)]}'
{"trove/backup/state.py":[{"author":{"_account_id":13157,"name":"Zhao Chao","email":"zhaochao1984@gmail.com","username":"zhaochao"},"change_message_id":"a834bded3073c11d6475f4b681e068c8360f3919","unresolved":false,"context_lines":[{"line_number":22,"context_line":"    COMPLETED \u003d \"COMPLETED\""},{"line_number":23,"context_line":"    FAILED \u003d \"FAILED\""},{"line_number":24,"context_line":"    DELETE_FAILED \u003d \"DELETE_FAILED\""},{"line_number":25,"context_line":"    RUNNING_STATES \u003d [BUILDING, SAVING]"},{"line_number":26,"context_line":"    END_STATES \u003d [COMPLETED, FAILED, DELETE_FAILED]"}],"source_content_type":"text/x-python","patch_set":1,"id":"bf659307_374f96be","line":25,"updated":"2018-03-26 04:54:53.000000000","message":"\u0027NEW\u0027 shouldn\u0027t be removed from \u0027RUNNING_STATES\u0027 if we\u0027re not setting the backup to \u0027BUILDING\u0027 immediately when guestagent received the RPC message.\n\nBy now, the \u0027NEW\u0027 state of a backup may last for the following periods:\n- at API side, the backup is created\n- API calls TaskManager to create it\n- TaskManager calls GuestAgent to create it\n- GuestAgent gets the RPC message, and still does not start actually \u0027stream_backup_to_storage\u0027\n\nSo according the bug report (and the suggestion by the reporter which I think it\u0027s correct), we should set backup state to \u0027BUILDING\u0027 immediately when GuestAgent received the RPC message, and at API side, more work should be done to correctly update the backup status and show instance status related to backing-up(the later one should be the reason why many Zuul check jobs failed).","commit_id":"4d12c017f946d1b484a481957b4dbfb15a87cd7f"}]}
