)]}'
{"test/probe/test_container_merge_policy_index.py":[{"author":{"_account_id":5189,"name":"Yuan Zhou","email":"dunk007@gmail.com","username":"yuan-zhou"},"change_message_id":"29f3aa34511fc9e37ff62ef4bee34e3c788975fe","unresolved":false,"context_lines":[{"line_number":87,"context_line":"        self.brain.put_container()"},{"line_number":88,"context_line":"        self.brain.start_first_half()"},{"line_number":89,"context_line":"        self.brain.stop_second_half()"},{"line_number":90,"context_line":"        self.brain.put_container()"},{"line_number":91,"context_line":"        self.brain.put_object()"},{"line_number":92,"context_line":"        self.brain.start_second_half()"},{"line_number":93,"context_line":"        # make sure we have some mannor of split brain"}],"source_content_type":"text/x-python","patch_set":5,"id":"AAAAXH%2F%2Fk60%3D","line":90,"updated":"2014-04-19 05:49:03.000000000","message":"I can understand the whole patch but in which scenario would this happen?\nI\u0027m a bit confused about the two put_container(). Generally when one container is created, there\u0027s only one put request. If some primary nodes are in failure, partial container databases will be created on hand-off nodes and then be replicated back later. So seems there will be only one policy in these container dbs.","commit_id":"c005d4e1a89fdb531eda332260ebfc8cb63d8698"},{"author":{"_account_id":1179,"name":"Clay Gerrard","email":"clay.gerrard@gmail.com","username":"clay-gerrard"},"change_message_id":"ca184c3995f37b46c5a7f24bcaaa3123628ca9bf","unresolved":false,"context_lines":[{"line_number":87,"context_line":"        self.brain.put_container()"},{"line_number":88,"context_line":"        self.brain.start_first_half()"},{"line_number":89,"context_line":"        self.brain.stop_second_half()"},{"line_number":90,"context_line":"        self.brain.put_container()"},{"line_number":91,"context_line":"        self.brain.put_object()"},{"line_number":92,"context_line":"        self.brain.start_second_half()"},{"line_number":93,"context_line":"        # make sure we have some mannor of split brain"}],"source_content_type":"text/x-python","patch_set":5,"id":"AAAAXH%2F%2FRjU%3D","line":90,"in_reply_to":"AAAAXH%2F%2Fk60%3D","updated":"2014-04-22 20:16:26.000000000","message":"only one policy once they\u0027re replicated, and you can see them sync up after the call to \"get_to_final_state\" - but any objects placed while there was disagreement may still be on disk in the wrong policy - that\u0027s why the reconciler has to find them and move them.","commit_id":"c005d4e1a89fdb531eda332260ebfc8cb63d8698"}]}