)]}'
{"doc/source/admin/support-compute.rst":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"915ad8df871352bafd5291be86ebed4c80f0ac27","unresolved":false,"context_lines":[{"line_number":15,"context_line":"   :maxdepth: 1"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"   troubleshooting/orphaned-allocations.rst"},{"line_number":18,"context_line":"   troubleshooting/rebuild-placement-db.rst"},{"line_number":19,"context_line":""},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"Compute service logging"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_238404d4","line":18,"updated":"2020-03-27 15:35:32.000000000","message":"I\u0027ve been trying to start a pattern of adding new troubleshooting entries as sub-entries to this bigger doc because the current page here is pretty huge when you\u0027re looking in a browser. It\u0027s more digestable with sub-pages/topics IMO.","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"28836656c10a8beeb9eb1bce00dcc476886cf25f","unresolved":false,"context_lines":[{"line_number":15,"context_line":"   :maxdepth: 1"},{"line_number":16,"context_line":""},{"line_number":17,"context_line":"   troubleshooting/orphaned-allocations.rst"},{"line_number":18,"context_line":"   troubleshooting/rebuild-placement-db.rst"},{"line_number":19,"context_line":""},{"line_number":20,"context_line":""},{"line_number":21,"context_line":"Compute service logging"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_4309482f","line":18,"in_reply_to":"df33271e_238404d4","updated":"2020-03-27 15:56:06.000000000","message":"Ah, thanks. I didn\u0027t catch those at the beginning of the page.","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"eaadaee67baa2e3395fe100a6ed28e45b56be429","unresolved":false,"context_lines":[{"line_number":542,"context_line":"  :oslo.config:option:`keystone_authtoken.service_token_roles` is configured"},{"line_number":543,"context_line":"  in cinder, glance and neutron)."},{"line_number":544,"context_line":""},{"line_number":545,"context_line":"Affinity policy not enforced with parallel requests"},{"line_number":546,"context_line":"---------------------------------------------------"},{"line_number":547,"context_line":""},{"line_number":548,"context_line":"Problem"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_4372e8e4","line":545,"range":{"start_line":545,"start_character":16,"end_line":545,"end_character":28},"updated":"2020-03-27 15:39:40.000000000","message":"I might use \"violated\" here. The scheduler tries to enforce affinity but isn\u0027t architected to today with parallel requests.","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"28836656c10a8beeb9eb1bce00dcc476886cf25f","unresolved":false,"context_lines":[{"line_number":542,"context_line":"  :oslo.config:option:`keystone_authtoken.service_token_roles` is configured"},{"line_number":543,"context_line":"  in cinder, glance and neutron)."},{"line_number":544,"context_line":""},{"line_number":545,"context_line":"Affinity policy not enforced with parallel requests"},{"line_number":546,"context_line":"---------------------------------------------------"},{"line_number":547,"context_line":""},{"line_number":548,"context_line":"Problem"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_e6286a87","line":545,"range":{"start_line":545,"start_character":16,"end_line":545,"end_character":28},"in_reply_to":"df33271e_4372e8e4","updated":"2020-03-27 15:56:06.000000000","message":"Makes sense, will change.","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"915ad8df871352bafd5291be86ebed4c80f0ac27","unresolved":false,"context_lines":[{"line_number":548,"context_line":"Problem"},{"line_number":549,"context_line":"~~~~~~~"},{"line_number":550,"context_line":""},{"line_number":551,"context_line":"Parallel server create requests for affinity land on the same host and servers"},{"line_number":552,"context_line":"go to the ``ACTIVE`` state even though the affinity policy was violated."},{"line_number":553,"context_line":""},{"line_number":554,"context_line":"Solution"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_43d668c5","line":551,"updated":"2020-03-27 15:35:32.000000000","message":"Could skin this a few ways - it\u0027s also a problem for an anti-affinity group, you could race and have servers land on the same host violating that kind of policy.\n\nIt\u0027s also more than server create - any kind of move operation could violate the policy if the servers are moved at the same time (evacuate, unshelve, cold/live migrate).\n\nThe only time we do the late affinity up-call check is during server create or evacuate though, and of those server create is the only one that reschedules to an alternate host.","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"28836656c10a8beeb9eb1bce00dcc476886cf25f","unresolved":false,"context_lines":[{"line_number":548,"context_line":"Problem"},{"line_number":549,"context_line":"~~~~~~~"},{"line_number":550,"context_line":""},{"line_number":551,"context_line":"Parallel server create requests for affinity land on the same host and servers"},{"line_number":552,"context_line":"go to the ``ACTIVE`` state even though the affinity policy was violated."},{"line_number":553,"context_line":""},{"line_number":554,"context_line":"Solution"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_2671b2aa","line":551,"in_reply_to":"df33271e_43d668c5","updated":"2020-03-27 15:56:06.000000000","message":"\u003e Could skin this a few ways - it\u0027s also a problem for an\n \u003e anti-affinity group, you could race and have servers land on the\n \u003e same host violating that kind of policy.\n\nAh, heh, I intentionally used \"affinity\" to mean both same host and different host not realizing that \"affinity\" only refers to the former as a named policy. Will add.\n\n \u003e It\u0027s also more than server create - any kind of move operation\n \u003e could violate the policy if the servers are moved at the same time\n \u003e (evacuate, unshelve, cold/live migrate).\n \u003e \n \u003e The only time we do the late affinity up-call check is during\n \u003e server create or evacuate though, and of those server create is the\n \u003e only one that reschedules to an alternate host.\n\nThat sounds like it means that there is only a \"Solution\" for the server create scenario. And the rest must wait until placement supports anti-/affinity. Were you thinking to mention those cases and then in the Solution section explain that the move requests can\u0027t be supported properly until placement supports anti-/affinity?","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"eaadaee67baa2e3395fe100a6ed28e45b56be429","unresolved":false,"context_lines":[{"line_number":551,"context_line":"Parallel server create requests for affinity land on the same host and servers"},{"line_number":552,"context_line":"go to the ``ACTIVE`` state even though the affinity policy was violated."},{"line_number":553,"context_line":""},{"line_number":554,"context_line":"Solution"},{"line_number":555,"context_line":"~~~~~~~~"},{"line_number":556,"context_line":""},{"line_number":557,"context_line":"Currently, affinity races are handled by what is called the \"late affinity"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_435b0853","line":554,"updated":"2020-03-27 15:39:40.000000000","message":"So the other solution you don\u0027t mention here, and what Tempest does, is when creating multiple servers do it in a multi-create request using the `min_count` parameter because in that case the scheduler can enforce affinity properly, it\u0027s the parallel request case that causes problems because the scheduler workers don\u0027t know what\u0027s going on in memory with each other.","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"28836656c10a8beeb9eb1bce00dcc476886cf25f","unresolved":false,"context_lines":[{"line_number":551,"context_line":"Parallel server create requests for affinity land on the same host and servers"},{"line_number":552,"context_line":"go to the ``ACTIVE`` state even though the affinity policy was violated."},{"line_number":553,"context_line":""},{"line_number":554,"context_line":"Solution"},{"line_number":555,"context_line":"~~~~~~~~"},{"line_number":556,"context_line":""},{"line_number":557,"context_line":"Currently, affinity races are handled by what is called the \"late affinity"}],"source_content_type":"text/x-rst","patch_set":3,"id":"df33271e_865fde0c","line":554,"in_reply_to":"df33271e_435b0853","updated":"2020-03-27 15:56:06.000000000","message":"Good point. I did know that but didn\u0027t think outside the box of separate requests. Will add.","commit_id":"bc647c29a0babb06d3a82fd3c0621c3f84db8dcd"}],"doc/source/admin/troubleshooting/affinity-policy-violated.rst":[{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"f13a7d1096e349a2c525649aaa3abb4e2de6db33","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Problem"},{"line_number":5,"context_line":"-------"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Parallel server create requests for affinity or anti-affinity land on the same"},{"line_number":8,"context_line":"host and servers go to the ``ACTIVE`` state even though the affinity or"},{"line_number":9,"context_line":"anti-affinity policy was violated."},{"line_number":10,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"df33271e_569d08ba","line":7,"range":{"start_line":7,"start_character":74,"end_line":7,"end_character":78},"updated":"2020-04-02 20:15:55.000000000","message":"nit: \"wrong\" since the same host is what you want for an `affinity` policy but not `anti-affinity`.","commit_id":"df216de6d9b195782be3cfc2d51296f3c4442b54"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"bc73ce6c93fca9c3b3ecfdd56ba06bc56187c426","unresolved":false,"context_lines":[{"line_number":4,"context_line":"Problem"},{"line_number":5,"context_line":"-------"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Parallel server create requests for affinity or anti-affinity land on the same"},{"line_number":8,"context_line":"host and servers go to the ``ACTIVE`` state even though the affinity or"},{"line_number":9,"context_line":"anti-affinity policy was violated."},{"line_number":10,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"df33271e_9ca32f00","line":7,"range":{"start_line":7,"start_character":74,"end_line":7,"end_character":78},"in_reply_to":"df33271e_569d08ba","updated":"2020-04-02 22:18:59.000000000","message":"Hm ... yeah this confused me because the late affinity check only checks things on one host, so I was thinking it can only check a violation where two servers ended up on the same host. But yet it also checks for a violation of affinity policy ... I don\u0027t quite understand that. I\u0027ll go look at it again.","commit_id":"df216de6d9b195782be3cfc2d51296f3c4442b54"},{"author":{"_account_id":6873,"name":"Matt Riedemann","email":"mriedem.os@gmail.com","username":"mriedem"},"change_message_id":"f13a7d1096e349a2c525649aaa3abb4e2de6db33","unresolved":false,"context_lines":[{"line_number":5,"context_line":"-------"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Parallel server create requests for affinity or anti-affinity land on the same"},{"line_number":8,"context_line":"host and servers go to the ``ACTIVE`` state even though the affinity or"},{"line_number":9,"context_line":"anti-affinity policy was violated."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Solution"},{"line_number":12,"context_line":"--------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"df33271e_36a044f4","line":9,"range":{"start_line":8,"start_character":60,"end_line":9,"end_character":14},"updated":"2020-04-02 20:15:55.000000000","message":"nit: simple shorthand I use is \"(anti-)affinity\" so you don\u0027t need to use an `or` everywhere.\n\n(later)\n\nI see you\u0027re using \"anti-/affinity\" below, same idea.","commit_id":"df216de6d9b195782be3cfc2d51296f3c4442b54"},{"author":{"_account_id":4690,"name":"melanie witt","display_name":"melwitt","email":"melwittt@gmail.com","username":"melwitt"},"change_message_id":"bc73ce6c93fca9c3b3ecfdd56ba06bc56187c426","unresolved":false,"context_lines":[{"line_number":5,"context_line":"-------"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"Parallel server create requests for affinity or anti-affinity land on the same"},{"line_number":8,"context_line":"host and servers go to the ``ACTIVE`` state even though the affinity or"},{"line_number":9,"context_line":"anti-affinity policy was violated."},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"Solution"},{"line_number":12,"context_line":"--------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"df33271e_bcab93d3","line":9,"range":{"start_line":8,"start_character":60,"end_line":9,"end_character":14},"in_reply_to":"df33271e_36a044f4","updated":"2020-04-02 22:18:59.000000000","message":"Ah yeah. I think your way looks better though. I\u0027ll respin this because of my OCD.","commit_id":"df216de6d9b195782be3cfc2d51296f3c4442b54"}]}