    \u0027\u0027\u0027"},{"line_number":117,"context_line":""},{"line_number":118,"context_line":"    def __init__(self, topic):"}],"source_content_type":"text/x-python","patch_set":2,"id":"9fdfeff1_3e420cad","line":115,"range":{"start_line":115,"start_character":8,"end_line":115,"end_character":61},"updated":"2019-03-05 07:06:59.000000000","message":"One question for this: since the \u0027update_device_list\u0027 now accept varying K-V args. In order to make this backportable to stable branch, IMO, I do not want to increase the RPC verion here. What do you think?","commit_id":"9a75d4cbf2edc7f609ffccbe120cb3b1555048f3"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"7cf2b35814e8ef06e7f5d60d7b4b20828f86b077","unresolved":false,"context_lines":[{"line_number":112,"context_line":"        1.4 - tunnel_sync rpc signature upgrade to obtain \u0027host\u0027"},{"line_number":113,"context_line":"        1.5 - Support update_device_list and"},{"line_number":114,"context_line":"              get_devices_details_list_and_failed_devices"},{"line_number":115,"context_line":"        1.6 - Add param agent_restarted to update_device_list"},{"line_number":116,"context_line":"    \u0027\u0027\u0027"},{"line_number":117,"context_line":""},{"line_number":118,"context_line":"    def __init__(self, topic):"}],"source_content_type":"text/x-python","patch_set":2,"id":"5fc1f717_8250a834","line":115,"range":{"start_line":115,"start_character":8,"end_line":115,"end_character":61},"in_reply_to":"9fdfeff1_3e420cad","updated":"2019-03-05 23:39:08.000000000","message":"Thanks for the feedbacks, this should be backported to all the stable branches.","commit_id":"9a75d4cbf2edc7f609ffccbe120cb3b1555048f3"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"67702b0bab1362822087d6a8c09f9eb449e46393","unresolved":false,"context_lines":[{"line_number":159,"context_line":"        return cctxt.call(context, \u0027update_device_list\u0027,"},{"line_number":160,"context_line":"                          devices_up\u003ddevices_up, devices_down\u003ddevices_down,"},{"line_number":161,"context_line":"                          agent_id\u003dagent_id, host\u003dhost,"},{"line_number":162,"context_line":"                          agent_restarted\u003dagent_restarted)"},{"line_number":163,"context_line":""},{"line_number":164,"context_line":"    def tunnel_sync(self, context, tunnel_ip, tunnel_type\u003dNone, host\u003dNone):"},{"line_number":165,"context_line":"        cctxt \u003d self.client.prepare(version\u003d\u00271.4\u0027)"}],"source_content_type":"text/x-python","patch_set":4,"id":"9fdfeff1_7ffc04b5","line":162,"updated":"2019-03-05 16:43:41.000000000","message":"I *think* we can do this and be backwards-compat, see 67c4c6d809e4c9e112d9fb848b5bdce9d5cd04ac for some info on catching unsupported version errors.  But then I guess the problem will be that if the agent doesn\u0027t send this info, the agent is now marked as \"not restarted\", so the time-based code might need to be retained as a backup?  I haven\u0027t thought through it completely.","commit_id":"a8053802250fa3cdeb76646db87f623284e61295"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"ef9d8ec3c00d22d33ac997276baa5fe2ecc76b42","unresolved":false,"context_lines":[{"line_number":159,"context_line":"        return cctxt.call(context, \u0027update_device_list\u0027,"},{"line_number":160,"context_line":"                          devices_up\u003ddevices_up, devices_down\u003ddevices_down,"},{"line_number":161,"context_line":"                          agent_id\u003dagent_id, host\u003dhost,"},{"line_number":162,"context_line":"                          agent_restarted\u003dagent_restarted)"},{"line_number":163,"context_line":""},{"line_number":164,"context_line":"    def tunnel_sync(self, context, tunnel_ip, tunnel_type\u003dNone, host\u003dNone):"},{"line_number":165,"context_line":"        cctxt \u003d self.client.prepare(version\u003d\u00271.4\u0027)"}],"source_content_type":"text/x-python","patch_set":4,"id":"5fc1f717_96cb0d1d","line":162,"in_reply_to":"9fdfeff1_7ffc04b5","updated":"2019-03-06 00:20:04.000000000","message":"Done","commit_id":"a8053802250fa3cdeb76646db87f623284e61295"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"cec93c783705f0f3cdd992fce2f229fdfcd2869a","unresolved":false,"context_lines":[{"line_number":153,"context_line":"                          agent_id\u003dagent_id, host\u003dhost)"},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"    def update_device_list(self, context, devices_up, devices_down,"},{"line_number":156,"context_line":"                           agent_id, host, agent_restarted\u003dNone):"},{"line_number":157,"context_line":"        cctxt \u003d self.client.prepare(version\u003d\u00271.5\u0027)"},{"line_number":158,"context_line":"        kwargs \u003d {\"devices_up\": devices_up,"},{"line_number":159,"context_line":"                  \"devices_down\": devices_down,"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_a27148aa","line":156,"range":{"start_line":156,"start_character":59,"end_line":156,"end_character":63},"updated":"2019-03-17 08:17:02.000000000","message":"Temporarily backward compatibility, refactor back to bool:\nhttps://review.openstack.org/#/c/641866/","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"4338b66995f6f8c055cb04671baf6cf66d431f3f","unresolved":false,"context_lines":[{"line_number":153,"context_line":"                          agent_id\u003dagent_id, host\u003dhost)"},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"    def update_device_list(self, context, devices_up, devices_down,"},{"line_number":156,"context_line":"                           agent_id, host, agent_restarted\u003dNone):"},{"line_number":157,"context_line":"        cctxt \u003d self.client.prepare(version\u003d\u00271.5\u0027)"},{"line_number":158,"context_line":"        kwargs \u003d {\"devices_up\": devices_up,"},{"line_number":159,"context_line":"                  \"devices_down\": devices_down,"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_421f3f07","line":156,"range":{"start_line":156,"start_character":59,"end_line":156,"end_character":63},"in_reply_to":"5fc1f717_42397fd4","updated":"2019-03-18 04:46:37.000000000","message":"Done","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"8366f5178fb82d5188c09a9f5aad9bc4ed85ebba","unresolved":false,"context_lines":[{"line_number":153,"context_line":"                          agent_id\u003dagent_id, host\u003dhost)"},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"    def update_device_list(self, context, devices_up, devices_down,"},{"line_number":156,"context_line":"                           agent_id, host, agent_restarted\u003dNone):"},{"line_number":157,"context_line":"        cctxt \u003d self.client.prepare(version\u003d\u00271.5\u0027)"},{"line_number":158,"context_line":"        kwargs \u003d {\"devices_up\": devices_up,"},{"line_number":159,"context_line":"                  \"devices_down\": devices_down,"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_42397fd4","line":156,"range":{"start_line":156,"start_character":59,"end_line":156,"end_character":63},"in_reply_to":"5fc1f717_a27148aa","updated":"2019-03-18 03:46:52.000000000","message":"Wait this is the agent side should be safe to always set to agent_restarted\u003dFalse otherwise by default None means the old logic. In this case your new agent code will also send None which will trigger the old logic in some cases.","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"}],"neutron/conf/plugins/ml2/drivers/l2pop.py":[{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"7979e494c8ba9cc523c23a92bd03c7b28291596f","unresolved":false,"context_lines":[{"line_number":19,"context_line":""},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":24,"context_line":"                      \u0027existing ports when it restarts\u0027)),"},{"line_number":25,"context_line":"]"}],"source_content_type":"text/x-python","patch_set":4,"id":"9fdfeff1_5fc6e06a","side":"PARENT","line":22,"updated":"2019-03-05 16:47:29.000000000","message":"one less config nice work","commit_id":"0755c97a2d26a6bb6fcd5c71215ade3f6b184d5a"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"67702b0bab1362822087d6a8c09f9eb449e46393","unresolved":false,"context_lines":[{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":24,"context_line":"                      \u0027existing ports when it restarts\u0027)),"},{"line_number":25,"context_line":"]"},{"line_number":26,"context_line":""},{"line_number":27,"context_line":""}],"source_content_type":"text/x-python","patch_set":4,"id":"9fdfeff1_dcb41e85","side":"PARENT","line":24,"updated":"2019-03-05 16:43:41.000000000","message":"I think to be safe we should mark this deprecated (deprecated_for_removal\u003dTrue) and change the message to say it\u0027s been removed, but then I only see it being used in the neutron repo and no others.  It at least needs a release note in the \"upgrade\" (or update?) section mentioning the value is gone.","commit_id":"0755c97a2d26a6bb6fcd5c71215ade3f6b184d5a"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"ef9d8ec3c00d22d33ac997276baa5fe2ecc76b42","unresolved":false,"context_lines":[{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":24,"context_line":"                      \u0027existing ports when it restarts\u0027)),"},{"line_number":25,"context_line":"]"},{"line_number":26,"context_line":""},{"line_number":27,"context_line":""}],"source_content_type":"text/x-python","patch_set":4,"id":"5fc1f717_36d019ca","side":"PARENT","line":24,"in_reply_to":"5fc1f717_d68d157c","updated":"2019-03-06 00:20:04.000000000","message":"Done","commit_id":"0755c97a2d26a6bb6fcd5c71215ade3f6b184d5a"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"b75c4d3e4d1b17a5c9622bbe64ab87275a2ac4b4","unresolved":false,"context_lines":[{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":24,"context_line":"                      \u0027existing ports when it restarts\u0027)),"},{"line_number":25,"context_line":"]"},{"line_number":26,"context_line":""},{"line_number":27,"context_line":""}],"source_content_type":"text/x-python","patch_set":4,"id":"5fc1f717_d68d157c","side":"PARENT","line":24,"in_reply_to":"9fdfeff1_dcb41e85","updated":"2019-03-05 23:57:46.000000000","message":"OK, I will add a note for this. IMO, let\u0027s directly remove this. Do not let the users to waste time on this. And this will be backport to all the stable branches, IMO, let\u0027s make it easy to backport.","commit_id":"0755c97a2d26a6bb6fcd5c71215ade3f6b184d5a"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"d29b30c023e8d7412371f6699086b3e63a838819","unresolved":false,"context_lines":[{"line_number":23,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":24,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":25,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":26,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":27,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":28,"context_line":"                      \u0027release.\u0027)),"},{"line_number":29,"context_line":"]"},{"line_number":30,"context_line":""}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_ee727b49","line":27,"range":{"start_line":26,"start_character":58,"end_line":27,"end_character":37},"updated":"2019-03-08 02:16:33.000000000","message":"s/directly reporting restart state on initial RPC message ?","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"926f549e7a485a19280d18fc069e46f10e745190","unresolved":false,"context_lines":[{"line_number":23,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":24,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":25,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":26,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":27,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":28,"context_line":"                      \u0027release.\u0027)),"},{"line_number":29,"context_line":"]"},{"line_number":30,"context_line":""}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_ee027b0b","line":27,"range":{"start_line":26,"start_character":58,"end_line":27,"end_character":37},"in_reply_to":"5fc1f717_ee727b49","updated":"2019-03-08 03:21:01.000000000","message":"\u0027initial RPC\u0027 looks not so much proper.","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"54e7b16c26b2b11006251cecfbe7b8d3d34a477c","unresolved":false,"context_lines":[{"line_number":24,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":25,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":26,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":27,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":28,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":29,"context_line":"                      \u0027release.\u0027)),"},{"line_number":30,"context_line":"]"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_926ad9b0","line":27,"range":{"start_line":27,"start_character":49,"end_line":27,"end_character":57},"updated":"2019-03-11 09:04:51.000000000","message":"nit: direct","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":24,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":25,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"},{"line_number":26,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":27,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":28,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":29,"context_line":"                      \u0027release.\u0027)),"},{"line_number":30,"context_line":"]"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_88c6269a","line":27,"range":{"start_line":27,"start_character":49,"end_line":27,"end_character":57},"in_reply_to":"5fc1f717_926ad9b0","updated":"2019-03-11 10:37:10.000000000","message":"Done","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"cab2bec0cc92bc821c77f902e7f9309077e94197","unresolved":false,"context_lines":[{"line_number":26,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":27,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":28,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":29,"context_line":"                      \u0027release.\u0027)),"},{"line_number":30,"context_line":"]"},{"line_number":31,"context_line":""},{"line_number":32,"context_line":""}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_f2b9dde8","line":29,"updated":"2019-03-11 08:55:50.000000000","message":"Please add info in which release it is deprecated. That will help us later to know when we should/can remove it :)\nYou can add \"deprecated_since\" argument for that, like e.g.: https://github.com/openstack/neutron/blob/f9a80d30ef69221961ca4ad7a8455a7e04b5e04c/neutron/cmd/sanity_check.py#L321","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":26,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":27,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":28,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":29,"context_line":"                      \u0027release.\u0027)),"},{"line_number":30,"context_line":"]"},{"line_number":31,"context_line":""},{"line_number":32,"context_line":""}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_48c03e7c","line":29,"in_reply_to":"5fc1f717_52779103","updated":"2019-03-11 10:37:10.000000000","message":"Done","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"54e7b16c26b2b11006251cecfbe7b8d3d34a477c","unresolved":false,"context_lines":[{"line_number":26,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":27,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":28,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":29,"context_line":"                      \u0027release.\u0027)),"},{"line_number":30,"context_line":"]"},{"line_number":31,"context_line":""},{"line_number":32,"context_line":""}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_52779103","line":29,"in_reply_to":"5fc1f717_f2b9dde8","updated":"2019-03-11 09:04:51.000000000","message":"+1 \nI think this is the relevant doc for deprecation policy (at least I found this one):\nhttps://governance.openstack.org/tc/reference/tags/assert_follows-standard-deprecation.html","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":26,"context_line":"                      \u0027existing ports when it restarts. This option \u0027"},{"line_number":27,"context_line":"                      \u0027is deprecated in favor of directly RPC restart \u0027"},{"line_number":28,"context_line":"                      \u0027state transfer and will be removed in a future \u0027"},{"line_number":29,"context_line":"                      \u0027release.\u0027)),"},{"line_number":30,"context_line":"]"},{"line_number":31,"context_line":""},{"line_number":32,"context_line":""}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_a8c1a282","line":29,"in_reply_to":"5fc1f717_f2b9dde8","updated":"2019-03-11 10:37:10.000000000","message":"Done","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"d80d89092e7a6d9c0d01d845c9b9376de3cc61eb","unresolved":false,"context_lines":[{"line_number":20,"context_line":""},{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               max\u003d180,"},{"line_number":24,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":25,"context_line":"               deprecated_since\u003d\u0027Stein\u0027,"},{"line_number":26,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"}],"source_content_type":"text/x-python","patch_set":11,"id":"5fc1f717_05ac4b32","line":23,"updated":"2019-03-12 07:54:46.000000000","message":"So, if You want to remove it, why setting max value?","commit_id":"964e526829644098c6e60e60cedfedc7f8377ac8"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"55e0e99110bc3f69dfdc124e19a10da84041b7dd","unresolved":false,"context_lines":[{"line_number":20,"context_line":""},{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               max\u003d180,"},{"line_number":24,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":25,"context_line":"               deprecated_since\u003d\u0027Stein\u0027,"},{"line_number":26,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"}],"source_content_type":"text/x-python","patch_set":11,"id":"5fc1f717_2817c053","line":23,"in_reply_to":"5fc1f717_05ac4b32","updated":"2019-03-13 11:01:58.000000000","message":"For backport purpose, because some user may already set this to a very very large value. We do not want this to take effect during this period of agent normal running.\nhttps://review.openstack.org/#/c/640797/11/neutron/plugins/ml2/drivers/l2pop/mech_driver.py@305\n\n\nYou can also see the context here:\nhttps://review.openstack.org/#/c/640797/8/neutron/plugins/ml2/drivers/l2pop/mech_driver.py@304\nor the release note for more detail.","commit_id":"964e526829644098c6e60e60cedfedc7f8377ac8"},{"author":{"_account_id":5948,"name":"Oleg Bondarev","email":"obondarev@mirantis.com","username":"obondarev"},"change_message_id":"087589f54ef95fb7213aac238bc7e8e882a50568","unresolved":false,"context_lines":[{"line_number":20,"context_line":""},{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               max\u003d180,"},{"line_number":24,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":25,"context_line":"               deprecated_since\u003d\u0027Stein\u0027,"},{"line_number":26,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"}],"source_content_type":"text/x-python","patch_set":11,"id":"5fc1f717_c104d61a","line":23,"in_reply_to":"5fc1f717_2817c053","updated":"2019-03-13 14:21:31.000000000","message":"this will cause neutron server restart failure if value in config is greater than 180 - kind of a rude way to force operators to change value. Consider lowering value in code with a warning to logs.","commit_id":"964e526829644098c6e60e60cedfedc7f8377ac8"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"f97da920a7e90efdf43f6bd835279e76c37629ed","unresolved":false,"context_lines":[{"line_number":20,"context_line":""},{"line_number":21,"context_line":"l2_population_options \u003d ["},{"line_number":22,"context_line":"    cfg.IntOpt(\u0027agent_boot_time\u0027, default\u003d180,"},{"line_number":23,"context_line":"               max\u003d180,"},{"line_number":24,"context_line":"               deprecated_for_removal\u003dTrue,"},{"line_number":25,"context_line":"               deprecated_since\u003d\u0027Stein\u0027,"},{"line_number":26,"context_line":"               help\u003d_(\u0027Delay within which agent is expected to update \u0027"}],"source_content_type":"text/x-python","patch_set":11,"id":"5fc1f717_c7a9467c","line":23,"in_reply_to":"5fc1f717_c104d61a","updated":"2019-03-13 15:17:30.000000000","message":"nice suggest, Done","commit_id":"964e526829644098c6e60e60cedfedc7f8377ac8"}],"neutron/plugins/ml2/drivers/l2pop/mech_driver.py":[{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"6fe1a646867b7c76d1f2bf34858c68958a1762e4","unresolved":false,"context_lines":[{"line_number":289,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":290,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":291,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":292,"context_line":"        if is_first_port or agent_restarted:"},{"line_number":293,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":294,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":295,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":7,"id":"5fc1f717_5c6352dd","line":292,"updated":"2019-03-07 20:01:23.000000000","message":"So this raises the question of how we will backport this, as if an agent doesn\u0027t report it restarted (because it\u0027s running old code) we will return here every time.","commit_id":"3e9ea13e265601fcd918a3e1b8e7d9d24b8021ed"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"3a5c4bf92fd2e7fa3538508a7506e56557f76409","unresolved":false,"context_lines":[{"line_number":289,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":290,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":291,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":292,"context_line":"        if is_first_port or agent_restarted:"},{"line_number":293,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":294,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":295,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":7,"id":"5fc1f717_b9d6ff26","line":292,"in_reply_to":"5fc1f717_05280ae0","updated":"2019-03-08 03:23:29.000000000","message":"Will there be such case as the agent is always local to neutron server no? I mean upgrade should be done to both.\n\nOne way would be to introduce a new config flag but I\u0027m not a big fan of yet another config. \n\nIdeal solution I could think of would be to extract the RPC version of the agent through the rpc context. The code would then be something like:\n\nagent_restarted \u003d agent_restared if rpc.version \u003e Rocky else (agent_boot_time logic for backward compatibility)","commit_id":"3e9ea13e265601fcd918a3e1b8e7d9d24b8021ed"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"3c1485d04b0a9c9ca1504b27596961c4834ba2b2","unresolved":false,"context_lines":[{"line_number":289,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":290,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":291,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":292,"context_line":"        if is_first_port or agent_restarted:"},{"line_number":293,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":294,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":295,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":7,"id":"5fc1f717_05280ae0","line":292,"in_reply_to":"5fc1f717_5c6352dd","updated":"2019-03-07 23:24:46.000000000","message":"Hmm, good question, our operators once said they sometimes met such situation that some ovs-agent were not updated successfully due to some repo(yum) issue. OK, that \u0027agent_boot_time\u0027 can not be removed directly here for this situation, we can add a \u0027or\u0027 condition here.","commit_id":"3e9ea13e265601fcd918a3e1b8e7d9d24b8021ed"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"ab516c9a732a2bdf98e80c7ed91519e7fc0eadfb","unresolved":false,"context_lines":[{"line_number":289,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":290,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":291,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":292,"context_line":"        if is_first_port or agent_restarted:"},{"line_number":293,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":294,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":295,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":7,"id":"5fc1f717_94050ac5","line":292,"in_reply_to":"5fc1f717_b9d6ff26","updated":"2019-03-08 04:41:04.000000000","message":"See my comments in patch set 8.","commit_id":"3e9ea13e265601fcd918a3e1b8e7d9d24b8021ed"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"59c6fde231ffb96825d1d3d94f74086a09f71fdb","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_b9927f4a","line":304,"updated":"2019-03-08 03:33:38.000000000","message":"I don\u0027t like this solution what if someone puts a really big agent_restarted boot time value, you could end up where agent_restarted from rpc completed but the agent_restarted from boot time not completed due to a misconfiguration.\n\nEither like I proposed earlier using the rpc version to check or somehow is it possible to make agent_restarted rpc flag as optional value so we have not only true or false but the following states (None, true, false)","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"d03cbaed2d51cc9d2a2c00e85f059aef31116c3f","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_98551d20","line":304,"in_reply_to":"5fc1f717_3890a943","updated":"2019-03-11 05:11:55.000000000","message":"What\u0027s the different? None as a condition of if-else, it\u0027s equal to False. The self.agent_restarted will be called anyway if it is None or False.\n\nAgent did not send `agent_restarted` which does not always mean it is old code. Agent side rpc_loop can call this RPC when setup port in a new code version.","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"81f2fe56018f95500a2efd7a60fd63a52e0bfa49","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_94d3ca38","line":304,"in_reply_to":"5fc1f717_74cf3eb0","updated":"2019-03-08 04:47:45.000000000","message":"OK, it is your result.... So, what\u0027s 1800s happen for a no pressure agent?","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"b0fdb368a3d33cd795463d0802b4e5f4b333e037","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_f4a7aea6","line":304,"in_reply_to":"5fc1f717_94d3ca38","updated":"2019-03-08 05:07:08.000000000","message":"Let\u0027s prevent your too robust config, a max value was added:\nhttps://review.openstack.org/#/c/640797/9/neutron/conf/plugins/ml2/drivers/l2pop.py@23\nhttps://review.openstack.org/#/c/640797/9/releasenotes/notes/precise-agent-state-transfer-67c771cb1ee04dd0.yaml@7","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"7a38d2f207f40c9333f443d7e26493460d341094","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_b3d2935b","line":304,"in_reply_to":"5fc1f717_94d3ca38","updated":"2019-03-08 16:07:52.000000000","message":"yes let me test it first, I remember when setting a too big number I had some side effects but I need to double check. The 1800s was perfect fit for our environment but may not work for others.\n\nMy concern is we made a fix with the RPC but mixing that with the old logic. Meaning if the agent_boot_time is big enough in an old config that only got code upgrade then the RPC state fix is not even applicable.","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"a62ab99958d12abc814c18b9aa6f0ebe8b9e0270","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_60b2e596","line":304,"in_reply_to":"5fc1f717_98551d20","updated":"2019-03-11 22:08:18.000000000","message":"My point is u handle None in specific way not by just checking if else. Like the example I mentioned you would rewrite the code as:\n\nNone \u003d\u003e old agent code\nTrue/False \u003d\u003e new agent code\n\nthe refactoring of the above code would be explicit check for None:\n\nif agent_restarted \u003d\u003d None:\n    #deprecated to be removed in the future\n    agent_restarted \u003d self.agent_restarted(context)\n\nif is_first_port or agent_restarted:\n    ......\n\n\nBut that would also mean you need to make the new agent code send False by default otherwise as you mentioned it would not be able to differentiate between new and old.\n\nIf it requires too much change on the agent, I\u0027m fine with the current change as well since I didn\u0027t notice any side effects of one overriding the others logic given a high agent_boot_time.","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"7e318f36af135137a3c95644bc5549c952e1dbb7","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_3890a943","line":304,"in_reply_to":"5fc1f717_b3d2935b","updated":"2019-03-11 03:31:56.000000000","message":"Yulong, I tested with a really big values for agent_boot_time and it seems to cause no side effect besides recreating the flows many times. Still I\u0027m not sure the initial motivation for this flag beside avoiding re-creating existing flows. But could be some concurrency issue we are not able to repro at this point.\n\nWhat do I suggest to be on the safe side and ensure that the fix you added cause no side effects with other logic would be to not sent agent_restarted to False by default. and here I would replace the logic as follow:\n\nrpc.py:\ninstead of agent_restarted \u003d kwargs.pop(\u0027agent_restarted\u0027, False)\nleave it as agent_restarted \u003d kwargs.pop(\u0027agent_restarted\u0027)\n\nmech_driver.py:\nif agent_restarted \u003d\u003d None:\n    agent_restarted \u003d self.agent_restarted(context)\n\nIn my proposed changed, if agent_restarted is not passed by the agent it means it\u0027s still the old code.","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"ab516c9a732a2bdf98e80c7ed91519e7fc0eadfb","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":8,"id":"5fc1f717_74cf3eb0","line":304,"in_reply_to":"5fc1f717_b9927f4a","updated":"2019-03-08 04:41:04.000000000","message":"1. rpc version upgrade is not acceptable for backport.\n2. self.agent_restarted(context) is for downward compatibility.\n3. if some neutron-servers does not upgrade successfully, they are the original code.\n4. if one upgraded agent send rpc to neutron server which is not updated, then still the original behavior.\n5. this is ONLY for an exception when there are inconsistent version in a sigle neutron deployment.\n6. Cloud operators should take responsibility for the config setting.\n7. IIRC, if this notification was triggered again, no bad thing happen in the agent side. We can re-send it. In this comment, they set it to 1800s\nhttps://bugs.launchpad.net/neutron/+bug/1794991/comments/41\nand they just say still missing flows, but no misconfiguration flows. Or maybe you can test this, and then give us a result.\n\nMake sense?","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"54e7b16c26b2b11006251cecfbe7b8d3d34a477c","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_3218c5d1","line":304,"range":{"start_line":304,"start_character":28,"end_line":304,"end_character":43},"updated":"2019-03-11 09:04:51.000000000","message":"Would be good to see a short note about what is the difference between in this context the agent_restarted arg and the result of self.agent_restarted call","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_e8b9eae8","line":304,"range":{"start_line":304,"start_character":28,"end_line":304,"end_character":43},"in_reply_to":"5fc1f717_3218c5d1","updated":"2019-03-11 10:37:10.000000000","message":"We have release note, commit message. It is simple, and will be removed soon, really need to make this verbose? I don\u0027t think so. : )","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"a94d32d756454dd8b49177efec7b1aa15215e5f6","unresolved":false,"context_lines":[{"line_number":301,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":302,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":303,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":304,"context_line":"        if is_first_port or agent_restarted or self.agent_restarted(context):"},{"line_number":305,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":306,"context_line":"            # we have to provide it with the whole list of fdb entries"},{"line_number":307,"context_line":"            agent_fdb_entries \u003d self._create_agent_fdb(port_context,"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_412fb9f5","line":304,"range":{"start_line":304,"start_character":28,"end_line":304,"end_character":43},"in_reply_to":"5fc1f717_e8b9eae8","updated":"2019-03-11 13:15:27.000000000","message":"My concern is that in the same condition there is agent_restarted (which is an argument) and agent_restarted (which is a method) and I suppose the single reader, like me will be confused which is which. This code will not run on an eniac like monster with limited memory capacity, and this will just help developers to understand the code when they need to touch it. \nBut I can live without it :-)","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":13995,"name":"Nate Johnston","email":"nate.johnston@redhat.com","username":"natejohnston"},"change_message_id":"a566f892f71e2bef30aaa590a01c6e47fa0ae252","unresolved":false,"context_lines":[{"line_number":248,"context_line":""},{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_45e3d7a4","line":251,"updated":"2019-03-15 21:36:25.000000000","message":"Bruca: I definitely understand where you are coming from, but do you think this risk is mitigated with release notes?  This could/should also be a neutron-upgrade check - check if there is a value set that is not 180, and warn the operator pre-upgrade.","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"5149148985fd4f8c42cf4c6077611df306df2008","unresolved":false,"context_lines":[{"line_number":248,"context_line":""},{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_ced94857","line":251,"updated":"2019-03-14 14:23:11.000000000","message":"This is where I\u0027m not a big fan of multiple logic. My philosophy is more going to the explicit way. I\u0027m still fine with this logic for the sake that it might be an edge case. But if we want to cover all scenarios, this will break existing logic.\n\nFor instance, one upgraded neutron-server but not neutron-agent. One updated agent_boot_time to very high number since that was the fix he did long time ago for the reboot issue. Now with this override basically the old neutron-agent will break. Thus it\u0027ll result in lots of investigation time for the operator.\n\nTo summarize, I\u0027m more a fan of If new agent code detected then use latest logic. Just my 2 cents.","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"51a29f7059a5956d9b8805257e379958a89b197b","unresolved":false,"context_lines":[{"line_number":248,"context_line":""},{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_e27bd088","line":251,"in_reply_to":"5fc1f717_0234dcdc","updated":"2019-03-17 08:13:57.000000000","message":"[1] https://review.openstack.org/#/c/641866/","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"daddaeb1132b4a157e2b8239c607e24b76028b61","unresolved":false,"context_lines":[{"line_number":248,"context_line":""},{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_76fb2bc0","line":251,"in_reply_to":"5fc1f717_45e3d7a4","updated":"2019-03-16 03:40:21.000000000","message":"We are now facing some conflicts:\n(1) we want to do the upgrade-check for the config, but stable branch does not have that function\n(2) we want to force (adding the max value) the operator to change the value of this config, but someone do not like that start failure.\n(3) we do not want the very very large value to take effect, but we can not ensure that all the neutron server and the agent can be upgraded with consistent version. Such inconsistent can happen.\n(4) we want to let the operators to read the release note and check the value, this is impossible for all the deployments all the users all over the global.","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"fbc6d4a992435a2d3adb8579e518c3a6160f5ac5","unresolved":false,"context_lines":[{"line_number":248,"context_line":""},{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_966a5717","line":251,"in_reply_to":"5fc1f717_76fb2bc0","updated":"2019-03-16 04:41:00.000000000","message":"To answer your question:\n(1) the upgrade-check is nice but u currently already have a way to check that as I proposed to check for the following states: None(old code), False/True (new code). This is guaranteed to work although not explicit and less code write. update_device_list in rpc.py, the value is always set to False for agent rpc code, thus u don\u0027t need to set to False everywhere if u don\u0027t pass any value.\n\nIf we don\u0027t do the above upgrade-check then I agree with Brian to leave it alone. The release note should be enough if u are interested to upgrade. It doesn\u0027t hurt anyways if the agent_boot_time is a very high number means the current operator is still running with the old logic which is his choice, but at least it doesn\u0027t break his current state, it\u0027s simply not optimal. And by the next release the agent_boot_time is gone anyways.\n\nSo by not doing anything here like Brian mentioned:\n(1) clients that upgraded to the latest version (both server/agent) gets the benefit of the fix given that their agent_boot_time is too low or eventually becomes too low with more computes.\n(2) clients that don\u0027t upgrade their agents still gets the benefit of their old working logic without this override breaking it. Happy customer till the agent_boot_time becomes too low. At least we don\u0027t break the existing logic. \n\nOne bonus solution would be to add a one time warning when agent_uptime \u003e agent_boot_time in this case they\u0027ve been warned to switch.","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"20bc8070e9bed8918caf534ed2e03968731d7402","unresolved":false,"context_lines":[{"line_number":248,"context_line":""},{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_0234dcdc","line":251,"in_reply_to":"5fc1f717_966a5717","updated":"2019-03-17 08:13:28.000000000","message":"@Bruca, thanks, and done.\nBut in order to prevent such inconsistent between bool and None, I refactor this back to default False here [1]. We should give the boolean param a default True/False value.","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":13995,"name":"Nate Johnston","email":"nate.johnston@redhat.com","username":"natejohnston"},"change_message_id":"a566f892f71e2bef30aaa590a01c6e47fa0ae252","unresolved":false,"context_lines":[{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"},{"line_number":255,"context_line":"        port_context \u003d context._plugin_context"},{"line_number":256,"context_line":"        agent \u003d l2pop_db.get_agent_by_host(port_context, agent_host)"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_65de13db","line":253,"updated":"2019-03-15 21:36:25.000000000","message":"I agree, something like:\n\n        if cfg.CONF.l2pop.agent_boot_time !\u003d 180:\n            cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)\n            LOG.debug(\"Override l2pop agent_boot_time config option to 180s.\")\n\nThe deprecation is handled in the option definition, no need to repeat it here.","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"93dea46a1ce670fa611f74268a06fe0afbaa2548","unresolved":false,"context_lines":[{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"},{"line_number":255,"context_line":"        port_context \u003d context._plugin_context"},{"line_number":256,"context_line":"        agent \u003d l2pop_db.get_agent_by_host(port_context, agent_host)"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_ec475076","line":253,"updated":"2019-03-15 18:45:09.000000000","message":"This isn\u0027t actually the best place to do this, since it will get printed every time this is called.\n\nAnd like Bruca, it could be aggravating for a user to debug.  Why can\u0027t we just leave it alone, but print out a single message at LOG.warning level saying it\u0027s higher than the default?","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"daddaeb1132b4a157e2b8239c607e24b76028b61","unresolved":false,"context_lines":[{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        cfg.CONF.set_override(\u0027agent_boot_time\u0027, 180, \u0027l2pop\u0027)"},{"line_number":252,"context_line":"        LOG.debug(\"Override l2pop agent_boot_time config option to 180s. \""},{"line_number":253,"context_line":"                  \"It is deprecated and will be removed in Train release\")"},{"line_number":254,"context_line":"        agent_host \u003d context.host"},{"line_number":255,"context_line":"        port_context \u003d context._plugin_context"},{"line_number":256,"context_line":"        agent \u003d l2pop_db.get_agent_by_host(port_context, agent_host)"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_76248b5e","line":253,"in_reply_to":"5fc1f717_ec475076","updated":"2019-03-16 03:40:21.000000000","message":"@Brian, see my upper comments (3).","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"cec93c783705f0f3cdd992fce2f229fdfcd2869a","unresolved":false,"context_lines":[{"line_number":280,"context_line":"            self.L2populationAgentNotify.remove_fdb_entries("},{"line_number":281,"context_line":"                self.rpc_ctx, fdb_entries)"},{"line_number":282,"context_line":""},{"line_number":283,"context_line":"    def update_port_up(self, context, agent_restarted\u003dNone):"},{"line_number":284,"context_line":"        port \u003d context.current"},{"line_number":285,"context_line":"        agent_host \u003d context.host"},{"line_number":286,"context_line":"        port_context \u003d context._plugin_context"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_426c24cd","line":283,"range":{"start_line":283,"start_character":54,"end_line":283,"end_character":58},"updated":"2019-03-17 08:17:02.000000000","message":"Temporarily backward compatibility, refactor back to bool:\nhttps://review.openstack.org/#/c/641866/","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"580acacd48abf683e046bd37dbd2c2c1802012c7","unresolved":false,"context_lines":[{"line_number":306,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":307,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":308,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":309,"context_line":"        if agent_restarted is None:"},{"line_number":310,"context_line":"            # Only for backport compatibility, will be removed."},{"line_number":311,"context_line":"            agent_restarted \u003d self.agent_restarted(context)"},{"line_number":312,"context_line":"        if is_first_port or agent_restarted:"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_ce8cbb37","line":309,"updated":"2019-03-18 02:48:16.000000000","message":"you can optimize it further to avoid extra processing by doing:\n\nif not is_first_port and agent_restarted is None:\n     agent_restarted \u003d self.agent_restarted(context)","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"4338b66995f6f8c055cb04671baf6cf66d431f3f","unresolved":false,"context_lines":[{"line_number":306,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":307,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":308,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":309,"context_line":"        if agent_restarted is None:"},{"line_number":310,"context_line":"            # Only for backport compatibility, will be removed."},{"line_number":311,"context_line":"            agent_restarted \u003d self.agent_restarted(context)"},{"line_number":312,"context_line":"        if is_first_port or agent_restarted:"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_a2d5c31d","line":309,"in_reply_to":"5fc1f717_ce8cbb37","updated":"2019-03-18 04:46:37.000000000","message":"Let\u0027s do not mix those two different conditions.","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"cec93c783705f0f3cdd992fce2f229fdfcd2869a","unresolved":false,"context_lines":[{"line_number":306,"context_line":"        # with high concurrency more than 1 port may be activated on an agent"},{"line_number":307,"context_line":"        # at the same time (like VM port + a DVR port) so checking for 1 or 2"},{"line_number":308,"context_line":"        is_first_port \u003d agent_active_ports in (1, 2)"},{"line_number":309,"context_line":"        if agent_restarted is None:"},{"line_number":310,"context_line":"            # Only for backport compatibility, will be removed."},{"line_number":311,"context_line":"            agent_restarted \u003d self.agent_restarted(context)"},{"line_number":312,"context_line":"        if is_first_port or agent_restarted:"},{"line_number":313,"context_line":"            # First port(s) activated on current agent in this network,"},{"line_number":314,"context_line":"            # we have to provide it with the whole list of fdb entries"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_6267e0ea","line":311,"range":{"start_line":309,"start_character":8,"end_line":311,"end_character":59},"updated":"2019-03-17 08:17:02.000000000","message":"Temporarily backward compatibility, refactor back to bool:\nhttps://review.openstack.org/#/c/641866/","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"c4426190ec8fd7fafefec71aa27ff6878a717b2b","unresolved":false,"context_lines":[{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        LOG.warning(\"l2pop \u0027agent_boot_time\u0027 config option is deprecated \""},{"line_number":252,"context_line":"                    \"and will be removed in Train release. Please follow \""},{"line_number":253,"context_line":"                    \"the fix https://review.openstack.org/#/c/640797/ to \""},{"line_number":254,"context_line":"                    \"ensure all the agent are upgraded.\")"},{"line_number":255,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":16,"id":"5fc1f717_473988ef","line":252,"range":{"start_line":252,"start_character":41,"end_line":252,"end_character":43},"updated":"2019-03-18 14:41:47.000000000","message":"s/in the","commit_id":"a9fed5528e925f880d42b22266d637bc52e24728"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"4cf66321b692e399f4fc9872d9e709bf57587c20","unresolved":false,"context_lines":[{"line_number":249,"context_line":"    # This will be removed in next T release"},{"line_number":250,"context_line":"    def agent_restarted(self, context):"},{"line_number":251,"context_line":"        LOG.warning(\"l2pop \u0027agent_boot_time\u0027 config option is deprecated \""},{"line_number":252,"context_line":"                    \"and will be removed in Train release. Please follow \""},{"line_number":253,"context_line":"                    \"the fix https://review.openstack.org/#/c/640797/ to \""},{"line_number":254,"context_line":"                    \"ensure all the agent are upgraded.\")"},{"line_number":255,"context_line":"        agent_host \u003d context.host"}],"source_content_type":"text/x-python","patch_set":16,"id":"5fc1f717_6d858306","line":252,"range":{"start_line":252,"start_character":41,"end_line":252,"end_character":43},"in_reply_to":"5fc1f717_473988ef","updated":"2019-03-18 14:53:31.000000000","message":"Done","commit_id":"a9fed5528e925f880d42b22266d637bc52e24728"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"c4426190ec8fd7fafefec71aa27ff6878a717b2b","unresolved":false,"context_lines":[{"line_number":251,"context_line":"        LOG.warning(\"l2pop \u0027agent_boot_time\u0027 config option is deprecated \""},{"line_number":252,"context_line":"                    \"and will be removed in Train release. Please follow \""},{"line_number":253,"context_line":"                    \"the fix https://review.openstack.org/#/c/640797/ to \""},{"line_number":254,"context_line":"                    \"ensure all the agent are upgraded.\")"},{"line_number":255,"context_line":"        agent_host \u003d context.host"},{"line_number":256,"context_line":"        port_context \u003d context._plugin_context"},{"line_number":257,"context_line":"        agent \u003d l2pop_db.get_agent_by_host(port_context, agent_host)"}],"source_content_type":"text/x-python","patch_set":16,"id":"5fc1f717_2ace8159","line":254,"updated":"2019-03-18 14:41:47.000000000","message":"So I understand what you\u0027re trying to do here - log a message when an agent sends the \u0027old\u0027 report without agent_restarted being set.\n\nOne problem is that this could get printed a lot until all the agents are updated, which could be an issue in larger deployments.\n\nI think we can maybe do one of two things:\n\n1) print the agent_host so the operator knows which host it needs to restart the agent on.  Printing once would be preferred, but that could be hard to track here using a dict.\n\n2) print a generic message once\n\nAnd I don\u0027t think adding the bug here is correct, that should be in the release note.","commit_id":"a9fed5528e925f880d42b22266d637bc52e24728"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"4cf66321b692e399f4fc9872d9e709bf57587c20","unresolved":false,"context_lines":[{"line_number":251,"context_line":"        LOG.warning(\"l2pop \u0027agent_boot_time\u0027 config option is deprecated \""},{"line_number":252,"context_line":"                    \"and will be removed in Train release. Please follow \""},{"line_number":253,"context_line":"                    \"the fix https://review.openstack.org/#/c/640797/ to \""},{"line_number":254,"context_line":"                    \"ensure all the agent are upgraded.\")"},{"line_number":255,"context_line":"        agent_host \u003d context.host"},{"line_number":256,"context_line":"        port_context \u003d context._plugin_context"},{"line_number":257,"context_line":"        agent \u003d l2pop_db.get_agent_by_host(port_context, agent_host)"}],"source_content_type":"text/x-python","patch_set":16,"id":"5fc1f717_0d807ff4","line":254,"in_reply_to":"5fc1f717_2ace8159","updated":"2019-03-18 14:53:31.000000000","message":"Done","commit_id":"a9fed5528e925f880d42b22266d637bc52e24728"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"4ffff3dd5f6ab3e92785e950b862324bc06fc41f","unresolved":false,"context_lines":[{"line_number":254,"context_line":"        if l2pop_db.get_agent_uptime(agent) \u003c cfg.CONF.l2pop.agent_boot_time:"},{"line_number":255,"context_line":"            LOG.warning(\"l2pop \u0027agent_boot_time\u0027 config option is deprecated \""},{"line_number":256,"context_line":"                        \"and will be removed in the Train release. \""},{"line_number":257,"context_line":"                        \"RPC request agent host: %s\", agent_host)"},{"line_number":258,"context_line":"            return True"},{"line_number":259,"context_line":"        return False"},{"line_number":260,"context_line":""}],"source_content_type":"text/x-python","patch_set":17,"id":"5fc1f717_8a743a9e","line":257,"updated":"2019-03-19 12:57:15.000000000","message":"Agent on host \u0027%s\u0027 did not supply \u0027agent_restarted\u0027 information in RPC message, determined it restarted based on deprecated \u0027agent_boot_time\u0027 config option.\n\nSorry, I know I am nitpicking.","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"68097a620af40401f56f8ea6f61e45400d147ae8","unresolved":false,"context_lines":[{"line_number":254,"context_line":"        if l2pop_db.get_agent_uptime(agent) \u003c cfg.CONF.l2pop.agent_boot_time:"},{"line_number":255,"context_line":"            LOG.warning(\"l2pop \u0027agent_boot_time\u0027 config option is deprecated \""},{"line_number":256,"context_line":"                        \"and will be removed in the Train release. \""},{"line_number":257,"context_line":"                        \"RPC request agent host: %s\", agent_host)"},{"line_number":258,"context_line":"            return True"},{"line_number":259,"context_line":"        return False"},{"line_number":260,"context_line":""}],"source_content_type":"text/x-python","patch_set":17,"id":"5fc1f717_c0c78364","line":257,"in_reply_to":"5fc1f717_8a743a9e","updated":"2019-03-19 14:25:16.000000000","message":"Done","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"}],"neutron/plugins/ml2/drivers/openvswitch/agent/ovs_neutron_agent.py":[{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"4d51587dda4b356274181614fea99c8f0ecc9525","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d False if self.iter_num else True"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":5,"id":"5fc1f717_8c93fa4f","line":952,"range":{"start_line":952,"start_character":12,"end_line":952,"end_character":62},"updated":"2019-03-06 02:15:08.000000000","message":"FYI: this looks a bit confused, but pep8 can not pass for the original style:\n\"\"\"\nagent_restarted \u003d True if self.iter_num \u003d\u003d 0 else False\n\"\"\"","commit_id":"40205ed127f90d31a49a703265fa1edf81442982"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"b7b2f0a4bc293fa8291e2740a729eeb419c043f3","unresolved":false,"context_lines":[{"line_number":349,"context_line":""},{"line_number":350,"context_line":"            # we only want to update resource versions on startup"},{"line_number":351,"context_line":"            self.agent_state.pop(\u0027resource_versions\u0027, None)"},{"line_number":352,"context_line":"            if self.agent_state.pop(\u0027start_flag\u0027, None) and self.iter_num \u003d\u003d 0:"},{"line_number":353,"context_line":"                # On initial start, we notify systemd after initialization"},{"line_number":354,"context_line":"                # is complete."},{"line_number":355,"context_line":"                systemd.notify_once()"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_6650cb5e","line":352,"range":{"start_line":352,"start_character":60,"end_line":352,"end_character":78},"updated":"2019-03-11 12:34:24.000000000","message":"So you want to save this check here?\nFor the readability, it is worth to. But seems we are going to mix two different scenarios.","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"cab2bec0cc92bc821c77f902e7f9309077e94197","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_d2b2a1e6","line":952,"updated":"2019-03-11 08:55:50.000000000","message":"There is also \"self.agent_state[\u0027start_flag\u0027] \" available and it basically shows that agent was restarted.\nAlso please keep remember that You should populate all fdb flows also in case when ovs is restarted. Currently we are doing it by setting this self.agent_state[\u0027start_flag\u0027]  to True in _handle_ovs_restart() method in this file.","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"dc84215bfa49ff80d85c563109c5fa4c6cd38da1","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_439a536e","line":952,"in_reply_to":"5fc1f717_0639c73f","updated":"2019-03-11 22:40:55.000000000","message":"I agree we are mixing two different things. When I first red the code I thought it was for ovs restart but at the end I realized it\u0027s ovsd restart and not neutron ovs. We should perhaps refactor it in the future to be more explicit. At the same time I don\u0027t see where we set the start_flag back to False?\n\nCould be interesting to know whether flows are repopulated properly on ovsd restart given that iter_num is not reset to 0 on ovsd restart, from my memory it seems to be ok.","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"e0bece4e3f63c5a6887b971f83c5840da497fca8","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_0b15509c","line":952,"in_reply_to":"5fc1f717_08ea360e","updated":"2019-03-11 11:24:13.000000000","message":"yes, but in case when ovs is restarted, we also need to populate all fdb flows when L2 population is used. That\u0027s why some time ago I added that flag \"agent_state[\u0027start_flag\u0027]\" is set there to true. It\u0027s kind of hack to force l2 population to send all flows, exactly in same way like it is done when neutron-ova-agent is restarted.\nSo now I think that maybe You should change name of this new option to e.g. \"full_fdb_sync\" or something like that and use it also when ovs was restarted :)","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"b7b2f0a4bc293fa8291e2740a729eeb419c043f3","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_0639c73f","line":952,"in_reply_to":"5fc1f717_0b15509c","updated":"2019-03-11 12:34:24.000000000","message":"When the first rpc_loop is done, the \u0027new option\u0027 should be set back to False. For a normal port processing, we can not add such option as True. Then ovs restart will not be able to count on it.","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":10,"id":"5fc1f717_08ea360e","line":952,"in_reply_to":"5fc1f717_d2b2a1e6","updated":"2019-03-11 10:37:10.000000000","message":"That _handle_ovs_restart is for handling the status of the openvswitch. Here is very specificlly for the first RPC loop of our ovs-agent to handle the bridge ports (neutron ports) during the restart of python neutron ovs-agent.\n\nYou have any specific suggestions?","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":5948,"name":"Oleg Bondarev","email":"obondarev@mirantis.com","username":"obondarev"},"change_message_id":"0ccb32eabac984f9f90005734adb8f8cd7822566","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":11,"id":"5fc1f717_814d6e9b","line":952,"range":{"start_line":952,"start_character":30,"end_line":952,"end_character":53},"updated":"2019-03-13 14:18:05.000000000","message":"why not just \"self.iter_num \u003d\u003d 0\"","commit_id":"964e526829644098c6e60e60cedfedc7f8377ac8"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"f97da920a7e90efdf43f6bd835279e76c37629ed","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":11,"id":"5fc1f717_07faeebb","line":952,"range":{"start_line":952,"start_character":30,"end_line":952,"end_character":53},"in_reply_to":"5fc1f717_814d6e9b","updated":"2019-03-13 15:17:30.000000000","message":"Py3 pep8 does not happy.","commit_id":"964e526829644098c6e60e60cedfedc7f8377ac8"},{"author":{"_account_id":13995,"name":"Nate Johnston","email":"nate.johnston@redhat.com","username":"natejohnston"},"change_message_id":"a566f892f71e2bef30aaa590a01c6e47fa0ae252","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_05f38f75","line":952,"updated":"2019-03-15 21:36:25.000000000","message":"This is non-intuitive to read.  Wouldn\u0027t it be more intuitive to say this:\n\n            agent_restarted \u003d self.iter_num \u003d\u003d 0","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"20bc8070e9bed8918caf534ed2e03968731d7402","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d not bool(self.iter_num)"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":13,"id":"5fc1f717_02493c5e","line":952,"in_reply_to":"5fc1f717_05f38f75","updated":"2019-03-17 08:13:28.000000000","message":"Pep8 once did not be happy, but let see what\u0027s the status now.","commit_id":"4463a9ce1ef2f8816d9171445ee38f5578c37a61"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"d083c9ede81fb86e361ae187d43814fa497fd6d2","unresolved":false,"context_lines":[{"line_number":349,"context_line":""},{"line_number":350,"context_line":"            # we only want to update resource versions on startup"},{"line_number":351,"context_line":"            self.agent_state.pop(\u0027resource_versions\u0027, None)"},{"line_number":352,"context_line":"            if self.agent_state.pop(\u0027start_flag\u0027, None) and self.iter_num \u003d\u003d 0:"},{"line_number":353,"context_line":"                # On initial start, we notify systemd after initialization"},{"line_number":354,"context_line":"                # is complete."},{"line_number":355,"context_line":"                systemd.notify_once()"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_96f1d298","line":352,"range":{"start_line":352,"start_character":15,"end_line":352,"end_character":55},"updated":"2019-03-19 04:58:13.000000000","message":"RPC loop and state report pool will run concurrently, \u0027start_flag\u0027 now seems still rely on that \u0027agent_boot_time\u0027. When \u0027start_flag\u0027 is true, then refresh the up_time of the ovs-agent.\n\nBut this patch still remain that boot time check. We are not giving aggression on this now.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"a8bca3d28452428fd91c5f0f06b5606b7ec4872e","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_b87bb489","line":952,"updated":"2019-03-18 08:01:48.000000000","message":"see my comment in L2112, there is more conditions where You should set this flag to True","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"d083c9ede81fb86e361ae187d43814fa497fd6d2","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_f35aa84e","line":952,"in_reply_to":"5fc1f717_53605c36","updated":"2019-03-19 04:58:13.000000000","message":"@Bruca, thanks for the test, what\u0027s the flow result without this patch?\nIMO, if there are large scale ports for ovs-agent, current \u0027start_flag\u0027 mechanism (without this fix) is not enough. ovs-agent report state will remove the \u0027start_flag\u0027 while ovs just added the port to the bridge according to the ovs-db. Then next rpc loop, some flow may still get missing. Plz see line 352. I will test it too.\n\nAnd by the way, the condition should be added here:\nhttps://review.openstack.org/#/c/640797/15/neutron/plugins/ml2/drivers/l2pop/mech_driver.py@312\nSo, here we are facing the new problem, when/on which conditon set the \u0027start_flag\u0027 to Flase. How neutron server side determain whether to update fdbs.\n\nAnyway, the ovs restart should not be handled in this fix IMO, maybe we should open another bug for it.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"ead3b38476c2274d3aecb260c475d68125c24ee6","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_53605c36","line":952,"in_reply_to":"5fc1f717_5b972a3b","updated":"2019-03-19 02:53:50.000000000","message":"I tested just doing ovsd restart, and Slawek is right we indeed loose the flows as well. wonder if it\u0027s enough to do:\n\nagent_restarted \u003d self.iter_num \u003d\u003d 0 or self.agent_state[\u0027start_flag\u0027]","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"0259d175794273aa9496c63558d3789af64e5810","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_8e819c4d","line":952,"in_reply_to":"5fc1f717_5cd1492f","updated":"2019-03-20 03:25:54.000000000","message":"Nice when u mean finally restored means u loose them for a while? There is a time window when flows are gone? I agree if indeed an issue we should probably open another issue and not mix fixes. Otherwise it\u0027ll become hard to track what code fix what in the future.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"4834829f1c03d1d2b3688a6f5c421e5e6f96170c","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_699d4a09","line":952,"in_reply_to":"5fc1f717_8e819c4d","updated":"2019-03-20 06:07:19.000000000","message":"Yes, flows are gone during the first rpc_loop port processing period. Actually all flow entries will get lost when ovs is down, ovs-agent will reinstall a basic flow structure, and then other mechanism flows. You may notice that in this output, there are some flows are younger (duration\u003d484.126s) than others:\nhttp://paste.openstack.org/show/747995/\nBut we do not lose them. This should be the normal service recovery time.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"678c5b3b9ad0d51d0ba20251760897e465bc7b7c","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_d84d60c4","line":952,"in_reply_to":"5fc1f717_b87bb489","updated":"2019-03-18 08:57:11.000000000","message":"So did you mean add check \"or self.agent_state[\u0027start_flag\u0027] \u003d\u003d True\" to also handle the ovs restart?","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"e01a15ce3b0c85dd97619f994d673d3820deea03","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_5b972a3b","line":952,"in_reply_to":"5fc1f717_d84d60c4","updated":"2019-03-18 09:42:43.000000000","message":"no, rather to check here something like \"or ovs_restarted\" and then set this new flag to true.\nBasically agent_state[\u0027start_flag\u0027] is used now to tell neutron server that agent was just started. And server uses this information to determine if agent was just restarted (what You want to change in this patch). We did this small hack to set this flag to true in case when ovs was restarted as this was the easiest way to fix issue which we had after ovs restart. But now I think this case should be also addressed in same way how You are addressing neutron-ovs-agent restart case.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":30047,"name":"Bruca Lock","email":"lockshi@gmail.com"},"change_message_id":"0259d175794273aa9496c63558d3789af64e5810","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_eeb6203f","line":952,"in_reply_to":"5fc1f717_f35aa84e","updated":"2019-03-20 03:25:54.000000000","message":"Like I mentioned without this patch I loose the flows.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"26e5d8335c8a9db2763b5addd3f9976fa8ab74bd","unresolved":false,"context_lines":[{"line_number":949,"context_line":"            # to notify the agent to refresh the tunnel related flows."},{"line_number":950,"context_line":"            # Otherwise, these flows will be cleaned as stale due to the"},{"line_number":951,"context_line":"            # different cookie id."},{"line_number":952,"context_line":"            agent_restarted \u003d self.iter_num \u003d\u003d 0"},{"line_number":953,"context_line":"            devices_set \u003d self.plugin_rpc.update_device_list("},{"line_number":954,"context_line":"                self.context, devices_up, devices_down, self.agent_id,"},{"line_number":955,"context_line":"                self.conf.host, agent_restarted\u003dagent_restarted)"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_5cd1492f","line":952,"in_reply_to":"5fc1f717_f35aa84e","updated":"2019-03-19 08:24:31.000000000","message":"This is my test result (with this fix) for only restart openvswith:\nhttp://paste.openstack.org/show/747994/\nYes, 400 ports and 1000+ sg rules.\n\nThe tunnel flow are finally restored, and connection to VM can be established. br-tun flow dump:\nhttp://paste.openstack.org/show/747995/","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"a8bca3d28452428fd91c5f0f06b5606b7ec4872e","unresolved":false,"context_lines":[{"line_number":2109,"context_line":"            self._reset_tunnel_ofports()"},{"line_number":2110,"context_line":"            self.setup_tunnel_br()"},{"line_number":2111,"context_line":"            self.setup_tunnel_br_flows()"},{"line_number":2112,"context_line":"            self.agent_state[\u0027start_flag\u0027] \u003d True"},{"line_number":2113,"context_line":"            # Force state report to avoid race condition"},{"line_number":2114,"context_line":"            # with l2pop fdb entries update"},{"line_number":2115,"context_line":"            self._report_state()"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_383ee4b7","line":2112,"updated":"2019-03-18 08:01:48.000000000","message":"I think I already said that but in such case You should also treat everything like agent would be restarted because all flows are cleaned after ovs restart on host.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"ee8041669632c044b8c135e7655b270dc9e8d894","unresolved":false,"context_lines":[{"line_number":2109,"context_line":"            self._reset_tunnel_ofports()"},{"line_number":2110,"context_line":"            self.setup_tunnel_br()"},{"line_number":2111,"context_line":"            self.setup_tunnel_br_flows()"},{"line_number":2112,"context_line":"            self.agent_state[\u0027start_flag\u0027] \u003d True"},{"line_number":2113,"context_line":"            # Force state report to avoid race condition"},{"line_number":2114,"context_line":"            # with l2pop fdb entries update"},{"line_number":2115,"context_line":"            self._report_state()"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_dbe23aa0","line":2112,"in_reply_to":"5fc1f717_1bfc426e","updated":"2019-03-18 09:57:24.000000000","message":"Thank you for your detailed clarifying. : )\n\nIf remove the \u0027agent_boot_time\u0027, this \u0027start_flag\u0027 may not work in neutron server side for ovs restart. OK, let\u0027s investigate more for next step.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"c230ec0c1f82c399840eb04e1696d479aff8a4b8","unresolved":false,"context_lines":[{"line_number":2109,"context_line":"            self._reset_tunnel_ofports()"},{"line_number":2110,"context_line":"            self.setup_tunnel_br()"},{"line_number":2111,"context_line":"            self.setup_tunnel_br_flows()"},{"line_number":2112,"context_line":"            self.agent_state[\u0027start_flag\u0027] \u003d True"},{"line_number":2113,"context_line":"            # Force state report to avoid race condition"},{"line_number":2114,"context_line":"            # with l2pop fdb entries update"},{"line_number":2115,"context_line":"            self._report_state()"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_7b5b06f5","line":2112,"in_reply_to":"5fc1f717_383ee4b7","updated":"2019-03-18 09:32:58.000000000","message":"After reading some code of this \u0027start_flag\u0027, I think we should better not mix this with ovs-agent restart for now. Although ovs restart may have the same processing condition/logical/fdbs with ovs-agent restart. IMO, we can fix this ovs-agent restart issue by current approach. And refactor or change code for the ovs restart. This \u0027start_flag\u0027 seems also have some coupled code we should take care in neutron server side. Maybe add a new param for update_device_list for ovs restart. Make sense?","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"d08f755bae0b1630e7f44322013d62be1e6d640a","unresolved":false,"context_lines":[{"line_number":2109,"context_line":"            self._reset_tunnel_ofports()"},{"line_number":2110,"context_line":"            self.setup_tunnel_br()"},{"line_number":2111,"context_line":"            self.setup_tunnel_br_flows()"},{"line_number":2112,"context_line":"            self.agent_state[\u0027start_flag\u0027] \u003d True"},{"line_number":2113,"context_line":"            # Force state report to avoid race condition"},{"line_number":2114,"context_line":"            # with l2pop fdb entries update"},{"line_number":2115,"context_line":"            self._report_state()"}],"source_content_type":"text/x-python","patch_set":15,"id":"5fc1f717_1bfc426e","line":2112,"in_reply_to":"5fc1f717_7b5b06f5","updated":"2019-03-18 09:47:10.000000000","message":"As I wrote in my comment above it is now kind of hack to make it working properly after ovs restart. Now, every time when ovs will be restarted on agent it will probably cause logging warning message about deprecated agent_down_time option. So IMO it should be addressed in same patch.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"}],"neutron/plugins/ml2/rpc.py":[{"author":{"_account_id":5948,"name":"Oleg Bondarev","email":"obondarev@mirantis.com","username":"obondarev"},"change_message_id":"6fe3c390deb22ef676391c40a8840d2f37780963","unresolved":false,"context_lines":[{"line_number":351,"context_line":"        else:"},{"line_number":352,"context_line":"            l2pop_driver.obj.update_port_down(port_context)"},{"line_number":353,"context_line":""},{"line_number":354,"context_line":"    def update_device_list(self, rpc_context, **kwargs):"},{"line_number":355,"context_line":"        devices_up \u003d []"},{"line_number":356,"context_line":"        failed_devices_up \u003d []"},{"line_number":357,"context_line":"        devices_down \u003d []"}],"source_content_type":"text/x-python","patch_set":6,"id":"5fc1f717_a067d725","line":354,"range":{"start_line":354,"start_character":8,"end_line":354,"end_character":26},"updated":"2019-03-06 06:58:41.000000000","message":"this was updated in rpc, not update_device_up","commit_id":"c4ea4b13d37e0a21fdef4584b4c20f0a4276f228"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"032f1e3853eb6b66de5f55a0b813736805c29c0f","unresolved":false,"context_lines":[{"line_number":351,"context_line":"        else:"},{"line_number":352,"context_line":"            l2pop_driver.obj.update_port_down(port_context)"},{"line_number":353,"context_line":""},{"line_number":354,"context_line":"    def update_device_list(self, rpc_context, **kwargs):"},{"line_number":355,"context_line":"        devices_up \u003d []"},{"line_number":356,"context_line":"        failed_devices_up \u003d []"},{"line_number":357,"context_line":"        devices_down \u003d []"}],"source_content_type":"text/x-python","patch_set":6,"id":"5fc1f717_3b5b1a88","line":354,"range":{"start_line":354,"start_character":8,"end_line":354,"end_character":26},"in_reply_to":"5fc1f717_1b3bbe38","updated":"2019-03-06 07:42:57.000000000","message":"This is the basic call stack:\nhttps://review.openstack.org/#/c/640797/6/neutron/plugins/ml2/drivers/openvswitch/agent/ovs_neutron_agent.py@953\nhttps://review.openstack.org/#/c/640797/6/neutron/agent/rpc.py@155\n\nand then neutron-server this line.","commit_id":"c4ea4b13d37e0a21fdef4584b4c20f0a4276f228"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"5a61353d43772828e97b06c3c42220917eb9d4bf","unresolved":false,"context_lines":[{"line_number":351,"context_line":"        else:"},{"line_number":352,"context_line":"            l2pop_driver.obj.update_port_down(port_context)"},{"line_number":353,"context_line":""},{"line_number":354,"context_line":"    def update_device_list(self, rpc_context, **kwargs):"},{"line_number":355,"context_line":"        devices_up \u003d []"},{"line_number":356,"context_line":"        failed_devices_up \u003d []"},{"line_number":357,"context_line":"        devices_down \u003d []"}],"source_content_type":"text/x-python","patch_set":6,"id":"5fc1f717_7b620219","line":354,"range":{"start_line":354,"start_character":8,"end_line":354,"end_character":26},"in_reply_to":"5fc1f717_3b5b1a88","updated":"2019-03-06 07:51:47.000000000","message":"Or, you mean we should add the following code to this function?\n\"\"\"\nagent_restarted \u003d kwargs.pop(\u0027agent_restarted\u0027, False)\n\"\"\"\nSince line 363 already send all the kwargs to the \u0027update_device_up\u0027, I don\u0027t think it is necessary to pop it here. Otherwise add  \u0027agent_restarted\u0027 to \u0027update_device_up\u0027 params list, then, you change it!\n\nIMO, the current style should be the smallest change.","commit_id":"c4ea4b13d37e0a21fdef4584b4c20f0a4276f228"},{"author":{"_account_id":5948,"name":"Oleg Bondarev","email":"obondarev@mirantis.com","username":"obondarev"},"change_message_id":"866306f57aba9aa0e8c22f014ed3a3868903a23a","unresolved":false,"context_lines":[{"line_number":351,"context_line":"        else:"},{"line_number":352,"context_line":"            l2pop_driver.obj.update_port_down(port_context)"},{"line_number":353,"context_line":""},{"line_number":354,"context_line":"    def update_device_list(self, rpc_context, **kwargs):"},{"line_number":355,"context_line":"        devices_up \u003d []"},{"line_number":356,"context_line":"        failed_devices_up \u003d []"},{"line_number":357,"context_line":"        devices_down \u003d []"}],"source_content_type":"text/x-python","patch_set":6,"id":"5fc1f717_61aa5086","line":354,"range":{"start_line":354,"start_character":8,"end_line":354,"end_character":26},"in_reply_to":"5fc1f717_7b620219","updated":"2019-03-06 09:50:51.000000000","message":"Right, I missed that all kwargs are passed to update_device_up, sorry.","commit_id":"c4ea4b13d37e0a21fdef4584b4c20f0a4276f228"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"afbd5f3b978f168ed4c96d75687b7ea7b27afd18","unresolved":false,"context_lines":[{"line_number":351,"context_line":"        else:"},{"line_number":352,"context_line":"            l2pop_driver.obj.update_port_down(port_context)"},{"line_number":353,"context_line":""},{"line_number":354,"context_line":"    def update_device_list(self, rpc_context, **kwargs):"},{"line_number":355,"context_line":"        devices_up \u003d []"},{"line_number":356,"context_line":"        failed_devices_up \u003d []"},{"line_number":357,"context_line":"        devices_down \u003d []"}],"source_content_type":"text/x-python","patch_set":6,"id":"5fc1f717_1b3bbe38","line":354,"range":{"start_line":354,"start_character":8,"end_line":354,"end_character":26},"in_reply_to":"5fc1f717_a067d725","updated":"2019-03-06 07:37:29.000000000","message":"Sorry I do not get it, you mean the ovs-agent does not call this?","commit_id":"c4ea4b13d37e0a21fdef4584b4c20f0a4276f228"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"cec93c783705f0f3cdd992fce2f229fdfcd2869a","unresolved":false,"context_lines":[{"line_number":254,"context_line":""},{"line_number":255,"context_line":"    def update_device_up(self, rpc_context, **kwargs):"},{"line_number":256,"context_line":"        \"\"\"Device is up on agent.\"\"\""},{"line_number":257,"context_line":"        agent_restarted \u003d kwargs.pop(\u0027agent_restarted\u0027, None)"},{"line_number":258,"context_line":"        agent_id, host, device \u003d self._get_request_details(kwargs)"},{"line_number":259,"context_line":"        LOG.debug(\"Device %(device)s up at agent %(agent_id)s\","},{"line_number":260,"context_line":"                  {\u0027device\u0027: device, \u0027agent_id\u0027: agent_id})"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_02629cd8","line":257,"range":{"start_line":257,"start_character":56,"end_line":257,"end_character":60},"updated":"2019-03-17 08:17:02.000000000","message":"Temporarily backward compatibility, refactor back to bool:\nhttps://review.openstack.org/#/c/641866/","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"cec93c783705f0f3cdd992fce2f229fdfcd2869a","unresolved":false,"context_lines":[{"line_number":307,"context_line":"                provisioning_blocks.L2_AGENT_ENTITY)"},{"line_number":308,"context_line":""},{"line_number":309,"context_line":"    def notify_l2pop_port_wiring(self, port_id, rpc_context,"},{"line_number":310,"context_line":"                                 status, host, agent_restarted\u003dNone):"},{"line_number":311,"context_line":"        \"\"\"Notify the L2pop driver that a port has been wired/unwired."},{"line_number":312,"context_line":""},{"line_number":313,"context_line":"        The L2pop driver uses this notification to broadcast forwarding"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_225d5814","line":310,"range":{"start_line":310,"start_character":63,"end_line":310,"end_character":67},"updated":"2019-03-17 08:17:02.000000000","message":"Temporarily backward compatibility, refactor back to bool:\nhttps://review.openstack.org/#/c/641866/","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"cec93c783705f0f3cdd992fce2f229fdfcd2869a","unresolved":false,"context_lines":[{"line_number":330,"context_line":"        # and so we don\u0027t need to update it again here. But, l2pop did not"},{"line_number":331,"context_line":"        # handle DVR ports while restart neutron-*-agent, we need to handle"},{"line_number":332,"context_line":"        # it here."},{"line_number":333,"context_line":"        if agent_restarted is None:"},{"line_number":334,"context_line":"            agent_restarted \u003d l2pop_driver.obj.agent_restarted(port_context)"},{"line_number":335,"context_line":"        if (port[\u0027device_owner\u0027] \u003d\u003d n_const.DEVICE_OWNER_DVR_INTERFACE and"},{"line_number":336,"context_line":"                not agent_restarted):"},{"line_number":337,"context_line":"            return"}],"source_content_type":"text/x-python","patch_set":14,"id":"5fc1f717_c29d74bc","line":334,"range":{"start_line":333,"start_character":7,"end_line":334,"end_character":76},"updated":"2019-03-17 08:17:02.000000000","message":"Temporarily backward compatibility, refactor back to bool:\nhttps://review.openstack.org/#/c/641866/","commit_id":"2de3118b47a05d29621978b6d13ec329a1b7df03"}],"releasenotes/notes/precise-agent-state-transfer-67c771cb1ee04dd0.yaml":[{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"3c1485d04b0a9c9ca1504b27596961c4834ba2b2","unresolved":false,"context_lines":[{"line_number":11,"context_line":"    for L2 population. Because of this, some flow update operations will not"},{"line_number":12,"context_line":"    be triggerred, resulting in lost flows during agent restart, especially"},{"line_number":13,"context_line":"    for host-to-host vxlan tunnel flows, causing the original tunnel flows to"},{"line_number":14,"context_line":"    be treated as stale due to the different cookie IDs. A subsequent agent"},{"line_number":15,"context_line":"    RPC loop will do a stale flow clean-up procedure and delete them, leading"},{"line_number":16,"context_line":"    to a loss of connectivity."},{"line_number":17,"context_line":"fixes:"}],"source_content_type":"text/x-yaml","patch_set":7,"id":"5fc1f717_85063a86","line":14,"range":{"start_line":14,"start_character":59,"end_line":14,"end_character":69},"updated":"2019-03-07 23:24:46.000000000","message":"This is not True, the clean procedure is running in the first loop.","commit_id":"3e9ea13e265601fcd918a3e1b8e7d9d24b8021ed"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"d29b30c023e8d7412371f6699086b3e63a838819","unresolved":false,"context_lines":[{"line_number":11,"context_line":"    for L2 population. Because of this, some flow update operations will not"},{"line_number":12,"context_line":"    be triggerred, resulting in lost flows during agent restart, especially"},{"line_number":13,"context_line":"    for host-to-host vxlan tunnel flows, causing the original tunnel flows to"},{"line_number":14,"context_line":"    be treated as stale due to the different cookie IDs. And agent first"},{"line_number":15,"context_line":"    RPC loop will also do a stale flow clean-up procedure and delete them,"},{"line_number":16,"context_line":"    leading to a loss of connectivity."},{"line_number":17,"context_line":"fixes:"}],"source_content_type":"text/x-yaml","patch_set":8,"id":"5fc1f717_aea473ba","line":14,"range":{"start_line":14,"start_character":57,"end_line":14,"end_character":66},"updated":"2019-03-08 02:16:33.000000000","message":"s/The agent\u0027s\n\nI tried hard on the re-wording :)","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"b0fdb368a3d33cd795463d0802b4e5f4b333e037","unresolved":false,"context_lines":[{"line_number":11,"context_line":"    for L2 population. Because of this, some flow update operations will not"},{"line_number":12,"context_line":"    be triggerred, resulting in lost flows during agent restart, especially"},{"line_number":13,"context_line":"    for host-to-host vxlan tunnel flows, causing the original tunnel flows to"},{"line_number":14,"context_line":"    be treated as stale due to the different cookie IDs. And agent first"},{"line_number":15,"context_line":"    RPC loop will also do a stale flow clean-up procedure and delete them,"},{"line_number":16,"context_line":"    leading to a loss of connectivity."},{"line_number":17,"context_line":"fixes:"}],"source_content_type":"text/x-yaml","patch_set":8,"id":"5fc1f717_14a35aa0","line":14,"range":{"start_line":14,"start_character":57,"end_line":14,"end_character":66},"in_reply_to":"5fc1f717_aea473ba","updated":"2019-03-08 05:07:08.000000000","message":"Done","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"926f549e7a485a19280d18fc069e46f10e745190","unresolved":false,"context_lines":[{"line_number":11,"context_line":"    for L2 population. Because of this, some flow update operations will not"},{"line_number":12,"context_line":"    be triggerred, resulting in lost flows during agent restart, especially"},{"line_number":13,"context_line":"    for host-to-host vxlan tunnel flows, causing the original tunnel flows to"},{"line_number":14,"context_line":"    be treated as stale due to the different cookie IDs. And agent first"},{"line_number":15,"context_line":"    RPC loop will also do a stale flow clean-up procedure and delete them,"},{"line_number":16,"context_line":"    leading to a loss of connectivity."},{"line_number":17,"context_line":"fixes:"}],"source_content_type":"text/x-yaml","patch_set":8,"id":"5fc1f717_d9b8ab2b","line":14,"range":{"start_line":14,"start_character":57,"end_line":14,"end_character":66},"in_reply_to":"5fc1f717_aea473ba","updated":"2019-03-08 03:21:01.000000000","message":"It hurts my brain too...","commit_id":"e1fc4581205f2e744452af8c169de941d9247c59"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"54e7b16c26b2b11006251cecfbe7b8d3d34a477c","unresolved":false,"context_lines":[{"line_number":2,"context_line":"deprecations:"},{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    The L2 population ``agent_boot_time`` config option is deprecated in"},{"line_number":5,"context_line":"    favor of the directly RPC agent restart state transfer. It will be"},{"line_number":6,"context_line":"    removed in a future release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_d2570102","line":5,"range":{"start_line":5,"start_character":17,"end_line":5,"end_character":25},"updated":"2019-03-11 09:04:51.000000000","message":"nit: direct","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":2,"context_line":"deprecations:"},{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    The L2 population ``agent_boot_time`` config option is deprecated in"},{"line_number":5,"context_line":"    favor of the directly RPC agent restart state transfer. It will be"},{"line_number":6,"context_line":"    removed in a future release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_a89e8251","line":5,"range":{"start_line":5,"start_character":17,"end_line":5,"end_character":25},"in_reply_to":"5fc1f717_d2570102","updated":"2019-03-11 10:37:10.000000000","message":"Done","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"cab2bec0cc92bc821c77f902e7f9309077e94197","unresolved":false,"context_lines":[{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    The L2 population ``agent_boot_time`` config option is deprecated in"},{"line_number":5,"context_line":"    favor of the directly RPC agent restart state transfer. It will be"},{"line_number":6,"context_line":"    removed in a future release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_32e1a5d1","line":6,"updated":"2019-03-11 08:55:50.000000000","message":"Please add info in which release it can/should be released","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    The L2 population ``agent_boot_time`` config option is deprecated in"},{"line_number":5,"context_line":"    favor of the directly RPC agent restart state transfer. It will be"},{"line_number":6,"context_line":"    removed in a future release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_48ad9e86","line":6,"in_reply_to":"5fc1f717_32e1a5d1","updated":"2019-03-11 10:37:10.000000000","message":"Done","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"cab2bec0cc92bc821c77f902e7f9309077e94197","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"},{"line_number":10,"context_line":"    prevent too much robust value for it, we set the max to 180s. This"},{"line_number":11,"context_line":"    may cause the neutron server or agent failed to restart, but give the"},{"line_number":12,"context_line":"    operators a chance to remove this config."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_b2ccb561","line":12,"updated":"2019-03-11 08:55:50.000000000","message":"maybe You can then add some upgrade-check for that and warn operator if they have higher value configured already.","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"b7b2f0a4bc293fa8291e2740a729eeb419c043f3","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"},{"line_number":10,"context_line":"    prevent too much robust value for it, we set the max to 180s. This"},{"line_number":11,"context_line":"    may cause the neutron server or agent failed to restart, but give the"},{"line_number":12,"context_line":"    operators a chance to remove this config."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_a6d0f320","line":12,"in_reply_to":"5fc1f717_6bd654f4","updated":"2019-03-11 12:34:24.000000000","message":"Instead of checking that, we can remove the use of this config, just left it there for that deprecation procedure. So it will be silently ignored no matter the value is. It will done in a independent patch:\nhttps://review.openstack.org/#/c/641866/\nMake sense?","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"3e1d47cdc451c62d60c2d64e8a13b6e7bc4af103","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"},{"line_number":10,"context_line":"    prevent too much robust value for it, we set the max to 180s. This"},{"line_number":11,"context_line":"    may cause the neutron server or agent failed to restart, but give the"},{"line_number":12,"context_line":"    operators a chance to remove this config."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_d962b43e","line":12,"in_reply_to":"5fc1f717_a6d0f320","updated":"2019-03-12 01:23:10.000000000","message":"\u003e Instead of checking that, we can remove the use of this config,\n \u003e just left it there for that deprecation procedure. So it will be\n \u003e silently ignored no matter the value is. It will done in a\n \u003e independent patch:\n \u003e https://review.openstack.org/#/c/641866/\n \u003e Make sense?","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"e0bece4e3f63c5a6887b971f83c5840da497fca8","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"},{"line_number":10,"context_line":"    prevent too much robust value for it, we set the max to 180s. This"},{"line_number":11,"context_line":"    may cause the neutron server or agent failed to restart, but give the"},{"line_number":12,"context_line":"    operators a chance to remove this config."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_6bd654f4","line":12,"in_reply_to":"5fc1f717_a83a0231","updated":"2019-03-11 11:24:13.000000000","message":"First of all we will not remove it for at least 2 cycles as IIRC that is correct deprecation procedure.\nSecond thing is that I was thinking about something like:\n\n    if cfg.agent_boot_time \u003e 180:\n        WARNING(\"Agent config option value is now limited to 180 seconds at max. Please correct Your config\")\n\nor something like that.\nYou should of course handle case if there is no such option there and not display this warning then.\n\nThird thing: we don\u0027t even have \"neutron-status upgrade check\" tool in stable branches yet. This tool is expected to be used after installing new packages but before restarting services. That way You can run checks from new version to verify if restart service will go fine.","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"dc00deed235e18e2f99678180c07a4b34c1480b8","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"},{"line_number":10,"context_line":"    prevent too much robust value for it, we set the max to 180s. This"},{"line_number":11,"context_line":"    may cause the neutron server or agent failed to restart, but give the"},{"line_number":12,"context_line":"    operators a chance to remove this config."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_a83a0231","line":12,"in_reply_to":"5fc1f717_b2ccb561","updated":"2019-03-11 10:37:10.000000000","message":"We will remove this config soon, then there will be a conflict, we will remove it, and we check it... And stable branch does not have that upgrade check functionities.","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"54b6d8d5298907d88158d3387a45d6e28bdb906d","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    For the L2 population ``agent_boot_time`` config option, in order to"},{"line_number":10,"context_line":"    prevent too much robust value for it, we set the max to 180s. This"},{"line_number":11,"context_line":"    may cause the neutron server or agent failed to restart, but give the"},{"line_number":12,"context_line":"    operators a chance to remove this config."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":10,"id":"5fc1f717_f9ca1018","line":12,"in_reply_to":"5fc1f717_d962b43e","updated":"2019-03-12 01:24:11.000000000","message":"Done here:\nhttps://review.openstack.org/#/c/641866/","commit_id":"235525f430a3e86fb79d3fa9b22fa695f6f958c1"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"a8bca3d28452428fd91c5f0f06b5606b7ec4872e","unresolved":false,"context_lines":[{"line_number":6,"context_line":"    removed in next ``Train`` release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all the neutron server and neutron-openswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version. After that, the L2 population"},{"line_number":11,"context_line":"    ``agent_boot_time`` config option will no longer take effect."},{"line_number":12,"context_line":"critical:"}],"source_content_type":"text/x-yaml","patch_set":15,"id":"5fc1f717_7871cca7","line":9,"range":{"start_line":9,"start_character":58,"end_line":9,"end_character":68},"updated":"2019-03-18 08:01:48.000000000","message":"nit: openvswitch (missing \"v\")","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"86b2dc58c7255e5354f0438b6725e5d21a2b503c","unresolved":false,"context_lines":[{"line_number":6,"context_line":"    removed in next ``Train`` release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all the neutron server and neutron-openswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version. After that, the L2 population"},{"line_number":11,"context_line":"    ``agent_boot_time`` config option will no longer take effect."},{"line_number":12,"context_line":"critical:"}],"source_content_type":"text/x-yaml","patch_set":15,"id":"5fc1f717_367915c7","line":9,"range":{"start_line":9,"start_character":58,"end_line":9,"end_character":68},"in_reply_to":"5fc1f717_7871cca7","updated":"2019-03-18 11:09:47.000000000","message":"Done","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"a8bca3d28452428fd91c5f0f06b5606b7ec4872e","unresolved":false,"context_lines":[{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all the neutron server and neutron-openswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version. After that, the L2 population"},{"line_number":11,"context_line":"    ``agent_boot_time`` config option will no longer take effect."},{"line_number":12,"context_line":"critical:"},{"line_number":13,"context_line":"  - |"},{"line_number":14,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":15,"id":"5fc1f717_f8b13c45","line":11,"updated":"2019-03-18 08:01:48.000000000","message":"You can upgrade agents first and then server, that should work and that is only supported way of upgrade AFAIK.\nSo IMHO this section is not necessary here. Upgrade process isn\u0027t any different in this case than in all other cases and that message may only confuse operators","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"7a53c4c95db7c085b63bc0cd4602550556930a2d","unresolved":false,"context_lines":[{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all the neutron server and neutron-openswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version. After that, the L2 population"},{"line_number":11,"context_line":"    ``agent_boot_time`` config option will no longer take effect."},{"line_number":12,"context_line":"critical:"},{"line_number":13,"context_line":"  - |"},{"line_number":14,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":15,"id":"5fc1f717_bbbdeeb7","line":11,"in_reply_to":"5fc1f717_b89ad410","updated":"2019-03-18 09:44:24.000000000","message":"IMO that is quite obvious and such information isn\u0027t necessary but I will not block this patch because of that :)","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"4b739b702e6a0d5ad006cb0e41bd3b5df5c3d663","unresolved":false,"context_lines":[{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all the neutron server and neutron-openswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version. After that, the L2 population"},{"line_number":11,"context_line":"    ``agent_boot_time`` config option will no longer take effect."},{"line_number":12,"context_line":"critical:"},{"line_number":13,"context_line":"  - |"},{"line_number":14,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":15,"id":"5fc1f717_b89ad410","line":11,"in_reply_to":"5fc1f717_f8b13c45","updated":"2019-03-18 09:20:47.000000000","message":"This section is for operators to ensure/check that there is no inconstent between server side and agent side.","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"a8bca3d28452428fd91c5f0f06b5606b7ec4872e","unresolved":false,"context_lines":[{"line_number":11,"context_line":"    ``agent_boot_time`` config option will no longer take effect."},{"line_number":12,"context_line":"critical:"},{"line_number":13,"context_line":"  - |"},{"line_number":14,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"},{"line_number":15,"context_line":"    a large number of ports, exceeding its timeout value, ``agent_boot_time``,"},{"line_number":16,"context_line":"    for L2 population. Because of this, some flow update operations will not"},{"line_number":17,"context_line":"    be triggerred, resulting in lost flows during agent restart, especially"}],"source_content_type":"text/x-yaml","patch_set":15,"id":"5fc1f717_b8897475","line":14,"range":{"start_line":14,"start_character":16,"end_line":14,"end_character":26},"updated":"2019-03-18 08:01:48.000000000","message":"nit: openvswitch","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"86b2dc58c7255e5354f0438b6725e5d21a2b503c","unresolved":false,"context_lines":[{"line_number":11,"context_line":"    ``agent_boot_time`` config option will no longer take effect."},{"line_number":12,"context_line":"critical:"},{"line_number":13,"context_line":"  - |"},{"line_number":14,"context_line":"    The neutron-openswitch-agent can sometimes spend too much time handling"},{"line_number":15,"context_line":"    a large number of ports, exceeding its timeout value, ``agent_boot_time``,"},{"line_number":16,"context_line":"    for L2 population. Because of this, some flow update operations will not"},{"line_number":17,"context_line":"    be triggerred, resulting in lost flows during agent restart, especially"}],"source_content_type":"text/x-yaml","patch_set":15,"id":"5fc1f717_5674e1ce","line":14,"range":{"start_line":14,"start_character":16,"end_line":14,"end_character":26},"in_reply_to":"5fc1f717_b8897475","updated":"2019-03-18 11:09:47.000000000","message":"Done","commit_id":"ca95e7470d6a969665e17eecf8c45b30d38f9c36"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"4ffff3dd5f6ab3e92785e950b862324bc06fc41f","unresolved":false,"context_lines":[{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    The L2 population ``agent_boot_time`` config option is deprecated in"},{"line_number":5,"context_line":"    favor of the direct RPC agent restart state transfer. It will be"},{"line_number":6,"context_line":"    removed in next ``Train`` release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_0f84dcb0","line":6,"range":{"start_line":6,"start_character":15,"end_line":6,"end_character":19},"updated":"2019-03-19 12:57:15.000000000","message":"s/the","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"68097a620af40401f56f8ea6f61e45400d147ae8","unresolved":false,"context_lines":[{"line_number":3,"context_line":"  - |"},{"line_number":4,"context_line":"    The L2 population ``agent_boot_time`` config option is deprecated in"},{"line_number":5,"context_line":"    favor of the direct RPC agent restart state transfer. It will be"},{"line_number":6,"context_line":"    removed in next ``Train`` release."},{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_40a9538e","line":6,"range":{"start_line":6,"start_character":15,"end_line":6,"end_character":19},"in_reply_to":"5fc1f717_0f84dcb0","updated":"2019-03-19 14:25:16.000000000","message":"Done","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"4ffff3dd5f6ab3e92785e950b862324bc06fc41f","unresolved":false,"context_lines":[{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version in one single deployment. After"},{"line_number":11,"context_line":"    that, the L2 population ``agent_boot_time`` config option will no longer"},{"line_number":12,"context_line":"    take effect."},{"line_number":13,"context_line":"critical:"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_8f344cfd","line":10,"range":{"start_line":10,"start_character":4,"end_line":10,"end_character":7},"updated":"2019-03-19 12:57:15.000000000","message":"s/binaries are (?)","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"4ffff3dd5f6ab3e92785e950b862324bc06fc41f","unresolved":false,"context_lines":[{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version in one single deployment. After"},{"line_number":11,"context_line":"    that, the L2 population ``agent_boot_time`` config option will no longer"},{"line_number":12,"context_line":"    take effect."},{"line_number":13,"context_line":"critical:"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_2f12f863","line":10,"range":{"start_line":10,"start_character":17,"end_line":10,"end_character":74},"updated":"2019-03-19 12:57:15.000000000","message":"s/for the changes to take effect, after which the L2 population...","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"68097a620af40401f56f8ea6f61e45400d147ae8","unresolved":false,"context_lines":[{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version in one single deployment. After"},{"line_number":11,"context_line":"    that, the L2 population ``agent_boot_time`` config option will no longer"},{"line_number":12,"context_line":"    take effect."},{"line_number":13,"context_line":"critical:"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_00a10b96","line":10,"range":{"start_line":10,"start_character":17,"end_line":10,"end_character":74},"in_reply_to":"5fc1f717_2f12f863","updated":"2019-03-19 14:25:16.000000000","message":"Done","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"68097a620af40401f56f8ea6f61e45400d147ae8","unresolved":false,"context_lines":[{"line_number":7,"context_line":"upgrade:"},{"line_number":8,"context_line":"  - |"},{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version in one single deployment. After"},{"line_number":11,"context_line":"    that, the L2 population ``agent_boot_time`` config option will no longer"},{"line_number":12,"context_line":"    take effect."},{"line_number":13,"context_line":"critical:"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_80889b22","line":10,"range":{"start_line":10,"start_character":4,"end_line":10,"end_character":7},"in_reply_to":"5fc1f717_8f344cfd","updated":"2019-03-19 14:25:16.000000000","message":"Done","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"4ffff3dd5f6ab3e92785e950b862324bc06fc41f","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version in one single deployment. After"},{"line_number":11,"context_line":"    that, the L2 population ``agent_boot_time`` config option will no longer"},{"line_number":12,"context_line":"    take effect."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openvswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_6f4a4084","line":12,"updated":"2019-03-19 12:57:15.000000000","message":"The strange thing about release notes is that this one note will be spread-out over 4 different areas, making each hard to relate to each other.  So they should be descriptive enough on a standalone basis.  For that reason I would move this piece down into the \u0027critical\u0027 section instead, with the slight re-wording.","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"},{"author":{"_account_id":9531,"name":"liuyulong","display_name":"LIU Yulong","email":"i@liuyulong.me","username":"LIU-Yulong"},"change_message_id":"68097a620af40401f56f8ea6f61e45400d147ae8","unresolved":false,"context_lines":[{"line_number":9,"context_line":"    Please ensure that all neutron-server and neutron-openvswitch-agent"},{"line_number":10,"context_line":"    are upgraded with same package version in one single deployment. After"},{"line_number":11,"context_line":"    that, the L2 population ``agent_boot_time`` config option will no longer"},{"line_number":12,"context_line":"    take effect."},{"line_number":13,"context_line":"critical:"},{"line_number":14,"context_line":"  - |"},{"line_number":15,"context_line":"    The neutron-openvswitch-agent can sometimes spend too much time handling"}],"source_content_type":"text/x-yaml","patch_set":17,"id":"5fc1f717_40e5d3db","line":12,"in_reply_to":"5fc1f717_6f4a4084","updated":"2019-03-19 14:25:16.000000000","message":"Done","commit_id":"f7343120e8e1a31e591849c457bfb504d276d518"}]}
