)]}'
{"neutron/agent/l3/agent.py":[{"author":{"_account_id":14957,"name":"Brad Behle","email":"behle@us.ibm.com","username":"behle"},"change_message_id":"682c5ccba9a7cdd52c4fcf236560ba288a5b333a","unresolved":false,"context_lines":[{"line_number":491,"context_line":"                removed \u003d self._safe_router_removed(update.id)"},{"line_number":492,"context_line":"                if not removed:"},{"line_number":493,"context_line":"                    self._resync_router(update.id, update.timestamp,"},{"line_number":494,"context_line":"                                        update.action)"},{"line_number":495,"context_line":"                else:"},{"line_number":496,"context_line":"                    # need to update timestamp of removed router in case"},{"line_number":497,"context_line":"                    # there are older events for the same router in the"}],"source_content_type":"text/x-python","patch_set":1,"id":"ba15a1d1_71a728bd","line":494,"updated":"2015-09-17 21:02:22.000000000","message":"I don\u0027t think just queueing up another RouterUpdate here is a good idea in this case where a router can not be found or removed.  If this fails the first time here, then when it is requeued won\u0027t it just fail again, and again, etc, in a cycle we will never break out of?  Setting fullsync\u003dTrue here has the advantage that there is some additional cleanup work that happens in fetch_and_sync_all_routers() in a response to fullsync\u003dTrue.","commit_id":"157f1c60d34e94ba19fdc36f94f30a82d7b07322"},{"author":{"_account_id":11825,"name":"Sudhakar Babu Gariganti","email":"sudhakar.gariganti@gmail.com","username":"sudhakar-gariganti"},"change_message_id":"ca1ef1787c40d8d4d681cbddc48b5ca6c1ec1389","unresolved":false,"context_lines":[{"line_number":491,"context_line":"                removed \u003d self._safe_router_removed(update.id)"},{"line_number":492,"context_line":"                if not removed:"},{"line_number":493,"context_line":"                    self._resync_router(update.id, update.timestamp,"},{"line_number":494,"context_line":"                                        update.action)"},{"line_number":495,"context_line":"                else:"},{"line_number":496,"context_line":"                    # need to update timestamp of removed router in case"},{"line_number":497,"context_line":"                    # there are older events for the same router in the"}],"source_content_type":"text/x-python","patch_set":1,"id":"ba15a1d1_8c997b7b","line":494,"in_reply_to":"ba15a1d1_71a728bd","updated":"2015-09-18 05:13:34.000000000","message":"Per my understanding, in the _safe_router_removed, it may so happen that runtime errors can happen during namespace cleanup. \nIf such kind of errors are bound to happen peristently, then I kind of agree with you.\n\nBut if a retry helps to process it successfully next time around, then its better to do this way than a full sync. full_sync is really terrible for scale setups. So its better to avoid it as much as possible.\n\n@Carl,\nYour thoughts would be helpful here.","commit_id":"157f1c60d34e94ba19fdc36f94f30a82d7b07322"},{"author":{"_account_id":7448,"name":"Carl Baldwin","email":"carl@ecbaldwin.net","username":"carl-baldwin"},"change_message_id":"f71c5cb84a917c49aca06803f4410660ee9e67cb","unresolved":false,"context_lines":[{"line_number":491,"context_line":"                removed \u003d self._safe_router_removed(update.id)"},{"line_number":492,"context_line":"                if not removed:"},{"line_number":493,"context_line":"                    self._resync_router(update.id, update.timestamp,"},{"line_number":494,"context_line":"                                        update.action)"},{"line_number":495,"context_line":"                else:"},{"line_number":496,"context_line":"                    # need to update timestamp of removed router in case"},{"line_number":497,"context_line":"                    # there are older events for the same router in the"}],"source_content_type":"text/x-python","patch_set":1,"id":"ba15a1d1_09e099f9","line":494,"in_reply_to":"ba15a1d1_8c997b7b","updated":"2015-09-21 21:27:40.000000000","message":"As it currently stands, each time a router fails it triggers an update of *all* of the routers.  It is already a cycle and, IMO, probably a worse one.  With this change, it will be more obvious in the logs that it is the same router(s) cycling which helps in debugging.  Also, the cycle is lower priority so the agent is using spare cycles.  It isn\u0027t ideal but I tend to think that this patch is a step in the right direction.\n\nMaybe we should have a periodic sync of *all* routers occasionally to \"take advantage of the additional cleanup work that happens in fetch_and_sync_all_routers\" but that should be independent of whether a given router sees some kind of exception.  We should have that discussion independently too.\n\nEven so, it is a bit late in the cycle to be changing behavior like this without a serious bug behind it.","commit_id":"157f1c60d34e94ba19fdc36f94f30a82d7b07322"},{"author":{"_account_id":1131,"name":"Brian Haley","email":"haleyb.dev@gmail.com","username":"brian-haley"},"change_message_id":"edc2b7ae1a50b68f8cb84cf9bbbd125b320672d0","unresolved":false,"context_lines":[{"line_number":461,"context_line":"                       priority\u003dqueue.PRIORITY_SYNC_ROUTERS_TASK):"},{"line_number":462,"context_line":"        update \u003d queue.RouterUpdate(router_id, priority,"},{"line_number":463,"context_line":"                                    timestamp\u003dtimestamp,"},{"line_number":464,"context_line":"                                    action\u003daction)"},{"line_number":465,"context_line":"        self._queue.add(update)"},{"line_number":466,"context_line":""},{"line_number":467,"context_line":"    def _process_router_update(self):"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa80f949_3a363304","line":464,"updated":"2015-11-03 17:31:01.000000000","message":"So why do we create a new object here and not just pass \u0027update\u0027 and set update.priority, then put it back in the queue ?\n\nAlso, it seems from below that sometimes we\u0027ll get here with an updated timestamp, sometimes we won\u0027t.  Do we always want to update that so this entry has a lower priority?","commit_id":"f83fc217ae1a4cf942703cf4e23e123a808631f2"},{"author":{"_account_id":11825,"name":"Sudhakar Babu Gariganti","email":"sudhakar.gariganti@gmail.com","username":"sudhakar-gariganti"},"change_message_id":"44e8f6fac8acb0e8c48143d8510bf719f850ee42","unresolved":false,"context_lines":[{"line_number":461,"context_line":"                       priority\u003dqueue.PRIORITY_SYNC_ROUTERS_TASK):"},{"line_number":462,"context_line":"        update \u003d queue.RouterUpdate(router_id, priority,"},{"line_number":463,"context_line":"                                    timestamp\u003dtimestamp,"},{"line_number":464,"context_line":"                                    action\u003daction)"},{"line_number":465,"context_line":"        self._queue.add(update)"},{"line_number":466,"context_line":""},{"line_number":467,"context_line":"    def _process_router_update(self):"}],"source_content_type":"text/x-python","patch_set":3,"id":"fa80f949_622394be","line":464,"in_reply_to":"fa80f949_3a363304","updated":"2015-11-04 10:59:34.000000000","message":"Good one. Will update the patch to reuse the existing object and update the timestamp as well.","commit_id":"f83fc217ae1a4cf942703cf4e23e123a808631f2"}],"neutron/tests/unit/agent/l3/test_agent.py":[{"author":{"_account_id":11682,"name":"Ryan Moats","email":"rmoats@us.ibm.com","username":"regXboi"},"change_message_id":"ac069ea7b4456284e6f04652c9d1dc1a4d761022","unresolved":false,"context_lines":[{"line_number":1780,"context_line":"        agent._process_router_update()"},{"line_number":1781,"context_line":"        self.assertFalse(agent.fullsync)"},{"line_number":1782,"context_line":"        self.assertFalse(agent._process_router_if_compatible.called)"},{"line_number":1783,"context_line":"        self.assertEqual(1, agent._queue.add.call_count)"},{"line_number":1784,"context_line":""},{"line_number":1785,"context_line":"    def test_process_routers_update_rpc_timeout_on_get_ext_net(self):"},{"line_number":1786,"context_line":"        agent \u003d l3_agent.L3NATAgent(HOSTNAME, self.conf)"}],"source_content_type":"text/x-python","patch_set":1,"id":"ba15a1d1_1f29279e","line":1783,"updated":"2015-09-17 14:57:44.000000000","message":"I would feel more comfortable here and at L1797 if we verified that the queue call that was made is what we expected.","commit_id":"157f1c60d34e94ba19fdc36f94f30a82d7b07322"},{"author":{"_account_id":11825,"name":"Sudhakar Babu Gariganti","email":"sudhakar.gariganti@gmail.com","username":"sudhakar-gariganti"},"change_message_id":"ca1ef1787c40d8d4d681cbddc48b5ca6c1ec1389","unresolved":false,"context_lines":[{"line_number":1780,"context_line":"        agent._process_router_update()"},{"line_number":1781,"context_line":"        self.assertFalse(agent.fullsync)"},{"line_number":1782,"context_line":"        self.assertFalse(agent._process_router_if_compatible.called)"},{"line_number":1783,"context_line":"        self.assertEqual(1, agent._queue.add.call_count)"},{"line_number":1784,"context_line":""},{"line_number":1785,"context_line":"    def test_process_routers_update_rpc_timeout_on_get_ext_net(self):"},{"line_number":1786,"context_line":"        agent \u003d l3_agent.L3NATAgent(HOSTNAME, self.conf)"}],"source_content_type":"text/x-python","patch_set":1,"id":"ba15a1d1_2cb94fcc","line":1783,"in_reply_to":"ba15a1d1_1f29279e","updated":"2015-09-18 05:13:34.000000000","message":"I doubt if I can do that, as _queue.add takes input an which is associated with a timestamp. And the timestamp is being calculated at runtime, which I cannot assign and assert in the UT.\n\nDo let me know if you have someway out, will be more than happy to change and cover that.","commit_id":"157f1c60d34e94ba19fdc36f94f30a82d7b07322"},{"author":{"_account_id":11682,"name":"Ryan Moats","email":"rmoats@us.ibm.com","username":"regXboi"},"change_message_id":"eeeee93deb312e2bf2e1c166beded2bfeb549a84","unresolved":false,"context_lines":[{"line_number":1780,"context_line":"        agent._process_router_update()"},{"line_number":1781,"context_line":"        self.assertFalse(agent.fullsync)"},{"line_number":1782,"context_line":"        self.assertFalse(agent._process_router_if_compatible.called)"},{"line_number":1783,"context_line":"        self.assertEqual(1, agent._queue.add.call_count)"},{"line_number":1784,"context_line":""},{"line_number":1785,"context_line":"    def test_process_routers_update_rpc_timeout_on_get_ext_net(self):"},{"line_number":1786,"context_line":"        agent \u003d l3_agent.L3NATAgent(HOSTNAME, self.conf)"}],"source_content_type":"text/x-python","patch_set":1,"id":"ba15a1d1_691ceac1","line":1783,"in_reply_to":"ba15a1d1_2cb94fcc","updated":"2015-09-19 18:56:29.000000000","message":"Can\u0027t you can mock the generation of the timestamp so that you have a known value to work with?","commit_id":"157f1c60d34e94ba19fdc36f94f30a82d7b07322"},{"author":{"_account_id":11825,"name":"Sudhakar Babu Gariganti","email":"sudhakar.gariganti@gmail.com","username":"sudhakar-gariganti"},"change_message_id":"32881b337bc6401ae1f58035fc7209125c9c34e1","unresolved":false,"context_lines":[{"line_number":1780,"context_line":"        agent._process_router_update()"},{"line_number":1781,"context_line":"        self.assertFalse(agent.fullsync)"},{"line_number":1782,"context_line":"        self.assertFalse(agent._process_router_if_compatible.called)"},{"line_number":1783,"context_line":"        self.assertEqual(1, agent._queue.add.call_count)"},{"line_number":1784,"context_line":""},{"line_number":1785,"context_line":"    def test_process_routers_update_rpc_timeout_on_get_ext_net(self):"},{"line_number":1786,"context_line":"        agent \u003d l3_agent.L3NATAgent(HOSTNAME, self.conf)"}],"source_content_type":"text/x-python","patch_set":1,"id":"ba15a1d1_6a092edd","line":1783,"in_reply_to":"ba15a1d1_691ceac1","updated":"2015-09-21 06:51:45.000000000","message":"Missed to see the point that \u0027update\u0027 is already mocked. I was able to use update.timestamp. Posted the new patch with the required check.","commit_id":"157f1c60d34e94ba19fdc36f94f30a82d7b07322"},{"author":{"_account_id":11159,"name":"Hong Hui Xiao","email":"honghui_xiao@yeah.net","username":"HongHuiXiao"},"change_message_id":"30347b92bf345705cd4d24123d032f9a156661f5","unresolved":false,"context_lines":[{"line_number":1812,"context_line":"        self._test_process_routers_update_rpc_timeout()"},{"line_number":1813,"context_line":""},{"line_number":1814,"context_line":"    def test_process_routers_update_rpc_timeout_on_get_ext_net(self):"},{"line_number":1815,"context_line":"        self._test_process_routers_update_rpc_timeout(ext_net_call\u003dTrue,"},{"line_number":1816,"context_line":"                                                      ext_net_call_failed\u003dTrue)"},{"line_number":1817,"context_line":""},{"line_number":1818,"context_line":"    def test_process_routers_update_router_deleted(self):"}],"source_content_type":"text/x-python","patch_set":5,"id":"fa80f949_f0e39931","line":1815,"updated":"2015-11-05 13:44:42.000000000","message":"It looks like you have modified 3 place in agent.py, but here only have 2 test case. router_remove fail is not covered here.","commit_id":"c59ae8300b1a77d244fdc47d10ba5962fae134df"},{"author":{"_account_id":11825,"name":"Sudhakar Babu Gariganti","email":"sudhakar.gariganti@gmail.com","username":"sudhakar-gariganti"},"change_message_id":"2067bf05521d001d8abc46f57d4574284b56e716","unresolved":false,"context_lines":[{"line_number":1812,"context_line":"        self._test_process_routers_update_rpc_timeout()"},{"line_number":1813,"context_line":""},{"line_number":1814,"context_line":"    def test_process_routers_update_rpc_timeout_on_get_ext_net(self):"},{"line_number":1815,"context_line":"        self._test_process_routers_update_rpc_timeout(ext_net_call\u003dTrue,"},{"line_number":1816,"context_line":"                                                      ext_net_call_failed\u003dTrue)"},{"line_number":1817,"context_line":""},{"line_number":1818,"context_line":"    def test_process_routers_update_router_deleted(self):"}],"source_content_type":"text/x-python","patch_set":5,"id":"7a740942_2d59b7fe","line":1815,"in_reply_to":"fa80f949_f0e39931","updated":"2015-12-07 09:02:19.000000000","message":"Covered the router_removal failure case also.","commit_id":"c59ae8300b1a77d244fdc47d10ba5962fae134df"}]}
