)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"be4cfc002d7e685994143dd09df20bda2688519f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"720a8b6d_33b5728d","updated":"2022-10-26 07:58:00.000000000","message":"Generally looks very good for me. -1 just because of the testing part.","commit_id":"0c8ccb5191a2ebb93ed7601a161461f283d85b46"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"5a8f4379b6eb627de3893e9228c913ff9cc0f916","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"3d87cef4_62ab80f6","updated":"2022-11-02 10:41:10.000000000","message":"Basically looks ok, I will check the wip patch","commit_id":"65451382993acd2b4328318c23fda57477c46ae5"},{"author":{"_account_id":19118,"name":"Eran Kuris","email":"ekuris@redhat.com","username":"ekuris"},"change_message_id":"ad8c78c4985b27702703a168fa66bd72fb825a8e","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"ec6860b0_85a8238b","updated":"2022-10-31 08:42:06.000000000","message":"For which version are we planning this RFE?","commit_id":"65451382993acd2b4328318c23fda57477c46ae5"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"ed8e7651ace191c945fe90afdf21a6fa7f0778e9","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"b49d3e5f_befb0886","in_reply_to":"ec6860b0_85a8238b","updated":"2022-11-02 08:22:27.000000000","message":"For Antelope","commit_id":"65451382993acd2b4328318c23fda57477c46ae5"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"33d11158209ff807bebe4c92903e0b098de0409b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":4,"id":"d9c5129b_ca28733d","updated":"2022-11-02 13:34:37.000000000","message":"I have some questions","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"52e34d769852bfef157c71118038c1526817244c","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":6,"id":"74c4ae07_39b0c3e3","updated":"2022-11-04 12:28:44.000000000","message":"New PS to be submitted","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f09e790d5c5776abf9e5716872f7e2e42df46366","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":6,"id":"2957f3a4_b5a9c3d9","in_reply_to":"74c4ae07_39b0c3e3","updated":"2022-11-04 12:40:32.000000000","message":"ack ill re reivew again later.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9832cbc4f62816398090f4821a4a87f85d87142a","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"e77b8e68_e0d08c6a","updated":"2022-11-07 17:41:26.000000000","message":"I\u0027ll push a new PS tomorrow. I\u0027ll remove the section describing the script to heal the inventories. We\u0027ll reduce the scope to a simple script that will inform about the ports with min-bw rules in overlay networks.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"041402c8454832c5cf80c437a805b17b704a0b0f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"25ce4c72_aab9dd1c","updated":"2022-11-11 12:21:14.000000000","message":"Wanted to ask one more question: Do you plan to cover both placement and data plane enforcement or only placement?","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"847c93f4ae154a7f0274f5f17c95d20cdf9faf76","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":7,"id":"9db9cc65_b88f3810","in_reply_to":"25ce4c72_aab9dd1c","updated":"2022-11-11 12:43:28.000000000","message":"Dataplane enforcement is not considered in this spec.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"19d278c01ae532eb578399cbcaaa9e39019bdfe5","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":8,"id":"004e9ed3_4f7051d0","updated":"2022-11-14 17:17:32.000000000","message":"+1 over all but one comment on the trait that proably shoudl be adressed before this is accepted","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"e3cf4ca58f256e2ee186cef0c758c8a079911144","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":8,"id":"ecbefdc4_d275d6c9","updated":"2022-11-14 13:47:29.000000000","message":"See my comment below, let\u0027s wait for others votes","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"4ee28263cbcd59ce0f2e2e72ce5d9c621296bb33","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":10,"id":"df020697_c081ff9a","updated":"2022-11-15 16:26:03.000000000","message":"The placement modeling looks OK to me as well as the resource_request. Nova should be able to handle that transparently. (If not then I would consider that as a bug on nova side). I\u0027m not super happy about the fact that we ignore the upgrade impact but I guess I\u0027m alone in that so I won\u0027t try to block on that.\n\n","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"36c05f944541da980ee2245b8117f30b9f8fafef","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":10,"id":"e58faf0d_f059092e","updated":"2022-11-17 12:22:40.000000000","message":"\\o/","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"410ad8698dd1762db3ee2f8c1ddaddb77e99dc4e","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":10,"id":"b8cdf3e5_39db388a","in_reply_to":"4b82006c_bf5076a9","updated":"2022-11-15 16:59:05.000000000","message":"Ahh cool. I missed the fact that OVS prevented the problem. Cool then.","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9fa8e9189112d2581e7b2db869706f6f818aed73","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":10,"id":"4b82006c_bf5076a9","in_reply_to":"df020697_c081ff9a","updated":"2022-11-15 16:53:31.000000000","message":"No no, we already considered that:\n* OVS: we currently prevent that. It could not be possible to be in this case.\n* OVN: recently implemented, not so many deployments with port in tunnelled network and min-bw rules.\n\nIn any case, if I have examples how to execute the allocations heal, that could be perfect.","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"}],"specs/2023.1/strict-minimum-bandwidth-tunnelled-networks.rst":[{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"5a8f4379b6eb627de3893e9228c913ff9cc0f916","unresolved":true,"context_lines":[{"line_number":153,"context_line":"* Unit/functional Tests."},{"line_number":154,"context_line":"* Fullstack tests: increase the current fullstack tests coverage to check"},{"line_number":155,"context_line":"  this new feature."},{"line_number":156,"context_line":""},{"line_number":157,"context_line":""},{"line_number":158,"context_line":"Documentation Impact"},{"line_number":159,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"aae8e4f5_41c3b2f1","line":156,"updated":"2022-11-02 10:41:10.000000000","message":"tempest? https://opendev.org/openstack/tempest/src/branch/master/tempest/scenario/test_network_qos_placement.py","commit_id":"65451382993acd2b4328318c23fda57477c46ae5"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"15af4a9e8030516203af09410e358d3e29afd2a0","unresolved":false,"context_lines":[{"line_number":153,"context_line":"* Unit/functional Tests."},{"line_number":154,"context_line":"* Fullstack tests: increase the current fullstack tests coverage to check"},{"line_number":155,"context_line":"  this new feature."},{"line_number":156,"context_line":""},{"line_number":157,"context_line":""},{"line_number":158,"context_line":"Documentation Impact"},{"line_number":159,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"502752aa_feb2fe3c","line":156,"in_reply_to":"aae8e4f5_41c3b2f1","updated":"2022-11-02 10:51:12.000000000","message":"Right, I really don\u0027t know why I missed this part.","commit_id":"65451382993acd2b4328318c23fda57477c46ae5"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b47655397408bbc50b81b9fb11ca7f6ae02f0d0f","unresolved":true,"context_lines":[{"line_number":30,"context_line":"(flat or VLAN) but overlay networs (VXLAN and Geneve). Of course those"},{"line_number":31,"context_line":"deployments use ML2/OVS and ML2/OVN; ML2/SR-IOV does not support tunnelled"},{"line_number":32,"context_line":"networks. That leads to an existing gap in the currently implemented feature:"},{"line_number":33,"context_line":"there is no way to model tunnelled networks."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"Proposed Change"}],"source_content_type":"text/x-rst","patch_set":4,"id":"3d0849c8_71c0c2a3","line":33,"updated":"2022-11-03 15:23:07.000000000","message":"tunnles were ment to be supproted by https://specs.openstack.org/openstack/neutron-specs/specs/rocky/minimum-bandwidth-allocation-placement-api.html\n so im really surpised to see that its not?","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":30,"context_line":"(flat or VLAN) but overlay networs (VXLAN and Geneve). Of course those"},{"line_number":31,"context_line":"deployments use ML2/OVS and ML2/OVN; ML2/SR-IOV does not support tunnelled"},{"line_number":32,"context_line":"networks. That leads to an existing gap in the currently implemented feature:"},{"line_number":33,"context_line":"there is no way to model tunnelled networks."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"Proposed Change"}],"source_content_type":"text/x-rst","patch_set":4,"id":"e9eba511_5f7a7c16","line":33,"in_reply_to":"03af39a3_927912c0","updated":"2022-11-04 10:23:55.000000000","message":"The API [1] is for the QoS min BW rule support. But this is not related to the strict minium bandiwht feature. We of course use it for BW scheduling, but nothing else.\n\nIn Neutron we didn\u0027t support scheduling using ports in overlay networks. This is why I\u0027m introducing this RFE.\n\n[1]https://specs.openstack.org/openstack/neutron-specs/specs/newton/ml2-qos-minimum-egress-bw-support.html","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"28145a7534ec366eaf900c70b59e46823012ee71","unresolved":true,"context_lines":[{"line_number":30,"context_line":"(flat or VLAN) but overlay networs (VXLAN and Geneve). Of course those"},{"line_number":31,"context_line":"deployments use ML2/OVS and ML2/OVN; ML2/SR-IOV does not support tunnelled"},{"line_number":32,"context_line":"networks. That leads to an existing gap in the currently implemented feature:"},{"line_number":33,"context_line":"there is no way to model tunnelled networks."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"Proposed Change"}],"source_content_type":"text/x-rst","patch_set":4,"id":"27273954_b9baf604","line":33,"in_reply_to":"03af39a3_927912c0","updated":"2022-11-04 09:50:33.000000000","message":"the minium bandwith qos policy predates the guarneteed strict minium bandiwht feature. it was intoduced in newton https://specs.openstack.org/openstack/neutron-specs/specs/newton/ml2-qos-minimum-egress-bw-support.html that api was intend for use wiht all netowrk backends\n\n\ntunneled support was in the original design requiremnt for the strict/guarenteeed bandwidth feature\n\nhttps://review.opendev.org/c/openstack/neutron-specs/+/396297/7/specs/pike/strict-minimum-bandwidth-support.rst#48\n\n\n\"\"\"\nthe resources will have the following form, please note INGRESS is\nprovided as a reference but it\u0027s out of the scope of this spec:\n  NIC_BW_EGRESS.\u003cphysical-network\u003e\n  NIC_BW_INGRESS.\u003cphysical-network\u003e\nphysical-network will be the \"physnet\" in the reference implementation,\nor \"tunneling\" in the case of requesting bandwidth on the tunneling\npath.\n\"\"\"\n\nthat comment was filed for this bug which rodolfo created https://bugs.launchpad.net/neutron/+bug/1913180\n\nits not a reflection of the orginal intent just of a limiation in neutron due to an incomplete implementation.\n\nsomewhere between the pike spec and the queens spec https://review.opendev.org/c/openstack/neutron-specs/+/508149/ the refence to tunneling were remvoed but it was never declared out of scope in the spec or on any conversation taht i recall.\n\nso not it was not just ment to work with physnets.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"8ec529e039adf279c257377760d5600de1977803","unresolved":true,"context_lines":[{"line_number":30,"context_line":"(flat or VLAN) but overlay networs (VXLAN and Geneve). Of course those"},{"line_number":31,"context_line":"deployments use ML2/OVS and ML2/OVN; ML2/SR-IOV does not support tunnelled"},{"line_number":32,"context_line":"networks. That leads to an existing gap in the currently implemented feature:"},{"line_number":33,"context_line":"there is no way to model tunnelled networks."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":""},{"line_number":36,"context_line":"Proposed Change"}],"source_content_type":"text/x-rst","patch_set":4,"id":"03af39a3_927912c0","line":33,"in_reply_to":"3d0849c8_71c0c2a3","updated":"2022-11-04 08:31:13.000000000","message":"the original feature was supposed to work with networks that are backed with a physical interface, like physnet for OVS and physical device for SRIOV, and that excludes tunneled nets.\nSee the answer from ralonsoh @l95\nhttps://opendev.org/openstack/neutron/src/branch/master/neutron/services/qos/drivers/openvswitch/driver.py#L79","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"33d11158209ff807bebe4c92903e0b098de0409b","unresolved":true,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"b7fad626_390c7a3a","line":46,"updated":"2022-11-02 13:34:37.000000000","message":"Is it always true that the tunneled traffic uses one single host interface on a given host?","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"46aa6240_37981679","line":46,"in_reply_to":"03052501_b4f8f9a6","updated":"2022-11-04 10:23:55.000000000","message":"In this spec I\u0027m not considering the idea of having a shared RP for two network types. This is out of scope. This spec is adding the support for tunnelled networks only, as an independent HW resource. This spec does not consider (as it wasn\u0027t the original implementation of the BW aware scheduling in Neutron) sharing BW resources between networks.\n\nIn any case I would insist that the BW assignation is defined by the administrator. He/she would know how the underlying HW is configured and how it should be model.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"28145a7534ec366eaf900c70b59e46823012ee71","unresolved":true,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"7248ea33_685d2544","line":46,"in_reply_to":"03052501_b4f8f9a6","updated":"2022-11-04 09:50:33.000000000","message":"There is a long history but ooo should implement this today or there was a regression since it was implmented many years ago.\n\nIt was the reference topology we implemented in networking-ovs-dpdk and kolla-ansible when using dpdk.\n\ni belive would be required for hardware offloaded ovs to fully offload tunneled networks in hardware but i have not tested that.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"52e34d769852bfef157c71118038c1526817244c","unresolved":false,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"c346ee21_ab19de16","line":46,"in_reply_to":"3c419499_1f27dad1","updated":"2022-11-04 12:28:44.000000000","message":"That depends on the deployment. If admin can\u0027t have multiple devices for both tunnelled and physical networks, we can\u0027t model that. Again, we can\u0027t model a shared resource, placement API does not have this concept yet.\n\nThis spec is improving the scope of the current placement support in Neutron. Now we don\u0027t have a way to support overlay networks. This is what this spec is introducing.\n\nWith your conclusion, we should remove any support for placement: what if two physical networks use the same bond? Please, review the spec with the scope defined.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f651e3a366e9b65ac2d00f98183b5ad1bf522a77","unresolved":true,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"3c419499_1f27dad1","line":46,"in_reply_to":"46aa6240_37981679","updated":"2022-11-04 12:19:23.000000000","message":"\"\" This spec is adding the support for tunnelled networks only, as an independent HW resource. \"\"\n\nif that is the scope fo the spec i dont think we shoudl proceed with the spec.\n\ni agree that the admin need to define the bandwith via the config options but we need to ensure they can model either having a indepetnt pool of bandwith for tunnles or a shared pool with one of the existing physnets.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"40e6941bcfdf019d6185864ace71dfd9d80d9cb7","unresolved":false,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"e1042e3f_91e61d4c","line":46,"in_reply_to":"b7fad626_390c7a3a","updated":"2022-11-03 14:24:48.000000000","message":"Tunnelled traffic is sent to the interface that uses this \"local_ip\".\n\nIn ML2/OVS, the OVS service uses the local_ip to send this traffic to the kernel interfave\n\nIn ML2/OVN [1], the OVN controller reads the OVS config \"external_ids:ovn-encap-ip\", that is (from this doc):\n\"The  IP  address  that a chassis should use to connect to this node using encapsulation types specified  by  external_ids:ovn-encap-type\"\n\n[1]https://www.openvswitch.org/support/dist-docs-2.5/ovn-controller.8.txt","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"9d219ce47bf1db639f477c9408a9da43da57f28a","unresolved":true,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"03052501_b4f8f9a6","line":46,"in_reply_to":"ba92c100_fe80fcac","updated":"2022-11-04 08:51:38.000000000","message":"@gibi: yes, the local_ip is set on one interface (it can be bond).\n\nRegarding the best practice mentioned by Sean (residing the local_ip on br-ex), would be good to know how common is that, for example knowing if deployment tools do it or not.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f09e790d5c5776abf9e5716872f7e2e42df46366","unresolved":false,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"53cf3969_efe452d8","line":46,"in_reply_to":"c346ee21_ab19de16","updated":"2022-11-04 12:40:32.000000000","message":"placemnt can support what im proposing today.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"49a3ab20eb154198167492f28d68475351be29c9","unresolved":true,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"e60dcef8_e2ad13c1","line":46,"in_reply_to":"c6d0d07f_ad1f89d9","updated":"2022-11-03 17:37:23.000000000","message":"If both the physnet and the tunneled traffic use the same interface then modelling that as two separate RPs with bandwidth resource is problematic. At least it forces the deployer to statically split the bw pool of the interface to two parts (two RPs) one for the physnet traffic and one for the tunneled traffic.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b47655397408bbc50b81b9fb11ca7f6ae02f0d0f","unresolved":false,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"c6d0d07f_ad1f89d9","line":46,"in_reply_to":"e1042e3f_91e61d4c","updated":"2022-11-03 15:23:07.000000000","message":"the best pratics is for the local tunnel endpoint ip to be assitned to the br-ex \nso it is normally shared with a physnet bandwith.\n\n\nml2/linux bridge support multicast vxlan so that will not required to be a singel interface.\n\nfor ml2/ovs ml2/ovn (hardwar offloaed , kernel, or ovs-dpdk) one itnerface will have the local endpoint ip (typically a ovs brige like br-ex or br-phy)\nbut it can just be eth0  if you want it can also be a bond.\n\nso the bandwith cna come form multipel phsical intercases but logically it comes form one interface. or one pool of bandwidth.\n\nand as i said the normal and most efficent deployment case has a sinel br-ex shared btween the tunneled traifc and the tenant physnet.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"1964c3afd1b95213d56c5c65fb2e9f4e34b7997c","unresolved":true,"context_lines":[{"line_number":43,"context_line":"The referred backends handle the overlay traffic sending and receiving this"},{"line_number":44,"context_line":"traffic from a host interface, that acts as a VTEP [3]_. This host interface"},{"line_number":45,"context_line":"is identified by an IP address, known as \"local_ip\" in the ML2 plugin"},{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ba92c100_fe80fcac","line":46,"in_reply_to":"e60dcef8_e2ad13c1","updated":"2022-11-03 18:04:38.000000000","message":"ya you woudl have to staticaly partion it via config but it would not be enforced at the dataplane level.\n\nneutron need to be able to share a singel RP/invenotry of bandwith in the case where its shared instead of partioning for this to make sense without hurting performacne or preventign this form workign with ovs-dpdk or hardwawre offloaded ovs.\n\n\n\nhttps://meetings.opendev.org/irclogs/%23openstack-nova/%23openstack-nova.2022-11-03.log.html#t2022-11-03T17:41:36\n\n17:41 \u003cgibi\u003e sean-k-mooney: could there be more than one interface for the tunneled traffic ?\n17:43 \u003copendevreview\u003e Alex Chan proposed openstack/nova master: record action log when deleting shelved instance Closes-Bug: #1993736 Change-Id: I9ce18cbba5083c55d15d9b7c2a89133d227754ea  https://review.opendev.org/c/openstack/nova/+/863547\n17:46 \u003csean-k-mooney\u003e gibi: there is only one local tunnel ip\n17:46 \u003c-- SDrozdov (~SDrozdov@2a02:c7d:5047:ff00:fde2:9c6d:c433:c556) has quit (Remote host closed the connection)\n17:46 \u003csean-k-mooney\u003e gibi: so yes but its defiend by the routing table which interface is used\n17:46 \u003csean-k-mooney\u003e there are two way to do this\n17:47 \u003csean-k-mooney\u003e either you have a singel intterface (can be a bond or bridge) that has the tunnel local ip\n17:47 --\u003e SDrozdov (~SDrozdov@2a02:c7d:5047:ff00:fde2:9c6d:c433:c556) has joined #openstack-nova\n17:47 \u003csean-k-mooney\u003e and you confiure the route to all remote host via that\n17:47 \u003csean-k-mooney\u003e the simples way is to use a single subnet for all compute/network nodes\n17:47 \u003csean-k-mooney\u003e or you can deviced your tunneld network in to multipel l3 subnets\n17:48 \u003csean-k-mooney\u003e in which case if you have a multi homed server the multipel interface can be used\n17:48 \u003csean-k-mooney\u003e i dont see a go way to suppor the multi homed case simply\n17:49 \u003csean-k-mooney\u003e how the tx path works is ovs pushes a vxlan/geneve header on the packet and looks at a cache of the host routing tabel to 1 determin the source interface to use to reach the destination tunnel endpoint\n17:50 \u003csean-k-mooney\u003e if the source interafeace is an ovs bridge it just uses the normal action to transmit the tunneld packet onto the physical netowrk via normal mac larning\n17:51 \u003csean-k-mooney\u003e if the tunnel local ip is not assocated with an ovs brige it passes the encpsulated packet to the host kernel netowrk stack and it routes it\n17:52 \u003csean-k-mooney\u003e passing it to the host kernel routign stack is best avoided becasue its slower and hard to do qos for if there are potitally mulitiple paths","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"d29bd54f8994c8dec0091a261eb6f19f4a9cbfdd","unresolved":true,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  [ovs]"},{"line_number":54,"context_line":"  resource_provider_bandwidths \u003d br0:EGRESS:INGRESS,rp_tunnelled:EGRESS:INGRESS"},{"line_number":55,"context_line":"  resource_provider_hypervisors \u003d br0:compute2,rp_tunnelled:compute1"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"}],"source_content_type":"text/x-rst","patch_set":4,"id":"d848e6ea_e2349932","line":55,"range":{"start_line":55,"start_character":2,"end_line":55,"end_character":68},"updated":"2022-11-02 15:10:57.000000000","message":"this is an agent cfg (at least for OVS and sriov), so this should be always compute1, but not mixed, am I right?","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"576e9980a9da96c7c5c45621dc8ac12203f6e342","unresolved":true,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  [ovs]"},{"line_number":54,"context_line":"  resource_provider_bandwidths \u003d br0:EGRESS:INGRESS,rp_tunnelled:EGRESS:INGRESS"},{"line_number":55,"context_line":"  resource_provider_hypervisors \u003d br0:compute2,rp_tunnelled:compute1"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"}],"source_content_type":"text/x-rst","patch_set":4,"id":"c0b28fbd_cf68a45c","line":55,"range":{"start_line":55,"start_character":2,"end_line":55,"end_character":68},"in_reply_to":"310f24c5_0784c5da","updated":"2022-11-03 18:05:28.000000000","message":"this is not resoved sorry","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"1964c3afd1b95213d56c5c65fb2e9f4e34b7997c","unresolved":false,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  [ovs]"},{"line_number":54,"context_line":"  resource_provider_bandwidths \u003d br0:EGRESS:INGRESS,rp_tunnelled:EGRESS:INGRESS"},{"line_number":55,"context_line":"  resource_provider_hypervisors \u003d br0:compute2,rp_tunnelled:compute1"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"}],"source_content_type":"text/x-rst","patch_set":4,"id":"310f24c5_0784c5da","line":55,"range":{"start_line":55,"start_character":2,"end_line":55,"end_character":68},"in_reply_to":"9ad2e320_a37be805","updated":"2022-11-03 18:04:38.000000000","message":"in stead of this i woudl do the following\n\nadd a new config option called tunnel_bandwidth_provider\n\nif you have a singel shared interface\n[ovs]\ntunnel_bandwidth_provider \u003d br0\nresource_provider_bandwidths \u003d br0:EGRESS:INGRESS\n\nif you have a dedicated interface you can do \n\n[ovs]\ntunnel_bandwidth_provider \u003d br1\nresource_provider_bandwidths \u003d br0:EGRESS:INGRESS,br1:EGRESS:INGRESS\n\nthe tunnel_bandwidth_provider must be the key in one of the resource_provider_bandwidths pairs.\n\ntunnel_bandwidth_provider shoudl default to None/unset\nwhen set the bandwidth provider refenced in tunnel_bandwidth_provider will have the new CUSTOM_TUNNELLED_NETWORKS reproted for that inventory.\n\nthat will allow us ot have a shared or split inventory.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  [ovs]"},{"line_number":54,"context_line":"  resource_provider_bandwidths \u003d br0:EGRESS:INGRESS,rp_tunnelled:EGRESS:INGRESS"},{"line_number":55,"context_line":"  resource_provider_hypervisors \u003d br0:compute2,rp_tunnelled:compute1"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"}],"source_content_type":"text/x-rst","patch_set":4,"id":"1cc0fb5c_9a944390","line":55,"range":{"start_line":55,"start_character":2,"end_line":55,"end_character":68},"in_reply_to":"c0b28fbd_cf68a45c","updated":"2022-11-04 10:23:55.000000000","message":"When vlan/flat networks will share BW with tunnelled ones?\n\nAbout \"tunnel_bandwidth_provider\", we have that in [1]. The RP name is configurable already. If this RP is not present in the \"resource_provider_bandwidths\" list, this host won\u0027t model the tunnelled networks (same as now).\n\n[1]https://review.opendev.org/c/openstack/neutron/+/860639/8/neutron/conf/plugins/ml2/config.py","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b47655397408bbc50b81b9fb11ca7f6ae02f0d0f","unresolved":false,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  [ovs]"},{"line_number":54,"context_line":"  resource_provider_bandwidths \u003d br0:EGRESS:INGRESS,rp_tunnelled:EGRESS:INGRESS"},{"line_number":55,"context_line":"  resource_provider_hypervisors \u003d br0:compute2,rp_tunnelled:compute1"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"}],"source_content_type":"text/x-rst","patch_set":4,"id":"9ad2e320_a37be805","line":55,"range":{"start_line":55,"start_character":2,"end_line":55,"end_character":68},"in_reply_to":"ca7112ba_97df5fd2","updated":"2022-11-03 15:23:07.000000000","message":"in general vlan/flat networks will share bandwith with the tunneled networks\nwe need to make sure its easy to moddel that.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"40e6941bcfdf019d6185864ace71dfd9d80d9cb7","unresolved":false,"context_lines":[{"line_number":52,"context_line":""},{"line_number":53,"context_line":"  [ovs]"},{"line_number":54,"context_line":"  resource_provider_bandwidths \u003d br0:EGRESS:INGRESS,rp_tunnelled:EGRESS:INGRESS"},{"line_number":55,"context_line":"  resource_provider_hypervisors \u003d br0:compute2,rp_tunnelled:compute1"},{"line_number":56,"context_line":""},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"}],"source_content_type":"text/x-rst","patch_set":4,"id":"ca7112ba_97df5fd2","line":55,"range":{"start_line":55,"start_character":2,"end_line":55,"end_character":68},"in_reply_to":"d848e6ea_e2349932","updated":"2022-11-03 14:24:48.000000000","message":"Right. I need to differentiate ML2/OVS and ML2/OVN.\n\nIn OVS the agent will read the hostname. Because this is a per compute node config, there is no need to specify it.\n\nIn OVN we\u0027ll need to define it due to how it is implemented in the OVN QoS extension.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b47655397408bbc50b81b9fb11ca7f6ae02f0d0f","unresolved":true,"context_lines":[{"line_number":69,"context_line":"  +--------------------------------------+------------------------------------------+------------+--------------------------------------+--------------------------------------+"},{"line_number":70,"context_line":"  | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | u20ovn                                   |         10 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | None                                 |"},{"line_number":71,"context_line":"  | cb101b60-527b-5264-8e7f-213c7b88e9e1 | u20ovn:OVN Controller agent              |          1 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 |"},{"line_number":72,"context_line":"  | 521f53a6-c8c0-583c-98da-7a47f39ff887 | u20ovn:OVN Controller agent:br-ex        |         20 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | cb101b60-527b-5264-8e7f-213c7b88e9e1 |"},{"line_number":73,"context_line":"  | dfdbf43f-f60b-577c-bae8-3dcea448c735 | u20ovn:OVN Controller agent:rp_tunnelled |          6 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | cb101b60-527b-5264-8e7f-213c7b88e9e1 |"},{"line_number":74,"context_line":"  +--------------------------------------+------------------------------------------+------------+--------------------------------------+--------------------------------------+"},{"line_number":75,"context_line":""},{"line_number":76,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"9bd4587c_61a6a631","line":73,"range":{"start_line":72,"start_character":5,"end_line":73,"end_character":176},"updated":"2022-11-03 15:23:07.000000000","message":"this would only work in some cases in general br-ex is used for tunnles especially for ovs-dpdk and hardware offloaded ovs but even for kernel ovs.\n\ni have not read over the spec in detail as im jumping between multiple things but we need to ensure that this can either be a shared inventory or seperate contoled by config so we cant assume it salwasy rp_tunnelled we need to ensure we can set it to br-ex","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":69,"context_line":"  +--------------------------------------+------------------------------------------+------------+--------------------------------------+--------------------------------------+"},{"line_number":70,"context_line":"  | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | u20ovn                                   |         10 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | None                                 |"},{"line_number":71,"context_line":"  | cb101b60-527b-5264-8e7f-213c7b88e9e1 | u20ovn:OVN Controller agent              |          1 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 |"},{"line_number":72,"context_line":"  | 521f53a6-c8c0-583c-98da-7a47f39ff887 | u20ovn:OVN Controller agent:br-ex        |         20 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | cb101b60-527b-5264-8e7f-213c7b88e9e1 |"},{"line_number":73,"context_line":"  | dfdbf43f-f60b-577c-bae8-3dcea448c735 | u20ovn:OVN Controller agent:rp_tunnelled |          6 | 8f0e060d-bf63-42a1-85e6-710c8b2fccc8 | cb101b60-527b-5264-8e7f-213c7b88e9e1 |"},{"line_number":74,"context_line":"  +--------------------------------------+------------------------------------------+------------+--------------------------------------+--------------------------------------+"},{"line_number":75,"context_line":""},{"line_number":76,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"193ce1a2_048dff97","line":73,"range":{"start_line":72,"start_character":5,"end_line":73,"end_character":176},"in_reply_to":"9bd4587c_61a6a631","updated":"2022-11-04 10:23:55.000000000","message":"Of course but that is on the administrator responsibility. The RP BW definitions is a model of the real hardware deployment. As you commented, in the case of DPDK, the traffic is redirected via the physical bridge. This case is not covered by this feature because that means we are sharing a single HW resource with two resource providers. This is not even considered in the placement API.\n\nI would consider this case as a corner case.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b47655397408bbc50b81b9fb11ca7f6ae02f0d0f","unresolved":true,"context_lines":[{"line_number":90,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":91,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"146b0fbe_c13de0aa","line":93,"updated":"2022-11-03 15:23:07.000000000","message":"i kind of feel like it woudl make more sense to convert theyse custom_ traits into real ones by the way.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":90,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":91,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"b57a58f0_8eeeca2f","line":93,"in_reply_to":"146b0fbe_c13de0aa","updated":"2022-11-04 10:23:55.000000000","message":"Yes, I would move the CUSTOM_VNIC_etc traits to constant ones in ostrait. That will require a migration process after that. But this is out of scope here.\n\nAbout the custom tunnelled networks trait, what I\u0027m thinking is about making it static, regardless of the RP name, and as you said adding it to ostraits.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f651e3a366e9b65ac2d00f98183b5ad1bf522a77","unresolved":true,"context_lines":[{"line_number":90,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":91,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"bb1f04c6_6887b7f7","line":93,"in_reply_to":"b57a58f0_8eeeca2f","updated":"2022-11-04 12:19:23.000000000","message":"you can just use both traits for a time or indefinetly but standard traits in os-traits would be better long term.\n\nits not a blocker for this spec if we use custom_ but it would be better to standarise them if we can","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"52e34d769852bfef157c71118038c1526817244c","unresolved":false,"context_lines":[{"line_number":90,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":91,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":4,"id":"d089dcab_818da4c3","line":93,"in_reply_to":"bb1f04c6_6887b7f7","updated":"2022-11-04 12:28:44.000000000","message":"Any migration is always a pain. I would prefer to add this trait in ostraits now.\n\nI\u0027ll push a patch.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"33d11158209ff807bebe4c92903e0b098de0409b","unresolved":true,"context_lines":[{"line_number":91,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"},{"line_number":97,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"e9abce30_4d291c77","line":94,"updated":"2022-11-02 13:34:37.000000000","message":"could you add an example how the port.resource_request will look like for such a tunnneled port with QoS min bw?","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"40e6941bcfdf019d6185864ace71dfd9d80d9cb7","unresolved":false,"context_lines":[{"line_number":91,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"},{"line_number":97,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":4,"id":"fe418b1a_80c19016","line":94,"in_reply_to":"e9abce30_4d291c77","updated":"2022-11-03 14:24:48.000000000","message":"Right now.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"33d11158209ff807bebe4c92903e0b098de0409b","unresolved":true,"context_lines":[{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"},{"line_number":97,"context_line":"---------------"},{"line_number":98,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"a7b18f5c_3d80c99b","line":95,"updated":"2022-11-02 13:34:37.000000000","message":"Did neutron rejected tunnelled ports with QoS min bw before? \nIf yes, then we can remove such limitation after this implementation. \nIf no, then we can have tunneled ports with QoS min bw that are missing resource allocation in placement and we have to create a solution that healing such missing allocations.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"3f9ae8803d57035ad73d49eafa6e8cc3c66cd14a","unresolved":false,"context_lines":[{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"},{"line_number":97,"context_line":"---------------"},{"line_number":98,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"37f1a1ea_ca876ca1","line":95,"in_reply_to":"6fcd2d6f_d8f2b260","updated":"2022-11-03 15:11:22.000000000","message":"So if there are tunneled ports with QoS min bw rules today that are missing allocation then you have to heal those missing allocations. In the past I implemented $ nova-manage placement heal_allocations [1][2] CLI for such healing of non tunneled ports. \n\nSo one way to fix this is to use that CLI. I\u0027m 70% confident that if the port.resource_request will reflect the needed resources for a tunneled port then the current CLI impl will do the healing out of the box for that port too. But you need to verify my believes. :)\n\n[1] https://docs.openstack.org/nova/latest/cli/nova-manage.html#placement-heal-allocations\n[2] https://github.com/openstack/nova/blob/5ff11fe6e2bb40cf5124db96a7326d6702118a66/nova/cmd/manage.py#L1924","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b47655397408bbc50b81b9fb11ca7f6ae02f0d0f","unresolved":false,"context_lines":[{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"},{"line_number":97,"context_line":"---------------"},{"line_number":98,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"f6c1b3ab_37e892c5","line":95,"in_reply_to":"6fcd2d6f_d8f2b260","updated":"2022-11-03 15:23:07.000000000","message":"this was ment to work why was that decision made.\n\nwhen we desiginted this orgianlly tunneled port were inted to work too.\n\nyou cannot update the used filed because it does not exist.\n\nit caluated form the allocations against an inventory.\n\ninventories track capasitry\nallocations track usage.","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"40e6941bcfdf019d6185864ace71dfd9d80d9cb7","unresolved":false,"context_lines":[{"line_number":92,"context_line":"  | CUSTOM_TUNNELLED_NETWORKS        |"},{"line_number":93,"context_line":"  +----------------------------------+"},{"line_number":94,"context_line":""},{"line_number":95,"context_line":""},{"line_number":96,"context_line":"REST API Impact"},{"line_number":97,"context_line":"---------------"},{"line_number":98,"context_line":""}],"source_content_type":"text/x-rst","patch_set":4,"id":"6fcd2d6f_d8f2b260","line":95,"in_reply_to":"a7b18f5c_3d80c99b","updated":"2022-11-03 14:24:48.000000000","message":"No, the min-bw rules where discarded. Any port without a physical network didn\u0027t generate a resource request.\n\nIn ML2/OVS, the qos plugin prevented the assignation of those QoS policies to any port bound to OVS. In ML2/OVN there was no limitation.\n\nIf we need to repair a deployment depending on the port min-bw placed on tunnelled networks, I would need to know how to do it. My question is how can we update the \"used\" field in a RP inventory? Checking [1] and the neutron-lib placement library, I can\u0027t find the way.\n\n[1]https://docs.openstack.org/api-ref/placement/?expanded\u003d#update-resource-provider-inventories","commit_id":"c9a9703febc0f67e23536af42331483c85fb137e"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"3f9ae8803d57035ad73d49eafa6e8cc3c66cd14a","unresolved":true,"context_lines":[{"line_number":110,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":111,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":112,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":113,"context_line":"  | CUSTOM_TUNNELLED_RP_TUNNELLED    |"},{"line_number":114,"context_line":"  +----------------------------------+"},{"line_number":115,"context_line":""},{"line_number":116,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"bf1c3090_17a457c7","line":113,"updated":"2022-11-03 15:11:22.000000000","message":"you can probably drop one of the TUNNELLED words from the trait name.","commit_id":"ad037f2c5d57565d4f88cf6094f9f9cf9204270f"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"890c6085cdc0527db396a01daaef7101ec5492f0","unresolved":false,"context_lines":[{"line_number":110,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":111,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":112,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":113,"context_line":"  | CUSTOM_TUNNELLED_RP_TUNNELLED    |"},{"line_number":114,"context_line":"  +----------------------------------+"},{"line_number":115,"context_line":""},{"line_number":116,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"724989c4_b010fed0","line":113,"in_reply_to":"5a6e7861_97475dc9","updated":"2022-11-04 12:26:12.000000000","message":"why? how do you plan to use this?\n\nthis trait need to be pass form neutorn ot nova before we have selected a host so you cannot depend on the RP name or the tunnelled_network_rp_name parmater unless you sete that to the same value on every host.\n\nso i dont see why this shoudl be dynmaic\n\nit shoudl jsut be CUSTOM_NETWORK_TUNNELED","commit_id":"ad037f2c5d57565d4f88cf6094f9f9cf9204270f"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":110,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":111,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":112,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":113,"context_line":"  | CUSTOM_TUNNELLED_RP_TUNNELLED    |"},{"line_number":114,"context_line":"  +----------------------------------+"},{"line_number":115,"context_line":""},{"line_number":116,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"5a6e7861_97475dc9","line":113,"in_reply_to":"bf1c3090_17a457c7","updated":"2022-11-04 10:23:55.000000000","message":"But I need a unique prefix for this trait. I\u0027m building it this way:\nCUSTOM_ + TUNNELLED_ + \u003cRP-name\u003e\n\nOf course, the RP name is \"rp_tunnelled\", this is why we have this duplication. But I prefer that rather than allowing a user to override another trait.","commit_id":"ad037f2c5d57565d4f88cf6094f9f9cf9204270f"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"eacd3757364c1c54c25ae6bb8efb4ca076b75216","unresolved":true,"context_lines":[{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Example of ML2/OVS configuration section::"},{"line_number":53,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"2755bbe9_1f98485e","line":50,"range":{"start_line":50,"start_character":62,"end_line":50,"end_character":74},"updated":"2022-11-04 10:00:52.000000000","message":"It may open up some options if we used something informative here (but with a syntax that makes it possible to tell it\u0027s not simply a bridge name), like:\n\nIFACE/LOCAL_IP, e.g.: br-ex/1.2.3.4, or\nIFACE/tunneled, e.g.: br-ex/tunneled\n\nAnd then we could detect for example the case when a physnet and a tunnel vtep uses the same physical interface.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"6d8db2b0ee03eeeb2311e6345c0e6d561c0dfe37","unresolved":true,"context_lines":[{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Example of ML2/OVS configuration section::"},{"line_number":53,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"3c583fa4_9d3f70d8","line":50,"range":{"start_line":50,"start_character":62,"end_line":50,"end_character":74},"in_reply_to":"2755bbe9_1f98485e","updated":"2022-11-04 10:09:11.000000000","message":"i orginaly comment on this in an olde revsion but i ill copy it here too\n\n\n\"\"\"\nin stead of this i woudl do the following\n\nadd a new config option called tunnel_bandwidth_provider\n\nif you have a singel shared interface\n[ovs]\ntunnel_bandwidth_provider \u003d br0\nresource_provider_bandwidths \u003d br0:EGRESS:INGRESS\n\nif you have a dedicated interface you can do \n\n[ovs]\ntunnel_bandwidth_provider \u003d br1\nresource_provider_bandwidths \u003d br0:EGRESS:INGRESS,br1:EGRESS:INGRESS\n\nthe tunnel_bandwidth_provider must be the key in one of the resource_provider_bandwidths pairs.\n\ntunnel_bandwidth_provider shoudl default to None/unset\nwhen set the bandwidth provider refenced in tunnel_bandwidth_provider will have the new CUSTOM_TUNNELLED_NETWORKS reproted for that inventory.\n\nthat will allow us ot have a shared or split inventory.\n\"\"\"\n\nim not sure i like the idea of makeing resource_provider_bandwidths more complicated by allowing\n\nresource_provider_bandwidths \u003d br-ex/tunneled:EGRESS:INGRESS\n\nif that is what you were suggesting?","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Example of ML2/OVS configuration section::"},{"line_number":53,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"cf72051d_f18608c1","line":50,"range":{"start_line":50,"start_character":62,"end_line":50,"end_character":74},"in_reply_to":"3c583fa4_9d3f70d8","updated":"2022-11-04 10:23:55.000000000","message":"As commented, I\u0027m proposing a config option [1] to define the RP name.\n\n[1]https://review.opendev.org/c/openstack/neutron/+/860639/8/neutron/conf/plugins/ml2/config.py","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"8714fb1b776fac14b601bd4056bccb71c956c59b","unresolved":false,"context_lines":[{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Example of ML2/OVS configuration section::"},{"line_number":53,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"dfcac144_378fef32","line":50,"range":{"start_line":50,"start_character":62,"end_line":50,"end_character":74},"in_reply_to":"3db2dbbc_ece0f506","updated":"2022-11-04 12:40:41.000000000","message":"It sounds like this can express the same - when the rp/physical nic is shared between a physnet and the tunneled networks. And it is syntactically simpler than what I proposed. So +1 from me for this part.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"52e34d769852bfef157c71118038c1526817244c","unresolved":false,"context_lines":[{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Example of ML2/OVS configuration section::"},{"line_number":53,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"3db2dbbc_ece0f506","line":50,"range":{"start_line":50,"start_character":62,"end_line":50,"end_character":74},"in_reply_to":"a2360ca4_ee52eabc","updated":"2022-11-04 12:28:44.000000000","message":"Right, I\u0027m also introducing a new config var. I\u0027ll provide this info in the spec.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f651e3a366e9b65ac2d00f98183b5ad1bf522a77","unresolved":true,"context_lines":[{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":""},{"line_number":52,"context_line":"Example of ML2/OVS configuration section::"},{"line_number":53,"context_line":""}],"source_content_type":"text/x-rst","patch_set":6,"id":"a2360ca4_ee52eabc","line":50,"range":{"start_line":50,"start_character":62,"end_line":50,"end_character":74},"in_reply_to":"cf72051d_f18608c1","updated":"2022-11-04 12:19:23.000000000","message":"so am i correct in understandign that if i want to use br0 for physnet0 and tunneled traffic  you are suggesting setting \n\n```\n[ml2]\ntunnelled_network_rp_name\u003dbr0\n\n[ovs]\nbridge_mappings \u003d physnet0:br0\nresource_provider_bandwidths \u003d br0:EGRESS:INGRESS\n```\n\nif so that is baseically the same as i was suggesting above \n\ntunnel_bandwidth_provider is basically the same as your tunnelled_network_rp_name\n\n\nif my example above will work and allow physnet0 and the tunneled netowks to share a singel rp then this should be viable\n\ncan we add this exampel explcitly to ensure that we test this and not just the usecase where the tunnels use a spereate independent inventory.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"eacd3757364c1c54c25ae6bb8efb4ca076b75216","unresolved":true,"context_lines":[{"line_number":73,"context_line":"                  system-id\u003d\"a55c8d85-2071-4452-92cb-95d15c29bde7\"}"},{"line_number":74,"context_line":""},{"line_number":75,"context_line":""},{"line_number":76,"context_line":"Note that in ML2/OVN, it is mandatory to define the tunnelled resource provider assignation to"},{"line_number":77,"context_line":"the host in the \"resource_provider_hypervisors\" list."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"},{"line_number":80,"context_line":"any possible clash, there will be a new check when parsing the physical"}],"source_content_type":"text/x-rst","patch_set":6,"id":"86c885c8_38338533","line":77,"range":{"start_line":76,"start_character":0,"end_line":77,"end_character":53},"updated":"2022-11-04 10:00:52.000000000","message":"I don\u0027t understand this sentence. Can you please explain?","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"1bd7c12da3e90be0f2f081eafba9080fadd05297","unresolved":false,"context_lines":[{"line_number":73,"context_line":"                  system-id\u003d\"a55c8d85-2071-4452-92cb-95d15c29bde7\"}"},{"line_number":74,"context_line":""},{"line_number":75,"context_line":""},{"line_number":76,"context_line":"Note that in ML2/OVN, it is mandatory to define the tunnelled resource provider assignation to"},{"line_number":77,"context_line":"the host in the \"resource_provider_hypervisors\" list."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"},{"line_number":80,"context_line":"any possible clash, there will be a new check when parsing the physical"}],"source_content_type":"text/x-rst","patch_set":6,"id":"feaa4fee_b46db5dd","line":77,"range":{"start_line":76,"start_character":0,"end_line":77,"end_character":53},"in_reply_to":"86c885c8_38338533","updated":"2022-11-04 10:23:55.000000000","message":"In OVN we need to provide the tunnelled RP host assignation. That means:\n  resource_provider_hypervisors\u003dbr-ex:u20ovn;rp_tunnelled:u20ovn\"\n  \nWhen the OVN mech driver receives the \"open_vswitch\" register event, it can associate this RP with the host and create the corresponding RP tree.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"8714fb1b776fac14b601bd4056bccb71c956c59b","unresolved":false,"context_lines":[{"line_number":73,"context_line":"                  system-id\u003d\"a55c8d85-2071-4452-92cb-95d15c29bde7\"}"},{"line_number":74,"context_line":""},{"line_number":75,"context_line":""},{"line_number":76,"context_line":"Note that in ML2/OVN, it is mandatory to define the tunnelled resource provider assignation to"},{"line_number":77,"context_line":"the host in the \"resource_provider_hypervisors\" list."},{"line_number":78,"context_line":""},{"line_number":79,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"},{"line_number":80,"context_line":"any possible clash, there will be a new check when parsing the physical"}],"source_content_type":"text/x-rst","patch_set":6,"id":"bbcd247d_49b33332","line":77,"range":{"start_line":76,"start_character":0,"end_line":77,"end_character":53},"in_reply_to":"feaa4fee_b46db5dd","updated":"2022-11-04 12:40:41.000000000","message":"I hope I\u0027m starting to get it. Thanks.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b2daed3070cde4dce5992e30a45aa17a0a05f0ad","unresolved":true,"context_lines":[{"line_number":96,"context_line":""},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"A new static trait is added to this resource provider:"},{"line_number":99,"context_line":"\"CUSTOM_TUNNELLED_NETWORKS\". This is what identify that this resource provider"},{"line_number":100,"context_line":"is for tunnelled networks. E.g.::"},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"  $ openstack resource provider trait list $rp_tun"}],"source_content_type":"text/x-rst","patch_set":6,"id":"5762cd5a_f91d3267","line":99,"range":{"start_line":99,"start_character":1,"end_line":99,"end_character":26},"updated":"2022-11-04 12:38:33.000000000","message":"this does not line up with the tabel by the way.\n\ni also prefer this name to what you have in the table and in the example below.\n\nif we stick withthe CUSTOM_ prefix can we just use this trait instead of \nCUSTOM_TUNNELLED_RP_TUNNELLED","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"19d278c01ae532eb578399cbcaaa9e39019bdfe5","unresolved":false,"context_lines":[{"line_number":96,"context_line":""},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"A new static trait is added to this resource provider:"},{"line_number":99,"context_line":"\"CUSTOM_TUNNELLED_NETWORKS\". This is what identify that this resource provider"},{"line_number":100,"context_line":"is for tunnelled networks. E.g.::"},{"line_number":101,"context_line":""},{"line_number":102,"context_line":"  $ openstack resource provider trait list $rp_tun"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9b267f2a_4bf817e2","line":99,"range":{"start_line":99,"start_character":1,"end_line":99,"end_character":26},"in_reply_to":"5762cd5a_f91d3267","updated":"2022-11-14 17:17:32.000000000","message":"Ack","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b2daed3070cde4dce5992e30a45aa17a0a05f0ad","unresolved":true,"context_lines":[{"line_number":162,"context_line":"from the network QoS policy). With this feature, the minimum bandwidth QoS"},{"line_number":163,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":164,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":165,"context_line":"network)."},{"line_number":166,"context_line":""},{"line_number":167,"context_line":"A new check will be added to inform about those ports located on"},{"line_number":168,"context_line":"tunnelled networks with minimum bandwidth QoS rules."}],"source_content_type":"text/x-rst","patch_set":6,"id":"676e4dd7_b58a9a08","line":165,"updated":"2022-11-04 12:38:33.000000000","message":"your going to have to reshape the placment allocation or nova will for any existing port that have a minbandwith qos rule applied to a tunneled port.\n\nthat will be non trivial.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9832cbc4f62816398090f4821a4a87f85d87142a","unresolved":false,"context_lines":[{"line_number":162,"context_line":"from the network QoS policy). With this feature, the minimum bandwidth QoS"},{"line_number":163,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":164,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":165,"context_line":"network)."},{"line_number":166,"context_line":""},{"line_number":167,"context_line":"A new check will be added to inform about those ports located on"},{"line_number":168,"context_line":"tunnelled networks with minimum bandwidth QoS rules."}],"source_content_type":"text/x-rst","patch_set":6,"id":"15a04950_bbc0d111","line":165,"in_reply_to":"151d6e60_26040d44","updated":"2022-11-07 17:41:26.000000000","message":"Ok, I\u0027m not going to fix what is not broken. What I\u0027ll propose is a tool that will inform about the ports that have QoS min-bw rules in overlay networks. If those ports are on hosts with an overlay network with a new RP, then we\u0027ll inform about this. Unless we have a tool to sync/heal an existing deployment, any port created before this new feature won\u0027t have an allocation.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f1f67b427a821f0c62500083c1a4674fdab17745","unresolved":true,"context_lines":[{"line_number":162,"context_line":"from the network QoS policy). With this feature, the minimum bandwidth QoS"},{"line_number":163,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":164,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":165,"context_line":"network)."},{"line_number":166,"context_line":""},{"line_number":167,"context_line":"A new check will be added to inform about those ports located on"},{"line_number":168,"context_line":"tunnelled networks with minimum bandwidth QoS rules."}],"source_content_type":"text/x-rst","patch_set":6,"id":"151d6e60_26040d44","line":165,"in_reply_to":"217bf906_b20c05df","updated":"2022-11-07 14:56:27.000000000","message":"nova does create the allcoations for all port resouce requests.\n\nneutron only manges the invetories/resouce providers which tracks capasity not usage.\n\nusage is track by the allocation which nova creates.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"6e708edfdb8dc043c42604a3a91ac306df906b75","unresolved":true,"context_lines":[{"line_number":162,"context_line":"from the network QoS policy). With this feature, the minimum bandwidth QoS"},{"line_number":163,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":164,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":165,"context_line":"network)."},{"line_number":166,"context_line":""},{"line_number":167,"context_line":"A new check will be added to inform about those ports located on"},{"line_number":168,"context_line":"tunnelled networks with minimum bandwidth QoS rules."}],"source_content_type":"text/x-rst","patch_set":6,"id":"217bf906_b20c05df","line":165,"in_reply_to":"676e4dd7_b58a9a08","updated":"2022-11-07 12:33:59.000000000","message":"Nova doesn\u0027t do this for an existing port, Neutron does.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"b2daed3070cde4dce5992e30a45aa17a0a05f0ad","unresolved":true,"context_lines":[{"line_number":177,"context_line":"* List all ports on tunnelled networks bound to ML2/OVN or ML2/OVS backends"},{"line_number":178,"context_line":"  with minimum bandwidth rules attached."},{"line_number":179,"context_line":"* Update the corresponding inventories."},{"line_number":180,"context_line":"* Log the errors in case of exceeding the maximum available bandwidth."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":""},{"line_number":183,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":6,"id":"1c7908f5_a5c4455e","line":180,"updated":"2022-11-04 12:38:33.000000000","message":"i think your are mixing ups how placemnt is ment to work\n\ninventories track total avalaibel capastyity and have a reserved value.\nthe reserved value tracks capatiy that is reserved for non openstack usage.\n\nthe consumtion of the resouces i.e. the usage is compute form the allcoations against the inventoies.\n\nso how this need to work is the reshape api need to be used to reshape the allocation when  this feature is enabled.\n\nthat reshape will likely need to be done by nova.\nit may be possibel to do it form neutron but we need to ensure that nova will generate the same allocations if we migrate or run the heall allcoations comand.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f1f67b427a821f0c62500083c1a4674fdab17745","unresolved":true,"context_lines":[{"line_number":177,"context_line":"* List all ports on tunnelled networks bound to ML2/OVN or ML2/OVS backends"},{"line_number":178,"context_line":"  with minimum bandwidth rules attached."},{"line_number":179,"context_line":"* Update the corresponding inventories."},{"line_number":180,"context_line":"* Log the errors in case of exceeding the maximum available bandwidth."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":""},{"line_number":183,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":6,"id":"81fbd140_206b81eb","line":180,"in_reply_to":"11840e74_ce293de3","updated":"2022-11-07 14:56:27.000000000","message":"there is no used value in placment.\n\nwhen you get the usage for a resouce provide we calualte it by adding up all the allocations for a resouce porvider.\n\nwe have a long standing know issue for how neutorn is currently mangaing the inventoies.\n\nonly nova should be modifying the allocations because the neutron port request are part of the allocation for the entire vm. nova merges the request form the flavor, neutron and cybrog into a sincel allcoation for the entire vm.\n\nso neutron cannot modify that unles nova agree with it. for nova to agree with it the port requests in the port would have to be correct so that if we move a vm or schedule a new vm the accounting is in sync.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"6e708edfdb8dc043c42604a3a91ac306df906b75","unresolved":true,"context_lines":[{"line_number":177,"context_line":"* List all ports on tunnelled networks bound to ML2/OVN or ML2/OVS backends"},{"line_number":178,"context_line":"  with minimum bandwidth rules attached."},{"line_number":179,"context_line":"* Update the corresponding inventories."},{"line_number":180,"context_line":"* Log the errors in case of exceeding the maximum available bandwidth."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":""},{"line_number":183,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":6,"id":"11840e74_ce293de3","line":180,"in_reply_to":"1c7908f5_a5c4455e","updated":"2022-11-07 12:33:59.000000000","message":"I never mentioned \"reserved\" value. I\u0027m talking about the \"used\" value. If we send a request to update the \"used\" value beyond the \"total\" value, that request will fail. This is what I\u0027ll log.\n\nWhat I\u0027m saying here is that once this feature is enabled, we\u0027ll need to recalculate the RP inventories for existing ports, and only if those ports have QoS policies with min-bw rules. In OVS this is not possible now for ports in overlay networks. OVN feature is too recent and most probably won\u0027t hit too many occurrences.\n\nWhy from Nova? Nova does not store this information. Nova is just a proxy for Placement. I don\u0027t see why Nova should handle this.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9832cbc4f62816398090f4821a4a87f85d87142a","unresolved":false,"context_lines":[{"line_number":177,"context_line":"* List all ports on tunnelled networks bound to ML2/OVN or ML2/OVS backends"},{"line_number":178,"context_line":"  with minimum bandwidth rules attached."},{"line_number":179,"context_line":"* Update the corresponding inventories."},{"line_number":180,"context_line":"* Log the errors in case of exceeding the maximum available bandwidth."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":""},{"line_number":183,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":6,"id":"63509a3e_889e8842","line":180,"in_reply_to":"81fbd140_206b81eb","updated":"2022-11-07 17:41:26.000000000","message":"\"we have a long standing know issue for how neutorn is currently mangaing the inventoies.\" What is this issue? From Neutron we currently can update the inventory of a RP when the port QoS change. If that is no allowed, please open a bug for Neutron and we\u0027ll discuss about how to change this.","commit_id":"09ee72205f522b16261a4ae919e03511386c0e9d"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"f5302378097905876073e77abe44c2864be010c6","unresolved":true,"context_lines":[{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":"This configuration variable will be stored in the ``[ml2]`` section and will be"},{"line_number":52,"context_line":"accesible from the Neutron server and the OVS agent::"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"  [ml2]"},{"line_number":55,"context_line":"  tunnelled_network_rp_name \u003d custome_rp_name  # \"rp_tunnelled\" by default."}],"source_content_type":"text/x-rst","patch_set":7,"id":"ee04b034_d60b4400","line":52,"updated":"2022-11-10 14:54:56.000000000","message":"1) So far we had kind of \"hierarchical\" rp names like \"devstack0:Open vSwitch agent:br-test\". Is the tunnelled_network_rp_name supposed to be the full name or only the last component? The sample rp listing below suggests the latter.\n\n2) \"accessible from the Neutron server and the OVS agent\" - Does this mean that we need to set this for each agent using this and then the agents report this to the server?","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"d0db798ddff8bb02beda1f35edb54abc0c59deb6","unresolved":false,"context_lines":[{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":"This configuration variable will be stored in the ``[ml2]`` section and will be"},{"line_number":52,"context_line":"accesible from the Neutron server and the OVS agent::"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"  [ml2]"},{"line_number":55,"context_line":"  tunnelled_network_rp_name \u003d custome_rp_name  # \"rp_tunnelled\" by default."}],"source_content_type":"text/x-rst","patch_set":7,"id":"47a04b4b_bc1acc08","line":52,"in_reply_to":"bebb4016_8ac1e4db","updated":"2022-11-11 12:41:34.000000000","message":"1) Good question. I\u0027ll add this comment to the spec. This is the suffix of the RP name. You can see an example in the output of \"$ openstack resource provider list\" L93. The host and agent type will be added by the mech driver (actually the placement extension of this driver). It will be the same as with other bridges.\n\n2) I need this configuration option being readable from the OVS agent and the mech driver. This is why I\u0027m placing it in the [ml2] section, instead of the specific agent section.\n\nWhy shouldn\u0027t have a default value? If you don\u0027t add it to the \"resource_provider_bandwidths\" config list, it won\u0027t matter the default value defined in \"tunnelled_network_rp_name\". The trigger for the reshape should be the presence of ports on those networks AND the enablement of those tunnelled network in a specific host/agent.\n\nWhat we are doing here is not a RP reshaping, we are creating a new one.\n\nAnd in any case, I\u0027m not going to consider the case of having pre-existing ports with min-bw rules, I\u0027ll let this out of scope until there is a clear process to execute it.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"335306518742a6e419d7366d50496293e2fa86c5","unresolved":true,"context_lines":[{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. The default value will be \"rp_tunnelled\"."},{"line_number":51,"context_line":"This configuration variable will be stored in the ``[ml2]`` section and will be"},{"line_number":52,"context_line":"accesible from the Neutron server and the OVS agent::"},{"line_number":53,"context_line":""},{"line_number":54,"context_line":"  [ml2]"},{"line_number":55,"context_line":"  tunnelled_network_rp_name \u003d custome_rp_name  # \"rp_tunnelled\" by default."}],"source_content_type":"text/x-rst","patch_set":7,"id":"bebb4016_8ac1e4db","line":52,"in_reply_to":"ee04b034_d60b4400","updated":"2022-11-10 16:02:47.000000000","message":"this will need to be set on each agent config file yes because it might not be the same in all network/compute nodes\n\ni would prefer if it was only the final part fo the name and it was used the generate the full name as that is eiaser for operators to use.\n\nlater\n-----\n\nfor upgrade reasons tunnelled_network_rp_name should not have a default value and shoudl be used to opt into the new beahivor and triggerign the reshape of the placment allcoations.\n\nhttps://docs.openstack.org/api-ref/placement/?expanded\u003did86-detail#id86","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"335306518742a6e419d7366d50496293e2fa86c5","unresolved":true,"context_lines":[{"line_number":102,"context_line":""},{"line_number":103,"context_line":""},{"line_number":104,"context_line":"A new static trait will be added to represent this resource provider:"},{"line_number":105,"context_line":"\"NETWORK_TUNNELLED_PROVIDER\". This trait will be added to ``os-traits``"},{"line_number":106,"context_line":"repository. This is what identify that this resource provider"},{"line_number":107,"context_line":"is for tunnelled networks. E.g.::"},{"line_number":108,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"71ed07ea_e8e4ab6c","line":105,"range":{"start_line":105,"start_character":1,"end_line":105,"end_character":27},"updated":"2022-11-10 16:02:47.000000000","message":"this works but NETWORK_TUNNEL_PROVIDER would be better.\nthis can be defered to the os-taits review and we can update the spec later.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"d0db798ddff8bb02beda1f35edb54abc0c59deb6","unresolved":false,"context_lines":[{"line_number":102,"context_line":""},{"line_number":103,"context_line":""},{"line_number":104,"context_line":"A new static trait will be added to represent this resource provider:"},{"line_number":105,"context_line":"\"NETWORK_TUNNELLED_PROVIDER\". This trait will be added to ``os-traits``"},{"line_number":106,"context_line":"repository. This is what identify that this resource provider"},{"line_number":107,"context_line":"is for tunnelled networks. E.g.::"},{"line_number":108,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"920651f9_9a503c53","line":105,"range":{"start_line":105,"start_character":1,"end_line":105,"end_character":27},"in_reply_to":"71ed07ea_e8e4ab6c","updated":"2022-11-11 12:41:34.000000000","message":"I\u0027ll push a patch for os-traits with this name.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"335306518742a6e419d7366d50496293e2fa86c5","unresolved":true,"context_lines":[{"line_number":110,"context_line":"  +----------------------------------+"},{"line_number":111,"context_line":"  | name                             |"},{"line_number":112,"context_line":"  +----------------------------------+"},{"line_number":113,"context_line":"  | CUSTOM_VNIC_TYPE_NORMAL          |"},{"line_number":114,"context_line":"  | CUSTOM_VNIC_TYPE_DIRECT          |"},{"line_number":115,"context_line":"  | CUSTOM_VNIC_TYPE_DIRECT_PHYSICAL |"},{"line_number":116,"context_line":"  | CUSTOM_VNIC_TYPE_MACVTAP         |"},{"line_number":117,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":118,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":119,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":120,"context_line":"  | NETWORK_TUNNELLED_PROVIDER       |"},{"line_number":121,"context_line":"  +----------------------------------+"},{"line_number":122,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"69a16c4f_9ff7de11","line":119,"range":{"start_line":113,"start_character":3,"end_line":119,"end_character":35},"updated":"2022-11-10 16:02:47.000000000","message":"so thses can continue to be used but i suspect we would want to expand the network namespace ass follows\n\n NETWORK_VNIC_TYPE_NORMAL \n NETWORK_VNIC_TYPE_DIRECT\n ...\n \n but we might also want to add\n \n NETWORK_TYPE_VLAN\n NETWORK_TYPE_VXLAN\n NETWORK_TYPE_FLAT\n NETWORK_TYPE_GENEVE\n NETWORK_TYPE_GRE\n \n so that we can eventually schedule ot hosts that have networks fo that type.\n \n that can be done speratly but that is how i would see the NETWORK tratis namespce being extended in the future.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"d0db798ddff8bb02beda1f35edb54abc0c59deb6","unresolved":false,"context_lines":[{"line_number":110,"context_line":"  +----------------------------------+"},{"line_number":111,"context_line":"  | name                             |"},{"line_number":112,"context_line":"  +----------------------------------+"},{"line_number":113,"context_line":"  | CUSTOM_VNIC_TYPE_NORMAL          |"},{"line_number":114,"context_line":"  | CUSTOM_VNIC_TYPE_DIRECT          |"},{"line_number":115,"context_line":"  | CUSTOM_VNIC_TYPE_DIRECT_PHYSICAL |"},{"line_number":116,"context_line":"  | CUSTOM_VNIC_TYPE_MACVTAP         |"},{"line_number":117,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":118,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":119,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":120,"context_line":"  | NETWORK_TUNNELLED_PROVIDER       |"},{"line_number":121,"context_line":"  +----------------------------------+"},{"line_number":122,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"ad125800_8d36285e","line":119,"range":{"start_line":113,"start_character":3,"end_line":119,"end_character":35},"in_reply_to":"69a16c4f_9ff7de11","updated":"2022-11-11 12:41:34.000000000","message":"This is not what we want to model with the RP traits. Each resource provider will have a list of VNIC types and the corresponding physical bridge or the tunnelled networks trait. It is out of scope to define the network type in the trait list.\n\nIf, in future features, we need to schedule by network type, we\u0027ll add it.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"c1527c6e44dc13c15f8f254fdb842ca3993febef","unresolved":false,"context_lines":[{"line_number":110,"context_line":"  +----------------------------------+"},{"line_number":111,"context_line":"  | name                             |"},{"line_number":112,"context_line":"  +----------------------------------+"},{"line_number":113,"context_line":"  | CUSTOM_VNIC_TYPE_NORMAL          |"},{"line_number":114,"context_line":"  | CUSTOM_VNIC_TYPE_DIRECT          |"},{"line_number":115,"context_line":"  | CUSTOM_VNIC_TYPE_DIRECT_PHYSICAL |"},{"line_number":116,"context_line":"  | CUSTOM_VNIC_TYPE_MACVTAP         |"},{"line_number":117,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":118,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":119,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":120,"context_line":"  | NETWORK_TUNNELLED_PROVIDER       |"},{"line_number":121,"context_line":"  +----------------------------------+"},{"line_number":122,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"96853f17_1fee4789","line":119,"range":{"start_line":113,"start_character":3,"end_line":119,"end_character":35},"in_reply_to":"ad125800_8d36285e","updated":"2022-11-11 12:47:39.000000000","message":"just to be clear schdulign based on network type is a request that predates the creation of placment.\n\nits always been a desire just never had people to work on it.\nbut yes it out of scope fo this but it is ideaaly somethign we coudl do in the future.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"f5302378097905876073e77abe44c2864be010c6","unresolved":true,"context_lines":[{"line_number":143,"context_line":"   and overlay network. What this spec is proposing is to provide the"},{"line_number":144,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":145,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":146,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":147,"context_line":"   provide a way to model a shared resource."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"6fb709a1_a85e273b","line":147,"range":{"start_line":146,"start_character":31,"end_line":147,"end_character":44},"updated":"2022-11-10 14:54:56.000000000","message":"Without debating the scope you set, I\u0027m not sure I agree with this. It seems possible to me to create a single inventory and then make multiple allocations (for different use cases like one for a physnet and another for a tunnelled net) against it, consuming the same inventory.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":8313,"name":"Lajos Katona","display_name":"lajoskatona","email":"katonalala@gmail.com","username":"elajkat","status":"Ericsson Software Technology"},"change_message_id":"e3cf4ca58f256e2ee186cef0c758c8a079911144","unresolved":false,"context_lines":[{"line_number":143,"context_line":"   and overlay network. What this spec is proposing is to provide the"},{"line_number":144,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":145,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":146,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":147,"context_line":"   provide a way to model a shared resource."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"3d08c5c4_18e84cc7","line":147,"range":{"start_line":146,"start_character":31,"end_line":147,"end_character":44},"in_reply_to":"19d37922_4714ff4a","updated":"2022-11-14 13:47:29.000000000","message":"Any further results of this discussion? I vote to do this with the limitation that the VIP of tunneled network is not sharing bw with physnet bridge.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"2a2fe856830468a9964a6f0b52b70447b67ebef0","unresolved":false,"context_lines":[{"line_number":143,"context_line":"   and overlay network. What this spec is proposing is to provide the"},{"line_number":144,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":145,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":146,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":147,"context_line":"   provide a way to model a shared resource."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"19d37922_4714ff4a","line":147,"range":{"start_line":146,"start_character":31,"end_line":147,"end_character":44},"in_reply_to":"2749b732_d9f8a1e3","updated":"2022-11-14 11:12:03.000000000","message":"Ok, ping me on IRC.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9c85d625de65351e5e4ad16c87ca446e17e80cef","unresolved":false,"context_lines":[{"line_number":143,"context_line":"   and overlay network. What this spec is proposing is to provide the"},{"line_number":144,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":145,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":146,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":147,"context_line":"   provide a way to model a shared resource."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"a2fbae7a_27fe879d","line":147,"range":{"start_line":146,"start_character":31,"end_line":147,"end_character":44},"in_reply_to":"3d08c5c4_18e84cc7","updated":"2022-11-15 08:36:45.000000000","message":"We agreed on keeping the 1:1 match between the network interface (physical, tunnelled) and the RP. Of course that will limit the possibility of sharing the same resource (or at least the way to model it in the Neutron placement implementation), but \"most\" of the operators usually use one network type (vlan, vxlan, geneve, etc).","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"c1527c6e44dc13c15f8f254fdb842ca3993febef","unresolved":false,"context_lines":[{"line_number":143,"context_line":"   and overlay network. What this spec is proposing is to provide the"},{"line_number":144,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":145,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":146,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":147,"context_line":"   provide a way to model a shared resource."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"2749b732_d9f8a1e3","line":147,"range":{"start_line":146,"start_character":31,"end_line":147,"end_character":44},"in_reply_to":"6aae2eed_3298ecca","updated":"2022-11-11 12:47:39.000000000","message":"rodolfo lets have a call to talk about htis becues you are not understanding what im saying.\n\nwhat im proposing can be done with placment","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"335306518742a6e419d7366d50496293e2fa86c5","unresolved":true,"context_lines":[{"line_number":143,"context_line":"   and overlay network. What this spec is proposing is to provide the"},{"line_number":144,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":145,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":146,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":147,"context_line":"   provide a way to model a shared resource."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"f1a41120_b46f0f38","line":147,"range":{"start_line":146,"start_character":31,"end_line":147,"end_character":44},"in_reply_to":"6fb709a1_a85e273b","updated":"2022-11-10 16:02:47.000000000","message":"it wrong. both in terminology and meaning.\n\nneutron is creating a resouce provider for the swtich bandwith.\n\nu20ovn:OVN Controller agent:br-ex\n\nthere is nothing preventing neutron form using that singel RP for both tunneled and untunneled networks.\n\nall you ahve to do is set the following on the compute node for ml2/ovs\n  [ml2]\n  tunnelled_network_rp_name\u003dbr-ex\n  [ovs]\n  resource_provider_bandwidths \u003d br0:EGRESS:INGRESS,\n  \n  neutron will create the RP based on the resource_provider_bandwidths paramater\n  and add the tunneled trait based on tunnelled_network_rp_name\n  \nThere is no technial reason form a placment or nova point of view to not do this.\n\nif you decied not to supprot this its a purly neutorn desicison to declar it out of scope.\n\ni do not think we shoudl implemnte this feature without supproting this configuration.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"d0db798ddff8bb02beda1f35edb54abc0c59deb6","unresolved":false,"context_lines":[{"line_number":143,"context_line":"   and overlay network. What this spec is proposing is to provide the"},{"line_number":144,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":145,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":146,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":147,"context_line":"   provide a way to model a shared resource."},{"line_number":148,"context_line":""},{"line_number":149,"context_line":""},{"line_number":150,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":7,"id":"6aae2eed_3298ecca","line":147,"range":{"start_line":146,"start_character":31,"end_line":147,"end_character":44},"in_reply_to":"f1a41120_b46f0f38","updated":"2022-11-11 12:41:34.000000000","message":"Bence, we can have a single inventory from one RP, of course. And we can place allocations from different requests. But let\u0027s keep to the current architecture: one RP for one physical bridge (this is what we have now). This spec is not considering sharing the RP inventory of a physical network with a tunnelled one.\n\nSean, we have discussed this several times. Placement API has no way to model a physical resource with two RP and share the inventory.\n\nWhat we decided in Neutron is to make a 1:1 assignation between a network and it\u0027s physical bridge. This is why currently we only support physical backed networks in Neutron.\n\n\"i do not think we shoudl implemnte this feature without supproting this configuration.\" That means we should also provide the ability of model one RP shared between two physical networks (without even considering adding the support for tunnelled networks). I would say it again: in Neutron we do a 1:1 between networks and physical devices (model as physical bridges, both in ML2/OVN and ML2/OVS). This RFE is extending this support for tunnelled network, considering that this traffic will go through the \"local_ip\", assigned to an independent network interface.\n\nIn other words, this spec is not considering creating a RP shared between two networks. Each RP will have a 1:1 match with a physical network or the \"local_ip\" interface. Each RP will have its own inventory. The case you are describing is a corner case, when the \"local_ip\" is using the same interface of a physical bridge.","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"335306518742a6e419d7366d50496293e2fa86c5","unresolved":true,"context_lines":[{"line_number":195,"context_line":"* List all ports on tunnelled networks bound to ML2/OVN or ML2/OVS backends"},{"line_number":196,"context_line":"  with minimum bandwidth rules attached."},{"line_number":197,"context_line":"* Update the corresponding inventories."},{"line_number":198,"context_line":"* Log the errors in case of exceeding the maximum available bandwidth."},{"line_number":199,"context_line":""},{"line_number":200,"context_line":""},{"line_number":201,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":7,"id":"2efd815a_918ac501","line":198,"updated":"2022-11-10 16:02:47.000000000","message":"i really dont think this is a good upgrade experince.\n\nneutron or nova should be enhanced to do a reshape of the allcotions\n\nneutron coudl do it automatically if it was opted into as a sideefect of defining tunnelled_network_rp_name\n\nto make that wort we shoudl not ahve a default value for tunnelled_network_rp_name","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"19d278c01ae532eb578399cbcaaa9e39019bdfe5","unresolved":false,"context_lines":[{"line_number":195,"context_line":"* List all ports on tunnelled networks bound to ML2/OVN or ML2/OVS backends"},{"line_number":196,"context_line":"  with minimum bandwidth rules attached."},{"line_number":197,"context_line":"* Update the corresponding inventories."},{"line_number":198,"context_line":"* Log the errors in case of exceeding the maximum available bandwidth."},{"line_number":199,"context_line":""},{"line_number":200,"context_line":""},{"line_number":201,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":7,"id":"0ff990d2_8a851649","line":198,"in_reply_to":"2efd815a_918ac501","updated":"2022-11-14 17:17:32.000000000","message":"Ack","commit_id":"7fdc74b912aa683acb27532c45a09c93dcf38a4a"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"dfe45caf0260ccaf07c4493cb7b7ee60a10f553e","unresolved":true,"context_lines":[{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. This string will be the suffix of the"},{"line_number":51,"context_line":"resource provider name, same as Neutron uses the physical bridge names to"},{"line_number":52,"context_line":"build their resource provider names. For example"}],"source_content_type":"text/x-rst","patch_set":8,"id":"962ee0cd_3a1cceeb","line":49,"range":{"start_line":49,"start_character":44,"end_line":49,"end_character":61},"updated":"2022-11-15 12:08:24.000000000","message":"a resource provider in Placement","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"67de389b993e8ee00138bccba74eb7379d670a33","unresolved":false,"context_lines":[{"line_number":46,"context_line":"configuration file [4]_."},{"line_number":47,"context_line":""},{"line_number":48,"context_line":"This RFE proposes to use the same configuration options provided in [1]_,"},{"line_number":49,"context_line":"adding a static string constant to define a tunnelled network that could be"},{"line_number":50,"context_line":"configurable by the administrator. This string will be the suffix of the"},{"line_number":51,"context_line":"resource provider name, same as Neutron uses the physical bridge names to"},{"line_number":52,"context_line":"build their resource provider names. For example"}],"source_content_type":"text/x-rst","patch_set":8,"id":"6a3bf816_963c2f61","line":49,"range":{"start_line":49,"start_character":44,"end_line":49,"end_character":61},"in_reply_to":"962ee0cd_3a1cceeb","updated":"2022-11-15 15:41:02.000000000","message":"Done","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"dfe45caf0260ccaf07c4493cb7b7ee60a10f553e","unresolved":true,"context_lines":[{"line_number":56,"context_line":"server and the OVS agent::"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"  [ml2]"},{"line_number":59,"context_line":"  tunnelled_network_rp_name \u003d custome_rp_name  # \"rp_tunnelled\" by default."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Example of ML2/OVS configuration section::"}],"source_content_type":"text/x-rst","patch_set":8,"id":"1b8de3e1_7c00436c","line":59,"range":{"start_line":59,"start_character":30,"end_line":59,"end_character":36},"updated":"2022-11-15 12:08:24.000000000","message":"nit: custom","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"67de389b993e8ee00138bccba74eb7379d670a33","unresolved":false,"context_lines":[{"line_number":56,"context_line":"server and the OVS agent::"},{"line_number":57,"context_line":""},{"line_number":58,"context_line":"  [ml2]"},{"line_number":59,"context_line":"  tunnelled_network_rp_name \u003d custome_rp_name  # \"rp_tunnelled\" by default."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"Example of ML2/OVS configuration section::"}],"source_content_type":"text/x-rst","patch_set":8,"id":"f30147f4_825059fc","line":59,"range":{"start_line":59,"start_character":30,"end_line":59,"end_character":36},"in_reply_to":"1b8de3e1_7c00436c","updated":"2022-11-15 15:41:02.000000000","message":"Done","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"19d278c01ae532eb578399cbcaaa9e39019bdfe5","unresolved":true,"context_lines":[{"line_number":121,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":122,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":123,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":124,"context_line":"  | NETWORK_TUNNEL_PROVIDER          |"},{"line_number":125,"context_line":"  +----------------------------------+"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":""}],"source_content_type":"text/x-rst","patch_set":8,"id":"39b5e4ea_d448821b","line":124,"range":{"start_line":124,"start_character":4,"end_line":124,"end_character":27},"updated":"2022-11-14 17:17:32.000000000","message":"given the intent to keep this minimal and since we are not going to standardise the rest of the trait let defer making this a standard trait for this release.\n\nsed \"s/NETWORK_TUNNEL_PROVIDER/CUSTOM_NETWORK_TUNNEL_PROVIDER/g\"\n\n\nwe can adress this in a sperate spec in the futre that focus on network aware schduleing.\n\ni.e. allowing nova to schdule to host that support the required vnic type, and network type.\n\nso my recomendation is to reduce thet scope fo the spec by using CUSTOM_NETWORK_TUNNEL_PROVIDER for now to avoid change to os-traits\nand we can come back to this in the B/C cycle.","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9c85d625de65351e5e4ad16c87ca446e17e80cef","unresolved":false,"context_lines":[{"line_number":121,"context_line":"  | CUSTOM_VNIC_TYPE_VDPA            |"},{"line_number":122,"context_line":"  | CUSTOM_VNIC_TYPE_REMOTE_MANAGED  |"},{"line_number":123,"context_line":"  | CUSTOM_VNIC_TYPE_BAREMETAL       |"},{"line_number":124,"context_line":"  | NETWORK_TUNNEL_PROVIDER          |"},{"line_number":125,"context_line":"  +----------------------------------+"},{"line_number":126,"context_line":""},{"line_number":127,"context_line":""}],"source_content_type":"text/x-rst","patch_set":8,"id":"c980ab3a_a26770db","line":124,"range":{"start_line":124,"start_character":4,"end_line":124,"end_character":27},"in_reply_to":"39b5e4ea_d448821b","updated":"2022-11-15 08:36:45.000000000","message":"Right.","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"19d278c01ae532eb578399cbcaaa9e39019bdfe5","unresolved":true,"context_lines":[{"line_number":184,"context_line":"from the network QoS policy). With this feature, the minimum bandwidth QoS"},{"line_number":185,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":186,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":187,"context_line":"network)."},{"line_number":188,"context_line":""},{"line_number":189,"context_line":"A new check will be added to inform about those ports located on"},{"line_number":190,"context_line":"tunnelled networks with minimum bandwidth QoS rules. The output of this check"}],"source_content_type":"text/x-rst","patch_set":8,"id":"83565dec_2f37dbb8","line":187,"updated":"2022-11-14 17:17:32.000000000","message":"i spoke to rodolfo about the direction here and i understand more the scope that they want to define for this.\n\n\nwe went over the history a bit and disucssed the current supprot in ml2/ovs and ml2/ovn\n\ntl;dr is for ml2/ovs applying min bandwith QOS policies to neutron network is blocked and ml2/ovn does not currently supprot the min bandwith qos.\n\nthat will change with this spec but we shoudl not have any existign usage of min bandwith qos with tunneled networks. as a result we do not need to provide an automated upgrade proceedure to reshape the allocations of exsiting ports.\n\nwe discussed that there is an existign gap when using min bandwith with phyent backed networks, notably that if you chang the bandwith qos policy config options \ntoday we do not suport reshapes today for the removal or addtion  of bandwith qos inventories if you have pre existing ports that shoudl NET_BW_EGR_KILOBIT_PER_SEC follow an upgrade.\n\nthis spec does not adress that but its not a new problem since it applies to vlan/flat networks today as well. as such i think its reasonable to declare that out of scope of this spec and track it as a bug/future improvment that can be done seperatly.","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9c85d625de65351e5e4ad16c87ca446e17e80cef","unresolved":false,"context_lines":[{"line_number":184,"context_line":"from the network QoS policy). With this feature, the minimum bandwidth QoS"},{"line_number":185,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":186,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":187,"context_line":"network)."},{"line_number":188,"context_line":""},{"line_number":189,"context_line":"A new check will be added to inform about those ports located on"},{"line_number":190,"context_line":"tunnelled networks with minimum bandwidth QoS rules. The output of this check"}],"source_content_type":"text/x-rst","patch_set":8,"id":"9c12a5e0_3b1f9478","line":187,"in_reply_to":"83565dec_2f37dbb8","updated":"2022-11-15 08:36:45.000000000","message":"This is something to be addressed, for sure, in future RFEs.","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"19d278c01ae532eb578399cbcaaa9e39019bdfe5","unresolved":true,"context_lines":[{"line_number":194,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":195,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":196,"context_line":"this action, a new spec/bug will be created to track this improvement, but"},{"line_number":197,"context_line":"this is out of scope in this spec."},{"line_number":198,"context_line":""},{"line_number":199,"context_line":""},{"line_number":200,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":8,"id":"c05933b6_ad6977ec","line":197,"updated":"2022-11-14 17:17:32.000000000","message":"one thing we discussed was wether nova\u0027s nova-manage heal-allocations  command might be able to fix things in place. if not that is perhaps an improment we coudl make in the future.\n\nnova would fix the allcoation on a cold/live migrate so that is an option albit a somewhat expensive one to adress this. this supprot was added in ussuri by \nhttps://specs.openstack.org/openstack/nova-specs/specs/ussuri/implemented/support-move-ops-with-qos-ports-ussuri.html\n\na less invaisve change would be to detach and attach the ports with qos policies\nwhich was added in wallaby.\n\nhttps://specs.openstack.org/openstack/nova-specs/specs/wallaby/implemented/support-interface-attach-with-qos-ports.html\n\nthe other option is to just remove teh min bandwith qos policy form the neutron ports before upgrading.\n\ni think documenting these options is likely the minium we need to do to enabel operators to determin the best parth forward for them in the A cycle.","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"9c85d625de65351e5e4ad16c87ca446e17e80cef","unresolved":false,"context_lines":[{"line_number":194,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":195,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":196,"context_line":"this action, a new spec/bug will be created to track this improvement, but"},{"line_number":197,"context_line":"this is out of scope in this spec."},{"line_number":198,"context_line":""},{"line_number":199,"context_line":""},{"line_number":200,"context_line":"Implementation"}],"source_content_type":"text/x-rst","patch_set":8,"id":"e19b21e9_f16c1c03","line":197,"in_reply_to":"c05933b6_ad6977ec","updated":"2022-11-15 08:36:45.000000000","message":"I\u0027ll refer to it in this spec and then I\u0027ll include these comments in the documentation.","commit_id":"1b2506fea06d524218886ae5079121155a9e51b2"},{"author":{"_account_id":15554,"name":"Bence Romsics","email":"bence.romsics@gmail.com","username":"ebenrom","status":"working for Ericsson, UTC+1 (+DST)"},"change_message_id":"dfe45caf0260ccaf07c4493cb7b7ee60a10f553e","unresolved":true,"context_lines":[{"line_number":198,"context_line":"  state (with ports located in overlay networks with QoS policies and minimum"},{"line_number":199,"context_line":"  bandwidth rules)."},{"line_number":200,"context_line":""},{"line_number":201,"context_line":"This spec does not consider the rebuilt of the current allocations. Any port"},{"line_number":202,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":203,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":204,"context_line":"this action, a new spec/bug will be created to track this improvement, but"}],"source_content_type":"text/x-rst","patch_set":9,"id":"87e3a8c0_d37fa0a1","line":201,"range":{"start_line":201,"start_character":32,"end_line":201,"end_character":39},"updated":"2022-11-15 12:08:24.000000000","message":"nit: rebuild","commit_id":"38d9b49b9547ec57be54c953b78ff04d3ed4049c"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"67de389b993e8ee00138bccba74eb7379d670a33","unresolved":false,"context_lines":[{"line_number":198,"context_line":"  state (with ports located in overlay networks with QoS policies and minimum"},{"line_number":199,"context_line":"  bandwidth rules)."},{"line_number":200,"context_line":""},{"line_number":201,"context_line":"This spec does not consider the rebuilt of the current allocations. Any port"},{"line_number":202,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":203,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":204,"context_line":"this action, a new spec/bug will be created to track this improvement, but"}],"source_content_type":"text/x-rst","patch_set":9,"id":"8c2d464a_28c7f939","line":201,"range":{"start_line":201,"start_character":32,"end_line":201,"end_character":39},"in_reply_to":"87e3a8c0_d37fa0a1","updated":"2022-11-15 15:41:02.000000000","message":"Done","commit_id":"38d9b49b9547ec57be54c953b78ff04d3ed4049c"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"4ee28263cbcd59ce0f2e2e72ce5d9c621296bb33","unresolved":true,"context_lines":[{"line_number":147,"context_line":"   scheduling functionality to ports in overlay networks. In case of having"},{"line_number":148,"context_line":"   shared resources, the administrator will need to split bandwidth assignation"},{"line_number":149,"context_line":"   between resource providers. Currently Placement API nor Neutron cannot"},{"line_number":150,"context_line":"   provide a way to model a shared resource."},{"line_number":151,"context_line":""},{"line_number":152,"context_line":""},{"line_number":153,"context_line":"REST API Impact"}],"source_content_type":"text/x-rst","patch_set":10,"id":"4a3bc969_ad410735","line":150,"updated":"2022-11-15 16:26:03.000000000","message":"Thanks. This is now stated explicitly so it is OK to me.","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"4ee28263cbcd59ce0f2e2e72ce5d9c621296bb33","unresolved":true,"context_lines":[{"line_number":202,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":203,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":204,"context_line":"this action, a new spec/bug will be created to track this improvement, but"},{"line_number":205,"context_line":"this is out of scope in this spec."},{"line_number":206,"context_line":""},{"line_number":207,"context_line":"Part of this RFE will be to document the alternatives the user has to, in"},{"line_number":208,"context_line":"case of having a port with minimum bandwidth rules before enabling this"}],"source_content_type":"text/x-rst","patch_set":10,"id":"bc9ae3cf_66209468","line":205,"updated":"2022-11-15 16:26:03.000000000","message":"This is pretty unfortunate. This means that on that host the resource tracking will be incorrect.\n\nI do believe that nova-manage placement heal_allocations could be used for this with limited extra effort.","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"764a35a694d07cbf235653814c7cbb10dd0f1484","unresolved":true,"context_lines":[{"line_number":202,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":203,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":204,"context_line":"this action, a new spec/bug will be created to track this improvement, but"},{"line_number":205,"context_line":"this is out of scope in this spec."},{"line_number":206,"context_line":""},{"line_number":207,"context_line":"Part of this RFE will be to document the alternatives the user has to, in"},{"line_number":208,"context_line":"case of having a port with minimum bandwidth rules before enabling this"}],"source_content_type":"text/x-rst","patch_set":10,"id":"e8a33049_cf77349f","line":205,"in_reply_to":"144fdec6_62e2e5f7","updated":"2022-11-16 09:12:03.000000000","message":"yep I\u0027m 75% sure it will work out of the box but testing is needed.","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"97421480ca21897803d7e59fc07526fb5324bd34","unresolved":true,"context_lines":[{"line_number":202,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":203,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":204,"context_line":"this action, a new spec/bug will be created to track this improvement, but"},{"line_number":205,"context_line":"this is out of scope in this spec."},{"line_number":206,"context_line":""},{"line_number":207,"context_line":"Part of this RFE will be to document the alternatives the user has to, in"},{"line_number":208,"context_line":"case of having a port with minimum bandwidth rules before enabling this"}],"source_content_type":"text/x-rst","patch_set":10,"id":"144fdec6_62e2e5f7","line":205,"in_reply_to":"55ad75fd_cdde1000","updated":"2022-11-15 17:32:09.000000000","message":"https://docs.openstack.org/nova/latest/cli/nova-manage.html#placement-heal-allocations\n\nnova-manage placement heal_allocations \n\nwill try to heal all allcotion for all instance automatically\n\n\"\"\"\nAlso if the instance has any port attached that has resource request (e.g. Quality of Service (QoS): Guaranteed Bandwidth) but the corresponding allocation is not found then the allocation is created against the network device resource providers according to the resource request of that port. It is possible that the missing allocation cannot be created either due to not having enough resource inventory on the host the instance resides on or because more than one resource provider could fulfill the request. In this case the instance needs to be manually deleted or the port needs to be detached. When nova supports migrating instances with guaranteed bandwidth ports, migration will heal missing allocations for these instances.\n\"\"\"\n\nso if neutorn has the correct info when this is run it should fix everything\nbut we would need to test that and confirm.","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"8f8ef376c527cf8432077d81065da3bb3e9bae9b","unresolved":true,"context_lines":[{"line_number":202,"context_line":"already present in a host that creates a new resource provider for tunnelled"},{"line_number":203,"context_line":"networks, won\u0027t be allocated. Once there is standard a procedure to perform"},{"line_number":204,"context_line":"this action, a new spec/bug will be created to track this improvement, but"},{"line_number":205,"context_line":"this is out of scope in this spec."},{"line_number":206,"context_line":""},{"line_number":207,"context_line":"Part of this RFE will be to document the alternatives the user has to, in"},{"line_number":208,"context_line":"case of having a port with minimum bandwidth rules before enabling this"}],"source_content_type":"text/x-rst","patch_set":10,"id":"55ad75fd_cdde1000","line":205,"in_reply_to":"bc9ae3cf_66209468","updated":"2022-11-15 16:50:59.000000000","message":"Do you have an example of use?","commit_id":"5c12c2b8b85aec899904defe43784367c11f5a46"}],"specs/2023.1/strict-minimum-bandwidth-tunnelled-networks.srt":[{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"a576d0852b311a6c88b3c31943626501e86de007","unresolved":true,"context_lines":[{"line_number":119,"context_line":"Other Impact"},{"line_number":120,"context_line":"------------"},{"line_number":121,"context_line":""},{"line_number":122,"context_line":"(talk about current ports in tunnelled networks with min-QoS)"},{"line_number":123,"context_line":""},{"line_number":124,"context_line":""},{"line_number":125,"context_line":""}],"source_content_type":"application/octet-stream","patch_set":1,"id":"fabba4ab_ae29e93f","line":122,"range":{"start_line":122,"start_character":0,"end_line":122,"end_character":21},"updated":"2022-10-10 17:04:31.000000000","message":"To be done.","commit_id":"f1bc36c00189f41fa2f8042f80836fd74bc27ff2"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"7906ba5a650c8fd33160b6aa4f9ab3bf69ed3919","unresolved":false,"context_lines":[{"line_number":119,"context_line":"Other Impact"},{"line_number":120,"context_line":"------------"},{"line_number":121,"context_line":""},{"line_number":122,"context_line":"(talk about current ports in tunnelled networks with min-QoS)"},{"line_number":123,"context_line":""},{"line_number":124,"context_line":""},{"line_number":125,"context_line":""}],"source_content_type":"application/octet-stream","patch_set":1,"id":"be9abba1_7c8f566a","line":122,"range":{"start_line":122,"start_character":0,"end_line":122,"end_character":21},"in_reply_to":"fabba4ab_ae29e93f","updated":"2022-10-11 09:16:02.000000000","message":"Done","commit_id":"f1bc36c00189f41fa2f8042f80836fd74bc27ff2"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"be4cfc002d7e685994143dd09df20bda2688519f","unresolved":true,"context_lines":[{"line_number":56,"context_line":""},{"line_number":57,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"},{"line_number":58,"context_line":"any possible clash, there will be a new check when parsing the physical"},{"line_number":59,"context_line":"network bridge mappings."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"A host with ML2/OVN backend with a physical network (mapped to the physical"},{"line_number":62,"context_line":"bridge \"br-ex\") and a tunnelled network will report the following resource"}],"source_content_type":"application/octet-stream","patch_set":2,"id":"03da1a37_bd9f52fe","line":59,"updated":"2022-10-26 07:58:00.000000000","message":"I don\u0027t think any operator will have clash here but maybe we should consider making  this name to be configurable in Neutron to avoid potential problems during upgrade process?","commit_id":"0c8ccb5191a2ebb93ed7601a161461f283d85b46"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"176d247e6df5f308110c802200f1ae616b85bdef","unresolved":false,"context_lines":[{"line_number":56,"context_line":""},{"line_number":57,"context_line":"This new string constant cannot be used as a physical bridge name. To avoid"},{"line_number":58,"context_line":"any possible clash, there will be a new check when parsing the physical"},{"line_number":59,"context_line":"network bridge mappings."},{"line_number":60,"context_line":""},{"line_number":61,"context_line":"A host with ML2/OVN backend with a physical network (mapped to the physical"},{"line_number":62,"context_line":"bridge \"br-ex\") and a tunnelled network will report the following resource"}],"source_content_type":"application/octet-stream","patch_set":2,"id":"f320097d_320f3536","line":59,"in_reply_to":"03da1a37_bd9f52fe","updated":"2022-10-27 10:54:49.000000000","message":"Ok, but it could be a problem if this configuration parameter is changed once the RP is created. In any case, I\u0027ll add a warning in the config variable description.","commit_id":"0c8ccb5191a2ebb93ed7601a161461f283d85b46"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"be4cfc002d7e685994143dd09df20bda2688519f","unresolved":true,"context_lines":[{"line_number":126,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":127,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":128,"context_line":"network). A new check will be added to inform about those ports located on"},{"line_number":129,"context_line":"tunnelled networks with minimum bandwidth QoS rules."},{"line_number":130,"context_line":""},{"line_number":131,"context_line":""},{"line_number":132,"context_line":"Implementation"}],"source_content_type":"application/octet-stream","patch_set":2,"id":"06f6d33e_2041d3ee","line":129,"updated":"2022-10-26 07:58:00.000000000","message":"+1 for new upgrade check","commit_id":"0c8ccb5191a2ebb93ed7601a161461f283d85b46"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"176d247e6df5f308110c802200f1ae616b85bdef","unresolved":false,"context_lines":[{"line_number":126,"context_line":"rules won\u0027t be discarded, like now, when the port resource request is built"},{"line_number":127,"context_line":"(that is the Placement blob to request a specific bandwidth in a specific"},{"line_number":128,"context_line":"network). A new check will be added to inform about those ports located on"},{"line_number":129,"context_line":"tunnelled networks with minimum bandwidth QoS rules."},{"line_number":130,"context_line":""},{"line_number":131,"context_line":""},{"line_number":132,"context_line":"Implementation"}],"source_content_type":"application/octet-stream","patch_set":2,"id":"29f6ae0a_67982845","line":129,"in_reply_to":"06f6d33e_2041d3ee","updated":"2022-10-27 10:54:49.000000000","message":"Done","commit_id":"0c8ccb5191a2ebb93ed7601a161461f283d85b46"},{"author":{"_account_id":11975,"name":"Slawek Kaplonski","email":"skaplons@redhat.com","username":"slaweq"},"change_message_id":"be4cfc002d7e685994143dd09df20bda2688519f","unresolved":true,"context_lines":[{"line_number":150,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":151,"context_line":""},{"line_number":152,"context_line":"* Unit/functional Tests."},{"line_number":153,"context_line":"* Tempest tests."},{"line_number":154,"context_line":""},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"Documentation Impact"}],"source_content_type":"application/octet-stream","patch_set":2,"id":"db6e1ca8_a189d10d","line":153,"updated":"2022-10-26 07:58:00.000000000","message":"As we recently discussed internally, can You maybe elaborate more about what kind of tempest tests do You think should be added to test it? Some API tests or scenario test?","commit_id":"0c8ccb5191a2ebb93ed7601a161461f283d85b46"},{"author":{"_account_id":16688,"name":"Rodolfo Alonso","email":"ralonsoh@redhat.com","username":"rodolfo-alonso-hernandez"},"change_message_id":"176d247e6df5f308110c802200f1ae616b85bdef","unresolved":false,"context_lines":[{"line_number":150,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":151,"context_line":""},{"line_number":152,"context_line":"* Unit/functional Tests."},{"line_number":153,"context_line":"* Tempest tests."},{"line_number":154,"context_line":""},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"Documentation Impact"}],"source_content_type":"application/octet-stream","patch_set":2,"id":"56bcf600_9b4f3570","line":153,"in_reply_to":"db6e1ca8_a189d10d","updated":"2022-10-27 10:54:49.000000000","message":"The API tests are already covered in tempest.\nInstead of tempest tests, what I should implement is fullstack tests, checking the mech driver agent (OVS, OVN) has correctly reported the requested BW.","commit_id":"0c8ccb5191a2ebb93ed7601a161461f283d85b46"}]}