)]}'
{"test/probe/test_sharder.py":[{"author":{"_account_id":15343,"name":"Tim Burke","email":"tburke@nvidia.com","username":"tburke"},"change_message_id":"a1db5387a0ff243583ecbe63310d0553337226bd","unresolved":false,"context_lines":[{"line_number":1943,"context_line":"            self.assertIn(\u0027X-Backend-Sharding-State\u0027, headers)"},{"line_number":1944,"context_line":"            self.assertEqual(\u0027sharded\u0027, headers[\u0027X-Backend-Sharding-State\u0027])"},{"line_number":1945,"context_line":"        self.assert_container_listing(all_obj_names,"},{"line_number":1946,"context_line":"                                      container_name\u003dcontainer_name)"}],"source_content_type":"text/x-python","patch_set":1,"id":"9f6a8fd7_28a5f3b5","line":1946,"updated":"2018-04-27 20:06:29.000000000","message":"It\u0027ll be fascinating to\n\n- run the sharder on policy_2 server\n- check that shards get created with policy_2\n- run sharder on policy_1 server\n- check that shards have gone back to policy_1\n- run reconciler on all shards, check that all objects are in policy_1\n\nI\u0027m skeptical about step 4, but maybe we\u0027re saved by the fact that there\u0027s a step 0 of getting shard ranges into the policy_2 server when only policy_1 servers have enough objects to know to shard -- so hopefully the replication that would be required will get the storage-policy sorted before we ever shard??","commit_id":"478a31d3166a44233e73ba6aa55c7093222c6fa7"}]}
