)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":6062,"name":"jichenjc","email":"jichenjc@cn.ibm.com","username":"jichenjc"},"change_message_id":"ccbc851f1e15666e9f1d0608289041caddfcc195","unresolved":false,"context_lines":[{"line_number":20,"context_line":"operation anyway."},{"line_number":21,"context_line":""},{"line_number":22,"context_line":"Also, the current implementation is buggy since a live migration is not"},{"line_number":23,"context_line":"considered as an io_ops (the MIGRATING task_state is missing in the conditional...)"},{"line_number":24,"context_line":"so that\u0027s a good opportunity for just saying bye-bye to the conditional."},{"line_number":25,"context_line":""},{"line_number":26,"context_line":"Partially-Implements: blueprint request-spec-object"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":14,"id":"9a41bdd9_cff940ca","line":23,"updated":"2015-07-17 20:02:57.000000000","message":"I would say not all live-migration are IO related\nor at least for some hypervisors if they are sharing the disk, three would be no IO at all\n\ncan we take into live-migration with some flags ?","commit_id":"2c4031b3a1581dd70106c086b4e70e9d82e36efc"}],"nova/scheduler/host_manager.py":[{"author":{"_account_id":5441,"name":"Andrew Laski","email":"andrew@lascii.com","username":"alaski"},"change_message_id":"b6bd36ca9fdd0aea252e33862529b28810468751","unresolved":false,"context_lines":[{"line_number":292,"context_line":"        # NOTE(sbauza): By considering all cases when the scheduler is called"},{"line_number":293,"context_line":"        # and when consume_from_instance() is run, we can safely say that there"},{"line_number":294,"context_line":"        # is always an IO operation because we want to move the instance, so we"},{"line_number":295,"context_line":"        # don\u0027t need to check neither vm_state nor task_state"},{"line_number":296,"context_line":"        self.num_io_ops +\u003d 1"},{"line_number":297,"context_line":""},{"line_number":298,"context_line":"    def __repr__(self):"}],"source_content_type":"text/x-python","patch_set":1,"id":"5afe65bd_8cae7496","line":295,"updated":"2015-06-04 16:47:13.000000000","message":"If this is only called from the scheduler it\u0027s fine.  Otherwise it would be good to add an \u0027increment_iops\u0027 parameter to consume_from_instance() to indicate when it should be incremented.","commit_id":"382bac5d56d6b089cfc00e80225727f7787cd24b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"f7a38ce8af58a416c973aac3b55c44ce24ddd4a5","unresolved":false,"context_lines":[{"line_number":292,"context_line":"        # NOTE(sbauza): By considering all cases when the scheduler is called"},{"line_number":293,"context_line":"        # and when consume_from_instance() is run, we can safely say that there"},{"line_number":294,"context_line":"        # is always an IO operation because we want to move the instance, so we"},{"line_number":295,"context_line":"        # don\u0027t need to check neither vm_state nor task_state"},{"line_number":296,"context_line":"        self.num_io_ops +\u003d 1"},{"line_number":297,"context_line":""},{"line_number":298,"context_line":"    def __repr__(self):"}],"source_content_type":"text/x-python","patch_set":1,"id":"3afb71cf_0a52ffdd","line":295,"in_reply_to":"5afe65bd_8cae7496","updated":"2015-06-04 19:58:39.000000000","message":"Yup, consume_from_instance() is only called by FilterScheduler once it finds the destination hosts for each instance.","commit_id":"382bac5d56d6b089cfc00e80225727f7787cd24b"},{"author":{"_account_id":5441,"name":"Andrew Laski","email":"andrew@lascii.com","username":"alaski"},"change_message_id":"b5ae8bbf077340d81b160c55e266853b182c6aab","unresolved":false,"context_lines":[{"line_number":292,"context_line":"        # NOTE(sbauza): By considering all cases when the scheduler is called"},{"line_number":293,"context_line":"        # and when consume_from_instance() is run, we can safely say that there"},{"line_number":294,"context_line":"        # is always an IO operation because we want to move the instance, so we"},{"line_number":295,"context_line":"        # don\u0027t need to check neither vm_state nor task_state"},{"line_number":296,"context_line":"        self.num_io_ops +\u003d 1"},{"line_number":297,"context_line":""},{"line_number":298,"context_line":"    def __repr__(self):"}],"source_content_type":"text/x-python","patch_set":3,"id":"3afb71cf_d5329fa6","line":295,"updated":"2015-06-09 21:17:53.000000000","message":"I would leave off the last part.  It makes sense in the context of the change but a year from now people may wonder why vm_state and task_state are being considered here.","commit_id":"df9aaa32e20a87b00c77146e97cd4d4e49dcc7e5"}]}