)]}'
{"specs/queens/approved/aggregate-affinity.rst":[{"author":{"_account_id":13692,"name":"Roman Dobosz","email":"gryf73@gmail.com","username":"gryf"},"change_message_id":"8edbe9e7f088674acaa28ef6a9364814d69f7f9a","unresolved":false,"context_lines":[{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":58,"context_line":"indicate to the nova scheduler that instances within the server group should be"},{"line_number":59,"context_line":"placed on compute nodes that are associated with the same aggregates. The"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bf8cb3f7_2c12f159","line":56,"updated":"2017-12-21 07:59:29.000000000","message":"How about adding `soft-aggregate-affinity` and `soft-aggregate-anti-affinity` polices?","commit_id":"dda575d7cade69d4aea8f216373dae68d0eabbc3"},{"author":{"_account_id":13692,"name":"Roman Dobosz","email":"gryf73@gmail.com","username":"gryf"},"change_message_id":"72a88764d756e78637173bfdcda3aba9f69c6c81","unresolved":false,"context_lines":[{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":58,"context_line":"indicate to the nova scheduler that instances within the server group should be"},{"line_number":59,"context_line":"placed on compute nodes that are associated with the same aggregates. The"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bf8cb3f7_f645a418","line":56,"in_reply_to":"bf8cb3f7_10be1065","updated":"2017-12-21 19:33:16.000000000","message":"Adding two additional polices will have a need for a spec, right? Why don\u0027t we add those into this spec and implement as separate patch? Also this spec seems like a perfect place for those polices :)","commit_id":"dda575d7cade69d4aea8f216373dae68d0eabbc3"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"374f5ea945295ab28bffb608c0b7411474aac2d4","unresolved":false,"context_lines":[{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":58,"context_line":"indicate to the nova scheduler that instances within the server group should be"},{"line_number":59,"context_line":"placed on compute nodes that are associated with the same aggregates. The"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bf8cb3f7_10be1065","line":56,"in_reply_to":"bf8cb3f7_2c12f159","updated":"2017-12-21 18:15:24.000000000","message":"Those can be added afterwards if need be. I don\u0027t want to stuff too much into this spec.","commit_id":"dda575d7cade69d4aea8f216373dae68d0eabbc3"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"564c244a52378c5dc3a6306d1bd3b62120f84a6c","unresolved":false,"context_lines":[{"line_number":136,"context_line":"  jaypipes"},{"line_number":137,"context_line":""},{"line_number":138,"context_line":"Other contributors:"},{"line_number":139,"context_line":"  jroll"},{"line_number":140,"context_line":""},{"line_number":141,"context_line":"Work Items"},{"line_number":142,"context_line":"----------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"bf8cb3f7_4012aeb3","line":139,"updated":"2017-12-20 13:43:06.000000000","message":"Fun!","commit_id":"dda575d7cade69d4aea8f216373dae68d0eabbc3"}],"specs/rocky/approved/aggregate-affinity.rst":[{"author":{"_account_id":11564,"name":"Chris Dent","email":"cdent@anticdent.org","username":"chdent"},"change_message_id":"0da913ee59e308c07e85b9c81bd1c4742bb9af95","unresolved":false,"context_lines":[{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"},{"line_number":61,"context_line":"placed on compute nodes that are associated with the same aggregates. The"},{"line_number":62,"context_line":"`aggregate-anti-affinity` policy indicates that instances in the server group"},{"line_number":63,"context_line":"should be placed on compute nodes that have *different* aggregate associations."},{"line_number":64,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9f91af0f_6adaf417","line":61,"updated":"2018-01-03 11:05:59.000000000","message":"Is \"associated with the same aggregates\" _all_ the same aggregates, a particular aggregate, complete subsets, incomplete subsets?\n\nThat is, what happens when some compute nodes are in more than set(N) aggregate, but others are in just set(N)?","commit_id":"25d5a744e45e396a4c317832e9e67cf593a38fef"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"cd262313985ef36842f8b028ee7d637324f800d1","unresolved":false,"context_lines":[{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"},{"line_number":61,"context_line":"placed on compute nodes that are associated with the same aggregates. The"},{"line_number":62,"context_line":"`aggregate-anti-affinity` policy indicates that instances in the server group"},{"line_number":63,"context_line":"should be placed on compute nodes that have *different* aggregate associations."},{"line_number":64,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9f91af0f_fde1e72b","line":61,"in_reply_to":"9f91af0f_6adaf417","updated":"2018-01-03 15:38:33.000000000","message":"excellent question. the way I was thinking of it to start with would be aggregate-affinity means \"place this instance on any compute node that is associated with ANY aggregate that is associated to compute nodes that ALL instances in this group belong to.\"\n\nAnd anti-aggregate-affinity would mean \"place this instance on a compute node that is associated with an aggregate that is NOT associated with a compute node that ANY instance in the group is on.\"\n\nMake sense?","commit_id":"25d5a744e45e396a4c317832e9e67cf593a38fef"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c2c258910c0fb81679ff80ccaa19fc1d6451ff72","unresolved":false,"context_lines":[{"line_number":28,"context_line":"nova-compute service, that means Ironic baremetal nodes cannot be adequately"},{"line_number":29,"context_line":"grouped together."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Secondly, if a user who spins up a baremetal instance wishes to ensure that"},{"line_number":32,"context_line":"another baremetal instance is either in or not in the same power failure domain"},{"line_number":33,"context_line":"as the other instance (for isolation, latency or redundancy purposes), there is"},{"line_number":34,"context_line":"currently no ability to use the server group functionality in Nova to ask for"},{"line_number":35,"context_line":"this because nova considers the nova-compute service to be equal to the compute"},{"line_number":36,"context_line":"node. Therefore a server group with a policy of affinity attempts to put all"},{"line_number":37,"context_line":"instances in the group on the same service host. With Ironic, clearly this"},{"line_number":38,"context_line":"doesn\u0027t make any sense."},{"line_number":39,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"df7087c5_bb5fefe3","line":36,"range":{"start_line":31,"start_character":0,"end_line":36,"end_character":5},"updated":"2018-03-20 16:31:45.000000000","message":"sure, but say we fix the aggregate problem accepting nodes (and not only services), operators could use AZs right?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"637dfb18a89392e2b409794bb823582c47f80e7c","unresolved":false,"context_lines":[{"line_number":28,"context_line":"nova-compute service, that means Ironic baremetal nodes cannot be adequately"},{"line_number":29,"context_line":"grouped together."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Secondly, if a user who spins up a baremetal instance wishes to ensure that"},{"line_number":32,"context_line":"another baremetal instance is either in or not in the same power failure domain"},{"line_number":33,"context_line":"as the other instance (for isolation, latency or redundancy purposes), there is"},{"line_number":34,"context_line":"currently no ability to use the server group functionality in Nova to ask for"},{"line_number":35,"context_line":"this because nova considers the nova-compute service to be equal to the compute"},{"line_number":36,"context_line":"node. Therefore a server group with a policy of affinity attempts to put all"},{"line_number":37,"context_line":"instances in the group on the same service host. With Ironic, clearly this"},{"line_number":38,"context_line":"doesn\u0027t make any sense."},{"line_number":39,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"5f7c97a3_d23940f2","line":36,"range":{"start_line":31,"start_character":0,"end_line":36,"end_character":5},"in_reply_to":"df7087c5_bb5fefe3","updated":"2018-06-07 22:31:23.000000000","message":"AZ doesn\u0027t work here, you have to explicitly request a named one, that is not what you as an operator want your users doing. You just want several groups of servers to be in different power domains, etc.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"253b58006a19e48cb62d4c171f533d77f336c062","unresolved":false,"context_lines":[{"line_number":50,"context_line":""},{"line_number":51,"context_line":"This proposal extends the server group functionality in nova to allow some new"},{"line_number":52,"context_line":"policies:"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":"* `soft-aggregate-affinity`"},{"line_number":57,"context_line":"* `soft-aggregate-anti-affinity`"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_cfd8f880","line":57,"range":{"start_line":53,"start_character":0,"end_line":57,"end_character":32},"updated":"2018-02-16 15:29:05.000000000","message":"Will these policies mutually exclusive with all the existing policies? (I server group can have more than one policy by the API). Todays policies are all mutually exclusive.\n\nI can imagine a server group that has both anti-affinity and aggregate-affinity policy. That would instruct the nova scheduler to place the servers in the group to different compute nodes inside the same aggregate.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c2c258910c0fb81679ff80ccaa19fc1d6451ff72","unresolved":false,"context_lines":[{"line_number":50,"context_line":""},{"line_number":51,"context_line":"This proposal extends the server group functionality in nova to allow some new"},{"line_number":52,"context_line":"policies:"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":"* `soft-aggregate-affinity`"},{"line_number":57,"context_line":"* `soft-aggregate-anti-affinity`"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"df7087c5_5bf333e8","line":57,"range":{"start_line":53,"start_character":0,"end_line":57,"end_character":32},"in_reply_to":"1f9dbf25_d391aff3","updated":"2018-03-20 16:31:45.000000000","message":"Like I said above, you already have a lot of ways to restrict groups of hosts using existing filters or AZs (if you want to have that isolation being known by users), or you can just have custom filters, right?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"e8e6fd8fb3666b54671bab6b29f7189a13c6c6fa","unresolved":false,"context_lines":[{"line_number":50,"context_line":""},{"line_number":51,"context_line":"This proposal extends the server group functionality in nova to allow some new"},{"line_number":52,"context_line":"policies:"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":"* `soft-aggregate-affinity`"},{"line_number":57,"context_line":"* `soft-aggregate-anti-affinity`"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"1f9dbf25_d391aff3","line":57,"range":{"start_line":53,"start_character":0,"end_line":57,"end_character":32},"in_reply_to":"3fa0c359_52d0c41f","updated":"2018-02-27 22:38:55.000000000","message":"Good to know that we restricted the API.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"86dfdf26e442d4ecd210cde64496549d78fe0fe5","unresolved":false,"context_lines":[{"line_number":50,"context_line":""},{"line_number":51,"context_line":"This proposal extends the server group functionality in nova to allow some new"},{"line_number":52,"context_line":"policies:"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":"* `soft-aggregate-affinity`"},{"line_number":57,"context_line":"* `soft-aggregate-anti-affinity`"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_52d0c41f","line":57,"range":{"start_line":53,"start_character":0,"end_line":57,"end_character":32},"in_reply_to":"3fa0c359_cfd8f880","updated":"2018-02-20 18:12:24.000000000","message":"\u003e Will these policies mutually exclusive with all the existing\n \u003e policies? (I server group can have more than one policy by the\n \u003e API).\n\nWe don\u0027t actually allow a server group to have \u003e1 policy. That was a non-sensical design choice to begin with and we now don\u0027t allow \u003e1 policy:\n\nhttps://github.com/openstack/nova/blob/master/nova/api/openstack/compute/schemas/server_groups.py#L29-L33","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"f923d8bfbb8734657d5d54dd8cafbf68cc744b92","unresolved":false,"context_lines":[{"line_number":50,"context_line":""},{"line_number":51,"context_line":"This proposal extends the server group functionality in nova to allow some new"},{"line_number":52,"context_line":"policies:"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":"* `soft-aggregate-affinity`"},{"line_number":57,"context_line":"* `soft-aggregate-anti-affinity`"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3f79a3b5_fa475679","line":57,"range":{"start_line":53,"start_character":0,"end_line":57,"end_character":32},"in_reply_to":"5f7c97a3_7516164c","updated":"2018-10-24 03:03:29.000000000","message":"@johnthetubaguy: no. an instance is \"added\" to a server group using the --group scheduler hint and only a single server group is allowed:\n\nhttps://github.com/openstack/nova/blob/master/nova/compute/api.py#L1099-L1108","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"637dfb18a89392e2b409794bb823582c47f80e7c","unresolved":false,"context_lines":[{"line_number":50,"context_line":""},{"line_number":51,"context_line":"This proposal extends the server group functionality in nova to allow some new"},{"line_number":52,"context_line":"policies:"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"* `aggregate-affinity`"},{"line_number":55,"context_line":"* `aggregate-anti-affinity`"},{"line_number":56,"context_line":"* `soft-aggregate-affinity`"},{"line_number":57,"context_line":"* `soft-aggregate-anti-affinity`"},{"line_number":58,"context_line":""},{"line_number":59,"context_line":"Server groups that are tagged with the `aggregate-affinity` policy will"},{"line_number":60,"context_line":"indicate to the nova scheduler that instances within the server group should be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"5f7c97a3_7516164c","line":57,"range":{"start_line":53,"start_character":0,"end_line":57,"end_character":32},"in_reply_to":"df7087c5_5bf333e8","updated":"2018-06-07 22:31:23.000000000","message":"Can a single server be in multiple groups?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"253b58006a19e48cb62d4c171f533d77f336c062","unresolved":false,"context_lines":[{"line_number":62,"context_line":"`aggregate-anti-affinity` policy indicates that instances in the server group"},{"line_number":63,"context_line":"should be placed on compute nodes that have *different* aggregate associations."},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The `soft-aggregate-*` variants will be best-effort placement and therefore"},{"line_number":66,"context_line":"will affect the weighing, not the filtering, of compute nodes."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"We also propose changing the implementation of the add-host-to-aggregate"},{"line_number":69,"context_line":"functionality to allow associating **compute nodes** (via the identifying"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_af3facb3","line":66,"range":{"start_line":65,"start_character":0,"end_line":66,"end_character":62},"updated":"2018-02-16 15:29:05.000000000","message":"Can we be a bit more specific here? Like:\nsoft-aggregate-affinity will minimize the number of aggregates the servers in the group are belonging to. For example when the initial aggregate is filled the next server will be booted to anther aggregate but then no server will be booted to a third aggregate until the second aggregate has place to host the newly requested servers.\n\nSimilar statement can be added for the soft-aggregate-anti-affinity","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"e8e6fd8fb3666b54671bab6b29f7189a13c6c6fa","unresolved":false,"context_lines":[{"line_number":62,"context_line":"`aggregate-anti-affinity` policy indicates that instances in the server group"},{"line_number":63,"context_line":"should be placed on compute nodes that have *different* aggregate associations."},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The `soft-aggregate-*` variants will be best-effort placement and therefore"},{"line_number":66,"context_line":"will affect the weighing, not the filtering, of compute nodes."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"We also propose changing the implementation of the add-host-to-aggregate"},{"line_number":69,"context_line":"functionality to allow associating **compute nodes** (via the identifying"}],"source_content_type":"text/x-rst","patch_set":6,"id":"1f9dbf25_7318db80","line":66,"range":{"start_line":65,"start_character":0,"end_line":66,"end_character":62},"in_reply_to":"3fa0c359_7292a059","updated":"2018-02-27 22:38:55.000000000","message":"Hm, after looking at the weigher code again it still seems to me that the returned weight of a host depends on the amount of members from the group on that host [1].\n\n[1] https://github.com/openstack/nova/blob/4bac420d85044b92df2be052b5795ae454d497e2/nova/scheduler/weights/affinity.py#L53","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"86dfdf26e442d4ecd210cde64496549d78fe0fe5","unresolved":false,"context_lines":[{"line_number":62,"context_line":"`aggregate-anti-affinity` policy indicates that instances in the server group"},{"line_number":63,"context_line":"should be placed on compute nodes that have *different* aggregate associations."},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The `soft-aggregate-*` variants will be best-effort placement and therefore"},{"line_number":66,"context_line":"will affect the weighing, not the filtering, of compute nodes."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"We also propose changing the implementation of the add-host-to-aggregate"},{"line_number":69,"context_line":"functionality to allow associating **compute nodes** (via the identifying"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_7292a059","line":66,"range":{"start_line":65,"start_character":0,"end_line":66,"end_character":62},"in_reply_to":"3fa0c359_af3facb3","updated":"2018-02-20 18:12:24.000000000","message":"\u003e Can we be a bit more specific here? Like:\n \u003e soft-aggregate-affinity will minimize the number of aggregates the\n \u003e servers in the group are belonging to. For example when the initial\n \u003e aggregate is filled the next server will be booted to anther\n \u003e aggregate but then no server will be booted to a third aggregate\n \u003e until the second aggregate has place to host the newly requested\n \u003e servers.\n\nThat\u0027s actually not what the weigher does, though. All the soft-affinity weighers do is give hosts that contain members of the instance group a weight of 1.0 and hosts that don\u0027t have a member of the instance group in it a weight of 0.0. Therefore, the effect is: place me on a host that contains other members of the instance group, but if there are no hosts with available space, put me elsewhere.\n\nhttps://github.com/openstack/nova/blob/master/nova/scheduler/weights/affinity.py#L72\n\nNote that CONF.scheduler_filters.soft_affinity_weight_multiplier is 1.0 by default...\n\nhttps://github.com/openstack/nova/blob/master/nova/conf/scheduler.py#L479\n\n \u003e Similar statement can be added for the soft-aggregate-anti-affinity","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c2c258910c0fb81679ff80ccaa19fc1d6451ff72","unresolved":false,"context_lines":[{"line_number":65,"context_line":"The `soft-aggregate-*` variants will be best-effort placement and therefore"},{"line_number":66,"context_line":"will affect the weighing, not the filtering, of compute nodes."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"We also propose changing the implementation of the add-host-to-aggregate"},{"line_number":69,"context_line":"functionality to allow associating **compute nodes** (via the identifying"},{"line_number":70,"context_line":"`hypervisor_hostname` field) to a host aggregate in addition to the current"},{"line_number":71,"context_line":"behaviour of only allowing nova-compute **service records** to be associated"},{"line_number":72,"context_line":"with a host aggregate."},{"line_number":73,"context_line":""},{"line_number":74,"context_line":"Alternatives"},{"line_number":75,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"df7087c5_fb1a4715","line":72,"range":{"start_line":68,"start_character":0,"end_line":72,"end_character":22},"updated":"2018-03-20 16:31:45.000000000","message":"that looks like the main problem to fix IMHO for the problem you said above.\n\nHaving server groups supporting aggregates is just a separate feature.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"f923d8bfbb8734657d5d54dd8cafbf68cc744b92","unresolved":false,"context_lines":[{"line_number":65,"context_line":"The `soft-aggregate-*` variants will be best-effort placement and therefore"},{"line_number":66,"context_line":"will affect the weighing, not the filtering, of compute nodes."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"We also propose changing the implementation of the add-host-to-aggregate"},{"line_number":69,"context_line":"functionality to allow associating **compute nodes** (via the identifying"},{"line_number":70,"context_line":"`hypervisor_hostname` field) to a host aggregate in addition to the current"},{"line_number":71,"context_line":"behaviour of only allowing nova-compute **service records** to be associated"},{"line_number":72,"context_line":"with a host aggregate."},{"line_number":73,"context_line":""},{"line_number":74,"context_line":"Alternatives"},{"line_number":75,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3f79a3b5_d5058b10","line":72,"range":{"start_line":68,"start_character":0,"end_line":72,"end_character":22},"in_reply_to":"5f7c97a3_f52306a6","updated":"2018-10-24 03:03:29.000000000","message":"I proposed this in a patch:\n\nhttps://review.openstack.org/#/c/526753/\n\nIt was shot down. Notes from Matt:\n\n\"The spec for this was never approved and there was disagreement at the Rocky PTG [1] on how this was being done (add a new microversion to note that both hosts and nodes can be in aggregates, or add a new node-specific aggregates API, or other). There was also discussion about modeling node aggregates in Placement and then using the placement pre-request filtering stuff for that which has landed in Rocky\"","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"637dfb18a89392e2b409794bb823582c47f80e7c","unresolved":false,"context_lines":[{"line_number":65,"context_line":"The `soft-aggregate-*` variants will be best-effort placement and therefore"},{"line_number":66,"context_line":"will affect the weighing, not the filtering, of compute nodes."},{"line_number":67,"context_line":""},{"line_number":68,"context_line":"We also propose changing the implementation of the add-host-to-aggregate"},{"line_number":69,"context_line":"functionality to allow associating **compute nodes** (via the identifying"},{"line_number":70,"context_line":"`hypervisor_hostname` field) to a host aggregate in addition to the current"},{"line_number":71,"context_line":"behaviour of only allowing nova-compute **service records** to be associated"},{"line_number":72,"context_line":"with a host aggregate."},{"line_number":73,"context_line":""},{"line_number":74,"context_line":"Alternatives"},{"line_number":75,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"5f7c97a3_f52306a6","line":72,"range":{"start_line":68,"start_character":0,"end_line":72,"end_character":22},"in_reply_to":"df7087c5_fb1a4715","updated":"2018-06-07 22:31:23.000000000","message":"Yeah, that\u0027s a good point. This bit we should do separately, and first. Then see what problems are left.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c2c258910c0fb81679ff80ccaa19fc1d6451ff72","unresolved":false,"context_lines":[{"line_number":93,"context_line":""},{"line_number":94,"context_line":"The only impact to the existing Compute API is the acceptance of the new"},{"line_number":95,"context_line":"server group policies `aggregate-affinity`, `aggregate-anti-affinity`,"},{"line_number":96,"context_line":"`soft-aggregate-affinity` and `soft-aggregate-anti-affinity`."},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"Security impact"},{"line_number":99,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"df7087c5_9bdf8b39","line":96,"updated":"2018-03-20 16:31:45.000000000","message":"maybe also an os-aggregates API modification, nope ?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"253b58006a19e48cb62d4c171f533d77f336c062","unresolved":false,"context_lines":[{"line_number":94,"context_line":"The only impact to the existing Compute API is the acceptance of the new"},{"line_number":95,"context_line":"server group policies `aggregate-affinity`, `aggregate-anti-affinity`,"},{"line_number":96,"context_line":"`soft-aggregate-affinity` and `soft-aggregate-anti-affinity`."},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"Security impact"},{"line_number":99,"context_line":"---------------"},{"line_number":100,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_6f9d2465","line":97,"updated":"2018-02-16 15:29:05.000000000","message":"Currently GET /os-aggregates/{aggregate_id} response contains a \u0027hosts\u0027 field. If we change the aggregates to contain nodes instead of hosts, then we have to evolve this API.\n\n\nhttps://developer.openstack.org/api-ref/compute/#show-aggregate-details","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":782,"name":"John Garbutt","email":"john@johngarbutt.com","username":"johngarbutt"},"change_message_id":"637dfb18a89392e2b409794bb823582c47f80e7c","unresolved":false,"context_lines":[{"line_number":94,"context_line":"The only impact to the existing Compute API is the acceptance of the new"},{"line_number":95,"context_line":"server group policies `aggregate-affinity`, `aggregate-anti-affinity`,"},{"line_number":96,"context_line":"`soft-aggregate-affinity` and `soft-aggregate-anti-affinity`."},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"Security impact"},{"line_number":99,"context_line":"---------------"},{"line_number":100,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"5f7c97a3_f558e631","line":97,"in_reply_to":"1f9dbf25_2e38aad1","updated":"2018-06-07 22:31:23.000000000","message":"Either way, we need to loose the validation, so I think that means a new microversion?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"86dfdf26e442d4ecd210cde64496549d78fe0fe5","unresolved":false,"context_lines":[{"line_number":94,"context_line":"The only impact to the existing Compute API is the acceptance of the new"},{"line_number":95,"context_line":"server group policies `aggregate-affinity`, `aggregate-anti-affinity`,"},{"line_number":96,"context_line":"`soft-aggregate-affinity` and `soft-aggregate-anti-affinity`."},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"Security impact"},{"line_number":99,"context_line":"---------------"},{"line_number":100,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_9295ec34","line":97,"in_reply_to":"3fa0c359_6f9d2465","updated":"2018-02-20 18:12:24.000000000","message":"Why?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"e8e6fd8fb3666b54671bab6b29f7189a13c6c6fa","unresolved":false,"context_lines":[{"line_number":94,"context_line":"The only impact to the existing Compute API is the acceptance of the new"},{"line_number":95,"context_line":"server group policies `aggregate-affinity`, `aggregate-anti-affinity`,"},{"line_number":96,"context_line":"`soft-aggregate-affinity` and `soft-aggregate-anti-affinity`."},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"Security impact"},{"line_number":99,"context_line":"---------------"},{"line_number":100,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"1f9dbf25_2e38aad1","line":97,"in_reply_to":"3fa0c359_9295ec34","updated":"2018-02-27 22:38:55.000000000","message":"I assume there is difference between compute host and compute node. So far we added only the compute hosts to the aggregates and now we will add the compute nodes instead. This feels like a semantic change about what the hosts field means in the response of the GET /os-aggregates/{aggregate_id} API call.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"f923d8bfbb8734657d5d54dd8cafbf68cc744b92","unresolved":false,"context_lines":[{"line_number":94,"context_line":"The only impact to the existing Compute API is the acceptance of the new"},{"line_number":95,"context_line":"server group policies `aggregate-affinity`, `aggregate-anti-affinity`,"},{"line_number":96,"context_line":"`soft-aggregate-affinity` and `soft-aggregate-anti-affinity`."},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"Security impact"},{"line_number":99,"context_line":"---------------"},{"line_number":100,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"3f79a3b5_f523e7a8","line":97,"in_reply_to":"5f7c97a3_f558e631","updated":"2018-10-24 03:03:29.000000000","message":"I added a new microversion in the patch that implemented this:\n\nhttps://review.openstack.org/#/c/529201/6/nova/api/openstack/compute/rest_api_version_history.rst","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"253b58006a19e48cb62d4c171f533d77f336c062","unresolved":false,"context_lines":[{"line_number":103,"context_line":"Notifications impact"},{"line_number":104,"context_line":"--------------------"},{"line_number":105,"context_line":""},{"line_number":106,"context_line":"None."},{"line_number":107,"context_line":""},{"line_number":108,"context_line":"Other end user impact"},{"line_number":109,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_2fa07ca5","line":106,"updated":"2018-02-16 15:29:05.000000000","message":"The current server_group.create and .delete versioned notifications contain a hosts field that is filled by the hosts that the server group has servers on. By adding aggregate policies we might want to add a similar field for aggregates.\n\nThe current aggregate.create .delete notifications also has a hosts field. If change the aggregate to contain nodes instead of hosts, then we have to evolve the notification payload as well.\n\nhttps://docs.openstack.org/nova/latest/reference/notifications.html#existing-versioned-notifications","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"e8e6fd8fb3666b54671bab6b29f7189a13c6c6fa","unresolved":false,"context_lines":[{"line_number":103,"context_line":"Notifications impact"},{"line_number":104,"context_line":"--------------------"},{"line_number":105,"context_line":""},{"line_number":106,"context_line":"None."},{"line_number":107,"context_line":""},{"line_number":108,"context_line":"Other end user impact"},{"line_number":109,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"1f9dbf25_ae511a11","line":106,"in_reply_to":"3fa0c359_12a9dc7a","updated":"2018-02-27 22:38:55.000000000","message":"It feels like the compute node is different thing from the compute host so we actually changing the semantics here.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"86dfdf26e442d4ecd210cde64496549d78fe0fe5","unresolved":false,"context_lines":[{"line_number":103,"context_line":"Notifications impact"},{"line_number":104,"context_line":"--------------------"},{"line_number":105,"context_line":""},{"line_number":106,"context_line":"None."},{"line_number":107,"context_line":""},{"line_number":108,"context_line":"Other end user impact"},{"line_number":109,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_12a9dc7a","line":106,"in_reply_to":"3fa0c359_2fa07ca5","updated":"2018-02-20 18:12:24.000000000","message":"I don\u0027t see why. We\u0027d just be putting compute nodes into the hosts field (along with compute *hosts* (nova-compute service workers).","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"253b58006a19e48cb62d4c171f533d77f336c062","unresolved":false,"context_lines":[{"line_number":113,"context_line":"Performance Impact"},{"line_number":114,"context_line":"------------------"},{"line_number":115,"context_line":""},{"line_number":116,"context_line":"When using the new aggregate affinity or anti-affinity policies for server"},{"line_number":117,"context_line":"groups, there will be an expensive check on the nova-compute service at spawn"},{"line_number":118,"context_line":"time to ensure the policy constraints were not violated. This check is even"},{"line_number":119,"context_line":"more expensive for aggregate affinity policies because additional DB/conductor"},{"line_number":120,"context_line":"round-trips are necessary from the nova-compute service to determine the"},{"line_number":121,"context_line":"aggregate associations for all hosts involved for all instances in the server"},{"line_number":122,"context_line":"group. As with the \"normal\" affinity policies, this pre-spawn check may be"},{"line_number":123,"context_line":"disabled with the `disable_group_policy_check_upcall` configuration option."},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"Upgrade impact"},{"line_number":126,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_2f05dc7a","line":123,"range":{"start_line":116,"start_character":0,"end_line":123,"end_character":75},"updated":"2018-02-16 15:29:05.000000000","message":"This feels like the wrong direction. As far as I understand we want to get rid of the upcalls as those are not working in multi cell environment. So introducing yet another set of upcalls are discouraged.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"86dfdf26e442d4ecd210cde64496549d78fe0fe5","unresolved":false,"context_lines":[{"line_number":113,"context_line":"Performance Impact"},{"line_number":114,"context_line":"------------------"},{"line_number":115,"context_line":""},{"line_number":116,"context_line":"When using the new aggregate affinity or anti-affinity policies for server"},{"line_number":117,"context_line":"groups, there will be an expensive check on the nova-compute service at spawn"},{"line_number":118,"context_line":"time to ensure the policy constraints were not violated. This check is even"},{"line_number":119,"context_line":"more expensive for aggregate affinity policies because additional DB/conductor"},{"line_number":120,"context_line":"round-trips are necessary from the nova-compute service to determine the"},{"line_number":121,"context_line":"aggregate associations for all hosts involved for all instances in the server"},{"line_number":122,"context_line":"group. As with the \"normal\" affinity policies, this pre-spawn check may be"},{"line_number":123,"context_line":"disabled with the `disable_group_policy_check_upcall` configuration option."},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"Upgrade impact"},{"line_number":126,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_92ae4c7f","line":123,"range":{"start_line":116,"start_character":0,"end_line":123,"end_character":75},"in_reply_to":"3fa0c359_2f05dc7a","updated":"2018-02-20 18:12:24.000000000","message":"This isn\u0027t new, though. This is how the server group functionality was designed, unfortunately. This spec doesn\u0027t change any of that.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"e8e6fd8fb3666b54671bab6b29f7189a13c6c6fa","unresolved":false,"context_lines":[{"line_number":113,"context_line":"Performance Impact"},{"line_number":114,"context_line":"------------------"},{"line_number":115,"context_line":""},{"line_number":116,"context_line":"When using the new aggregate affinity or anti-affinity policies for server"},{"line_number":117,"context_line":"groups, there will be an expensive check on the nova-compute service at spawn"},{"line_number":118,"context_line":"time to ensure the policy constraints were not violated. This check is even"},{"line_number":119,"context_line":"more expensive for aggregate affinity policies because additional DB/conductor"},{"line_number":120,"context_line":"round-trips are necessary from the nova-compute service to determine the"},{"line_number":121,"context_line":"aggregate associations for all hosts involved for all instances in the server"},{"line_number":122,"context_line":"group. As with the \"normal\" affinity policies, this pre-spawn check may be"},{"line_number":123,"context_line":"disabled with the `disable_group_policy_check_upcall` configuration option."},{"line_number":124,"context_line":""},{"line_number":125,"context_line":"Upgrade impact"},{"line_number":126,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"1f9dbf25_6e5b122e","line":123,"range":{"start_line":116,"start_character":0,"end_line":123,"end_character":75},"in_reply_to":"3fa0c359_92ae4c7f","updated":"2018-02-27 22:38:55.000000000","message":"Do you mean that this change does not need a new upcall it just use the same upcall that already exists for late policy check in compute manager?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"253b58006a19e48cb62d4c171f533d77f336c062","unresolved":false,"context_lines":[{"line_number":125,"context_line":"Upgrade impact"},{"line_number":126,"context_line":"--------------"},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"None."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":"Other deployer impact"},{"line_number":131,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_afad6ceb","line":128,"updated":"2018-02-16 15:29:05.000000000","message":"What will happen with the already existing aggregates that are containing nova-compute service records? Will those be converted during the upgrade to compute node records?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"86dfdf26e442d4ecd210cde64496549d78fe0fe5","unresolved":false,"context_lines":[{"line_number":125,"context_line":"Upgrade impact"},{"line_number":126,"context_line":"--------------"},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"None."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":"Other deployer impact"},{"line_number":131,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_d2ddb414","line":128,"in_reply_to":"3fa0c359_afad6ceb","updated":"2018-02-20 18:12:24.000000000","message":"no, nothing will happen to them. We will just be storing compute node names (the compute_nodes.hypervisor_hostname value technically) in the API database\u0027s aggregate_hosts table.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"e8e6fd8fb3666b54671bab6b29f7189a13c6c6fa","unresolved":false,"context_lines":[{"line_number":125,"context_line":"Upgrade impact"},{"line_number":126,"context_line":"--------------"},{"line_number":127,"context_line":""},{"line_number":128,"context_line":"None."},{"line_number":129,"context_line":""},{"line_number":130,"context_line":"Other deployer impact"},{"line_number":131,"context_line":"---------------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"1f9dbf25_ce490661","line":128,"in_reply_to":"3fa0c359_d2ddb414","updated":"2018-02-27 22:38:55.000000000","message":"So if there was an aggregate with compute hosts in it before the upgrade then this will not be changed to compute nodes during the upgrade. But after the upgrade the admin can go and add compute nodes to this aggregate as well. Is this the case?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":1063,"name":"Ed Leafe","email":"ed@leafe.com","username":"ed-leafe"},"change_message_id":"f09e4382442b6d0d5de99d83f84011353d8321ec","unresolved":false,"context_lines":[{"line_number":152,"context_line":"Work Items"},{"line_number":153,"context_line":"----------"},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"* Allow compute nodes, not just services, to be associated with an aggregate"},{"line_number":156,"context_line":"* Add new server group aggregate affinity policies to REST API"},{"line_number":157,"context_line":"* Add new scheduler filters that assess whether a host\u0027s aggregates comply with"},{"line_number":158,"context_line":"  the affinity constraints of the server group aggregate policy"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_f2d5c5e6","line":155,"range":{"start_line":155,"start_character":2,"end_line":155,"end_character":76},"updated":"2018-02-14 22:40:39.000000000","message":"would there need to be any disambiguation needed? Or is there not an issue with overlapping between host names and hypervisor_hostname?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"86dfdf26e442d4ecd210cde64496549d78fe0fe5","unresolved":false,"context_lines":[{"line_number":152,"context_line":"Work Items"},{"line_number":153,"context_line":"----------"},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"* Allow compute nodes, not just services, to be associated with an aggregate"},{"line_number":156,"context_line":"* Add new server group aggregate affinity policies to REST API"},{"line_number":157,"context_line":"* Add new scheduler filters that assess whether a host\u0027s aggregates comply with"},{"line_number":158,"context_line":"  the affinity constraints of the server group aggregate policy"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3fa0c359_d2061487","line":155,"range":{"start_line":155,"start_character":2,"end_line":155,"end_character":76},"in_reply_to":"3fa0c359_f2d5c5e6","updated":"2018-02-20 18:12:24.000000000","message":"there should not be overlap, no. if there is, it\u0027s the deployers problem. I have no knowledge of any Ironic *and* Nova deployments that name their Ironic nodes the same as their Nova hypervisor hosts.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c2c258910c0fb81679ff80ccaa19fc1d6451ff72","unresolved":false,"context_lines":[{"line_number":153,"context_line":"----------"},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"* Allow compute nodes, not just services, to be associated with an aggregate"},{"line_number":156,"context_line":"* Add new server group aggregate affinity policies to REST API"},{"line_number":157,"context_line":"* Add new scheduler filters that assess whether a host\u0027s aggregates comply with"},{"line_number":158,"context_line":"  the affinity constraints of the server group aggregate policy"},{"line_number":159,"context_line":"* Add new scheduler weighers that sort potential compute nodes based on whether"},{"line_number":160,"context_line":"  they are or are not in the same aggregate as other instances in the server"},{"line_number":161,"context_line":"  group"},{"line_number":162,"context_line":""},{"line_number":163,"context_line":"Dependencies"},{"line_number":164,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":6,"id":"df7087c5_bbbb6fe3","line":161,"range":{"start_line":156,"start_character":0,"end_line":161,"end_character":7},"updated":"2018-03-20 16:31:45.000000000","message":"the above items don\u0027t see to help the problem you said.","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"},{"author":{"_account_id":7,"name":"Jay Pipes","email":"jaypipes@gmail.com","username":"jaypipes"},"change_message_id":"f923d8bfbb8734657d5d54dd8cafbf68cc744b92","unresolved":false,"context_lines":[{"line_number":153,"context_line":"----------"},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"* Allow compute nodes, not just services, to be associated with an aggregate"},{"line_number":156,"context_line":"* Add new server group aggregate affinity policies to REST API"},{"line_number":157,"context_line":"* Add new scheduler filters that assess whether a host\u0027s aggregates comply with"},{"line_number":158,"context_line":"  the affinity constraints of the server group aggregate policy"},{"line_number":159,"context_line":"* Add new scheduler weighers that sort potential compute nodes based on whether"},{"line_number":160,"context_line":"  they are or are not in the same aggregate as other instances in the server"},{"line_number":161,"context_line":"  group"},{"line_number":162,"context_line":""},{"line_number":163,"context_line":"Dependencies"},{"line_number":164,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3f79a3b5_7537d7e5","line":161,"range":{"start_line":156,"start_character":0,"end_line":161,"end_character":7},"in_reply_to":"df7087c5_bbbb6fe3","updated":"2018-10-24 03:03:29.000000000","message":"How so?","commit_id":"413d26fa253d27b152f7feb45f87593e8b01314a"}]}