)]}'
{"specs/rocky/approved/nova_scheduler_logger.rst":[{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"b5a432d7f9bd5da1dbda264a783cb5d3456777d2","unresolved":false,"context_lines":[{"line_number":30,"context_line":"Proposed change"},{"line_number":31,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"This patch should introduce a new logging component (*SchedulerLogger*), which"},{"line_number":34,"context_line":"should be called by the FilterScheduler. Additionally all activated filter and"},{"line_number":35,"context_line":"weigher classes should be wrapped with callback functions to this logger"},{"line_number":36,"context_line":"(to keep track of the consecutive filter/weigher results for an instance)."}],"source_content_type":"text/x-rst","patch_set":2,"id":"5f7c97a3_7de86fb2","line":33,"range":{"start_line":33,"start_character":52,"end_line":33,"end_character":70},"updated":"2018-06-06 09:24:36.000000000","message":"what would help the new logger instead of the existing behaviour ?\nWhat\u0027s the main problem with the existing logs ? Why can\u0027t we just modify the logs and rather need a new logging component ?","commit_id":"1c993e03e9d7d86f17bf636c2cc10f09249fa107"},{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"bea3416cf86306505b2acf0e6c986e61ed95854e","unresolved":false,"context_lines":[{"line_number":30,"context_line":"Proposed change"},{"line_number":31,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":32,"context_line":""},{"line_number":33,"context_line":"This patch should introduce a new logging component (*SchedulerLogger*), which"},{"line_number":34,"context_line":"should be called by the FilterScheduler. Additionally all activated filter and"},{"line_number":35,"context_line":"weigher classes should be wrapped with callback functions to this logger"},{"line_number":36,"context_line":"(to keep track of the consecutive filter/weigher results for an instance)."}],"source_content_type":"text/x-rst","patch_set":2,"id":"5f7c97a3_dc773c9f","line":33,"range":{"start_line":33,"start_character":52,"end_line":33,"end_character":70},"in_reply_to":"5f7c97a3_7de86fb2","updated":"2018-06-06 15:46:20.000000000","message":"To me (and a bunch of operators at a previous Summit session) the biggest issue is that on failure it can be difficult to figure out exactly what filter failed and for what reason (especially if you have NUMA and PCI and topology restrictions and such), but if you turn on debug logs then you end up with too many logs.\n\nThis was the impetus behind https://review.openstack.org/#/c/331306 where it\u0027s quiet during the success path and emits detailed logs on failure.","commit_id":"1c993e03e9d7d86f17bf636c2cc10f09249fa107"},{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"49ce6f266cba72b9d13b0672b189ed950a15b8b6","unresolved":false,"context_lines":[{"line_number":38,"context_line":"configuration file (*log_scheduling_decisions*). The output for an scheduled"},{"line_number":39,"context_line":"instance should consider 3 possible cases:"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"+ when the scheduling process was **successful** AND there were"},{"line_number":42,"context_line":"  **at least 2 hosts** returned after filtering - the scheduler calculates"},{"line_number":43,"context_line":"  weights for those hosts"},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"+ when the scheduling process was **successful** AND there was"},{"line_number":46,"context_line":"  **only one host** returned after filtering - no weighing will happen"},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"+ when the scheduling process **failed**"},{"line_number":49,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"1f9dbf25_1ecba9a5","line":46,"range":{"start_line":41,"start_character":0,"end_line":46,"end_character":70},"updated":"2018-02-27 11:29:09.000000000","message":"Could you add some an explanation of why you think it would be useful to log information for the case where the scheduler was successful?  In the discussions around the spec linked by Matt the operators were concerned about the size of the logs resulting from logging all scheduler operations, and so the decision was made to log detailed information only on failure.\n\nAt a minimum, if you want detailed information for the success path then I would suggest a config option to turn it on/off (with it defaulting to off).","commit_id":"1c993e03e9d7d86f17bf636c2cc10f09249fa107"},{"author":{"_account_id":26936,"name":"Surya Seetharaman","email":"suryaseetharaman.9@gmail.com","username":"tssurya"},"change_message_id":"0b9dcb2227f55082edc0c2a252989b25cd415600","unresolved":false,"context_lines":[{"line_number":38,"context_line":"configuration file (*log_scheduling_decisions*). The output for an scheduled"},{"line_number":39,"context_line":"instance should consider 3 possible cases:"},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"+ when the scheduling process was **successful** AND there were"},{"line_number":42,"context_line":"  **at least 2 hosts** returned after filtering - the scheduler calculates"},{"line_number":43,"context_line":"  weights for those hosts"},{"line_number":44,"context_line":""},{"line_number":45,"context_line":"+ when the scheduling process was **successful** AND there was"},{"line_number":46,"context_line":"  **only one host** returned after filtering - no weighing will happen"},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"+ when the scheduling process **failed**"},{"line_number":49,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"5f7c97a3_663f7c87","line":46,"range":{"start_line":41,"start_character":0,"end_line":46,"end_character":70},"in_reply_to":"1f9dbf25_1ecba9a5","updated":"2018-06-06 11:33:29.000000000","message":"\u003e At a minimum, if you want detailed information for the success path\n \u003e then I would suggest a config option to turn it on/off (with it\n \u003e defaulting to off).\n\n+1","commit_id":"1c993e03e9d7d86f17bf636c2cc10f09249fa107"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"8607d067569cf2884d58b474e9ec8ed81ed54f4f","unresolved":false,"context_lines":[{"line_number":51,"context_line":"Alternatives"},{"line_number":52,"context_line":"------------"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"None."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"Data model impact"},{"line_number":57,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"3fa0c359_8f18f3d2","line":54,"updated":"2018-02-26 15:52:49.000000000","message":"You may want to read and understand what was proposed, and previously approved, in this other spec:\n\nhttps://specs.openstack.org/openstack/nova-specs/specs/newton/approved/improve-sched-logging.html\n\nAnd then reference that along with why your proposal is better than this alternative.","commit_id":"1c993e03e9d7d86f17bf636c2cc10f09249fa107"},{"author":{"_account_id":8768,"name":"Chris Friesen","email":"chris.friesen@windriver.com","username":"cbf123"},"change_message_id":"49ce6f266cba72b9d13b0672b189ed950a15b8b6","unresolved":false,"context_lines":[{"line_number":51,"context_line":"Alternatives"},{"line_number":52,"context_line":"------------"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"None."},{"line_number":55,"context_line":""},{"line_number":56,"context_line":"Data model impact"},{"line_number":57,"context_line":"-----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"1f9dbf25_be9ab56b","line":54,"in_reply_to":"3fa0c359_8f18f3d2","updated":"2018-02-27 11:29:09.000000000","message":"The previous code submission was here:\n\nhttps://review.openstack.org/#/c/331306\n\nIt sort of stalled out when I didn\u0027t have the time to pursue some of the other alternatives.  I think the idea of storing the logs in the RequestContext is worth looking at more closely.","commit_id":"1c993e03e9d7d86f17bf636c2cc10f09249fa107"}]}