)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":31542,"name":"Andrew Bonney","email":"andrew.bonney@bbc.co.uk","username":"andrewbonney"},"change_message_id":"f0163dc35403abbd8a0dcf3dfdca2a4f40d3608b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"b558b528_261b6277","updated":"2025-01-15 11:04:46.000000000","message":"I believe I would need some feedback/pointers to move this forward in an acceptable way. The segments_to_bind need filtering for routed segments. This can be achieved by checking the port for fixed IPs associated with segments, and checking host agents for physnet bindings, but I assume referencing the \u0027Baremetal Node\u0027 agent type may be too Ironic-specific to be hard coded here.","commit_id":"a5a564a4f33b6a2493c028710ff63472d9238a8a"},{"author":{"_account_id":31542,"name":"Andrew Bonney","email":"andrew.bonney@bbc.co.uk","username":"andrewbonney"},"change_message_id":"8d8c8cee485dfdda9716c287990f1cfc15ac7ad9","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"a9094b0e_8209ecf9","updated":"2025-01-14 12:33:59.000000000","message":"I believe the tox error may be a transient one, but I\u0027ll hold off re-running until someone familiar can confirm.","commit_id":"a5a564a4f33b6a2493c028710ff63472d9238a8a"},{"author":{"_account_id":31542,"name":"Andrew Bonney","email":"andrew.bonney@bbc.co.uk","username":"andrewbonney"},"change_message_id":"8e0b8b828863ba21559f68d6b7da419e7bf5bb05","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"769344b0_643bb9a8","updated":"2025-01-14 10:18:18.000000000","message":"I hope this is of interest. We\u0027re using it in a 2024.1 deployment and have tested it against standard VLAN networks, and VLANs within routed segment networks. I don\u0027t have local capability to check against a flat network to ensure there are no regressions.","commit_id":"a5a564a4f33b6a2493c028710ff63472d9238a8a"},{"author":{"_account_id":31542,"name":"Andrew Bonney","email":"andrew.bonney@bbc.co.uk","username":"andrewbonney"},"change_message_id":"13dbd223958df5fb73d082c9e08a092f7235265d","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"74803fb5_85cc38a5","updated":"2025-01-15 08:31:06.000000000","message":"Turns out \u0027context.segments_to_bind\u0027 in \u0027bind_port\u0027 isn\u0027t behaving quite as I thought. Just looking into this.","commit_id":"a5a564a4f33b6a2493c028710ff63472d9238a8a"},{"author":{"_account_id":31542,"name":"Andrew Bonney","email":"andrew.bonney@bbc.co.uk","username":"andrewbonney"},"change_message_id":"1e60b449881a35b835f8c6c6a240c96bc78326fd","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"51307b1d_64c89567","updated":"2025-02-12 09:13:33.000000000","message":"I\u0027ve added filtering of segments by subnet which we\u0027re using on Caracal/Dalmation deployments. If desirable, _is_link_valid could also be looped over for the case where a port is using deferred IP allocation.\n\nIf this looks useful and needs an RFE ticket or similar let me know, I just needed to prioritise getting something working for our deployments before refining things upstream.","commit_id":"d665d68d358072888d891b23882a46236ae7a6f7"},{"author":{"_account_id":5890,"name":"Doug Goldstein","email":"cardoe@cardoe.com","username":"cardoe"},"change_message_id":"d0ca8ccad79af1468fcc2307c61abdc59fe72777","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"25015f78_d2efd0be","updated":"2025-02-12 13:29:11.000000000","message":"I suspect this wouldn\u0027t interact well with https://specs.openstack.org/openstack/neutron-specs/specs/newton/routed-networks.html#l2-adjacency if you\u0027re attempting L2VNI by using VXLAN with VLAN segments. But VXLAN is completely out of the window on NGS so this isn\u0027t a negative review of your change. Just something of note for @james.denton@outlook.com and myself.","commit_id":"d8261fd3ec78d9b232e2e1e3769ea7d326174f5b"},{"author":{"_account_id":11655,"name":"Julia Kreger","email":"juliaashleykreger@gmail.com","username":"jkreger","status":"Flying to the moon with a Jetpack!"},"change_message_id":"2f33011c990d6635c6113dc86069813bb1719382","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"3571e8ea_40884783","updated":"2025-02-14 17:06:10.000000000","message":"Overall, looks good to me and the explanation makes sense.\n\nOne bit of feedback, lets try to do better about doing inline notes of what is going on and why. Given the influx of contribution for NGS, we likely need to make sure things are as readable and easy to follow as possible.","commit_id":"d8261fd3ec78d9b232e2e1e3769ea7d326174f5b"},{"author":{"_account_id":5890,"name":"Doug Goldstein","email":"cardoe@cardoe.com","username":"cardoe"},"change_message_id":"dc585dd562553029efdb92b67ddbb1f73234e193","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"ee9f6bef_ca259de0","updated":"2025-02-14 22:50:05.000000000","message":"Yeah this makes sense with how neutron operates today. I spoke with James about the change the other day and I\u0027ll have a forth coming bug written up about the behavior I was point out.","commit_id":"d8261fd3ec78d9b232e2e1e3769ea7d326174f5b"},{"author":{"_account_id":31542,"name":"Andrew Bonney","email":"andrew.bonney@bbc.co.uk","username":"andrewbonney"},"change_message_id":"9c244f768291eedffdd76e939b6f8970b9c1fac5","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"6f0e04c8_133a7d1c","in_reply_to":"25015f78_d2efd0be","updated":"2025-02-13 17:11:40.000000000","message":"If I\u0027m interpreting that correctly, this is a case where the network has segments defined, but subnets aren\u0027t mapped to those segments?\n\nProvided the port in question has no IP associated with it, or the IP is in a subnet which doesn\u0027t map to a segment directly, then the code at line 480 won\u0027t filter the segments and it should behave as it did before, just not hitting KeyErrors for networks with segments, although it would always pick the first segment as things stand.\n\nAs noted, one improvement which may be relevant to this use case would be to loop over _is_link_valid. As such, the segments could be filtered based on the physical networks which the switch in question has access to in cases where a port doesn\u0027t have a segment/subnet directly associated to it.","commit_id":"d8261fd3ec78d9b232e2e1e3769ea7d326174f5b"},{"author":{"_account_id":31542,"name":"Andrew Bonney","email":"andrew.bonney@bbc.co.uk","username":"andrewbonney"},"change_message_id":"6dacbe55aa771324099a7493ce88ce935ebd199b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"5a2b2c16_aae11705","in_reply_to":"6f0e04c8_133a7d1c","updated":"2025-02-14 08:44:54.000000000","message":"The corresponding Ironic patch was lacking in this regard. I\u0027ve just pushed a change which should handle this case.","commit_id":"d8261fd3ec78d9b232e2e1e3769ea7d326174f5b"}]}