)]}'
{"specs/liberty/ironic-ml2-integration.rst":[{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":24,"context_line":"virtual machines are."},{"line_number":25,"context_line":""},{"line_number":26,"context_line":"In order to provide the required network isolation, Ironic should be able to"},{"line_number":27,"context_line":"provide the requisite connectivity information (using LLDP) to the Neutron"},{"line_number":28,"context_line":"ML2 plugin to allow drivers to provision the top-of-rack (ToR) switch for the"},{"line_number":29,"context_line":"baremetal server."},{"line_number":30,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_2c0a5ed3","line":27,"updated":"2015-06-05 12:52:57.000000000","message":"s/requisite/required/ ?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":24,"context_line":"virtual machines are."},{"line_number":25,"context_line":""},{"line_number":26,"context_line":"In order to provide the required network isolation, Ironic should be able to"},{"line_number":27,"context_line":"provide the requisite connectivity information (using LLDP) to the Neutron"},{"line_number":28,"context_line":"ML2 plugin to allow drivers to provision the top-of-rack (ToR) switch for the"},{"line_number":29,"context_line":"baremetal server."},{"line_number":30,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_278dd731","line":27,"in_reply_to":"3afb71cf_2c0a5ed3","updated":"2015-06-05 13:27:26.000000000","message":"Either is ok. I\u0027m happy to change if you think it makes it more accessible for a wider range of people.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13719,"name":"Naohiro Tamura","email":"naohirot.openstack@gmail.com","username":"nao"},"change_message_id":"fc602fda271ec18492aac6fa2c216f1ebb09193b","unresolved":false,"context_lines":[{"line_number":31,"context_line":"Ironic also poses new challenges for Neutron, for example, the concept of"},{"line_number":32,"context_line":"aggregating links, however, we believe a complementary specification is not"},{"line_number":33,"context_line":"required as the current functionality in Neutron allows for updates to be"},{"line_number":34,"context_line":"confined to ML2 drivers."},{"line_number":35,"context_line":""},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_69025a0d","line":34,"updated":"2015-06-08 07:24:18.000000000","message":"It would not be easy to understand what this paragraph (L.28 - L.34) is trying to say in \"Problem description\"?\n\nI took that this spec doesn\u0027t cover the concept of aggregation, am I correct?\n\nThe concept of aggregation links is one of new challenges Ironic poses, I wondered what are the rest of challenges? I know this is the first draft, we need to state clearly if those challenges are important in the subsequent updates.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13997,"name":"SHIGEMATSU Mitsuhiro","email":"shigematsu.mitsuhiro@lab.ntt.co.jp","username":"pshige"},"change_message_id":"4443c825b1bf21613997ce60f768db3ac9326798","unresolved":false,"context_lines":[{"line_number":31,"context_line":"Ironic also poses new challenges for Neutron, for example, the concept of"},{"line_number":32,"context_line":"aggregating links, however, we believe a complementary specification is not"},{"line_number":33,"context_line":"required as the current functionality in Neutron allows for updates to be"},{"line_number":34,"context_line":"confined to ML2 drivers."},{"line_number":35,"context_line":""},{"line_number":36,"context_line":""},{"line_number":37,"context_line":"Proposed change"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_8fc5848f","line":34,"in_reply_to":"3afb71cf_69025a0d","updated":"2015-06-08 14:52:14.000000000","message":"At the same time, even if we will not cover aggregation problems in a few months, we\u0027d like to solve the others by the way we can relatively easily extend function of aggregations in the near future.  We should make a deeper discussion at this point.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10375,"name":"Om Kumar","email":"om.kumar@hpe.com","username":"lazy_prince"},"change_message_id":"9cabf56b2f6790f451072fed6ef11794e5e61381","unresolved":false,"context_lines":[{"line_number":45,"context_line":"Ironic should then send the local link information to Neutron in the form of a"},{"line_number":46,"context_line":"port binding profile. If there are port groups, they should be created as a"},{"line_number":47,"context_line":"Neutron port and have a binding profile stating that it is an aggregate. Each"},{"line_number":48,"context_line":"port within a part group should have its local link information registered in"},{"line_number":49,"context_line":"the group\u0027s binding profile. Ironic ports that are not part of a port"},{"line_number":50,"context_line":"group will have a corresponding Neutron port where their local link information"},{"line_number":51,"context_line":"will be registered."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_564ac784","line":48,"updated":"2015-06-05 12:27:44.000000000","message":"s/part/port/","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":45,"context_line":"Ironic should then send the local link information to Neutron in the form of a"},{"line_number":46,"context_line":"port binding profile. If there are port groups, they should be created as a"},{"line_number":47,"context_line":"Neutron port and have a binding profile stating that it is an aggregate. Each"},{"line_number":48,"context_line":"port within a part group should have its local link information registered in"},{"line_number":49,"context_line":"the group\u0027s binding profile. Ironic ports that are not part of a port"},{"line_number":50,"context_line":"group will have a corresponding Neutron port where their local link information"},{"line_number":51,"context_line":"will be registered."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_a779c731","line":48,"in_reply_to":"3afb71cf_564ac784","updated":"2015-06-05 13:27:26.000000000","message":"Thanks","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10375,"name":"Om Kumar","email":"om.kumar@hpe.com","username":"lazy_prince"},"change_message_id":"9cabf56b2f6790f451072fed6ef11794e5e61381","unresolved":false,"context_lines":[{"line_number":50,"context_line":"group will have a corresponding Neutron port where their local link information"},{"line_number":51,"context_line":"will be registered."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"The network should first be placed on a provisioning network and then onto the"},{"line_number":54,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":55,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":56,"context_line":"facilitate this there will be a variable stored in the binding profile to"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_1109f92c","line":53,"updated":"2015-06-05 12:27:44.000000000","message":"How do we plan to support cinder volumes for the baremetal when it becomes available? My understanding is that cinder volumes will not be accessible from tenant network.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13997,"name":"SHIGEMATSU Mitsuhiro","email":"shigematsu.mitsuhiro@lab.ntt.co.jp","username":"pshige"},"change_message_id":"4443c825b1bf21613997ce60f768db3ac9326798","unresolved":false,"context_lines":[{"line_number":50,"context_line":"group will have a corresponding Neutron port where their local link information"},{"line_number":51,"context_line":"will be registered."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"The network should first be placed on a provisioning network and then onto the"},{"line_number":54,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":55,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":56,"context_line":"facilitate this there will be a variable stored in the binding profile to"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_efd63826","line":53,"in_reply_to":"3afb71cf_1109f92c","updated":"2015-06-08 14:52:14.000000000","message":"I am also concerned about it.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":50,"context_line":"group will have a corresponding Neutron port where their local link information"},{"line_number":51,"context_line":"will be registered."},{"line_number":52,"context_line":""},{"line_number":53,"context_line":"The network should first be placed on a provisioning network and then onto the"},{"line_number":54,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":55,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":56,"context_line":"facilitate this there will be a variable stored in the binding profile to"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_c702833e","line":53,"in_reply_to":"3afb71cf_1109f92c","updated":"2015-06-05 13:27:26.000000000","message":"Thanks. We\u0027ll need to look into that.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":57,"context_line":"record whether a port binding has been requested. This will allow Ironic to"},{"line_number":58,"context_line":"defer binding until Ironic is able to populate the extra information that is"},{"line_number":59,"context_line":"required by Neutron. Note also, there will be no support for PXE booting after"},{"line_number":60,"context_line":"deploy (i.e. local disk installation only)."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"A supportive ML2 driver can then use the information supplied to provision the"},{"line_number":63,"context_line":"port."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_0c0fa2e2","line":60,"updated":"2015-06-05 12:52:57.000000000","message":"So I guess we should explicitly deny instance netboot, when this feature is used... maybe even make it the default","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13719,"name":"Naohiro Tamura","email":"naohirot.openstack@gmail.com","username":"nao"},"change_message_id":"fc602fda271ec18492aac6fa2c216f1ebb09193b","unresolved":false,"context_lines":[{"line_number":57,"context_line":"record whether a port binding has been requested. This will allow Ironic to"},{"line_number":58,"context_line":"defer binding until Ironic is able to populate the extra information that is"},{"line_number":59,"context_line":"required by Neutron. Note also, there will be no support for PXE booting after"},{"line_number":60,"context_line":"deploy (i.e. local disk installation only)."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"A supportive ML2 driver can then use the information supplied to provision the"},{"line_number":63,"context_line":"port."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_8cc5a43d","line":60,"in_reply_to":"3afb71cf_0c0fa2e2","updated":"2015-06-08 07:24:18.000000000","message":"If \"netboot\" is not supported after deploy, some brief explanation about why not supporting pxe boot after deploy would be helpful for those who didn\u0027t attend the discussion.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13997,"name":"SHIGEMATSU Mitsuhiro","email":"shigematsu.mitsuhiro@lab.ntt.co.jp","username":"pshige"},"change_message_id":"4443c825b1bf21613997ce60f768db3ac9326798","unresolved":false,"context_lines":[{"line_number":57,"context_line":"record whether a port binding has been requested. This will allow Ironic to"},{"line_number":58,"context_line":"defer binding until Ironic is able to populate the extra information that is"},{"line_number":59,"context_line":"required by Neutron. Note also, there will be no support for PXE booting after"},{"line_number":60,"context_line":"deploy (i.e. local disk installation only)."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"A supportive ML2 driver can then use the information supplied to provision the"},{"line_number":63,"context_line":"port."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_afd4c0bc","line":60,"in_reply_to":"3afb71cf_0d4ac7ac","updated":"2015-06-08 14:52:14.000000000","message":"Um, localboot only, just as I expected...","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":57,"context_line":"record whether a port binding has been requested. This will allow Ironic to"},{"line_number":58,"context_line":"defer binding until Ironic is able to populate the extra information that is"},{"line_number":59,"context_line":"required by Neutron. Note also, there will be no support for PXE booting after"},{"line_number":60,"context_line":"deploy (i.e. local disk installation only)."},{"line_number":61,"context_line":""},{"line_number":62,"context_line":"A supportive ML2 driver can then use the information supplied to provision the"},{"line_number":63,"context_line":"port."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_0d4ac7ac","line":60,"in_reply_to":"3afb71cf_8cc5a43d","updated":"2015-06-08 11:31:41.000000000","message":"Ok, will add.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The Ironic port object will be updated with a new local_link_connection field"},{"line_number":66,"context_line":"dict supplying the following values:"},{"line_number":67,"context_line":"* chassis_id"},{"line_number":68,"context_line":"* port_id"},{"line_number":69,"context_line":"* system_name"},{"line_number":70,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_ec13260c","line":67,"updated":"2015-06-05 12:52:57.000000000","message":"can\u0027t we reuse existing Ironic chassis object?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The Ironic port object will be updated with a new local_link_connection field"},{"line_number":66,"context_line":"dict supplying the following values:"},{"line_number":67,"context_line":"* chassis_id"},{"line_number":68,"context_line":"* port_id"},{"line_number":69,"context_line":"* system_name"},{"line_number":70,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_e7b45fd2","line":67,"in_reply_to":"3afb71cf_ec13260c","updated":"2015-06-05 13:27:26.000000000","message":"Not the same thing I\u0027m afraid. This is the chassis of the connected switch. I\u0027ve derived the keys from the LLDP TLVs. If this is confusing we could change but I would rather not.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"357a97f9b2e0911c529c59eea2ac805b5f575906","unresolved":false,"context_lines":[{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The Ironic port object will be updated with a new local_link_connection field"},{"line_number":66,"context_line":"dict supplying the following values:"},{"line_number":67,"context_line":"* chassis_id"},{"line_number":68,"context_line":"* port_id"},{"line_number":69,"context_line":"* system_name"},{"line_number":70,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_27a037f4","line":67,"in_reply_to":"3afb71cf_ec13260c","updated":"2015-06-05 13:35:45.000000000","message":"This would refer to the switch chassis_id","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10358,"name":"Brad Morgan","email":"brad@morgabra.com","username":"morgabra"},"change_message_id":"7a1211bac2291ef1f51cbd1e7ac3ced44eefcd6c","unresolved":false,"context_lines":[{"line_number":71,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":72,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":73,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":74,"context_line":"expect."},{"line_number":75,"context_line":""},{"line_number":76,"context_line":"To facilitate link aggregation a new port_group object will be created. In"},{"line_number":77,"context_line":"addition to the base object it will have the following fields:"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_8a3f0fd6","line":74,"updated":"2015-06-04 20:21:47.000000000","message":"Should we validate this field internally, to explicitly disallow having multiple ironic nodes with the same link_connection?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":71,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":72,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":73,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":74,"context_line":"expect."},{"line_number":75,"context_line":""},{"line_number":76,"context_line":"To facilitate link aggregation a new port_group object will be created. In"},{"line_number":77,"context_line":"addition to the base object it will have the following fields:"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_ccf98aca","line":74,"updated":"2015-06-05 12:52:57.000000000","message":"Sounds good. Which fields/combination of fields should be unique? maybe we should add seperate fields with UNIQUE constraint instead of a dict?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"2a514b81d51b8c459cc1583a54701ea62d5b2afb","unresolved":false,"context_lines":[{"line_number":71,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":72,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":73,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":74,"context_line":"expect."},{"line_number":75,"context_line":""},{"line_number":76,"context_line":"To facilitate link aggregation a new port_group object will be created. In"},{"line_number":77,"context_line":"addition to the base object it will have the following fields:"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_f34c3a78","line":74,"in_reply_to":"3afb71cf_8a3f0fd6","updated":"2015-06-04 21:26:31.000000000","message":"I think this is a good idea - this will filter out any accidental misconfigurations.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":71,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":72,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":73,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":74,"context_line":"expect."},{"line_number":75,"context_line":""},{"line_number":76,"context_line":"To facilitate link aggregation a new port_group object will be created. In"},{"line_number":77,"context_line":"addition to the base object it will have the following fields:"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_27303756","line":74,"in_reply_to":"3afb71cf_ccf98aca","updated":"2015-06-05 13:27:26.000000000","message":"I think the problem with adding separate fields is that different drivers may need different TLVs. Validating the field internally could be good. I think the bigger issue is when they go to Neutron. There may be a security issue in being able to change these values in the binding:profile.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"fa0c4536151fb78972a643a3f877abb6dcd96792","unresolved":false,"context_lines":[{"line_number":79,"context_line":"* uuid"},{"line_number":80,"context_line":"* name"},{"line_number":81,"context_line":"* node_id"},{"line_number":82,"context_line":"* address"},{"line_number":83,"context_line":"* extra"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The Ironic port object will then have the following fields added to support"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_3a0dd03e","line":82,"updated":"2015-06-06 19:01:08.000000000","message":"Better to name this field as mac_address as it actually represents MAC address for bonded NICs of baremetal server.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":79,"context_line":"* uuid"},{"line_number":80,"context_line":"* name"},{"line_number":81,"context_line":"* node_id"},{"line_number":82,"context_line":"* address"},{"line_number":83,"context_line":"* extra"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The Ironic port object will then have the following fields added to support"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_acf64ed5","line":82,"updated":"2015-06-05 12:52:57.000000000","message":"what is this address field?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":79,"context_line":"* uuid"},{"line_number":80,"context_line":"* name"},{"line_number":81,"context_line":"* node_id"},{"line_number":82,"context_line":"* address"},{"line_number":83,"context_line":"* extra"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The Ironic port object will then have the following fields added to support"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_ad5853fb","line":82,"in_reply_to":"3afb71cf_3a0dd03e","updated":"2015-06-08 11:31:41.000000000","message":"The current Ironic port object uses \u0027address\u0027 for MAC address hence the approach taken here. I don\u0027t have a strong opinion either way but if we are not being consistent my opinion would be that we would have to bring this up in an Ironic meeting and get consensus to change this.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":79,"context_line":"* uuid"},{"line_number":80,"context_line":"* name"},{"line_number":81,"context_line":"* node_id"},{"line_number":82,"context_line":"* address"},{"line_number":83,"context_line":"* extra"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The Ironic port object will then have the following fields added to support"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_e710dfaa","line":82,"in_reply_to":"3afb71cf_acf64ed5","updated":"2015-06-05 13:27:26.000000000","message":"A bond needs a MAC address. If we\u0027re going to provision it from Neutron we need to set it before.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"357a97f9b2e0911c529c59eea2ac805b5f575906","unresolved":false,"context_lines":[{"line_number":79,"context_line":"* uuid"},{"line_number":80,"context_line":"* name"},{"line_number":81,"context_line":"* node_id"},{"line_number":82,"context_line":"* address"},{"line_number":83,"context_line":"* extra"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The Ironic port object will then have the following fields added to support"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_a7dae776","line":82,"in_reply_to":"3afb71cf_acf64ed5","updated":"2015-06-05 13:35:45.000000000","message":"This would be the address associated with the port-group. I was thinking it was the pxe address. More comments on pxe below.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13997,"name":"SHIGEMATSU Mitsuhiro","email":"shigematsu.mitsuhiro@lab.ntt.co.jp","username":"pshige"},"change_message_id":"4443c825b1bf21613997ce60f768db3ac9326798","unresolved":false,"context_lines":[{"line_number":80,"context_line":"* name"},{"line_number":81,"context_line":"* node_id"},{"line_number":82,"context_line":"* address"},{"line_number":83,"context_line":"* extra"},{"line_number":84,"context_line":""},{"line_number":85,"context_line":"The Ironic port object will then have the following fields added to support"},{"line_number":86,"context_line":"new functionality:"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_6fac8835","line":83,"updated":"2015-06-08 14:52:14.000000000","message":"what do we use this extra field, for example?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":91,"context_line":"The pxe_enabled field is required to allow selection of the PXE address in"},{"line_number":92,"context_line":"the scenario where there are multiple ports (e.g. in the case of port groups)."},{"line_number":93,"context_line":"It is thought that introspection methods could automatically set this value"},{"line_number":94,"context_line":"e.g. discoverd could enable the interface that it booted from. In the case"},{"line_number":95,"context_line":"where there are multiple ports enabled for PXE booting the selection process"},{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_8c0392b8","line":94,"updated":"2015-06-05 12:52:57.000000000","message":"s/discoverd/inspector/  - we renamed it recently","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":91,"context_line":"The pxe_enabled field is required to allow selection of the PXE address in"},{"line_number":92,"context_line":"the scenario where there are multiple ports (e.g. in the case of port groups)."},{"line_number":93,"context_line":"It is thought that introspection methods could automatically set this value"},{"line_number":94,"context_line":"e.g. discoverd could enable the interface that it booted from. In the case"},{"line_number":95,"context_line":"where there are multiple ports enabled for PXE booting the selection process"},{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_87200bf9","line":94,"in_reply_to":"3afb71cf_8c0392b8","updated":"2015-06-05 13:27:26.000000000","message":"Thanks","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":92,"context_line":"the scenario where there are multiple ports (e.g. in the case of port groups)."},{"line_number":93,"context_line":"It is thought that introspection methods could automatically set this value"},{"line_number":94,"context_line":"e.g. discoverd could enable the interface that it booted from. In the case"},{"line_number":95,"context_line":"where there are multiple ports enabled for PXE booting the selection process"},{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"* Select port groups with an address that corresponds to an enabled port\u0027s"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_6c0056ab","line":95,"updated":"2015-06-05 12:52:57.000000000","message":"not sure how it can figure out it: currently I use kernel cmdline variables supplied by PXE boot loader, and it\u0027s usually only one MAC","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":95,"context_line":"where there are multiple ports enabled for PXE booting the selection process"},{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"* Select port groups with an address that corresponds to an enabled port\u0027s"},{"line_number":99,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Select enabled ports from port groups. If there are multiple ports choose"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_4c2c7a4a","line":98,"updated":"2015-06-05 12:52:57.000000000","message":"\"enabled port\" - do you mean Ironic or Neutron port here?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":95,"context_line":"where there are multiple ports enabled for PXE booting the selection process"},{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"* Select port groups with an address that corresponds to an enabled port\u0027s"},{"line_number":99,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Select enabled ports from port groups. If there are multiple ports choose"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_a76ae718","line":98,"in_reply_to":"3afb71cf_4c2c7a4a","updated":"2015-06-05 13:27:26.000000000","message":"Will add pxe_ to the front to prevent confusion","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10358,"name":"Brad Morgan","email":"brad@morgabra.com","username":"morgabra"},"change_message_id":"7a1211bac2291ef1f51cbd1e7ac3ced44eefcd6c","unresolved":false,"context_lines":[{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"* Select port groups with an address that corresponds to an enabled port\u0027s"},{"line_number":99,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Select enabled ports from port groups. If there are multiple ports choose"},{"line_number":102,"context_line":"  one at random."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_2ab93b2f","line":99,"updated":"2015-06-04 20:21:47.000000000","message":"Could we simplify this and pass all ports with pxe_enabled\u003dTrue and set is_link_aggregate\u003dFalse when creating the neutron port? There might be value in allowing multiple ports to be configured for PXE.\n\nAlternatively, if we are going to only choose one port to configure, explicitly disallow having more than one port with pxe_enabled\u003dTrue on a node. I could see this getting confusing as an operator if I didn\u0027t read the choosing code.\n\nOr, are we going to allow link aggregation for pxe? How does that work?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"37a49993d21cebb84b8961937cf49c229adb142c","unresolved":false,"context_lines":[{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"* Select port groups with an address that corresponds to an enabled port\u0027s"},{"line_number":99,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Select enabled ports from port groups. If there are multiple ports choose"},{"line_number":102,"context_line":"  one at random."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_36f3cb75","line":99,"in_reply_to":"3afb71cf_2ab93b2f","updated":"2015-06-05 12:15:00.000000000","message":"Link aggregation won\u0027t work for PXE as far as I know, however you can bring up the bond in the RAMdisk for faster connectivity after.\n\nHaving multiple PXE ports might be considered as a lot of overhead and I would need to search through the code to check what difficulties would arise with callbacks etc., however, I don\u0027t have a problem with it in principle.\n\nI\u0027m also happy with only having one port as pxe_enabled as long as we are aware that will change the current behavior on migration. Agree what is proposed could be confusing though I suppose operators could set one port only to remove the confusion but forcing them down that route could be the easiest to avoid queries.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"357a97f9b2e0911c529c59eea2ac805b5f575906","unresolved":false,"context_lines":[{"line_number":96,"context_line":"will be as follows:"},{"line_number":97,"context_line":""},{"line_number":98,"context_line":"* Select port groups with an address that corresponds to an enabled port\u0027s"},{"line_number":99,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":100,"context_line":""},{"line_number":101,"context_line":"* Select enabled ports from port groups. If there are multiple ports choose"},{"line_number":102,"context_line":"  one at random."}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_8c2f12a1","line":99,"in_reply_to":"3afb71cf_36f3cb75","updated":"2015-06-05 13:35:45.000000000","message":"I think the discussion on pxe_enabled is veering towards differentiation between physical and logical. At hardware level, there could be multiple interfaces enabled for PXE (bios level settings), but PXE should happen over just one. \n\nMy understanding of use of PXE address is that it\u0027s used for setting DHCP options against the Neutron port. There would be one Neutron port created per one Ironic port-group. I\u0027m not sure how 1 Neutron port would handle multiple addresses, but it sounds like it would need more work at Neutron side? If so, my suggestion would be that this spec focus on 1 PXE address per port-group. \n\nI\u0027ll think on the migration question.\n\nMy understanding is link aggregation for PXE is possible but the switch should be configured such that traffic flows through a single port until LACP PDUs are received - different switch vendors implement this in different ways. For example, on an Arista switch, lacp fallback can be enabled and this results in one active port (the active port is decided on priority settings). I\u0027m not sure of other vendor implementations.  Given the data passed in the binding profile, maybe the ML2 driver could ensure alignment but I think this would require the local_link_info in the binding profile to flag which port local_link_conenction dict is associated with the pxe address.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10358,"name":"Brad Morgan","email":"brad@morgabra.com","username":"morgabra"},"change_message_id":"7a1211bac2291ef1f51cbd1e7ac3ced44eefcd6c","unresolved":false,"context_lines":[{"line_number":111,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":112,"context_line":"| local_link_information | A list of the local link information from Ironic.  |"},{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_eae8c318","line":114,"updated":"2015-06-04 20:21:47.000000000","message":"Is there a plan for selecting which port/port group to use when attaching to a network?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"357a97f9b2e0911c529c59eea2ac805b5f575906","unresolved":false,"context_lines":[{"line_number":111,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":112,"context_line":"| local_link_information | A list of the local link information from Ironic.  |"},{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_e7869f12","line":114,"in_reply_to":"3afb71cf_96815f92","updated":"2015-06-05 13:35:45.000000000","message":"I was thinking this could also handle MLAG, which could allow 4 ports in the list.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"e5730504c2681cca84f07656d1600c4da8f70b3b","unresolved":false,"context_lines":[{"line_number":111,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":112,"context_line":"| local_link_information | A list of the local link information from Ironic.  |"},{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_49bb1ee6","line":114,"in_reply_to":"3afb71cf_e7869f12","updated":"2015-06-08 06:04:07.000000000","message":"I guess we could pass it as a list of local_link_informaiton (as suggested above). A given local_link_information dict will contain a list of 1 or 2 ports. \nThe list of local_link_information can contain 1 or 2 items. This cover a LAG as well MLAG situation - i.e. 2 port or 4 ports","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"29183bd5985586473e1d823de6e102b9be8efe84","unresolved":false,"context_lines":[{"line_number":111,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":112,"context_line":"| local_link_information | A list of the local link information from Ironic.  |"},{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_939d065c","line":114,"in_reply_to":"3afb71cf_eae8c318","updated":"2015-06-04 21:39:16.000000000","message":"I think it is assumed that all ports in the list are assumed to be part of LAG - which implies there will be either one or two ports in the list.\nIf there are more than 2 ports in the list, then you are correct, a mechanism is need to identify which two ports are part of a given LAG","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"37a49993d21cebb84b8961937cf49c229adb142c","unresolved":false,"context_lines":[{"line_number":111,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":112,"context_line":"| local_link_information | A list of the local link information from Ironic.  |"},{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_96815f92","line":114,"in_reply_to":"3afb71cf_eae8c318","updated":"2015-06-05 12:15:00.000000000","message":"I was hoping someone would suggest a good one :). At the minute you could pre-create the neutron ports but it isn\u0027t exactly nice.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10358,"name":"Brad Morgan","email":"brad@morgabra.com","username":"morgabra"},"change_message_id":"7a1211bac2291ef1f51cbd1e7ac3ced44eefcd6c","unresolved":false,"context_lines":[{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"},{"line_number":118,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":119,"context_line":"| bind_requested         | A boolean to specify whether to bind the port or   |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_0abb1f10","line":116,"updated":"2015-06-04 20:21:47.000000000","message":"This can be inferred if len(local_link_information) \u003e 1 right? Does it make sense to configure multiple ports for the same network in a non-aggregated fashion?\n\nThere\u0027s also a difference between link aggregation and trunking. It\u0027s totally okay to attach the same port/port group to multiple networks as long as those networks are not flat. Should we leave it up to neutron to fail if we try to do something that doesn\u0027t make sense?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"37a49993d21cebb84b8961937cf49c229adb142c","unresolved":false,"context_lines":[{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"},{"line_number":118,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":119,"context_line":"| bind_requested         | A boolean to specify whether to bind the port or   |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_96939ffa","line":116,"in_reply_to":"3afb71cf_0abb1f10","updated":"2015-06-05 12:15:00.000000000","message":"Agree, I was supposed to remove this before posting. Mea culpa.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"29183bd5985586473e1d823de6e102b9be8efe84","unresolved":false,"context_lines":[{"line_number":113,"context_line":"|                        | In the case of a single port it will have only one |"},{"line_number":114,"context_line":"|                        | element, the Ironic\u0027s port corresponding dict.     |"},{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"},{"line_number":118,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":119,"context_line":"| bind_requested         | A boolean to specify whether to bind the port or   |"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_13eeb6ec","line":116,"in_reply_to":"3afb71cf_0abb1f10","updated":"2015-06-04 21:39:16.000000000","message":"Yes, on len(local_link_information) \u003e 1 - its good idea. \n\nI think it will be assumed if multiple ports are connecting to same network, they are LAG\u0027ed ports (i.e. part of port channel)\n\nI do not believe (have to check to verify) that same port (or port group) can be attached to multiple networks. However, an instance can be connected to multiple networks by using multiple nics/ports.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"fa0c4536151fb78972a643a3f877abb6dcd96792","unresolved":false,"context_lines":[{"line_number":115,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":116,"context_line":"| is_link_aggregate      | A boolean to specify that this port describes an   |"},{"line_number":117,"context_line":"|                        | aggregate                                          |"},{"line_number":118,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":119,"context_line":"| bind_requested         | A boolean to specify whether to bind the port or   |"},{"line_number":120,"context_line":"|                        | defer binding."},{"line_number":121,"context_line":"+------------------------+----------------------------------------------------+"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_b5ff9100","line":118,"updated":"2015-06-06 19:01:08.000000000","message":"Neutron does not have a facility yet either to represent (or) to bind for trunk ports. So a single neutron port cannot be part of multiple networks.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13719,"name":"Naohiro Tamura","email":"naohirot.openstack@gmail.com","username":"nao"},"change_message_id":"fc602fda271ec18492aac6fa2c216f1ebb09193b","unresolved":false,"context_lines":[{"line_number":124,"context_line":"------------"},{"line_number":125,"context_line":""},{"line_number":126,"context_line":"The current model of prescribing flat networks could be maintained with the"},{"line_number":127,"context_line":"same flat network being used for everything."},{"line_number":128,"context_line":""},{"line_number":129,"context_line":"The in-progress Ironic-Neutron plugin could be updated with support from"},{"line_number":130,"context_line":"networking vendors. This would mean supporting a minimum of two layer 2"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_27f489f0","line":127,"updated":"2015-06-08 07:24:18.000000000","message":"I couldn\u0027t understand if there are alternatives or not.\n\nIt would be easy to understand for readers if the first paragraph of \"Alternatives\" simply begins with stating Yes or No.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":124,"context_line":"------------"},{"line_number":125,"context_line":""},{"line_number":126,"context_line":"The current model of prescribing flat networks could be maintained with the"},{"line_number":127,"context_line":"same flat network being used for everything."},{"line_number":128,"context_line":""},{"line_number":129,"context_line":"The in-progress Ironic-Neutron plugin could be updated with support from"},{"line_number":130,"context_line":"networking vendors. This would mean supporting a minimum of two layer 2"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_6db13b8f","line":127,"in_reply_to":"3afb71cf_27f489f0","updated":"2015-06-08 11:31:41.000000000","message":"This is following the current convention. I think this is more of a template discussion. If you believe that the template should prescribe a definitive statement at the start I think we should focus that on the template and not on a single specification.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"fa0c4536151fb78972a643a3f877abb6dcd96792","unresolved":false,"context_lines":[{"line_number":128,"context_line":""},{"line_number":129,"context_line":"The in-progress Ironic-Neutron plugin could be updated with support from"},{"line_number":130,"context_line":"networking vendors. This would mean supporting a minimum of two layer 2"},{"line_number":131,"context_line":"plugins and multiple plugins being required for a baremetal and virtual"},{"line_number":132,"context_line":"machine cloud."},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"Data model impact"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_d51a0d6e","line":131,"updated":"2015-06-06 19:01:08.000000000","message":"Can you please clarify what are the two layer 2 ironic-neutron plugins?\n\nAre they\na. flat network plugin\nb. flavor of rackspace plugin","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":128,"context_line":""},{"line_number":129,"context_line":"The in-progress Ironic-Neutron plugin could be updated with support from"},{"line_number":130,"context_line":"networking vendors. This would mean supporting a minimum of two layer 2"},{"line_number":131,"context_line":"plugins and multiple plugins being required for a baremetal and virtual"},{"line_number":132,"context_line":"machine cloud."},{"line_number":133,"context_line":""},{"line_number":134,"context_line":"Data model impact"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_ad8293fe","line":131,"in_reply_to":"3afb71cf_d51a0d6e","updated":"2015-06-08 11:31:41.000000000","message":"There aren\u0027t. This is taking the view of a networking vendor while writing drivers for a Neutron plugin. ML2 is a layer two plugin as is the Ironic-Neutron plugin. I don\u0027t think it is ideal if vendors have to write separate drivers for both. I\u0027ll update to make this clearer.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":13719,"name":"Naohiro Tamura","email":"naohirot.openstack@gmail.com","username":"nao"},"change_message_id":"fc602fda271ec18492aac6fa2c216f1ebb09193b","unresolved":false,"context_lines":[{"line_number":135,"context_line":"-----------------"},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed change will be to add the following fields to the port object with"},{"line_number":138,"context_line":"their data type and default value for migrations:"},{"line_number":139,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":140,"context_line":"| Field Name            | Field Type   | Migration Value |"},{"line_number":141,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\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":1,"id":"3afb71cf_67904144","line":138,"updated":"2015-06-08 07:24:18.000000000","message":"It seems the word \"migrations\" is appeard here at the first time.\nI just wondored what migrations are they? and from which and to which does it migrate what?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":135,"context_line":"-----------------"},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed change will be to add the following fields to the port object with"},{"line_number":138,"context_line":"their data type and default value for migrations:"},{"line_number":139,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":140,"context_line":"| Field Name            | Field Type   | Migration Value |"},{"line_number":141,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\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":1,"id":"3afb71cf_eddd2b15","line":138,"in_reply_to":"3afb71cf_67904144","updated":"2015-06-08 11:31:41.000000000","message":"Any update to the current data model requires a migration. Current ports may need to have values set for new columns hence the migration values.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"fa0c4536151fb78972a643a3f877abb6dcd96792","unresolved":false,"context_lines":[{"line_number":162,"context_line":"+-----------------------+-------------------------+"},{"line_number":163,"context_line":"| node_id               | str_or_none             |"},{"line_number":164,"context_line":"+-----------------------+-------------------------+"},{"line_number":165,"context_line":"| address               | str_or_none             |"},{"line_number":166,"context_line":"+-----------------------+-------------------------+"},{"line_number":167,"context_line":"| extra                 | dict_or_none            |"},{"line_number":168,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_b52671b8","line":165,"updated":"2015-06-06 19:01:08.000000000","message":"As earlier, good to explicitly use mac_address\u0027 here.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":162,"context_line":"+-----------------------+-------------------------+"},{"line_number":163,"context_line":"| node_id               | str_or_none             |"},{"line_number":164,"context_line":"+-----------------------+-------------------------+"},{"line_number":165,"context_line":"| address               | str_or_none             |"},{"line_number":166,"context_line":"+-----------------------+-------------------------+"},{"line_number":167,"context_line":"| extra                 | dict_or_none            |"},{"line_number":168,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_4d3b1f39","line":165,"in_reply_to":"3afb71cf_b52671b8","updated":"2015-06-08 11:31:41.000000000","message":"As earlier, we\u0027ll need to bring this up at a meeting if we want to diverge from the current port object.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"fa0c4536151fb78972a643a3f877abb6dcd96792","unresolved":false,"context_lines":[{"line_number":164,"context_line":"+-----------------------+-------------------------+"},{"line_number":165,"context_line":"| address               | str_or_none             |"},{"line_number":166,"context_line":"+-----------------------+-------------------------+"},{"line_number":167,"context_line":"| extra                 | dict_or_none            |"},{"line_number":168,"context_line":"+-----------------------+-------------------------+"},{"line_number":169,"context_line":"| created_at            | datetime_or_str_or_none |"},{"line_number":170,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_7520f99b","line":167,"updated":"2015-06-06 19:01:08.000000000","message":"can you please let us know what the \u0027extra\u0027 will hold?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":164,"context_line":"+-----------------------+-------------------------+"},{"line_number":165,"context_line":"| address               | str_or_none             |"},{"line_number":166,"context_line":"+-----------------------+-------------------------+"},{"line_number":167,"context_line":"| extra                 | dict_or_none            |"},{"line_number":168,"context_line":"+-----------------------+-------------------------+"},{"line_number":169,"context_line":"| created_at            | datetime_or_str_or_none |"},{"line_number":170,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_ad35f346","line":167,"in_reply_to":"3afb71cf_7520f99b","updated":"2015-06-08 11:31:41.000000000","message":"Following convention from the port object. Extra is used as it is there for extra information that operators, developers etc want to store. As an example, our PoC used the port\u0027s extra field.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"fa0c4536151fb78972a643a3f877abb6dcd96792","unresolved":false,"context_lines":[{"line_number":175,"context_line":"--------------------"},{"line_number":176,"context_line":""},{"line_number":177,"context_line":"The state machine will not be directly impacted, however, when changes can be"},{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_f534e9de","line":178,"updated":"2015-06-06 19:01:08.000000000","message":"Can we please rephrase the sentence \u0027when changes can be made will be impacted by the state machine\u0027 .\n\nIs it proposed that \u0027changes will be made without impacting the existing state machine\u0027","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"4edd8565426af27643d15eb22f9a6551c2495371","unresolved":false,"context_lines":[{"line_number":175,"context_line":"--------------------"},{"line_number":176,"context_line":""},{"line_number":177,"context_line":"The state machine will not be directly impacted, however, when changes can be"},{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_b8f3ef09","line":178,"in_reply_to":"3afb71cf_f534e9de","updated":"2015-06-08 11:31:41.000000000","message":"The sentence is to refer to the impact the state machine has on the proposed changes. Changes to the new port group object and additions of port groups will only be able to made when a node is in a particular set of states. I\u0027ll see if I can rephrase.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"f9e96662540012ff05c9bc69d3071424e6ac0f42","unresolved":false,"context_lines":[{"line_number":176,"context_line":""},{"line_number":177,"context_line":"The state machine will not be directly impacted, however, when changes can be"},{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_2c313ea3","line":179,"updated":"2015-06-05 12:52:57.000000000","message":"MANAGEABLE, INSPECTING and ENROLL (to be introduced soon)","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"e3991e900ca0e392db893c79f36ab5ac86f3fb7a","unresolved":false,"context_lines":[{"line_number":176,"context_line":""},{"line_number":177,"context_line":"The state machine will not be directly impacted, however, when changes can be"},{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_e75e9f75","line":179,"in_reply_to":"3afb71cf_2c313ea3","updated":"2015-06-05 13:27:26.000000000","message":"Thanks","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10358,"name":"Brad Morgan","email":"brad@morgabra.com","username":"morgabra"},"change_message_id":"7a1211bac2291ef1f51cbd1e7ac3ced44eefcd6c","unresolved":false,"context_lines":[{"line_number":177,"context_line":"The state machine will not be directly impacted, however, when changes can be"},{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":"REST API impact"},{"line_number":183,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_aae76bab","line":180,"updated":"2015-06-04 20:21:47.000000000","message":"What happens when you update, for example, port information on a pxe_enabled port? We\u0027d need a hook there to move/update the neutron port right?","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"2a514b81d51b8c459cc1583a54701ea62d5b2afb","unresolved":false,"context_lines":[{"line_number":177,"context_line":"The state machine will not be directly impacted, however, when changes can be"},{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":"REST API impact"},{"line_number":183,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_b3f24205","line":180,"in_reply_to":"3afb71cf_aae76bab","updated":"2015-06-04 21:26:31.000000000","message":"Yes, I think that is correct","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"fa0c4536151fb78972a643a3f877abb6dcd96792","unresolved":false,"context_lines":[{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":"REST API impact"},{"line_number":183,"context_line":"---------------"},{"line_number":184,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_95525512","line":181,"updated":"2015-06-06 19:01:08.000000000","message":"It will be good to capture movement of baremetal servers, which would be realized as delete_port and then a create_port for Neutron from Ironic.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"e5730504c2681cca84f07656d1600c4da8f70b3b","unresolved":false,"context_lines":[{"line_number":178,"context_line":"made will be impacted by the state machine. Any add/update/delete to port_group"},{"line_number":179,"context_line":"can only be made when the node is in a MANAGEABLE state or in maintenance mode"},{"line_number":180,"context_line":"when in the AVAILABLE/ACTIVE/RESCUE states."},{"line_number":181,"context_line":""},{"line_number":182,"context_line":"REST API impact"},{"line_number":183,"context_line":"---------------"},{"line_number":184,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_69a93a7b","line":181,"in_reply_to":"3afb71cf_95525512","updated":"2015-06-08 06:04:07.000000000","message":"Good idea - Also, we should explicitly state that move of BM from one switch/rack to another will be delete followed by re-discover","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"e5730504c2681cca84f07656d1600c4da8f70b3b","unresolved":false,"context_lines":[{"line_number":182,"context_line":"REST API impact"},{"line_number":183,"context_line":"---------------"},{"line_number":184,"context_line":""},{"line_number":185,"context_line":"TODO:"},{"line_number":186,"context_line":""},{"line_number":187,"context_line":"RPC API impact"},{"line_number":188,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_098406d2","line":185,"updated":"2015-06-08 06:04:07.000000000","message":"Two modifications are needed: \nFollowing is a proposed suggestion:\n\n1) ironic port-create -m \u003caddress\u003e -c chassis-id\u003d\u003cchassis_id\u003e, port-id\u003d\u003cport-id\u003e, system-name\u003d\u003csystem-name\u003e ....\nNote: equivalent change is needed to ironic port-update\n\n2) ironic port-group-create  -n \u003cnode-id\u003e -p \u003cport-id\u003e -m \u003caddress\u003e port-group-name \nNote: equivalent change is needed to ironic port-group-update","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"29d2c63de9dde4649d4049d10207d9e0ad8c495e","unresolved":false,"context_lines":[{"line_number":182,"context_line":"REST API impact"},{"line_number":183,"context_line":"---------------"},{"line_number":184,"context_line":""},{"line_number":185,"context_line":"TODO:"},{"line_number":186,"context_line":""},{"line_number":187,"context_line":"RPC API impact"},{"line_number":188,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3afb71cf_df5fa3e3","line":185,"in_reply_to":"3afb71cf_098406d2","updated":"2015-06-08 16:43:06.000000000","message":"API changes should be straightforward -- add the new keys to the JSON body for port-create/update. Add new endpoints for port-group.\n\nLet\u0027s be sure to specify this in terms of API endpoints and JSON, not client commands.","commit_id":"5fede799f39ad05d8918e906b12263d9b33cf3d4"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":8,"context_line":"Ironic Neutron ML2 Integration"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"https://blueprints.launchpad.net/ironic/+spec/ironic-ml2-integration"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_c1184a87","line":8,"updated":"2015-06-15 15:54:28.000000000","message":"Precisely speaking, it is not an integration with Neutron ML2 but also a generic Neutron integration. Even though the first backend support may be implemented as ML2 driver, it is a generic Neutron integration.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":8,"context_line":"Ironic Neutron ML2 Integration"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"https://blueprints.launchpad.net/ironic/+spec/ironic-ml2-integration"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_16b35aea","line":8,"in_reply_to":"fa32b979_93b7c7d7","updated":"2015-06-18 03:20:31.000000000","message":"I think calling it Ironic Neutron is reasonable. ML2 is a core plugin in Neutron. We do not need to explicitly call it out.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"69e10de8575e9c6ff964c9e116b0640551e5736b","unresolved":false,"context_lines":[{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":8,"context_line":"Ironic Neutron ML2 Integration"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"https://blueprints.launchpad.net/ironic/+spec/ironic-ml2-integration"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_93b7c7d7","line":8,"in_reply_to":"fa32b979_c1184a87","updated":"2015-06-16 16:14:46.000000000","message":"I\u0027m not sure this spec provides anything more than layer 2 integration with Neutron. I think we could change it from ML2 to Layer 2 but my understanding was that ML2 is the recommended way forward for Layer 2 connectivity in Neutron hence the current name.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":5,"context_line":" http://creativecommons.org/licenses/by/3.0/legalcode"},{"line_number":6,"context_line":""},{"line_number":7,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":8,"context_line":"Ironic Neutron ML2 Integration"},{"line_number":9,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":10,"context_line":""},{"line_number":11,"context_line":"https://blueprints.launchpad.net/ironic/+spec/ironic-ml2-integration"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_22045f48","line":8,"in_reply_to":"fa32b979_c1184a87","updated":"2015-06-16 16:10:21.000000000","message":"OK, happy to go with Ironic-Neutron integration","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks - this may require the capability to specify which NICs belong"},{"line_number":43,"context_line":"  to which network and is out of scope for this initial Ironic - Neutron ML2"},{"line_number":44,"context_line":"  integration."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_6075bcc9","line":44,"updated":"2015-06-15 14:59:58.000000000","message":"Is it reasonable to put this ability in scope, but not the method of deciding which interface gets which network?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10375,"name":"Om Kumar","email":"om.kumar@hpe.com","username":"lazy_prince"},"change_message_id":"fc70971104395ad849c45b38be7eb188b5090c65","unresolved":false,"context_lines":[{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks - this may require the capability to specify which NICs belong"},{"line_number":43,"context_line":"  to which network and is out of scope for this initial Ironic - Neutron ML2"},{"line_number":44,"context_line":"  integration."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_a94dec73","line":44,"in_reply_to":"fa32b979_138e8978","updated":"2015-06-15 16:05:11.000000000","message":"Probably, this needs to handled at Nova level as nova allocates network to interfaces.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks - this may require the capability to specify which NICs belong"},{"line_number":43,"context_line":"  to which network and is out of scope for this initial Ironic - Neutron ML2"},{"line_number":44,"context_line":"  integration."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_138e8978","line":44,"in_reply_to":"fa32b979_6075bcc9","updated":"2015-06-15 15:54:28.000000000","message":"Could you clarify what you suggest more?\n\nI think we can support some limited cases. For example, we support a case where an Ironic instance has multiple neutron ports (one ironic port corresponds to one neutron port) but users cannot specify which ironic port corresponds to which neutron port.)","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks - this may require the capability to specify which NICs belong"},{"line_number":43,"context_line":"  to which network and is out of scope for this initial Ironic - Neutron ML2"},{"line_number":44,"context_line":"  integration."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_f6da764a","line":44,"in_reply_to":"fa32b979_a94dec73","updated":"2015-06-18 03:20:31.000000000","message":"I have added it under the \"Long Term Items\" on our wiki to track it. See here - https://wiki.openstack.org/wiki/Meetings/Ironic-neutron.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks - this may require the capability to specify which NICs belong"},{"line_number":43,"context_line":"  to which network and is out of scope for this initial Ironic - Neutron ML2"},{"line_number":44,"context_line":"  integration."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_62b7c786","line":44,"in_reply_to":"fa32b979_a94dec73","updated":"2015-06-16 16:10:21.000000000","message":"OK, I\u0027ll reword this section to say this is in scope but the method of deciding which interface gets which network will be decided elsewhere","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":44,"context_line":"  integration."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":48,"context_line":""},{"line_number":49,"context_line":"When nodes are enrolled/introspected in Ironic, local link (e.g. LLDP)"},{"line_number":50,"context_line":"information should be recorded against each port that is enrolled. The option"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_93a4d957","line":47,"updated":"2015-06-15 15:54:28.000000000","message":"(nit) In this spec proposal, we have many similar terms: (ironic) port, port group, neutron port. It is worth adding some terminology definitions.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":64,"context_line":"LAG for which all member switch ports will belong to the same network and to the same"},{"line_number":65,"context_line":"network segment."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"The network should first be placed on a provisioning network and then onto the"},{"line_number":68,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":69,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":70,"context_line":"facilitate this there will be a variable stored in the binding profile to"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_9b7395b1","line":67,"updated":"2015-06-15 14:59:58.000000000","message":"\"The network should\"... this should be node, right? Or maybe interface?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10375,"name":"Om Kumar","email":"om.kumar@hpe.com","username":"lazy_prince"},"change_message_id":"fc70971104395ad849c45b38be7eb188b5090c65","unresolved":false,"context_lines":[{"line_number":64,"context_line":"LAG for which all member switch ports will belong to the same network and to the same"},{"line_number":65,"context_line":"network segment."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"The network should first be placed on a provisioning network and then onto the"},{"line_number":68,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":69,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":70,"context_line":"facilitate this there will be a variable stored in the binding profile to"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_c9871060","line":67,"in_reply_to":"fa32b979_9b7395b1","updated":"2015-06-15 16:05:11.000000000","message":"+1\n\nShould be Node or Interface","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":64,"context_line":"LAG for which all member switch ports will belong to the same network and to the same"},{"line_number":65,"context_line":"network segment."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"The network should first be placed on a provisioning network and then onto the"},{"line_number":68,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":69,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":70,"context_line":"facilitate this there will be a variable stored in the binding profile to"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_2260df00","line":67,"in_reply_to":"fa32b979_c9871060","updated":"2015-06-16 16:10:21.000000000","message":"Will reword to \u0027Node\u0027","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":94,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":95,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":96,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":97,"context_line":"expect."},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"To facilitate link aggregation a new port_group object will be created. In"},{"line_number":100,"context_line":"addition to the base object it will have the following fields:"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_f690b397","line":97,"updated":"2015-06-15 15:54:28.000000000","message":"In my understanding, the most important point is that Ironic and Neutron (back-end) must share the same interpretation of this information. If they interpret it differently, we cannot connect Ironic instance to appropriate neutron network. It is worth mentioned here.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"afc958f4c448591912edf4041a4754afcb8ab41f","unresolved":false,"context_lines":[{"line_number":94,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":95,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":96,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":97,"context_line":"expect."},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"To facilitate link aggregation a new port_group object will be created. In"},{"line_number":100,"context_line":"addition to the base object it will have the following fields:"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_8e6335dd","line":97,"in_reply_to":"fa32b979_f373b313","updated":"2015-06-22 15:58:46.000000000","message":"Thanks for clarifying the intention. I agree that we don\u0027t need to define strict parameters here. What I would like to mention is just it is important to declare the bottom-line.\nI believe we are in the same page.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"69e10de8575e9c6ff964c9e116b0640551e5736b","unresolved":false,"context_lines":[{"line_number":94,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":95,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":96,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":97,"context_line":"expect."},{"line_number":98,"context_line":""},{"line_number":99,"context_line":"To facilitate link aggregation a new port_group object will be created. In"},{"line_number":100,"context_line":"addition to the base object it will have the following fields:"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_f373b313","line":97,"in_reply_to":"fa32b979_f690b397","updated":"2015-06-16 16:14:46.000000000","message":"Correct. My wish with the paragraph was for a debate on whether to standardize on LLDP or not. Standardizing on LLDP makes writing the inspector RAMdisk portion of this a lot easier. However, there will be nothing to stop people implementing their own key/value pairs to send between Ironic and Neutron.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_b3ba3de0","line":137,"updated":"2015-06-15 15:54:28.000000000","message":"In Neutron API point of view, binding profile in Neutron is a field which vendor driver can use freely and there is no schema definition for binding:profile. Ironic-Neutron integration defines a generic way to specify physical port information and we define some schema, so my suggestion is to define a new field to pass physical port information for a neutron port. By doing so, we can easily define a schema for these information. Though my comment may be beyond the scope of this spec, it needs to be discussed in Neutron side.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"dfadca0d68b13432dbe5c1fbcf7d3aaa00bd6519","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_64ce42b6","line":137,"in_reply_to":"fa32b979_1e5b0956","updated":"2015-06-19 21:04:28.000000000","message":"Correct - that is implied here - bind_requested will be used for binding purposes, and will remain as part of the dict (as per described in the spec). \nAdditionally, we will set the host-id to \"-Ironic-\" to reflect this port belongs to Ironic. \n\nI think this satisfies the requirement.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"6233c2ba1352474ea290398fff95e60da1331cfa","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_8dfb7447","line":137,"in_reply_to":"fa32b979_257b7bbf","updated":"2015-06-19 17:38:55.000000000","message":"Vivekanandan, \n\nI do not believe filtering should be the reason to add a new table. We have in access of 15 ML2 drivers and many projects interact/use neutron, if we added new tables to facilitate filtering for all these, neutron DB will be cluttered.  Having said that, we need to address the issue that some ML2 drivers would want to filter for Ironic ports, therefore, we can use host-id (as suggested by Bob) to specify anything we like, and possibly, can pull the bind_required flag out of dict and put in the ml2_port_binding. I am not aware of virtualized environments setting host-id to NULL - not to say that it is not possible. We can set host-id to something more specific e.g. \"Ironinc\", \"Ironic-Neutron\", \"Ironic-bm\", etc. etc. \n\nthoughts?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"afc958f4c448591912edf4041a4754afcb8ab41f","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_d1ea4074","line":137,"in_reply_to":"fa32b979_64ce42b6","updated":"2015-06-22 15:58:46.000000000","message":"In my understanding of the last meeting, we will define generic framework in this spec and the detail of Neutron API define needs to be clarified in the counterpart Neutron spec.\n(we can defer it to further release, but it would be great if it lands in Liberty).\n\nRegarding host_id vs binding_requested, we need to consider carefully. What does actually binding:host_id mean? The consistency between virtual env and bare-metal env is important. For virtual env, host_id means a compute node. I am not sure host_id for baremetal should point a compute-node for baremetal or each baremetal server.\n\nTo decide the way, we need to check carefully how binding:host_id is handled in ML2 plugin.\nI think we need to discuss/clarify this in Neutron side spec.\nIn my understanding, this spec discusses what procedure and what information are required for Ironic-Neutron integration.\n\nbinding_requested needs to be considered with network flipping and take into account nova procedure too.\nI am in the way of understanding the full picture.\n\nPlease correct me if I am wrong or out of the context.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"59741525e9a34b1b73578604744a46dece6f5ccd","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_1e5b0956","line":137,"in_reply_to":"fa32b979_8dfb7447","updated":"2015-06-19 20:20:26.000000000","message":"Sukhdev,\n\nSetting binding:host-id with an identifiable string for Ironic ports would be good path (atleast it will allow basic filtering for neutron users).\n\nIf we consider that, it looks like we may need to ensure in neutron ML2 plugin, that binding does not happen just because the availability of a string in binding:host-id, because today it works that way.  It should work only based on bind_requested\u003dTrue (atleast for Ironic ports).\n\nAlso, we believe it would be good if Ironic would always pass binding:host-id (as of today\u0027s Ironic behaviour) in port-create and port-update requests to Neutron, regardless of bind_requested value.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_b3cdd232","line":137,"in_reply_to":"fa32b979_9da4dfa6","updated":"2015-06-18 03:20:31.000000000","message":"Correct. I will capture it under Long Term Items at - https://wiki.openstack.org/wiki/Meetings/Ironic-neutron.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"1186e72977b445e06113d1bd881778ead1b0fc95","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_9da4dfa6","line":137,"in_reply_to":"fa32b979_b3ba3de0","updated":"2015-06-15 18:45:22.000000000","message":"As we discussed in IRC meeting today 6/15, we will initially roll out with binding-profile (though it has no schema validation) and expand in next release as by that time trunk port abstraction would have also would have landed into the Port model of  Neutron :)","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"d5d44b361249931549da3d2c0b953c6f5332aae6","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_be4fcb32","line":137,"in_reply_to":"fa32b979_b3cdd232","updated":"2015-06-18 11:48:09.000000000","message":"Sukhdev / Jim,\n\nWe have been developing mech-drivers based on the binding_profile assumption in this Spec and we have been having increasing difficulties to just do even basic operations like filter only ironic-ports from neutron port-table.\n\nWith the current neutron code , filtering by binding-profile is not available.  Further even if we fix to make it available (via a core ML2 fix), user still needs to provide the entire profile dict content to pick the right Ironic port , which becomes infeasible.\n\nI request that it will be good if we go with a new table called physical_port_table in the Neutron-side.  This physical_port_table need not be exposed directly by Neutron APIs to be managed.   It will serve the purpose of relating neutron_port_id to physical_port_information that is passed by Ironic which includes \"local_link_connection\" available at L338 below .\n\nAnd the new table will serve the purpose of querying for Ironic-only ports using neutron extension filters \nfor eg., CLI \"neutron port-list --physical-ports\".\n\nIf the suggestion is to have this physical_port_table table within each vendor mech-driver, it becomes substantial duplication.\n\nAlso, tables maintained in mech-drivers are not exposable to openstack neutron tenants.  They have to be accessed via backdoor entries that would become vendor proprietary.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"91872704098cd56023e0a33727fe377460847438","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_257b7bbf","line":137,"in_reply_to":"fa32b979_bac817a9","updated":"2015-06-19 07:12:00.000000000","message":"Sukhdev:  Thanks for your view points.  Couple of answers.\n\n 1) profile is available in m2_port_binding, and hence can be filtered. Having said that, you make a good point that filtering has to be done on the dict, which leads to the second point. \n\n[Vivek]  binding_profile is a dict, and if we have to filter on that, we have to give the entire dict value which would include binding_type , local_link_connection and bind_requested.  Now that we might not be considering as a filter, because \nit is abstract, difficult and not useful either to the openstack users (or) even for a programmatic extraction.\n\nWe have been advocating a framework for providing extension filters for all neutron resource in general, because we have so many extensions to neutron resources today, none of which can be filtered (this is on other side :) ).\n\n2) There are two proposals on the table and both are listed in the spec. One is to use Bind_request flag - which will be set only by Ironic, and second, proposed by Bob Kukura in one of the previous meetings, to use host-id and set it to empty (NULL? ) in case of Ironic. If need be, we could use both of these together.\n\n[Vivek] We are not saying that bind_requested and type should be removed from the binding_profile..\nIn fact both would continue to reside in binding_profile only because they are related to binding of the port.\nHowever, the Ironic information we feed should become an Ironic-specific port extension in neutron and so it need not be carried inside binding_profile.\n\nWe could not filter the ironic ports in neutron based on just bind_requested, because we have to give the full dict as answered in 1) above to filter anything based on binding_profile.\n\nComing to binding:host-id, there are implementations which donot populate host-id  at all during create-port for even virtual ports (and so NULL) and so we cannot rely on that to assume that is an Ironic port.\n\nDoing quicker is agreeable for us, but it should work for multiple vendors because cloud underlay fabric is generally multi-vendor for customers. And our mech-drivers for Ironic should work alongside with other reference mech-drivers from ovs, linuxbridge, l2pop etc.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"f3e58ff3a689a56e4664fe59b25675546b9929c2","unresolved":false,"context_lines":[{"line_number":134,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":135,"context_line":""},{"line_number":136,"context_line":""},{"line_number":137,"context_line":"The proposed binding profile for Neutron is as follows:"},{"line_number":138,"context_line":""},{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_bac817a9","line":137,"in_reply_to":"fa32b979_be4fcb32","updated":"2015-06-18 21:34:11.000000000","message":"Hi Vivekanandan, \n\nTwo points: \n1) profile is available in m2_port_binding, and hence can be filtered. Having said that, you make a good point that filtering has to be done on the dict, which leads to the second point. \n\n2) There are two proposals on the table and both are listed in the spec. One is to use Bind_request flag - which will be set only by Ironic, and second, proposed by Bob Kukura in one of the previous meetings, to use host-id and set it to empty (NULL? ) in case of Ironic. If need be, we could use both of these together.\n\nIf you need to filter, you could use it to filter all Ironic ports.\n\nThis gives us the ability to have initial version of implementation in a much simpler and quicker manner. Hope this helps.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"},{"line_number":141,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":142,"context_line":"| type                   | Type of the profile (\u0027Ironic\u0027 in this case)        |"},{"line_number":143,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":144,"context_line":"| local_link_information | A list of local link information either from all   |"},{"line_number":145,"context_line":"|                        | Ironic ports in a particular port group or from a  |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_5bc2ed86","line":142,"updated":"2015-06-15 14:59:58.000000000","message":"I think this should be \"bearmetal\" or something, don\u0027t limit it to Ironic IMO.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"},{"line_number":141,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":142,"context_line":"| type                   | Type of the profile (\u0027Ironic\u0027 in this case)        |"},{"line_number":143,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":144,"context_line":"| local_link_information | A list of local link information either from all   |"},{"line_number":145,"context_line":"|                        | Ironic ports in a particular port group or from a  |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_5393b15b","line":142,"in_reply_to":"fa32b979_5bc2ed86","updated":"2015-06-15 15:54:28.000000000","message":"Agree. As I commented above, I think it is better to define a new field to pass the physical port information described here","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":139,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":140,"context_line":"| Field Name             | Description                                        |"},{"line_number":141,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":142,"context_line":"| type                   | Type of the profile (\u0027Ironic\u0027 in this case)        |"},{"line_number":143,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":144,"context_line":"| local_link_information | A list of local link information either from all   |"},{"line_number":145,"context_line":"|                        | Ironic ports in a particular port group or from a  |"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_a2194f8a","line":142,"in_reply_to":"fa32b979_5bc2ed86","updated":"2015-06-16 16:10:21.000000000","message":"OK, happy to go with \u0027baremetal\u0027","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"1186e72977b445e06113d1bd881778ead1b0fc95","unresolved":false,"context_lines":[{"line_number":150,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":151,"context_line":""},{"line_number":152,"context_line":"As an alternative to ``bind_requested``, it should also be possible to use ``host-id`` because"},{"line_number":153,"context_line":"if it is left empty then the resulting ``binding:vif_type`` should be ``unbound``."},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"Alternatives"},{"line_number":156,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_7dd79bed","line":153,"updated":"2015-06-15 18:45:22.000000000","message":"Agree.  Please revisit if bind_requested is necessary.\r\n\r\nML2 mech drivers for Ironic can tag on binding-type to claim a port as jroll put it in earlier comment.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":150,"context_line":"+------------------------+----------------------------------------------------+"},{"line_number":151,"context_line":""},{"line_number":152,"context_line":"As an alternative to ``bind_requested``, it should also be possible to use ``host-id`` because"},{"line_number":153,"context_line":"if it is left empty then the resulting ``binding:vif_type`` should be ``unbound``."},{"line_number":154,"context_line":""},{"line_number":155,"context_line":"Alternatives"},{"line_number":156,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_043cc8c6","line":153,"in_reply_to":"fa32b979_7dd79bed","updated":"2015-06-16 16:10:21.000000000","message":"I think we can keep bind_requested flag as an attribute.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":164,"context_line":"of two layer 2 plugins (ML2 and Ironic-Neutron, ignoring vendor plugins) and"},{"line_number":165,"context_line":"multiple plugins being required for a baremetal and virtual machine cloud."},{"line_number":166,"context_line":"Requiring networking vendors to write and maintain separate drivers for both"},{"line_number":167,"context_line":"plugins is not ideal."},{"line_number":168,"context_line":""},{"line_number":169,"context_line":""},{"line_number":170,"context_line":"Data model impact"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_3b736982","line":167,"updated":"2015-06-15 14:59:58.000000000","message":"This plugin doesn\u0027t work with upstream Ironic anyway, so I don\u0027t think this is an alternative at all :)","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":176,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":177,"context_line":"| Field Name            | Field Type   | Migration Value |"},{"line_number":178,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":179,"context_line":"| local_link_connection | dict_or_none | None            |"},{"line_number":180,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":181,"context_line":"| port_group_id         | int_or_none  | None            |"},{"line_number":182,"context_line":"+-----------------------+--------------+-----------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_13f0c98f","line":179,"updated":"2015-06-15 15:54:28.000000000","message":"How can we represent \"dict\" in DB model?\nDo we need to break it down it several field, or do we store it as a string with JSON representation?\nMy current vote is the first option.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"64ccd5b93c6b325dab56a922295bef7c82be8840","unresolved":false,"context_lines":[{"line_number":176,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":177,"context_line":"| Field Name            | Field Type   | Migration Value |"},{"line_number":178,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":179,"context_line":"| local_link_connection | dict_or_none | None            |"},{"line_number":180,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":181,"context_line":"| port_group_id         | int_or_none  | None            |"},{"line_number":182,"context_line":"+-----------------------+--------------+-----------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_ccc15ee6","line":179,"in_reply_to":"fa32b979_13f0c98f","updated":"2015-06-16 01:42:16.000000000","message":"We have JSON fields in ironic already.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"afc958f4c448591912edf4041a4754afcb8ab41f","unresolved":false,"context_lines":[{"line_number":176,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":177,"context_line":"| Field Name            | Field Type   | Migration Value |"},{"line_number":178,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":179,"context_line":"| local_link_connection | dict_or_none | None            |"},{"line_number":180,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":181,"context_line":"| port_group_id         | int_or_none  | None            |"},{"line_number":182,"context_line":"+-----------------------+--------------+-----------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_eea0c1a7","line":179,"in_reply_to":"fa32b979_ccc15ee6","updated":"2015-06-22 15:58:46.000000000","message":"I just would like to clarify if what schema is applied to the field. JSON representation is okay.\n\nIf a content of JSON fields is validated in Ironic API layer, I don\u0027t see no problem. If no check, it is troublesome.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":196,"context_line":"+-----------------------+-------------------------+"},{"line_number":197,"context_line":"| uuid                  | str_or_none             |"},{"line_number":198,"context_line":"+-----------------------+-------------------------+"},{"line_number":199,"context_line":"| name                  | str_or_none             |"},{"line_number":200,"context_line":"+-----------------------+-------------------------+"},{"line_number":201,"context_line":"| node_id               | str_or_none             |"},{"line_number":202,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_c62620ba","line":199,"updated":"2015-06-15 14:59:58.000000000","message":"Does it make sense for a port group to have a name? I\u0027m not sure.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":196,"context_line":"+-----------------------+-------------------------+"},{"line_number":197,"context_line":"| uuid                  | str_or_none             |"},{"line_number":198,"context_line":"+-----------------------+-------------------------+"},{"line_number":199,"context_line":"| name                  | str_or_none             |"},{"line_number":200,"context_line":"+-----------------------+-------------------------+"},{"line_number":201,"context_line":"| node_id               | str_or_none             |"},{"line_number":202,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_f3dfdad3","line":199,"in_reply_to":"fa32b979_336e2d30","updated":"2015-06-18 03:20:31.000000000","message":"Akihiro +1","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":196,"context_line":"+-----------------------+-------------------------+"},{"line_number":197,"context_line":"| uuid                  | str_or_none             |"},{"line_number":198,"context_line":"+-----------------------+-------------------------+"},{"line_number":199,"context_line":"| name                  | str_or_none             |"},{"line_number":200,"context_line":"+-----------------------+-------------------------+"},{"line_number":201,"context_line":"| node_id               | str_or_none             |"},{"line_number":202,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_336e2d30","line":199,"in_reply_to":"fa32b979_c62620ba","updated":"2015-06-15 15:54:28.000000000","message":"It is not required but I think it is worth having some field for description.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":198,"context_line":"+-----------------------+-------------------------+"},{"line_number":199,"context_line":"| name                  | str_or_none             |"},{"line_number":200,"context_line":"+-----------------------+-------------------------+"},{"line_number":201,"context_line":"| node_id               | str_or_none             |"},{"line_number":202,"context_line":"+-----------------------+-------------------------+"},{"line_number":203,"context_line":"| address               | str_or_none             |"},{"line_number":204,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_d6d56f43","line":201,"updated":"2015-06-15 15:54:28.000000000","message":"Is it a foreign key to an entry in Node table?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":198,"context_line":"+-----------------------+-------------------------+"},{"line_number":199,"context_line":"| name                  | str_or_none             |"},{"line_number":200,"context_line":"+-----------------------+-------------------------+"},{"line_number":201,"context_line":"| node_id               | str_or_none             |"},{"line_number":202,"context_line":"+-----------------------+-------------------------+"},{"line_number":203,"context_line":"| address               | str_or_none             |"},{"line_number":204,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_246f8c93","line":201,"in_reply_to":"fa32b979_d6d56f43","updated":"2015-06-16 16:10:21.000000000","message":"Yes","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":216,"context_line":"group object and additions of port groups will only be allowed when a node is in a"},{"line_number":217,"context_line":"particular set of states. Any add/update/delete to port_group can only be made when"},{"line_number":218,"context_line":"the node is in a MANAGEABLE/INSPECTING/ENROLL state or if the node is in maintenance"},{"line_number":219,"context_line":"mode whilst in an AVAILABLE/ACTIVE/RESCUE state."},{"line_number":220,"context_line":""},{"line_number":221,"context_line":""},{"line_number":222,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_9bb955ec","line":219,"updated":"2015-06-15 14:59:58.000000000","message":"I don\u0027t believe this is true for the port object today... Is this a requirement? Would something break if I added a port group to an ACTIVE node? (I\u0027m not saying this is wrong, just curious)","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":216,"context_line":"group object and additions of port groups will only be allowed when a node is in a"},{"line_number":217,"context_line":"particular set of states. Any add/update/delete to port_group can only be made when"},{"line_number":218,"context_line":"the node is in a MANAGEABLE/INSPECTING/ENROLL state or if the node is in maintenance"},{"line_number":219,"context_line":"mode whilst in an AVAILABLE/ACTIVE/RESCUE state."},{"line_number":220,"context_line":""},{"line_number":221,"context_line":""},{"line_number":222,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_33af822b","line":219,"in_reply_to":"fa32b979_969c95c1","updated":"2015-06-18 03:20:31.000000000","message":"Yes, this makes sense","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":216,"context_line":"group object and additions of port groups will only be allowed when a node is in a"},{"line_number":217,"context_line":"particular set of states. Any add/update/delete to port_group can only be made when"},{"line_number":218,"context_line":"the node is in a MANAGEABLE/INSPECTING/ENROLL state or if the node is in maintenance"},{"line_number":219,"context_line":"mode whilst in an AVAILABLE/ACTIVE/RESCUE state."},{"line_number":220,"context_line":""},{"line_number":221,"context_line":""},{"line_number":222,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_842c5803","line":219,"in_reply_to":"fa32b979_9bb955ec","updated":"2015-06-16 16:10:21.000000000","message":"It might help to check what scenarios and requirements could be present. I read this as a question about how to move existing deployments to using this new model. If there was a different intention please could you specify?\n\nPossible scenarios:\n\n1. Node is in ACTIVE and landscape (servers/switches) already set up in a way that ought to be compatible with the port_group concept. There is a need to update the OpenStack port representations but not carry out any actions.\n- if we add port_group to Ironic node, and add ports into the port_group, then unless Neutron ports are also updated the Ironic-Neutron port mapping will not be consistent with using the port_group. Could impact the methods for updates and tear down process.\n\n2. Node is in ACTIVE. There is a need to update the OpenStack port representations and also the way the physical landscape is configured.\n- I don\u0027t see a reason for adding port_groups at this stage? \n\nDo you have other thoughts/suggestions?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"69e10de8575e9c6ff964c9e116b0640551e5736b","unresolved":false,"context_lines":[{"line_number":216,"context_line":"group object and additions of port groups will only be allowed when a node is in a"},{"line_number":217,"context_line":"particular set of states. Any add/update/delete to port_group can only be made when"},{"line_number":218,"context_line":"the node is in a MANAGEABLE/INSPECTING/ENROLL state or if the node is in maintenance"},{"line_number":219,"context_line":"mode whilst in an AVAILABLE/ACTIVE/RESCUE state."},{"line_number":220,"context_line":""},{"line_number":221,"context_line":""},{"line_number":222,"context_line":"REST API impact"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_969c95c1","line":219,"in_reply_to":"fa32b979_9bb955ec","updated":"2015-06-16 16:14:46.000000000","message":"When Laura, Sukhdev and I initially discussed this our thoughts were that on a change (e.g. port addition/deletion to the port-group) we would update the local_link_information in the binding_profile and trigger an update in Neutron. We thought it best to only allow this when the node was not in a state where uninterrupted connectivity was expected.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":291,"context_line":"      * JSON schema definition of Port"},{"line_number":292,"context_line":""},{"line_number":293,"context_line":""},{"line_number":294,"context_line":"* ``/v1/ports``"},{"line_number":295,"context_line":""},{"line_number":296,"context_line":"  * Update an existing port."},{"line_number":297,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_9b9e3555","line":294,"updated":"2015-06-15 14:59:58.000000000","message":"This should be ``/v1/ports/(port_uuid)``","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":291,"context_line":"      * JSON schema definition of Port"},{"line_number":292,"context_line":""},{"line_number":293,"context_line":""},{"line_number":294,"context_line":"* ``/v1/ports``"},{"line_number":295,"context_line":""},{"line_number":296,"context_line":"  * Update an existing port."},{"line_number":297,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_a47ddce8","line":294,"in_reply_to":"fa32b979_9b9e3555","updated":"2015-06-16 16:10:21.000000000","message":"The API was written in a way similar to http://docs.openstack.org/developer/ironic/webapi/v1.html#ports but also happy to modify as you\u0027ve suggested","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":361,"context_line":"  * Expected error http response code(s):"},{"line_number":362,"context_line":""},{"line_number":363,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":364,"context_line":"      * 404 for resource (e.g. node) not found"},{"line_number":365,"context_line":""},{"line_number":366,"context_line":"  * Parameters:"},{"line_number":367,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_9bb5d5ce","line":364,"updated":"2015-06-15 14:59:58.000000000","message":"When would this endpoint not be found? If there\u0027s no port groups in the DB it should just return an empty list, right?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":361,"context_line":"  * Expected error http response code(s):"},{"line_number":362,"context_line":""},{"line_number":363,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":364,"context_line":"      * 404 for resource (e.g. node) not found"},{"line_number":365,"context_line":""},{"line_number":366,"context_line":"  * Parameters:"},{"line_number":367,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_0456085c","line":364,"in_reply_to":"fa32b979_9bb5d5ce","updated":"2015-06-16 16:10:21.000000000","message":"Examples would be:\n\n404 if node not found. Empty list if address not found. \n\n400 if address is not mac-address format.\n\nShould be similar to current behaviour of ports and nodes.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":363,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":364,"context_line":"      * 404 for resource (e.g. node) not found"},{"line_number":365,"context_line":""},{"line_number":366,"context_line":"  * Parameters:"},{"line_number":367,"context_line":""},{"line_number":368,"context_line":"       * ``node (uuid_or_name)`` - UUID or name of a node, to only get"},{"line_number":369,"context_line":"         port_groups for that node."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_b6d30b38","line":366,"updated":"2015-06-15 15:54:28.000000000","message":"\"name\" and \"extras\" need to be documented here.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"afc958f4c448591912edf4041a4754afcb8ab41f","unresolved":false,"context_lines":[{"line_number":363,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":364,"context_line":"      * 404 for resource (e.g. node) not found"},{"line_number":365,"context_line":""},{"line_number":366,"context_line":"  * Parameters:"},{"line_number":367,"context_line":""},{"line_number":368,"context_line":"       * ``node (uuid_or_name)`` - UUID or name of a node, to only get"},{"line_number":369,"context_line":"         port_groups for that node."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_ee11a1b0","line":366,"in_reply_to":"fa32b979_28ba2124","updated":"2015-06-22 15:58:46.000000000","message":"Ah... i got it. It describes filter parameters for listing.\nI now understood we can\u0027t index on extras.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"2f5301431b43800162bb583165d149ee595ea091","unresolved":false,"context_lines":[{"line_number":363,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":364,"context_line":"      * 404 for resource (e.g. node) not found"},{"line_number":365,"context_line":""},{"line_number":366,"context_line":"  * Parameters:"},{"line_number":367,"context_line":""},{"line_number":368,"context_line":"       * ``node (uuid_or_name)`` - UUID or name of a node, to only get"},{"line_number":369,"context_line":"         port_groups for that node."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_28ba2124","line":366,"in_reply_to":"fa32b979_b6d30b38","updated":"2015-06-16 01:45:05.000000000","message":"We can\u0027t index on extra, so no way to filter by that.\n\nName is a valid port_group_ident for the endpoint below this one.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":426,"context_line":"  * Expected error http response code(s):"},{"line_number":427,"context_line":""},{"line_number":428,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":429,"context_line":"      * 409 for resource conflict"},{"line_number":430,"context_line":""},{"line_number":431,"context_line":"  * Parameters:"},{"line_number":432,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_5bcf4d56","line":429,"updated":"2015-06-15 14:59:58.000000000","message":"When would this happen, when a unique constraint is violated? Something else?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":426,"context_line":"  * Expected error http response code(s):"},{"line_number":427,"context_line":""},{"line_number":428,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":429,"context_line":"      * 409 for resource conflict"},{"line_number":430,"context_line":""},{"line_number":431,"context_line":"  * Parameters:"},{"line_number":432,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_245b0c56","line":429,"in_reply_to":"fa32b979_5bcf4d56","updated":"2015-06-16 16:10:21.000000000","message":"port_group name should be unique","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":441,"context_line":"      * JSON schema definition of Port_group"},{"line_number":442,"context_line":""},{"line_number":443,"context_line":""},{"line_number":444,"context_line":"* ``/v1/port_groups``"},{"line_number":445,"context_line":""},{"line_number":446,"context_line":"  * Delete a port_group."},{"line_number":447,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_7bfed16c","line":444,"updated":"2015-06-15 14:59:58.000000000","message":"Should be ``/v1/port_groups/{port_group_ident}``","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":467,"context_line":"      * N/A"},{"line_number":468,"context_line":""},{"line_number":469,"context_line":""},{"line_number":470,"context_line":"* ``/v1/port_groups``"},{"line_number":471,"context_line":""},{"line_number":472,"context_line":"  * Update an existing port_group."},{"line_number":473,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_fbf2e199","line":470,"updated":"2015-06-15 14:59:58.000000000","message":"Should be ``/v1/port_groups/{port_group_ident}``","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":478,"context_line":"  * Expected error http response code(s):"},{"line_number":479,"context_line":""},{"line_number":480,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":481,"context_line":"      * 404 for resource (e.g. port_group) not found"},{"line_number":482,"context_line":""},{"line_number":483,"context_line":"  * Parameters:"},{"line_number":484,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_db303d59","line":481,"updated":"2015-06-15 14:59:58.000000000","message":"Maybe add 409 assuming that means a unique constraint is violated?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":478,"context_line":"  * Expected error http response code(s):"},{"line_number":479,"context_line":""},{"line_number":480,"context_line":"      * 400 for bad query or malformed syntax"},{"line_number":481,"context_line":"      * 404 for resource (e.g. port_group) not found"},{"line_number":482,"context_line":""},{"line_number":483,"context_line":"  * Parameters:"},{"line_number":484,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_24222cba","line":481,"in_reply_to":"fa32b979_db303d59","updated":"2015-06-16 16:10:21.000000000","message":"ok","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":529,"context_line":""},{"line_number":530,"context_line":"  * Response:"},{"line_number":531,"context_line":""},{"line_number":532,"context_line":"      * JSON schema definition of Port_groupCollection"},{"line_number":533,"context_line":""},{"line_number":534,"context_line":""},{"line_number":535,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_fb20a1f9","line":532,"updated":"2015-06-15 14:59:58.000000000","message":"What\u0027s different about the response in this endpoint vs the regular ``/v1/port_groups``","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":529,"context_line":""},{"line_number":530,"context_line":"  * Response:"},{"line_number":531,"context_line":""},{"line_number":532,"context_line":"      * JSON schema definition of Port_groupCollection"},{"line_number":533,"context_line":""},{"line_number":534,"context_line":""},{"line_number":535,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_675b562f","line":532,"in_reply_to":"fa32b979_fb20a1f9","updated":"2015-06-16 16:10:21.000000000","message":"Was going with similar spec as for port at http://docs.openstack.org/developer/ironic/webapi/v1.html#ports The additional \u0027detail\u0027 option would return all fields. Without it only a subset of fields would be returned, say uuid and address.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"93a0100b71d47e964d56aad2137c3c790ddd8117","unresolved":false,"context_lines":[{"line_number":650,"context_line":"        \"address\": \"fe:54:00:77:07:d9\","},{"line_number":651,"context_line":"        \"node_uuid\": \"e7a6f1e2-7176-4fe8-b8e9-ed71c77d74dd\","},{"line_number":652,"context_line":"        \"name\": \"node1_portgroup1\""},{"line_number":653,"context_line":"        }"},{"line_number":654,"context_line":""},{"line_number":655,"context_line":"     * Response 201 with body:"},{"line_number":656,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_be069d1e","line":653,"updated":"2015-06-15 06:54:06.000000000","message":"Shouldn\u0027t this have port-id as well?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":650,"context_line":"        \"address\": \"fe:54:00:77:07:d9\","},{"line_number":651,"context_line":"        \"node_uuid\": \"e7a6f1e2-7176-4fe8-b8e9-ed71c77d74dd\","},{"line_number":652,"context_line":"        \"name\": \"node1_portgroup1\""},{"line_number":653,"context_line":"        }"},{"line_number":654,"context_line":""},{"line_number":655,"context_line":"     * Response 201 with body:"},{"line_number":656,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_db659d45","line":653,"in_reply_to":"fa32b979_be069d1e","updated":"2015-06-15 14:59:58.000000000","message":"No, ports have a port_group_uuid, not the other way around","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":650,"context_line":"        \"address\": \"fe:54:00:77:07:d9\","},{"line_number":651,"context_line":"        \"node_uuid\": \"e7a6f1e2-7176-4fe8-b8e9-ed71c77d74dd\","},{"line_number":652,"context_line":"        \"name\": \"node1_portgroup1\""},{"line_number":653,"context_line":"        }"},{"line_number":654,"context_line":""},{"line_number":655,"context_line":"     * Response 201 with body:"},{"line_number":656,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_335d4247","line":653,"in_reply_to":"fa32b979_db659d45","updated":"2015-06-18 03:20:31.000000000","message":"Got it.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"93a0100b71d47e964d56aad2137c3c790ddd8117","unresolved":false,"context_lines":[{"line_number":677,"context_line":"        \"uuid\": \"6eb02b44-18a3-4659-8c0b-8d2802581ae4\","},{"line_number":678,"context_line":"        }"},{"line_number":679,"context_line":""},{"line_number":680,"context_line":"*  Example of port update."},{"line_number":681,"context_line":""},{"line_number":682,"context_line":"     * Parameter \"port_uuid\"\u003d\"1004e542-2f9f-4d9b-b8b9-5b719fa6613f\""},{"line_number":683,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_610e3436","line":680,"updated":"2015-06-15 06:54:06.000000000","message":"I think you mean port-group update","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":677,"context_line":"        \"uuid\": \"6eb02b44-18a3-4659-8c0b-8d2802581ae4\","},{"line_number":678,"context_line":"        }"},{"line_number":679,"context_line":""},{"line_number":680,"context_line":"*  Example of port update."},{"line_number":681,"context_line":""},{"line_number":682,"context_line":"     * Parameter \"port_uuid\"\u003d\"1004e542-2f9f-4d9b-b8b9-5b719fa6613f\""},{"line_number":683,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_3bdcc9e3","line":680,"in_reply_to":"fa32b979_610e3436","updated":"2015-06-15 14:59:58.000000000","message":"This is adding a port to a port group by adding port_group_uuid.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"93a0100b71d47e964d56aad2137c3c790ddd8117","unresolved":false,"context_lines":[{"line_number":687,"context_line":""},{"line_number":688,"context_line":"        {"},{"line_number":689,"context_line":"        \"port_group_uuid\": 6eb02b44-18a3-4659-8c0b-8d2802581ae4,"},{"line_number":690,"context_line":"        }"},{"line_number":691,"context_line":""},{"line_number":692,"context_line":"     * Response 200 with body:"},{"line_number":693,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_21110c53","line":690,"updated":"2015-06-15 06:54:06.000000000","message":"Should this have port-id as well?","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":687,"context_line":""},{"line_number":688,"context_line":"        {"},{"line_number":689,"context_line":"        \"port_group_uuid\": 6eb02b44-18a3-4659-8c0b-8d2802581ae4,"},{"line_number":690,"context_line":"        }"},{"line_number":691,"context_line":""},{"line_number":692,"context_line":"     * Response 200 with body:"},{"line_number":693,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_7b6d7103","line":690,"in_reply_to":"fa32b979_21110c53","updated":"2015-06-15 14:59:58.000000000","message":"No.\n\nAs a note, Ironic uses JSON PATCH syntax, so this would look something like:\n\n    [{\"path\": \"/port_group_uuid\", \"value\": \"6eb02b44-18a3-4659-8c0b-8d2802581ae4\", \"op\": \"add\"}]\n\n\"op\" can be add/replace/delete.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":687,"context_line":""},{"line_number":688,"context_line":"        {"},{"line_number":689,"context_line":"        \"port_group_uuid\": 6eb02b44-18a3-4659-8c0b-8d2802581ae4,"},{"line_number":690,"context_line":"        }"},{"line_number":691,"context_line":""},{"line_number":692,"context_line":"     * Response 200 with body:"},{"line_number":693,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_f3291a9c","line":690,"in_reply_to":"fa32b979_7b6d7103","updated":"2015-06-18 03:20:31.000000000","message":"Got it. Thanks for clarification","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":753,"context_line":""},{"line_number":754,"context_line":""},{"line_number":755,"context_line":"Discuss any policy changes, and discuss what things a deployer needs to"},{"line_number":756,"context_line":"think about when defining their policy."},{"line_number":757,"context_line":""},{"line_number":758,"context_line":"* A deployer should be aware of the capabilities of the particular ML2 driver"},{"line_number":759,"context_line":"  for supporting use of the new local_link_information that will be passed to it"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_5b23cdbd","line":756,"updated":"2015-06-15 14:59:58.000000000","message":"Ironic is an admin-only API, our policy is very simple as it is only one rule :)","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"93a0100b71d47e964d56aad2137c3c790ddd8117","unresolved":false,"context_lines":[{"line_number":776,"context_line":""},{"line_number":777,"context_line":"Example usage of the new methods:"},{"line_number":778,"context_line":""},{"line_number":779,"context_line":"  * ironic port_group-create -node_uuid nodeUUID -name portgroupname"},{"line_number":780,"context_line":"  * ironic port_group-update port_group_uuid -address AA:BB:CC:DD:EE:FF"},{"line_number":781,"context_line":"  * ironic port_group-delete port_group_uuid"},{"line_number":782,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_0196484c","line":779,"updated":"2015-06-15 06:54:06.000000000","message":"I think we need to add port-group members here - i.e. port-id\u0027s of the ports that are members of the port group","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":776,"context_line":""},{"line_number":777,"context_line":"Example usage of the new methods:"},{"line_number":778,"context_line":""},{"line_number":779,"context_line":"  * ironic port_group-create -node_uuid nodeUUID -name portgroupname"},{"line_number":780,"context_line":"  * ironic port_group-update port_group_uuid -address AA:BB:CC:DD:EE:FF"},{"line_number":781,"context_line":"  * ironic port_group-delete port_group_uuid"},{"line_number":782,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_8657b801","line":779,"in_reply_to":"fa32b979_0196484c","updated":"2015-06-15 14:59:58.000000000","message":"I\u0027m kind of indifferent to that. I guess it can\u0027t hurt.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":776,"context_line":""},{"line_number":777,"context_line":"Example usage of the new methods:"},{"line_number":778,"context_line":""},{"line_number":779,"context_line":"  * ironic port_group-create -node_uuid nodeUUID -name portgroupname"},{"line_number":780,"context_line":"  * ironic port_group-update port_group_uuid -address AA:BB:CC:DD:EE:FF"},{"line_number":781,"context_line":"  * ironic port_group-delete port_group_uuid"},{"line_number":782,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_67459615","line":779,"in_reply_to":"fa32b979_8657b801","updated":"2015-06-16 16:10:21.000000000","message":"ok, port_ids can also optionally be passed - this would allow ports to be updated either through the port_group update/create or through the port update. I think this should be OK.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"93a0100b71d47e964d56aad2137c3c790ddd8117","unresolved":false,"context_lines":[{"line_number":777,"context_line":"Example usage of the new methods:"},{"line_number":778,"context_line":""},{"line_number":779,"context_line":"  * ironic port_group-create -node_uuid nodeUUID -name portgroupname"},{"line_number":780,"context_line":"  * ironic port_group-update port_group_uuid -address AA:BB:CC:DD:EE:FF"},{"line_number":781,"context_line":"  * ironic port_group-delete port_group_uuid"},{"line_number":782,"context_line":""},{"line_number":783,"context_line":"The python-ironicclient would also need the Port detailed resource extended"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_c19f005e","line":780,"updated":"2015-06-15 06:54:06.000000000","message":"Ditto here","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"1186e72977b445e06113d1bd881778ead1b0fc95","unresolved":false,"context_lines":[{"line_number":778,"context_line":""},{"line_number":779,"context_line":"  * ironic port_group-create -node_uuid nodeUUID -name portgroupname"},{"line_number":780,"context_line":"  * ironic port_group-update port_group_uuid -address AA:BB:CC:DD:EE:FF"},{"line_number":781,"context_line":"  * ironic port_group-delete port_group_uuid"},{"line_number":782,"context_line":""},{"line_number":783,"context_line":"The python-ironicclient would also need the Port detailed resource extended"},{"line_number":784,"context_line":"to include the new port attributes."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_5d309701","line":781,"updated":"2015-06-15 18:45:22.000000000","message":"In general, is the workflow that Ironic ports that need to be part of port-groups shall be created only after a port-group is first created and is available?\r\n\r\nOr is it possible to allow an Ironic port to be associated to an already available port-group?\r\n\r\nThere aren\u0027t CLIs listed here if it is possible pursue the latter.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"2f5301431b43800162bb583165d149ee595ea091","unresolved":false,"context_lines":[{"line_number":778,"context_line":""},{"line_number":779,"context_line":"  * ironic port_group-create -node_uuid nodeUUID -name portgroupname"},{"line_number":780,"context_line":"  * ironic port_group-update port_group_uuid -address AA:BB:CC:DD:EE:FF"},{"line_number":781,"context_line":"  * ironic port_group-delete port_group_uuid"},{"line_number":782,"context_line":""},{"line_number":783,"context_line":"The python-ironicclient would also need the Port detailed resource extended"},{"line_number":784,"context_line":"to include the new port attributes."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_c87ff564","line":781,"in_reply_to":"fa32b979_5d309701","updated":"2015-06-16 01:45:05.000000000","message":"ironic port-update $uuid replace port_group_id\u003d$id","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":778,"context_line":""},{"line_number":779,"context_line":"  * ironic port_group-create -node_uuid nodeUUID -name portgroupname"},{"line_number":780,"context_line":"  * ironic port_group-update port_group_uuid -address AA:BB:CC:DD:EE:FF"},{"line_number":781,"context_line":"  * ironic port_group-delete port_group_uuid"},{"line_number":782,"context_line":""},{"line_number":783,"context_line":"The python-ironicclient would also need the Port detailed resource extended"},{"line_number":784,"context_line":"to include the new port attributes."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_87a21aad","line":781,"in_reply_to":"fa32b979_c87ff564","updated":"2015-06-16 16:10:21.000000000","message":"I will add more CLIs and options in the next patch.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":844,"context_line":"  Nova.  The complementary `network-provider spec"},{"line_number":845,"context_line":"  \u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_ should provide"},{"line_number":846,"context_line":"  a more concrete proposal for the mechanism"},{"line_number":847,"context_line":"  to allow the flip between the provisioning and tenant networks."},{"line_number":848,"context_line":""},{"line_number":849,"context_line":""},{"line_number":850,"context_line":"* To enable the mapping between Neutron ports and Ironic ports and port_groups."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_865818fc","line":847,"updated":"2015-06-15 14:59:58.000000000","message":"So, I thought the way we discussed this is that Nova does the initial creation of the port, with the binding information it knows. Ironic then fetches this profile and updates Neutron with the additional information it knows about like switch port and chassis id etc. This is the way it\u0027s laid out in the network provider spec.\n\nIf we\u0027re going this route, you can likely remove all of the above Nova impact.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"1ce1f9e20a10ad1aae2617bc3b12cb168d7cb1f4","unresolved":false,"context_lines":[{"line_number":844,"context_line":"  Nova.  The complementary `network-provider spec"},{"line_number":845,"context_line":"  \u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_ should provide"},{"line_number":846,"context_line":"  a more concrete proposal for the mechanism"},{"line_number":847,"context_line":"  to allow the flip between the provisioning and tenant networks."},{"line_number":848,"context_line":""},{"line_number":849,"context_line":""},{"line_number":850,"context_line":"* To enable the mapping between Neutron ports and Ironic ports and port_groups."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_f6aaf318","line":847,"in_reply_to":"fa32b979_865818fc","updated":"2015-06-15 15:54:28.000000000","message":"I wonder what is the known binding information when nova does the initial port creation. What kind of information needs to be passed to \"nova boot\"? There is no additional information if an Ironic instance has a single Neutron port, but if an ironic instance has multiple neutron ports I think we need to pass additional information to \"nova boot\" to specify which Ironic port (physical interface) corresponds to which neutron network (port). I am in the middle way of learning Ironic-Nova interaction, so I might be missing very basic parts.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":844,"context_line":"  Nova.  The complementary `network-provider spec"},{"line_number":845,"context_line":"  \u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_ should provide"},{"line_number":846,"context_line":"  a more concrete proposal for the mechanism"},{"line_number":847,"context_line":"  to allow the flip between the provisioning and tenant networks."},{"line_number":848,"context_line":""},{"line_number":849,"context_line":""},{"line_number":850,"context_line":"* To enable the mapping between Neutron ports and Ironic ports and port_groups."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_e7ee86b3","line":847,"in_reply_to":"fa32b979_865818fc","updated":"2015-06-16 16:10:21.000000000","message":"OK, I will remove this from this spec and if I have comments/questions will leave on your spec.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9cb29757a39ce7c330a36f2fdcba1653d34933fe","unresolved":false,"context_lines":[{"line_number":844,"context_line":"  Nova.  The complementary `network-provider spec"},{"line_number":845,"context_line":"  \u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_ should provide"},{"line_number":846,"context_line":"  a more concrete proposal for the mechanism"},{"line_number":847,"context_line":"  to allow the flip between the provisioning and tenant networks."},{"line_number":848,"context_line":""},{"line_number":849,"context_line":""},{"line_number":850,"context_line":"* To enable the mapping between Neutron ports and Ironic ports and port_groups."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_de33bdf4","line":847,"in_reply_to":"fa32b979_865818fc","updated":"2015-06-18 03:20:31.000000000","message":"Yes, I agree. I think we should remove this from here as it is addressed in the network provider spec. This way we will not risk inconsistency between these two specs.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":871,"context_line":"The new REST API calls for port_groups should not be usable by the end user."},{"line_number":872,"context_line":"Only operators and administrators should be able to manage port_groups and"},{"line_number":873,"context_line":"local_link_connection data of ports, because these settings are used to"},{"line_number":874,"context_line":"configure the network."},{"line_number":875,"context_line":""},{"line_number":876,"context_line":""},{"line_number":877,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_060408be","line":874,"updated":"2015-06-15 14:59:58.000000000","message":"As said before, Ironic is an admin-only API.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"93a0100b71d47e964d56aad2137c3c790ddd8117","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"Using the binding profile to enable flipping between provisioning and tenant"},{"line_number":890,"context_line":"networks means there will be no support for PXE booting after deploy (i.e. local"},{"line_number":891,"context_line":"disk installation only). How to allow operators to deploy instances using either"},{"line_number":892,"context_line":"net-boot or local boot using the same Ironic conductor should be discussed in the"},{"line_number":893,"context_line":"complementary `network-provider spec"},{"line_number":894,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_44577e02","line":891,"updated":"2015-06-15 06:54:06.000000000","message":"I am not sure if this belongs under scalability impact section","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"Using the binding profile to enable flipping between provisioning and tenant"},{"line_number":890,"context_line":"networks means there will be no support for PXE booting after deploy (i.e. local"},{"line_number":891,"context_line":"disk installation only). How to allow operators to deploy instances using either"},{"line_number":892,"context_line":"net-boot or local boot using the same Ironic conductor should be discussed in the"},{"line_number":893,"context_line":"complementary `network-provider spec"},{"line_number":894,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_66ccf489","line":891,"in_reply_to":"fa32b979_44577e02","updated":"2015-06-15 14:59:58.000000000","message":"Agree, I think this is end-user or operator impact.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"Using the binding profile to enable flipping between provisioning and tenant"},{"line_number":890,"context_line":"networks means there will be no support for PXE booting after deploy (i.e. local"},{"line_number":891,"context_line":"disk installation only). How to allow operators to deploy instances using either"},{"line_number":892,"context_line":"net-boot or local boot using the same Ironic conductor should be discussed in the"},{"line_number":893,"context_line":"complementary `network-provider spec"},{"line_number":894,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_87c8da23","line":891,"in_reply_to":"fa32b979_66ccf489","updated":"2015-06-16 16:10:21.000000000","message":"ok","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"8e7ae55b44ad915f6885c264327b4670d359f34f","unresolved":false,"context_lines":[{"line_number":907,"context_line":"is introduced, so this will require a database migration."},{"line_number":908,"context_line":""},{"line_number":909,"context_line":"Deployers will need to deploy an ML2 mechanism driver that supports connecting"},{"line_number":910,"context_line":"baremetal resources to Neutron networks."},{"line_number":911,"context_line":""},{"line_number":912,"context_line":"Developer impact"},{"line_number":913,"context_line":"----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_660ab428","line":910,"updated":"2015-06-15 14:59:58.000000000","message":"Also the nova config option.","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"fb5bdad29ef5899842183eae6e3675242eafbf96","unresolved":false,"context_lines":[{"line_number":907,"context_line":"is introduced, so this will require a database migration."},{"line_number":908,"context_line":""},{"line_number":909,"context_line":"Deployers will need to deploy an ML2 mechanism driver that supports connecting"},{"line_number":910,"context_line":"baremetal resources to Neutron networks."},{"line_number":911,"context_line":""},{"line_number":912,"context_line":"Developer impact"},{"line_number":913,"context_line":"----------------"}],"source_content_type":"text/x-rst","patch_set":2,"id":"fa32b979_07ddaadf","line":910,"in_reply_to":"fa32b979_660ab428","updated":"2015-06-16 16:10:21.000000000","message":"ok, thanks","commit_id":"ce21142cf93c84c2e8b70db2d5fac16fa170f726"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":38,"context_line":"  separate spec and implementation in Neutron, see `vlan-aware-vms"},{"line_number":39,"context_line":"  \u003chttps://blueprints.launchpad.net/neutron/+spec/vlan-aware-vms\u003e`_."},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks.  This may require the capability to specify which NICs"},{"line_number":43,"context_line":"  belong to which network. It is in scope but the method of deciding which"},{"line_number":44,"context_line":"  interface gets which network will be decided elsewhere."}],"source_content_type":"text/x-rst","patch_set":3,"id":"fa32b979_318e5a7a","line":41,"updated":"2015-06-30 00:33:58.000000000","message":"I am not sure I appreciate the reason why this cannot be achieved in the context of this effort; can you elaborate?","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":38,"context_line":"  separate spec and implementation in Neutron, see `vlan-aware-vms"},{"line_number":39,"context_line":"  \u003chttps://blueprints.launchpad.net/neutron/+spec/vlan-aware-vms\u003e`_."},{"line_number":40,"context_line":""},{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks.  This may require the capability to specify which NICs"},{"line_number":43,"context_line":"  belong to which network. It is in scope but the method of deciding which"},{"line_number":44,"context_line":"  interface gets which network will be decided elsewhere."}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_65333263","line":41,"in_reply_to":"fa32b979_318e5a7a","updated":"2015-07-01 16:30:53.000000000","message":"This requires the capability to specify which NICs belong to which network. This is being addressed elsewhere.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":67,"context_line":"The node should first be placed on a provisioning network and then onto the"},{"line_number":68,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":69,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":70,"context_line":"facilitate this there will be a variable stored in the binding profile to"},{"line_number":71,"context_line":"record whether a port binding has been requested. This will allow Ironic to"},{"line_number":72,"context_line":"defer binding until Ironic is able to populate the extra information that is"},{"line_number":73,"context_line":"required by Neutron."}],"source_content_type":"text/x-rst","patch_set":3,"id":"fa32b979_a94f5ebd","line":70,"updated":"2015-06-30 00:33:58.000000000","message":"Does this mean that the existing binding logic needs to be modified to introspect the value of this variable? I think this might be a bit problematic, as we\u0027d be relying on an unstructure","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":67,"context_line":"The node should first be placed on a provisioning network and then onto the"},{"line_number":68,"context_line":"tenant network as specified in the `network-provider spec"},{"line_number":69,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_. To help"},{"line_number":70,"context_line":"facilitate this there will be a variable stored in the binding profile to"},{"line_number":71,"context_line":"record whether a port binding has been requested. This will allow Ironic to"},{"line_number":72,"context_line":"defer binding until Ironic is able to populate the extra information that is"},{"line_number":73,"context_line":"required by Neutron."}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_0bd7d7ed","line":70,"in_reply_to":"fa32b979_a94f5ebd","updated":"2015-07-01 16:30:53.000000000","message":"The new variable bind_requested would be set as True/False. Supportive ML2 drivers would use this to know when to do the binding. If a driver doesn\u0027t support this then the behaviour would be the same as of now (effectively defaults to true).","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":94,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":95,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":96,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":97,"context_line":"expect. The key point is that Ironic and Neutron should share the same"},{"line_number":98,"context_line":"understanding of this information in order to connect the Ironic instance to"},{"line_number":99,"context_line":"the appropriate Neutron network, and standardizing on LLDP makes this easier."},{"line_number":100,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_f17e84e3","line":97,"updated":"2015-06-30 00:33:58.000000000","message":"If the key point is to share the same understanding of this information, then using binding:profile is clearly problematic, as no validation can be done of what gets passed around.\n\nI wonder if a more viable/less problematic/more portable approach is to provide a mechanism driver extension (available to all drivers) to be used to attach well-known bare-metal port attributes, like the ones mentioned above.\n\nPerhaps this isn\u0027t something we do right now, but it\u0027s worth exploring or at least making the point here.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":94,"context_line":"These fields are based on their LLDP TLV counterparts. It is not strictly"},{"line_number":95,"context_line":"required to populate these with LLDP values but it is recommended to use LLDP"},{"line_number":96,"context_line":"to make it easier to avoid mismatches in the values a Neutron ML2 driver will"},{"line_number":97,"context_line":"expect. The key point is that Ironic and Neutron should share the same"},{"line_number":98,"context_line":"understanding of this information in order to connect the Ironic instance to"},{"line_number":99,"context_line":"the appropriate Neutron network, and standardizing on LLDP makes this easier."},{"line_number":100,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_6b6b03f2","line":97,"in_reply_to":"ba3cc151_f17e84e3","updated":"2015-07-01 16:30:53.000000000","message":"Let\u0027s go with binding_profile for now and standardize on LLDP values. The point you bring up could be added to the long term considerations on https://wiki.openstack.org/wiki/Meetings/Ironic-neutron","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":154,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":155,"context_line":""},{"line_number":156,"context_line":"In addition to using ``bind_requested``, the ``host-id`` attribute should be"},{"line_number":157,"context_line":"set to \u0027-Ironic-\u0027. This will be used to reflect that the port belongs to"},{"line_number":158,"context_line":"Ironic. This would allow at least basic filtering for Ironic ports by ML2"},{"line_number":159,"context_line":"drivers."},{"line_number":160,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_31a86c1f","line":157,"updated":"2015-06-30 00:33:58.000000000","message":"this is a bit of an abuse; \u0027ownership\u0027 of port is usually done with the device_owner attribute. Compute ports are tagged with a compute:\u003czone\u003e value. Host-id should be the identifier of the bare-metal node, to avoid potential confusion.\n\nIf, from the api standpoint nova and ironic ports cannot be made distinguishable, then we may need to mark ports after the fact. For instance we could consider the extension mechanism suggested above to keep track of the fact that certain points are bare metals, and hence provide optimized filtering capabilities that leverage that.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"e4dedaa237c2fa5f20ba1418f87fbc9dbe1e81ec","unresolved":false,"context_lines":[{"line_number":156,"context_line":"In addition to using ``bind_requested``, the ``host-id`` attribute should be"},{"line_number":157,"context_line":"set to \u0027-Ironic-\u0027. This will be used to reflect that the port belongs to"},{"line_number":158,"context_line":"Ironic. This would allow at least basic filtering for Ironic ports by ML2"},{"line_number":159,"context_line":"drivers."},{"line_number":160,"context_line":""},{"line_number":161,"context_line":"Alternatives"},{"line_number":162,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_f7cc3437","line":159,"updated":"2015-06-30 01:18:52.000000000","message":"So two things here:\n\n1) I think that we should set host_id to the ironic node uuid. That\u0027s the real host here.\n\n2) For filtering and to allow ML2 things to recognize that it\u0027s a baremetal port, we should use something like binding_profile:type","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":156,"context_line":"In addition to using ``bind_requested``, the ``host-id`` attribute should be"},{"line_number":157,"context_line":"set to \u0027-Ironic-\u0027. This will be used to reflect that the port belongs to"},{"line_number":158,"context_line":"Ironic. This would allow at least basic filtering for Ironic ports by ML2"},{"line_number":159,"context_line":"drivers."},{"line_number":160,"context_line":""},{"line_number":161,"context_line":"Alternatives"},{"line_number":162,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_d632f2f1","line":159,"in_reply_to":"ba3cc151_c487f55f","updated":"2015-07-01 16:30:53.000000000","message":"Let\u0027s go with:\n\n1) binding_profile: type \u003d \u0027baremetal\u0027 can be used for filtering on ML2 side. This is why it\u0027s in the binding_profile.\n2) binding_profile: host_id \u003d ironic node uuid as Jim has suggested.\n\nI will update the spec with these.\n\nLet\u0027s go with this for first version. We are capturing longer term items at https://wiki.openstack.org/wiki/Meetings/Ironic-neutron","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"84df11d0697edc230d9d2166aedffab4c8b9b671","unresolved":false,"context_lines":[{"line_number":156,"context_line":"In addition to using ``bind_requested``, the ``host-id`` attribute should be"},{"line_number":157,"context_line":"set to \u0027-Ironic-\u0027. This will be used to reflect that the port belongs to"},{"line_number":158,"context_line":"Ironic. This would allow at least basic filtering for Ironic ports by ML2"},{"line_number":159,"context_line":"drivers."},{"line_number":160,"context_line":""},{"line_number":161,"context_line":"Alternatives"},{"line_number":162,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_c487f55f","line":159,"in_reply_to":"ba3cc151_e30a36d2","updated":"2015-06-30 07:33:28.000000000","message":"One more thing:\n\nIt\u0027s worth keeping in mind here, that this is a piece of contract between Ironic and Neutron. Say we settle with this for now, should we feel the need to move to something more robust (like it\u0027s being hinted here), then we\u0027d have to pay the price to preserve bw compatibility across unpaired releases between the two systems...unless we accept this first run as \u0027experimental\u0027.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"4f725e2f1087e028ddde8b3608f7c7b9b097a175","unresolved":false,"context_lines":[{"line_number":156,"context_line":"In addition to using ``bind_requested``, the ``host-id`` attribute should be"},{"line_number":157,"context_line":"set to \u0027-Ironic-\u0027. This will be used to reflect that the port belongs to"},{"line_number":158,"context_line":"Ironic. This would allow at least basic filtering for Ironic ports by ML2"},{"line_number":159,"context_line":"drivers."},{"line_number":160,"context_line":""},{"line_number":161,"context_line":"Alternatives"},{"line_number":162,"context_line":"------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_e30a36d2","line":159,"in_reply_to":"ba3cc151_f7cc3437","updated":"2015-06-30 03:46:01.000000000","message":"I thought about the vif type, but then I would imagine that amongst bare metal ports we\u0027d like to have different types being tracked...and if we pad this field with \u0027ironic\u0027, we end up going back to square one by conflating two things into one, yuk ;(","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":168,"context_line":""},{"line_number":169,"context_line":""},{"line_number":170,"context_line":""},{"line_number":171,"context_line":"Data model impact"},{"line_number":172,"context_line":"-----------------"},{"line_number":173,"context_line":""},{"line_number":174,"context_line":"The proposed change will be to add the following fields to the port object"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_d44d6e73","line":171,"updated":"2015-06-30 00:33:58.000000000","message":"If we are planning to expose these fields explicitly, does this mean that we can filter ports that are bare metal based on the fact that they do have an extension stored in here?","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":177,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":178,"context_line":"| Field Name            | Field Type   | Migration Value |"},{"line_number":179,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":180,"context_line":"| local_link_connection | dict_or_none | None            |"},{"line_number":181,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":182,"context_line":"| port_group_id         | int_or_none  | None            |"},{"line_number":183,"context_line":"+-----------------------+--------------+-----------------+"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_d426ceb9","line":180,"updated":"2015-06-30 00:33:58.000000000","message":"please, please...refrain to store data as json blob, it\u0027s really really bad.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":177,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":178,"context_line":"| Field Name            | Field Type   | Migration Value |"},{"line_number":179,"context_line":"+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d+"},{"line_number":180,"context_line":"| local_link_connection | dict_or_none | None            |"},{"line_number":181,"context_line":"+-----------------------+--------------+-----------------+"},{"line_number":182,"context_line":"| port_group_id         | int_or_none  | None            |"},{"line_number":183,"context_line":"+-----------------------+--------------+-----------------+"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_3668feb8","line":180,"in_reply_to":"ba3cc151_d426ceb9","updated":"2015-07-01 16:30:53.000000000","message":"JSONEncodedDict is already used and supported in Ironic.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":203,"context_line":"+-----------------------+-------------------------+"},{"line_number":204,"context_line":"| address               | str_or_none             |"},{"line_number":205,"context_line":"+-----------------------+-------------------------+"},{"line_number":206,"context_line":"| extra                 | dict_or_none            |"},{"line_number":207,"context_line":"+-----------------------+-------------------------+"},{"line_number":208,"context_line":"| created_at            | datetime_or_str_or_none |"},{"line_number":209,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_34493a60","line":206,"updated":"2015-06-30 00:33:58.000000000","message":"This has portability (or lack thereof) written all over it.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":203,"context_line":"+-----------------------+-------------------------+"},{"line_number":204,"context_line":"| address               | str_or_none             |"},{"line_number":205,"context_line":"+-----------------------+-------------------------+"},{"line_number":206,"context_line":"| extra                 | dict_or_none            |"},{"line_number":207,"context_line":"+-----------------------+-------------------------+"},{"line_number":208,"context_line":"| created_at            | datetime_or_str_or_none |"},{"line_number":209,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_b6896e77","line":206,"in_reply_to":"ba3cc151_34493a60","updated":"2015-07-01 16:30:53.000000000","message":"This is consistent with other data models in Ironic.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":799,"context_line":""},{"line_number":800,"context_line":"     * Note that port_group_uuid is now returned in the response."},{"line_number":801,"context_line":""},{"line_number":802,"context_line":""},{"line_number":803,"context_line":"Discuss any policy changes, and discuss what things a deployer needs to"},{"line_number":804,"context_line":"think about when defining their policy."},{"line_number":805,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_94b046df","line":802,"updated":"2015-06-30 00:33:58.000000000","message":"After reading the Data Model and API impact I am somewhat confused how the existing binding:profile extension is going to be used and what for. My initial understanding was that binding:profile was all that was required to pass data to/from Ironic/Neutron, albeit not ideal. This was a reasonable first step. \n\nBut if there are data and API extensions on top of binding:profile, what\u0027s binding:profile used for?","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":799,"context_line":""},{"line_number":800,"context_line":"     * Note that port_group_uuid is now returned in the response."},{"line_number":801,"context_line":""},{"line_number":802,"context_line":""},{"line_number":803,"context_line":"Discuss any policy changes, and discuss what things a deployer needs to"},{"line_number":804,"context_line":"think about when defining their policy."},{"line_number":805,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_962eaa41","line":802,"in_reply_to":"ba3cc151_94b046df","updated":"2015-07-01 16:30:53.000000000","message":"Ironic DB holds the local link information - this represents the physical connections between the node and the switch. Any LAG\u0027ed ports are represented by port_groups. This information should be held in Ironic DB. To support Ironic holding this new data, data model and API updates are needed. The data is then passed from Ironic to Neutron via the binding_profile. Hopefully this helps.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":931,"context_line":""},{"line_number":932,"context_line":"  This modified functionality could be implemented using a new config flag in"},{"line_number":933,"context_line":"  Nova to allow toggling between the old and the new methods. The flag could"},{"line_number":934,"context_line":"  help de-couple the upgrading of Nova and of Ironic."},{"line_number":935,"context_line":""},{"line_number":936,"context_line":""},{"line_number":937,"context_line":"Security impact"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_d42b2eea","line":934,"updated":"2015-06-30 00:33:58.000000000","message":"So if we\u0027re tracking all changes in this spec, does this mean we\u0027re getting enough visibility in the other projects?","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":931,"context_line":""},{"line_number":932,"context_line":"  This modified functionality could be implemented using a new config flag in"},{"line_number":933,"context_line":"  Nova to allow toggling between the old and the new methods. The flag could"},{"line_number":934,"context_line":"  help de-couple the upgrading of Nova and of Ironic."},{"line_number":935,"context_line":""},{"line_number":936,"context_line":""},{"line_number":937,"context_line":"Security impact"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_167ffa15","line":934,"in_reply_to":"ba3cc151_d42b2eea","updated":"2015-07-01 16:30:53.000000000","message":"This was discussed in the Neutron Design Summit as well as the Ironic design Summit. Neutron community seemed to agree to this. Nova community is also aware of the specs and proposed workflow.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":748,"name":"Armando Migliaccio","email":"armamig@gmail.com","username":"armando-migliaccio"},"change_message_id":"2ed25a5b73d157f8bd33eee4d243e53b21cb5aa0","unresolved":false,"context_lines":[{"line_number":1035,"context_line":""},{"line_number":1036,"context_line":"* Implement the extension to the RPC API."},{"line_number":1037,"context_line":""},{"line_number":1038,"context_line":"* Implement the changes to the Nova driver to get and use the binding profile."},{"line_number":1039,"context_line":""},{"line_number":1040,"context_line":"* Implement the changes needed to get vifs for updating Neutron port DHCP"},{"line_number":1041,"context_line":"  options."}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_d479eefd","line":1038,"updated":"2015-06-30 00:33:58.000000000","message":"So if I understand this correctly, binding:profile is used as a vehicle to get stuff to/from Neutron, but then this information will be stored in separate table(s)? Did I get it right? If so, then why do we allow this data duplication? If not, why do we need binding:profile at all?","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":1035,"context_line":""},{"line_number":1036,"context_line":"* Implement the extension to the RPC API."},{"line_number":1037,"context_line":""},{"line_number":1038,"context_line":"* Implement the changes to the Nova driver to get and use the binding profile."},{"line_number":1039,"context_line":""},{"line_number":1040,"context_line":"* Implement the changes needed to get vifs for updating Neutron port DHCP"},{"line_number":1041,"context_line":"  options."}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_56aee281","line":1038,"in_reply_to":"ba3cc151_d479eefd","updated":"2015-07-01 16:30:53.000000000","message":"Please see my comment above. Ironic should know about the server connections to the switch.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":12315,"name":"Bertie Fulton","email":"bertiefulton@yahoo.co.uk","username":"bertie.fultonsap.com"},"change_message_id":"c26314d0318c083683d154772243d75d163f198d","unresolved":false,"context_lines":[{"line_number":1079,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":1080,"context_line":""},{"line_number":1081,"context_line":"This feature will be fully documented."},{"line_number":1082,"context_line":""},{"line_number":1083,"context_line":""},{"line_number":1084,"context_line":"References"},{"line_number":1085,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"fa32b979_f5fb1e79","line":1082,"updated":"2015-06-25 09:38:25.000000000","message":"Do we need the documentation in the work items?","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"87b03b972f83af67ada7bbb186b99746eae3c853","unresolved":false,"context_lines":[{"line_number":1079,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":1080,"context_line":""},{"line_number":1081,"context_line":"This feature will be fully documented."},{"line_number":1082,"context_line":""},{"line_number":1083,"context_line":""},{"line_number":1084,"context_line":"References"},{"line_number":1085,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"ba3cc151_76ea06ae","line":1082,"in_reply_to":"fa32b979_f5fb1e79","updated":"2015-07-01 16:30:53.000000000","message":"Good point, thanks Bertie. I\u0027ll add this.","commit_id":"a401796f84acd9bf6baa35bda13d4ae978d17ecc"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"acc92ee8a8c665c0c84ecd4867ca5f20b3a7bd10","unresolved":false,"context_lines":[{"line_number":137,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":138,"context_line":""},{"line_number":139,"context_line":""},{"line_number":140,"context_line":"The binding profile for Neutron should include the following fields:"},{"line_number":141,"context_line":""},{"line_number":142,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":143,"context_line":"| Field Name             | Description                                      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_94c3c237","line":140,"updated":"2015-07-02 19:35:23.000000000","message":"Precisely speaking, vnic_type and host_id are part of port binding extension in Neutron, but are not part of binding profile.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"e9a681a2640f46fb51da75f311d8be5a18ce4cb5","unresolved":false,"context_lines":[{"line_number":137,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":138,"context_line":""},{"line_number":139,"context_line":""},{"line_number":140,"context_line":"The binding profile for Neutron should include the following fields:"},{"line_number":141,"context_line":""},{"line_number":142,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":143,"context_line":"| Field Name             | Description                                      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_1fad4331","line":140,"in_reply_to":"ba3cc151_38665d2a","updated":"2015-07-06 10:46:11.000000000","message":"Akihiro, the JSON you have put is right information, except that the local_link_information is a list rather than a dict.    The list can have one or more switch-port in them. So with that ratification:\n\n {\"port\":\n   { \u003call other fields\u003e,\n     \"vnic_type\": \"baremetal\",\n     \"host_id\": \u003cIronic node UUID\u003e,\n     \"binding:profile\": {\n        \"local_link_information\": [ \n              {\n                  \"switch_id\": xxx,\n                  \"port_id\": xxx,\n                  \u003coptional more information\u003e\n              },\n              {\n                  \"switch_id\": xxx,\n                  \"port_id\": yyy,\n                  \u003coptional more information\u003e\n              }]\n        \"bind_requested\": true/false,\n        \u003csome more profile fields\u003e\n     }\n   }\n }","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"d807c66563f836ddae70c9b77c4d921ce18d82fc","unresolved":false,"context_lines":[{"line_number":137,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":138,"context_line":""},{"line_number":139,"context_line":""},{"line_number":140,"context_line":"The binding profile for Neutron should include the following fields:"},{"line_number":141,"context_line":""},{"line_number":142,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":143,"context_line":"| Field Name             | Description                                      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_c1050b14","line":140,"in_reply_to":"ba3cc151_94c3c237","updated":"2015-07-02 20:21:14.000000000","message":"I have pasted the columns of ml2_port_bindings here - http://paste.openstack.org/show/336262/\n\nhost, vnic_type, and profile are all separate columns . Local link information will go into the profile.\n\nNova uses these fields frequently, so, the precedence is already in place. We have to make sure that Ironic driver uses the fields the same way. The information will be correctly stored in appropriate DB correctly. \n\n@Laura - I think instead of saying \"The binding profile for Neutron should include the folloing\", perhaps you can remove binding profile and say something like - the information sent to neutron should include.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"8fc862e2a14f7b259976f86f87f9e53aadf77d21","unresolved":false,"context_lines":[{"line_number":137,"context_line":"* Choose a pxe_enabled port at random."},{"line_number":138,"context_line":""},{"line_number":139,"context_line":""},{"line_number":140,"context_line":"The binding profile for Neutron should include the following fields:"},{"line_number":141,"context_line":""},{"line_number":142,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":143,"context_line":"| Field Name             | Description                                      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_38665d2a","line":140,"in_reply_to":"ba3cc151_c1050b14","updated":"2015-07-03 14:03:28.000000000","message":"I think we are discussing not ML2 DB models related port binding but Neutron API here. This is the only reason I commented above.\n\nAs the implementation level, all port binding related information are stored in ml2_port_bindings table, and this is called \"binding profile\" in the spec.\nNow I understand why I felt a bit confused.\n\nI can just say \"The following information needs to be passed to Neutron port. \"\n\nA JSON example describes the structure clearly.\nIn my understanding, we are proposing the following.\n\n {\"port\":\n   { \u003call other fields\u003e,\n     \"vnic_type\": \"baremetal\",\n     \"host_id\": \u003cIronic node UUID\u003e,\n     \"binding:profile\": {\n        \"local_link_information\": {\n            \"switch_id\": xxx,\n            \"port_id\": xxx,\n            \u003coptional more information\u003e\n        },\n        \"bind_requested\": true/false,\n        \u003csome other fields\u003e\n     },\n   }\n }\n\nbinding:profile is already a free area in Neutron and the content can be stored without any change in the implementation, but adding non-vendor-specific fields (like Neutron-Ironic integration) would bring more chaos situation.\n\nWe can improve later, so I just would like to see what fields are expected here.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"acc92ee8a8c665c0c84ecd4867ca5f20b3a7bd10","unresolved":false,"context_lines":[{"line_number":146,"context_line":"|                        | This would allow at least basic filtering for    |"},{"line_number":147,"context_line":"|                        | Ironic ports by ML2 drivers.                     |"},{"line_number":148,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_6f53cb17","line":149,"updated":"2015-07-02 19:35:23.000000000","message":"I think we don\u0027t need to have this attribute in binding:profile. We can simply define this in a port resource.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":7787,"name":"Kevin Benton","email":"kevin@benton.pub","username":"blak111"},"change_message_id":"29f4487240e0dfcf9f8ec88a32361fed28b23f4f","unresolved":false,"context_lines":[{"line_number":146,"context_line":"|                        | This would allow at least basic filtering for    |"},{"line_number":147,"context_line":"|                        | Ironic ports by ML2 drivers.                     |"},{"line_number":148,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_c44af954","line":149,"updated":"2015-07-02 20:10:41.000000000","message":"Isn\u0027t this needed to convey bonding information?","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"d807c66563f836ddae70c9b77c4d921ce18d82fc","unresolved":false,"context_lines":[{"line_number":146,"context_line":"|                        | This would allow at least basic filtering for    |"},{"line_number":147,"context_line":"|                        | Ironic ports by ML2 drivers.                     |"},{"line_number":148,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_2412c557","line":149,"in_reply_to":"ba3cc151_6f53cb17","updated":"2015-07-02 20:21:14.000000000","message":"see my above comment.\nIn general ml2 port binding is associated with a given port-id. As long as we use the port-binding fields, we do not need to put it under port resource.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"e8ea8d6027c366e395365504a76613b08b5f659b","unresolved":false,"context_lines":[{"line_number":146,"context_line":"|                        | This would allow at least basic filtering for    |"},{"line_number":147,"context_line":"|                        | Ironic ports by ML2 drivers.                     |"},{"line_number":148,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_84e071ce","line":149,"in_reply_to":"ba3cc151_c44af954","updated":"2015-07-02 20:23:49.000000000","message":"yes.. that is correct","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":7787,"name":"Kevin Benton","email":"kevin@benton.pub","username":"blak111"},"change_message_id":"29f4487240e0dfcf9f8ec88a32361fed28b23f4f","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_e445bd42","line":152,"updated":"2015-07-02 20:10:41.000000000","message":"+1. In a monolithic vendor deployment, I can see where the structure of the LLDP messages doesn\u0027t matter too much since the ML2 driver knows what to expect. However, if someone had heterogeneous switches with multiple ML2 drivers, each driver will need to be able to inspect the LLDP payload to determine if it is responsible for binding the port.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"acc92ee8a8c665c0c84ecd4867ca5f20b3a7bd10","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_0f579fca","line":152,"updated":"2015-07-02 19:35:23.000000000","message":"Please define more concrete schema for local_link_information (as Neutron API).\nlocal_link_information is a dict field.\nWhat fields are defined? Though I am okay with using blob in Ironic DB model, we strongly would like to have a schema definition as Neutron API.\n\nIn Ironic API example below, I see \"chassis_id\", \"port_id\", \"system_name\" in link_local_information. Do all of these definition depend on vendors? If so, it is not a defined Neutron API. AFAIK, \"system name\" is optional in LLDP (though most impls support) and \"chassis_id\" and \"port_id\" are mandatory.\n\nCan we define more concrete API as Neutron API? For example, \"switch_id\" and \"port_id\" looks enough to me.\nThe interpretation of values for these fields can defer to vendor drivers, but we need to define what fields are available. If not, it is not an API but a framework.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"d807c66563f836ddae70c9b77c4d921ce18d82fc","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_6459ad34","line":152,"in_reply_to":"ba3cc151_0f579fca","updated":"2015-07-02 20:21:14.000000000","message":"If I understand it correctly, you are suggesting we define - switch_id, and port_id as mandatory fields and others as optional fields?","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":7787,"name":"Kevin Benton","email":"kevin@benton.pub","username":"blak111"},"change_message_id":"04f4a3630112645a1922592bdf534fe36fb12dda","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_7a2f043f","line":152,"in_reply_to":"ba3cc151_5746997b","updated":"2015-07-02 21:18:21.000000000","message":"For a slightly more concrete example. Let\u0027s say I want to make a UI that shows all ports attached to the same physical switch. Unless we agree on which identifier corresponds to the switch, there won\u0027t be a way for a tool to know which field to group things by.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"795c8a0a49f01f90e9a26ee93d7680b8a7077b57","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_7ad62429","line":152,"in_reply_to":"ba3cc151_5746997b","updated":"2015-07-02 21:22:16.000000000","message":"I think the minimal set of fields to define a switch port are already defined - but, perhaps can be called out specifically with suggested format - e.g. \n\nSwitch_port {\n       switch_id: \u003cswitch-id\u003e format for switch id is mac-address format\n       port_id: \u003cport-id\u003e   format for port-id is string \n       switch_info: \u003cswitch_info\u003e  format for switch info is string\n} \n\nOnce we get past this spec process, I am thinking about writing a document for Ironic-neutron integration which can be referenced from both Ironic as well Neutron documents - to describe step-by-step process for the vendors to follow to make the overall experience as easy as possible","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"8fc862e2a14f7b259976f86f87f9e53aadf77d21","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_b60a671e","line":152,"in_reply_to":"ba3cc151_7ad62429","updated":"2015-07-03 14:03:28.000000000","message":"Thanks for clarification! I understand why we need more information than switch_id and port_id, and it makes sense.\nI also understand why switch_info, or vendor identifier, is required.\n\nswitch_id, port_id and switch_info sounds good to me.\nAs Kevin says, by having this kind of defined fields,\nwe can build thing on top of it and we allow multi-vendor environments.\n\nNote that switch_id does not need to be a mac-address format. LLDP use mac address to identify a switch. OpenFlow uses datapath_id which is a bit different format, for example.\n\nI am a bit worried about deep nested data structure, but it is not a thing we discuss here.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":9361,"name":"Vivekanandan Narasimhan","email":"n.vivekanandan@ericsson.com","username":"viveknarasimhan"},"change_message_id":"e9a681a2640f46fb51da75f311d8be5a18ce4cb5","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_bf80d78c","line":152,"in_reply_to":"ba3cc151_8647ee1c","updated":"2015-07-06 10:46:11.000000000","message":"+1 for Sukhdev\u0027s point here.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"c6375657e2eff61bd14f0bb991213253e49ad55a","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_8647ee1c","line":152,"in_reply_to":"ba3cc151_b60a671e","updated":"2015-07-03 17:22:50.000000000","message":"I guess we can use the format for switch_id to be a string as well - this will gives flexibility to vendors. They can use  LLDP based mac-address, OpenFlow based datapath_id, or anything in between.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":7787,"name":"Kevin Benton","email":"kevin@benton.pub","username":"blak111"},"change_message_id":"bec951e71f1ecc9e38b0da3419898ca88c452eeb","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_5746997b","line":152,"in_reply_to":"ba3cc151_d74f0932","updated":"2015-07-02 21:05:52.000000000","message":"It makes sense, but I think we should agree on some minimum fields provided in this dict that uniquely identify a switch port. That way we can build user interfaces that are able to display switch information in a uniform way for all ports. \n\nIf we don\u0027t have that, the format will always be vendor-dependent so it will be difficult to build things on top of it.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"57ab0b4be5ece2f7c0bd3fbc88f2bd463e68980e","unresolved":false,"context_lines":[{"line_number":149,"context_line":"| local_link_information | A list of local link information either from all |"},{"line_number":150,"context_line":"|                        | Ironic ports in a particular port group or from a|"},{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_d74f0932","line":152,"in_reply_to":"ba3cc151_e445bd42","updated":"2015-07-02 20:54:35.000000000","message":"We should not confuse the issue with LLDP. I think author implies that this information can be gathered by using LLDP or should follow the format presented by LLDP - so that automation can be achieved by anybody who wishes to.\n\nThe port binding information is presented to all ML2 drivers by core plugin. The drivers are suppose to act on the information that they can deal with. \nThis is how ML2 drivers work presently. When we run multiple ML2 drivers, not all of them act on all the information in the context (or binding information). \n\nOne example comes to mind - say a data center has multiple clusters with different models of switches from the same vendor - (say 7510 and 7575). Each model of switch uses different set of config commands to plumb the same network. So, if the binding information has switch-info, the driver can then process the same request, but generate the correct configuration for the model of switch being configured. The vendor could implement one ML2 driver and act based upon the model number in the binding profile or implement multiple ML2 drivers and the relevant driver will act upon relevant binding information.\n\nHence, it makes sense to keep the profile information in the dict form. This gives the vendors flexibility, rather than creating multiple vendor specific extensions. \n\nAnother point, this is no way changing or impacting the Neutron API. All of this information is already present (or available) in the port context to ML2 drivers - most of the drivers do not look into binding profile. The drivers who are interested in Ironic integration will utilize and act upon it.\n\nHope this makes sense","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":7787,"name":"Kevin Benton","email":"kevin@benton.pub","username":"blak111"},"change_message_id":"29f4487240e0dfcf9f8ec88a32361fed28b23f4f","unresolved":false,"context_lines":[{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"},{"line_number":156,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":157,"context_line":"| host_id                | This should be set to the Ironic node uuid.      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_8477119f","line":154,"updated":"2015-07-02 20:10:41.000000000","message":"+1. This sounds like a good use of admin_state_up","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"acc92ee8a8c665c0c84ecd4867ca5f20b3a7bd10","unresolved":false,"context_lines":[{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"},{"line_number":156,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":157,"context_line":"| host_id                | This should be set to the Ironic node uuid.      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_1443326d","line":154,"updated":"2015-07-02 19:35:23.000000000","message":"The idea to use admin_state_up hits me now.\nIn Neutron context, admin_state_up\u003dFalse for a port means it does not allow any traffic to/from the logical port.\nI think what bind_requested tries to do is same.\nCan we just send admin_state_up \u003d False when creating a port?\nIf so, we don\u0027t need to introduce no additional attribute. Does it make sense?","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"d807c66563f836ddae70c9b77c4d921ce18d82fc","unresolved":false,"context_lines":[{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"},{"line_number":156,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":157,"context_line":"| host_id                | This should be set to the Ironic node uuid.      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_44a4a906","line":154,"in_reply_to":"ba3cc151_1443326d","updated":"2015-07-02 20:21:14.000000000","message":"This is interesting idea. I think we had discussed about using this in the past - but, do not recall why we did not choose to follow that path.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"8fc862e2a14f7b259976f86f87f9e53aadf77d21","unresolved":false,"context_lines":[{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"},{"line_number":156,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":157,"context_line":"| host_id                | This should be set to the Ironic node uuid.      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_f881d5c3","line":154,"in_reply_to":"ba3cc151_44a4a906","updated":"2015-07-03 14:03:28.000000000","message":"I noticed a reason that perhaps admin_state_up is not used.\nThe port is owned by a user who booted Ironic instance.\nIt means the user can change admin_state_up to True.\nThis information needs to be controlled only by an operator.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"c6375657e2eff61bd14f0bb991213253e49ad55a","unresolved":false,"context_lines":[{"line_number":151,"context_line":"|                        | single Ironic port not belonging to any port     |"},{"line_number":152,"context_line":"|                        | group.                                           |"},{"line_number":153,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"},{"line_number":156,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":157,"context_line":"| host_id                | This should be set to the Ironic node uuid.      |"}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_c65bb6f5","line":154,"in_reply_to":"ba3cc151_f881d5c3","updated":"2015-07-03 17:22:50.000000000","message":"Thanks for clarification - this makes sense. So, we will stay with the proposal at hand.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"acc92ee8a8c665c0c84ecd4867ca5f20b3a7bd10","unresolved":false,"context_lines":[{"line_number":154,"context_line":"| bind_requested         | A way to specify whether to bind the port or     |"},{"line_number":155,"context_line":"|                        | defer binding.                                   |"},{"line_number":156,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":157,"context_line":"| host_id                | This should be set to the Ironic node uuid.      |"},{"line_number":158,"context_line":"+------------------------+--------------------------------------------------+"},{"line_number":159,"context_line":""},{"line_number":160,"context_line":""}],"source_content_type":"text/x-rst","patch_set":5,"id":"ba3cc151_8f1c0f6b","line":157,"updated":"2015-07-02 19:35:23.000000000","message":"(just a comment) Now we have a different vnic_type, we can use a different interpretation of host_id. Good idea.","commit_id":"6b51f4440ef993ac8b61d87cbd396629924a6064"},{"author":{"_account_id":8106,"name":"Haomeng,Wang","email":"wanghaomeng@gmail.com","username":"whaom"},"change_message_id":"1e50327d2040f6ef219659c5418fd39e597f49c3","unresolved":false,"context_lines":[{"line_number":26,"context_line":"In order to provide the required network isolation, Ironic should be able to"},{"line_number":27,"context_line":"provide the requisite connectivity information (using LLDP) to the Neutron"},{"line_number":28,"context_line":"ML2 plugin to allow drivers to provision the top-of-rack (ToR) switch for the"},{"line_number":29,"context_line":"baremetal server."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Ironic also poses new challenges for Neutron, for example, the concepts of"},{"line_number":32,"context_line":"link aggregation and multiple networks per node. This spec covers the concept"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_3cda4c79","line":29,"updated":"2015-07-14 03:24:34.000000000","message":"I have concern here, if we can support multi-racks, which means our bare metal will run with different physical servers in differenets racks, so maybe user need such requirements to support not only one single rack switch management, all ToRs which used by user should be managed by ML2 drivers, so the rack and switch information should be retrieved via NMS(network management systems) interface services, then we have whole picture of all racks and ToRs and physical servers cable connections information. \n\nI think we just propose the first ML2 plugin version to support single rack, and will support multi-racks in the future.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"bbcd4a63e1ba566b4eb882b68e934dd3cbbcb34f","unresolved":false,"context_lines":[{"line_number":26,"context_line":"In order to provide the required network isolation, Ironic should be able to"},{"line_number":27,"context_line":"provide the requisite connectivity information (using LLDP) to the Neutron"},{"line_number":28,"context_line":"ML2 plugin to allow drivers to provision the top-of-rack (ToR) switch for the"},{"line_number":29,"context_line":"baremetal server."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Ironic also poses new challenges for Neutron, for example, the concepts of"},{"line_number":32,"context_line":"link aggregation and multiple networks per node. This spec covers the concept"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_685ec977","line":29,"in_reply_to":"9a41bdd9_3cda4c79","updated":"2015-07-14 16:06:45.000000000","message":"The server should know which switch(es) it is connected to. This switch information can then be passed to Neutron. I\u0027m not sure I understand the reason against this?","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"12d12bb50db4d4749f0687c79a8ab023de7e77ff","unresolved":false,"context_lines":[{"line_number":26,"context_line":"In order to provide the required network isolation, Ironic should be able to"},{"line_number":27,"context_line":"provide the requisite connectivity information (using LLDP) to the Neutron"},{"line_number":28,"context_line":"ML2 plugin to allow drivers to provision the top-of-rack (ToR) switch for the"},{"line_number":29,"context_line":"baremetal server."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Ironic also poses new challenges for Neutron, for example, the concepts of"},{"line_number":32,"context_line":"link aggregation and multiple networks per node. This spec covers the concept"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_a0127455","line":29,"in_reply_to":"9a41bdd9_3cda4c79","updated":"2015-07-15 04:58:23.000000000","message":"This spec is not intended to deal with the discovery of ToRs or the connectivity information - which could be discovered by using discoverd or any other mechanism. This spec assumes that such an information is available and populated in Ironic DB.\nThe automated discovery of the physical connectivity/topology will be addressed separately and if out of scope for this spec","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"425ebfa69547c50bf06a2029a5bcec43621f0067","unresolved":false,"context_lines":[{"line_number":26,"context_line":"In order to provide the required network isolation, Ironic should be able to"},{"line_number":27,"context_line":"provide the requisite connectivity information (using LLDP) to the Neutron"},{"line_number":28,"context_line":"ML2 plugin to allow drivers to provision the top-of-rack (ToR) switch for the"},{"line_number":29,"context_line":"baremetal server."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Ironic also poses new challenges for Neutron, for example, the concepts of"},{"line_number":32,"context_line":"link aggregation and multiple networks per node. This spec covers the concept"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_b4439512","line":29,"in_reply_to":"9a41bdd9_a0127455","updated":"2015-07-15 09:17:29.000000000","message":"As Laura and Sukhdev comment, the scope of this spec is to share adjacent switch (ToR) information between ironic and neutron.\nHow to collect ToR switch information connected to bare-metal servers is out of scope.\n\nTopic related to single rack vs multi-rack, switches from single vendor or multi vendors depend on how ToR switches are managed.\nIf ML2 drivers directly manage switches directly (or through switch vendor management layer) we can only support a single switch vendor. If ML2 driver talks with a backend SDN controller (for example OpenFlow controller), we can support multi vendor switches. It all depends on how network management system is configured and what kind of ML2 drivers/neutron plugins are available. In any cases, switch information discussed in this spec is always required.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10910,"name":"Stig Telfer","email":"stig.openstack@telfer.org","username":"stigtelfer"},"change_message_id":"955f5fbac5c4062346b8e332bc2afe962b543b71","unresolved":false,"context_lines":[{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks.  This may require the capability to specify which NICs"},{"line_number":43,"context_line":"  belong to which network. It is in scope but the method of deciding which"},{"line_number":44,"context_line":"  interface gets which network will be decided elsewhere."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":6,"id":"ba3cc151_c269d4fa","line":44,"updated":"2015-07-08 09:46:26.000000000","message":"For reference there is a patch set in review here https://review.openstack.org/#/c/153230/ which addresses this precise issue, and which we use in our environment to run bare metal instances on nodes with multiple physical networks.\n\nI saw you\u0027re referencing the associated bug below.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"a4c110dba3da36fce93c2fbfd2424d28c3d6ee6e","unresolved":false,"context_lines":[{"line_number":41,"context_line":"* the case where the bare metal server has multiple interfaces that belong to"},{"line_number":42,"context_line":"  different networks.  This may require the capability to specify which NICs"},{"line_number":43,"context_line":"  belong to which network. It is in scope but the method of deciding which"},{"line_number":44,"context_line":"  interface gets which network will be decided elsewhere."},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"Proposed change"},{"line_number":47,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_56e14dcf","line":44,"in_reply_to":"ba3cc151_c269d4fa","updated":"2015-07-13 15:11:44.000000000","message":"That patch doesn\u0027t allow the nova user to specify this, though.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10910,"name":"Stig Telfer","email":"stig.openstack@telfer.org","username":"stigtelfer"},"change_message_id":"955f5fbac5c4062346b8e332bc2afe962b543b71","unresolved":false,"context_lines":[{"line_number":56,"context_line":"contain a list of the local link information for each port in that port group."},{"line_number":57,"context_line":"For each port not belonging to a port group, the port binding profile will"},{"line_number":58,"context_line":"contain only the local link information for that particular port. Therefore"},{"line_number":59,"context_line":"the length of the list of local link information can be used by Neutron to"},{"line_number":60,"context_line":"determine whether to plumb the network for a single port or for a port group"},{"line_number":61,"context_line":"(in the case of a port group there may be additional switch configurations"},{"line_number":62,"context_line":"necessary to manage the associated LAG).  Each port binding profile sent to"}],"source_content_type":"text/x-rst","patch_set":6,"id":"ba3cc151_0553463a","line":59,"updated":"2015-07-08 09:46:26.000000000","message":"This overloads the meaning of a LAG, I\u0027m aware of single-member LAGs, whether due to temporary link failure or space reserved for future expansion.  Is there another way of distinguishing between the two?","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9bed81f124685174204e4e3781c60492e15faa08","unresolved":false,"context_lines":[{"line_number":56,"context_line":"contain a list of the local link information for each port in that port group."},{"line_number":57,"context_line":"For each port not belonging to a port group, the port binding profile will"},{"line_number":58,"context_line":"contain only the local link information for that particular port. Therefore"},{"line_number":59,"context_line":"the length of the list of local link information can be used by Neutron to"},{"line_number":60,"context_line":"determine whether to plumb the network for a single port or for a port group"},{"line_number":61,"context_line":"(in the case of a port group there may be additional switch configurations"},{"line_number":62,"context_line":"necessary to manage the associated LAG).  Each port binding profile sent to"}],"source_content_type":"text/x-rst","patch_set":6,"id":"ba3cc151_bbe5b209","line":59,"in_reply_to":"ba3cc151_0553463a","updated":"2015-07-08 22:55:29.000000000","message":"I am not sure if I understand your comment about overloading. You are making two very important points: \n1) LAG being a single member LAG (because of failure) - for the purpose of this interface, it is still a multi-member LAG, and the operator is expected to present both ports in the port-group definition (otherwise, it is not LAG). Moreover, it is expected that both sides match - i.e. the switch side must be pre-configured in LAG mode and this interface present it as a group of ports. Failures are part of life. The switches know how to deal with when one link of a LAG\u0027ed pair goes down.\n2) LAG being a single member (by design for future expansion). This is not supported in this implementation. When ever such physical connectivity information is changed (as is the case in scenario), the admin action is required. This means the server is removed from the network - i.e. neutron delete_port(), configure the LAG members, rediscover the sever, and issue a new neutron port_create() with port-group which will carry the LAG\u0027ed group information. \n\nI realize that I am brief in my explanation, but, this is well discussed in the Ironic_neutron meetings.\n\nHope this helps.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10910,"name":"Stig Telfer","email":"stig.openstack@telfer.org","username":"stigtelfer"},"change_message_id":"955f5fbac5c4062346b8e332bc2afe962b543b71","unresolved":false,"context_lines":[{"line_number":62,"context_line":"necessary to manage the associated LAG).  Each port binding profile sent to"},{"line_number":63,"context_line":"the Neutron port create API will result in the creation of a single Neutron"},{"line_number":64,"context_line":"port. In this way, each port group can represent a LAG for which all member"},{"line_number":65,"context_line":"switch ports will belong to the same network and to the same network segment."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"The node should first be placed on a provisioning network and then onto the"},{"line_number":68,"context_line":"tenant network as specified in the `network-provider spec"}],"source_content_type":"text/x-rst","patch_set":6,"id":"ba3cc151_8514364d","line":65,"updated":"2015-07-08 09:46:26.000000000","message":"The additional work for handling LAGs creates interesting problems here.  Switches operate on LAGs at a logical level.\n\nLLDP partner state is maintained on the physical ports that are members of the LAG.  \n\nAll other port state and configuration - including VLAN membership - is not applied to the member ports.  It is applied to a logical interface created to represent the LAG as a single entity.  When changing the VLAN membership of a LAG, the port group of physical LAG member ports is not the correct entity to address.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"9bed81f124685174204e4e3781c60492e15faa08","unresolved":false,"context_lines":[{"line_number":62,"context_line":"necessary to manage the associated LAG).  Each port binding profile sent to"},{"line_number":63,"context_line":"the Neutron port create API will result in the creation of a single Neutron"},{"line_number":64,"context_line":"port. In this way, each port group can represent a LAG for which all member"},{"line_number":65,"context_line":"switch ports will belong to the same network and to the same network segment."},{"line_number":66,"context_line":""},{"line_number":67,"context_line":"The node should first be placed on a provisioning network and then onto the"},{"line_number":68,"context_line":"tenant network as specified in the `network-provider spec"}],"source_content_type":"text/x-rst","patch_set":6,"id":"ba3cc151_1e004c00","line":65,"in_reply_to":"ba3cc151_8514364d","updated":"2015-07-08 22:55:29.000000000","message":"This is correct. The switches which support LAG configurations know how to deal with this. From the point of view of Ironic/Neutron integration and configuration, this is really irrelevant. I am assuming the operators will select appropriate switches when they are using LAG\u0027ed configuration. \n\nAnother point, we are not making any assumption about LLDP running on the servers. Not to say that it can not run, but, we are not relying on any of this configuration based upon LLDP. \n\nHope this helps","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":2889,"name":"Aeva Black","email":"aeva.online@gmail.com","username":"tenbrae"},"change_message_id":"89966e90245c86af5de18c65884c7e8c5a9af321","unresolved":false,"context_lines":[{"line_number":133,"context_line":"follows:"},{"line_number":134,"context_line":""},{"line_number":135,"context_line":"* Select port groups with an address that corresponds to a pxe_enabled port\u0027s"},{"line_number":136,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":137,"context_line":""},{"line_number":138,"context_line":"* Select pxe_enabled ports from port groups. If there are multiple ports"},{"line_number":139,"context_line":"  choose one at random."}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_75cea7b5","line":136,"updated":"2015-07-14 05:59:55.000000000","message":"I don\u0027t think there is any benefit to introducing nondeterministic behavior intentionally within the boot process.\n\nIs there any reason not to support \u003e1 NIC for PXE booting, whereby Neutron prepares the DHCPBOOT config for all ports that have this option set?","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"bbcd4a63e1ba566b4eb882b68e934dd3cbbcb34f","unresolved":false,"context_lines":[{"line_number":133,"context_line":"follows:"},{"line_number":134,"context_line":""},{"line_number":135,"context_line":"* Select port groups with an address that corresponds to a pxe_enabled port\u0027s"},{"line_number":136,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":137,"context_line":""},{"line_number":138,"context_line":"* Select pxe_enabled ports from port groups. If there are multiple ports"},{"line_number":139,"context_line":"  choose one at random."}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_7865201c","line":136,"in_reply_to":"9a41bdd9_75cea7b5","updated":"2015-07-14 16:06:45.000000000","message":"The intention is to have 1 Neutron port per Ironic port_group. For PXE, the port_group should be pxe_enabled and have pxe_enabled port members. This is trying to convey a methodology for determining the pxe_address for the port_group. \n\nNeutron would prepare this for all port_groups and all ports not belonging to port_groups. If Neutron was to prepare this for all ports in a given port_group, then wouldn\u0027t this mean the 1 Neutron port would have multiple addresses? Also, my understanding is that the LAG will only boot over 1 interface, but I\u0027d like to know more about the use case behind your suggestion. Thanks.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"425ebfa69547c50bf06a2029a5bcec43621f0067","unresolved":false,"context_lines":[{"line_number":133,"context_line":"follows:"},{"line_number":134,"context_line":""},{"line_number":135,"context_line":"* Select port groups with an address that corresponds to a pxe_enabled port\u0027s"},{"line_number":136,"context_line":"  address. If there are multiple ports choose one at random."},{"line_number":137,"context_line":""},{"line_number":138,"context_line":"* Select pxe_enabled ports from port groups. If there are multiple ports"},{"line_number":139,"context_line":"  choose one at random."}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_9da8ed78","line":136,"in_reply_to":"9a41bdd9_7865201c","updated":"2015-07-15 09:17:29.000000000","message":"I am not sure I understood Laura\u0027s comment correctly.\n\nPXE boot is used during the provisioning phase.\nEven if LAG is used for baremetal instance, I think Ironic can still use physical interfaces for PXE boot. It is used only during the provisioning phase, the decision is up to the network provider in Ironic.\n\nI don\u0027t fully understand why we cannot have mulitple pxe_enabled port.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":2889,"name":"Aeva Black","email":"aeva.online@gmail.com","username":"tenbrae"},"change_message_id":"89966e90245c86af5de18c65884c7e8c5a9af321","unresolved":false,"context_lines":[{"line_number":249,"context_line":"+-----------------------+-------------------------+"},{"line_number":250,"context_line":"| name                  | str_or_none             |"},{"line_number":251,"context_line":"+-----------------------+-------------------------+"},{"line_number":252,"context_line":"| node_id               | str_or_none             |"},{"line_number":253,"context_line":"+-----------------------+-------------------------+"},{"line_number":254,"context_line":"| address               | str_or_none             |"},{"line_number":255,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_e8998320","line":252,"updated":"2015-07-14 05:59:55.000000000","message":"node_id should be int_or_none\n\nin the REST API, we\u0027ll accept/return the UUID, but internally (in the db / data model) we\u0027ll store the INT ID","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"bbcd4a63e1ba566b4eb882b68e934dd3cbbcb34f","unresolved":false,"context_lines":[{"line_number":249,"context_line":"+-----------------------+-------------------------+"},{"line_number":250,"context_line":"| name                  | str_or_none             |"},{"line_number":251,"context_line":"+-----------------------+-------------------------+"},{"line_number":252,"context_line":"| node_id               | str_or_none             |"},{"line_number":253,"context_line":"+-----------------------+-------------------------+"},{"line_number":254,"context_line":"| address               | str_or_none             |"},{"line_number":255,"context_line":"+-----------------------+-------------------------+"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_cbc58f3c","line":252,"in_reply_to":"9a41bdd9_e8998320","updated":"2015-07-14 16:06:45.000000000","message":"Yes, agree, will update this","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"a4c110dba3da36fce93c2fbfd2424d28c3d6ee6e","unresolved":false,"context_lines":[{"line_number":276,"context_line":"splitting into two port_groups, or two Ironic ports associated with two"},{"line_number":277,"context_line":"Neutron ports merging into one port_group. The default Ironic update behavior"},{"line_number":278,"context_line":"would be to associate Neutron vif_ids with port_groups rather than ports, and"},{"line_number":279,"context_line":"to update Neutron if local_link_information in the binding profile changes."},{"line_number":280,"context_line":""},{"line_number":281,"context_line":""},{"line_number":282,"context_line":"Any port updates that update local_link_connection or pxe_enabled can only be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_39bb5ae7","line":279,"updated":"2015-07-13 15:11:44.000000000","message":"I think I\u0027d rather not allow port (group) updates while the node is in a state that has neutron ports associated. I tend to think Neutron port updates should only happen during a state change, not automatically with any port-update call. Just for the sake of code sanity. :)\n\nThis means we should allow it in MANAGEABLE/INSPECTING/ENROLL, and perhaps AVAILABLE + maintenance.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":2889,"name":"Aeva Black","email":"aeva.online@gmail.com","username":"tenbrae"},"change_message_id":"89966e90245c86af5de18c65884c7e8c5a9af321","unresolved":false,"context_lines":[{"line_number":276,"context_line":"splitting into two port_groups, or two Ironic ports associated with two"},{"line_number":277,"context_line":"Neutron ports merging into one port_group. The default Ironic update behavior"},{"line_number":278,"context_line":"would be to associate Neutron vif_ids with port_groups rather than ports, and"},{"line_number":279,"context_line":"to update Neutron if local_link_information in the binding profile changes."},{"line_number":280,"context_line":""},{"line_number":281,"context_line":""},{"line_number":282,"context_line":"Any port updates that update local_link_connection or pxe_enabled can only be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_68ad7330","line":279,"in_reply_to":"9a41bdd9_39bb5ae7","updated":"2015-07-14 05:59:55.000000000","message":"jroll: ++\n\nAt least in the initial rev of this, I think this is acceptable. (trying to change ports for a deployed Nova instance is .... eh, not necessarily fun)","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"bbcd4a63e1ba566b4eb882b68e934dd3cbbcb34f","unresolved":false,"context_lines":[{"line_number":276,"context_line":"splitting into two port_groups, or two Ironic ports associated with two"},{"line_number":277,"context_line":"Neutron ports merging into one port_group. The default Ironic update behavior"},{"line_number":278,"context_line":"would be to associate Neutron vif_ids with port_groups rather than ports, and"},{"line_number":279,"context_line":"to update Neutron if local_link_information in the binding profile changes."},{"line_number":280,"context_line":""},{"line_number":281,"context_line":""},{"line_number":282,"context_line":"Any port updates that update local_link_connection or pxe_enabled can only be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_6e9d7158","line":279,"in_reply_to":"9a41bdd9_68ad7330","updated":"2015-07-14 16:06:45.000000000","message":"Happy to go with this, so to confirm, we\u0027ll only allow port (group) updates in MANAGEABLE/INSPECTING/ENROLL and not ACTIVE/RESCUE+maintenance. Open question over AVAILABLE+maintenance. In first rev it might be easiest to simply disallow this as well. Operator could move the node from AVAILABLE to MANAGEABLE first, then do the port_group update, then move back to AVAILABLE (this is the recommended approach proposed for the upgrade process - see deployer impact section below).","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"12d12bb50db4d4749f0687c79a8ab023de7e77ff","unresolved":false,"context_lines":[{"line_number":276,"context_line":"splitting into two port_groups, or two Ironic ports associated with two"},{"line_number":277,"context_line":"Neutron ports merging into one port_group. The default Ironic update behavior"},{"line_number":278,"context_line":"would be to associate Neutron vif_ids with port_groups rather than ports, and"},{"line_number":279,"context_line":"to update Neutron if local_link_information in the binding profile changes."},{"line_number":280,"context_line":""},{"line_number":281,"context_line":""},{"line_number":282,"context_line":"Any port updates that update local_link_connection or pxe_enabled can only be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_a0c35488","line":279,"in_reply_to":"9a41bdd9_6e9d7158","updated":"2015-07-15 04:58:23.000000000","message":"I think this seems reasonable for the first version (and most likely in the future as well)","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"425ebfa69547c50bf06a2029a5bcec43621f0067","unresolved":false,"context_lines":[{"line_number":276,"context_line":"splitting into two port_groups, or two Ironic ports associated with two"},{"line_number":277,"context_line":"Neutron ports merging into one port_group. The default Ironic update behavior"},{"line_number":278,"context_line":"would be to associate Neutron vif_ids with port_groups rather than ports, and"},{"line_number":279,"context_line":"to update Neutron if local_link_information in the binding profile changes."},{"line_number":280,"context_line":""},{"line_number":281,"context_line":""},{"line_number":282,"context_line":"Any port updates that update local_link_connection or pxe_enabled can only be"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_1d5dbd78","line":279,"in_reply_to":"9a41bdd9_a0c35488","updated":"2015-07-15 09:17:29.000000000","message":"+1 for Laura.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"a4c110dba3da36fce93c2fbfd2424d28c3d6ee6e","unresolved":false,"context_lines":[{"line_number":286,"context_line":"should result in an update of local_link_information in the binding_profile,"},{"line_number":287,"context_line":"which would trigger an update in Neutron. It might be safest to only allow"},{"line_number":288,"context_line":"this when the node is not in a state where uninterrupted connectivity is"},{"line_number":289,"context_line":"expected."},{"line_number":290,"context_line":""},{"line_number":291,"context_line":"REST API impact"},{"line_number":292,"context_line":"---------------"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_b9cf6a7e","line":289,"updated":"2015-07-13 15:11:44.000000000","message":"Ditto.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":2889,"name":"Aeva Black","email":"aeva.online@gmail.com","username":"tenbrae"},"change_message_id":"89966e90245c86af5de18c65884c7e8c5a9af321","unresolved":false,"context_lines":[{"line_number":407,"context_line":"    \"updated_at\": \"2015-05-15T09:04:12.011844+00:00\","},{"line_number":408,"context_line":"    \"uuid\": \"1004e542-2f9f-4d9b-b8b9-5b719fa6613f\","},{"line_number":409,"context_line":"    \"local_link_connection\": {"},{"line_number":410,"context_line":"      \"chassis_id\": \"0a:1b:2c:3d:4e:5f\","},{"line_number":411,"context_line":"      \"port_id\": \"Ethernet3/1\","},{"line_number":412,"context_line":"      \"system_name\": \"switch1\","},{"line_number":413,"context_line":"    },"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_f5b5d730","line":410,"updated":"2015-07-14 05:59:55.000000000","message":"what is this chassis_id representing or referring to? I don\u0027t see the phrase any where else in this document","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"425ebfa69547c50bf06a2029a5bcec43621f0067","unresolved":false,"context_lines":[{"line_number":407,"context_line":"    \"updated_at\": \"2015-05-15T09:04:12.011844+00:00\","},{"line_number":408,"context_line":"    \"uuid\": \"1004e542-2f9f-4d9b-b8b9-5b719fa6613f\","},{"line_number":409,"context_line":"    \"local_link_connection\": {"},{"line_number":410,"context_line":"      \"chassis_id\": \"0a:1b:2c:3d:4e:5f\","},{"line_number":411,"context_line":"      \"port_id\": \"Ethernet3/1\","},{"line_number":412,"context_line":"      \"system_name\": \"switch1\","},{"line_number":413,"context_line":"    },"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_5df58554","line":410,"in_reply_to":"9a41bdd9_cef35d9d","updated":"2015-07-15 09:17:29.000000000","message":"switch_id and port_id are required for minimum case.\nswitch_info is optional for various cases including a case with multiple vendor ML2 drivers.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"bbcd4a63e1ba566b4eb882b68e934dd3cbbcb34f","unresolved":false,"context_lines":[{"line_number":407,"context_line":"    \"updated_at\": \"2015-05-15T09:04:12.011844+00:00\","},{"line_number":408,"context_line":"    \"uuid\": \"1004e542-2f9f-4d9b-b8b9-5b719fa6613f\","},{"line_number":409,"context_line":"    \"local_link_connection\": {"},{"line_number":410,"context_line":"      \"chassis_id\": \"0a:1b:2c:3d:4e:5f\","},{"line_number":411,"context_line":"      \"port_id\": \"Ethernet3/1\","},{"line_number":412,"context_line":"      \"system_name\": \"switch1\","},{"line_number":413,"context_line":"    },"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_cef35d9d","line":410,"in_reply_to":"9a41bdd9_f5b5d730","updated":"2015-07-14 16:06:45.000000000","message":"Good catch, sorry this was from an earlier version and I missed to update this. Will update in next patch. Should be switch_id, port_id, switch_info.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"a4c110dba3da36fce93c2fbfd2424d28c3d6ee6e","unresolved":false,"context_lines":[{"line_number":851,"context_line":""},{"line_number":852,"context_line":""},{"line_number":853,"context_line":"Discuss any policy changes, and discuss what things a deployer needs to"},{"line_number":854,"context_line":"think about when defining their policy."},{"line_number":855,"context_line":""},{"line_number":856,"context_line":"* A deployer should be aware of the capabilities of the particular ML2 driver"},{"line_number":857,"context_line":"  for supporting use of the new local_link_information that will be passed to"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_59e17ecf","line":854,"updated":"2015-07-13 15:11:44.000000000","message":"Ironic has a blanket admin-only policy, shouldn\u0027t need to worry about this.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10239,"name":"Dmitry Tantsur","email":"dtantsur@protonmail.com","username":"dtantsur"},"change_message_id":"9cb0f25ffadfeee4970edda29a44535222ce69d3","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"  * For port_groups, the CLI would support the following new methods:"},{"line_number":890,"context_line":""},{"line_number":891,"context_line":"    * ironic port_group-create --node \u003cnode\u003e [--name \u003cportgroupname\u003e]"},{"line_number":892,"context_line":"      [--address \u003cmac-address\u003e] [-e \u003ckey\u003dvalue\u003e]"},{"line_number":893,"context_line":""},{"line_number":894,"context_line":"        * To add ports to a port_group, the port_group should first"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3a50d1a3_a8e74911","line":891,"updated":"2015-07-20 08:05:24.000000000","message":"++ for not using _","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":2889,"name":"Aeva Black","email":"aeva.online@gmail.com","username":"tenbrae"},"change_message_id":"89966e90245c86af5de18c65884c7e8c5a9af321","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"  * For port_groups, the CLI would support the following new methods:"},{"line_number":890,"context_line":""},{"line_number":891,"context_line":"    * ironic port_group-create --node \u003cnode\u003e [--name \u003cportgroupname\u003e]"},{"line_number":892,"context_line":"      [--address \u003cmac-address\u003e] [-e \u003ckey\u003dvalue\u003e]"},{"line_number":893,"context_line":""},{"line_number":894,"context_line":"        * To add ports to a port_group, the port_group should first"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_f5471749","line":891,"updated":"2015-07-14 05:59:55.000000000","message":"mixing of \"_\" and \"-\" is very bad for usability. Why not just use \"-\" ?","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"425ebfa69547c50bf06a2029a5bcec43621f0067","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"  * For port_groups, the CLI would support the following new methods:"},{"line_number":890,"context_line":""},{"line_number":891,"context_line":"    * ironic port_group-create --node \u003cnode\u003e [--name \u003cportgroupname\u003e]"},{"line_number":892,"context_line":"      [--address \u003cmac-address\u003e] [-e \u003ckey\u003dvalue\u003e]"},{"line_number":893,"context_line":""},{"line_number":894,"context_line":"        * To add ports to a port_group, the port_group should first"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_ddaf5528","line":891,"in_reply_to":"9a41bdd9_002ec042","updated":"2015-07-15 09:17:29.000000000","message":"In the convention on CLI option, we don\u0027t use \"_\" at all.\nPlease use \"-\" instead.\n\nThis kind of fix was done in neutronclient several release ago,\nbut we still have compatibility option for \"_\" and it is really a tech debt. I would suggest not to use \"_\" in the command line option.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"12d12bb50db4d4749f0687c79a8ab023de7e77ff","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"  * For port_groups, the CLI would support the following new methods:"},{"line_number":890,"context_line":""},{"line_number":891,"context_line":"    * ironic port_group-create --node \u003cnode\u003e [--name \u003cportgroupname\u003e]"},{"line_number":892,"context_line":"      [--address \u003cmac-address\u003e] [-e \u003ckey\u003dvalue\u003e]"},{"line_number":893,"context_line":""},{"line_number":894,"context_line":"        * To add ports to a port_group, the port_group should first"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_002ec042","line":891,"in_reply_to":"9a41bdd9_04e7d411","updated":"2015-07-15 04:58:23.000000000","message":"In the interest of avoiding confusion, I think going with portgroup is a reasonable option.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"bbcd4a63e1ba566b4eb882b68e934dd3cbbcb34f","unresolved":false,"context_lines":[{"line_number":888,"context_line":""},{"line_number":889,"context_line":"  * For port_groups, the CLI would support the following new methods:"},{"line_number":890,"context_line":""},{"line_number":891,"context_line":"    * ironic port_group-create --node \u003cnode\u003e [--name \u003cportgroupname\u003e]"},{"line_number":892,"context_line":"      [--address \u003cmac-address\u003e] [-e \u003ckey\u003dvalue\u003e]"},{"line_number":893,"context_line":""},{"line_number":894,"context_line":"        * To add ports to a port_group, the port_group should first"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_04e7d411","line":891,"in_reply_to":"9a41bdd9_f5471749","updated":"2015-07-14 16:06:45.000000000","message":"This was discriminating between the object (port_group) and the action (create). I\u0027m also fine to change to port-group-create, but would like to clarify if this comment only refers to the cli command or does it also refer to the object? If it\u0027s consistently changed throughout to port-group then there\u0027ll be the same issue with eg port-group_uuid assuming we want to keep the notation \u0027_uuid\u0027. Or do you prefer port_group for Ironic model and REST API but port-group for the CLI? This looks a bit inconsistent, which is why I ask. Another option might be portgroup removing the need for either. What do you think? Thanks.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":2889,"name":"Aeva Black","email":"aeva.online@gmail.com","username":"tenbrae"},"change_message_id":"89966e90245c86af5de18c65884c7e8c5a9af321","unresolved":false,"context_lines":[{"line_number":896,"context_line":""},{"line_number":897,"context_line":"    * ironic port_group-delete port_group_uuid"},{"line_number":898,"context_line":""},{"line_number":899,"context_line":"    * ironic port_group-list [--detail] [--node \u003cnode\u003e]"},{"line_number":900,"context_line":"      [--address \u003cmac-address\u003e]"},{"line_number":901,"context_line":"      [--limit \u003climit\u003e]  [--marker \u003cport_group_uuid] [--sort-key \u003cfield\u003e]"},{"line_number":902,"context_line":"      [--sort-dir \u003cdirection\u003e]"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_95756b99","line":899,"updated":"2015-07-14 05:59:55.000000000","message":"ditto for the remainder of these commands","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":2889,"name":"Aeva Black","email":"aeva.online@gmail.com","username":"tenbrae"},"change_message_id":"89966e90245c86af5de18c65884c7e8c5a9af321","unresolved":false,"context_lines":[{"line_number":1032,"context_line":"turn this new feature on or off, which would be important when upgrading"},{"line_number":1033,"context_line":"Nova and Ironic."},{"line_number":1034,"context_line":""},{"line_number":1035,"context_line":"Deployers should follow this recommendation for upgrading a node in an"},{"line_number":1036,"context_line":"existing deployment to use this new feature:"},{"line_number":1037,"context_line":""},{"line_number":1038,"context_line":"* Upgrade the OpenStack services."}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_fcff04a8","line":1035,"updated":"2015-07-14 05:59:55.000000000","message":"how does this upgrade process differ for already-provisioned bare metal instances? Is it possible to do?","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"12d12bb50db4d4749f0687c79a8ab023de7e77ff","unresolved":false,"context_lines":[{"line_number":1032,"context_line":"turn this new feature on or off, which would be important when upgrading"},{"line_number":1033,"context_line":"Nova and Ironic."},{"line_number":1034,"context_line":""},{"line_number":1035,"context_line":"Deployers should follow this recommendation for upgrading a node in an"},{"line_number":1036,"context_line":"existing deployment to use this new feature:"},{"line_number":1037,"context_line":""},{"line_number":1038,"context_line":"* Upgrade the OpenStack services."}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_c032185e","line":1035,"in_reply_to":"9a41bdd9_fcff04a8","updated":"2015-07-15 04:58:23.000000000","message":"I think supporting automated upgrade or migration for the exiting nodes may be a problematic. DB migration alone may not do the trick. \nPerhaps we should recommend that operators follow the same process as is described for adding/modifying port-groups.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":16115,"name":"Laura Moore","email":"lauramoore1403@gmail.com","username":"lauramoore"},"change_message_id":"bbcd4a63e1ba566b4eb882b68e934dd3cbbcb34f","unresolved":false,"context_lines":[{"line_number":1032,"context_line":"turn this new feature on or off, which would be important when upgrading"},{"line_number":1033,"context_line":"Nova and Ironic."},{"line_number":1034,"context_line":""},{"line_number":1035,"context_line":"Deployers should follow this recommendation for upgrading a node in an"},{"line_number":1036,"context_line":"existing deployment to use this new feature:"},{"line_number":1037,"context_line":""},{"line_number":1038,"context_line":"* Upgrade the OpenStack services."}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_c4f7ac57","line":1035,"in_reply_to":"9a41bdd9_fcff04a8","updated":"2015-07-14 16:06:45.000000000","message":"So I was thinking this would be the recommended approach to start using this feature. I think the question ties in with the discussion above regarding the state machine and what states the node can be for the port_group introduction/update.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":8106,"name":"Haomeng,Wang","email":"wanghaomeng@gmail.com","username":"whaom"},"change_message_id":"9446441ae7e30cd5129740494271b62edc2dc1bf","unresolved":false,"context_lines":[{"line_number":1104,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."},{"line_number":1105,"context_line":""},{"line_number":1106,"context_line":"VLAN provisioning on switch(es) is dependent on ML2 driver functionality"},{"line_number":1107,"context_line":"being developed to support this feature."},{"line_number":1108,"context_line":""},{"line_number":1109,"context_line":""},{"line_number":1110,"context_line":"Testing"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_0674d3fa","line":1107,"updated":"2015-07-14 07:47:29.000000000","message":"Is there such VLAN ML2 driver spec/bp, or other references?","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":841,"name":"Akihiro Motoki","email":"amotoki@gmail.com","username":"amotoki"},"change_message_id":"425ebfa69547c50bf06a2029a5bcec43621f0067","unresolved":false,"context_lines":[{"line_number":1104,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."},{"line_number":1105,"context_line":""},{"line_number":1106,"context_line":"VLAN provisioning on switch(es) is dependent on ML2 driver functionality"},{"line_number":1107,"context_line":"being developed to support this feature."},{"line_number":1108,"context_line":""},{"line_number":1109,"context_line":""},{"line_number":1110,"context_line":"Testing"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_000a3679","line":1107,"in_reply_to":"9a41bdd9_007b20f6","updated":"2015-07-15 09:17:29.000000000","message":"I did not noticed this paragraph so far.\n\nIs \"VLAN provisioning\" a requirement?\nI know we need \"virtual network\"-aware switches (which means such switches isolate access ports connected to baremetal servers from different neutron networks), but I believe we can use any network isolation mechanism: VLAN, VXLAN and so on.\n\nMy above comment may be too strict. I am okay with the current sentence.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":8106,"name":"Haomeng,Wang","email":"wanghaomeng@gmail.com","username":"whaom"},"change_message_id":"f6abfa91c223029b12f285b64e372a935c079fb8","unresolved":false,"context_lines":[{"line_number":1104,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."},{"line_number":1105,"context_line":""},{"line_number":1106,"context_line":"VLAN provisioning on switch(es) is dependent on ML2 driver functionality"},{"line_number":1107,"context_line":"being developed to support this feature."},{"line_number":1108,"context_line":""},{"line_number":1109,"context_line":""},{"line_number":1110,"context_line":"Testing"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3a50d1a3_538603dc","line":1107,"in_reply_to":"9a41bdd9_007b20f6","updated":"2015-07-22 02:03:33.000000000","message":"Yes, I understand it should be virtual VLAN done by ovs I think, do you mean it is the vlan driver which can control and configurate the physical siwtches?","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":8106,"name":"Haomeng,Wang","email":"wanghaomeng@gmail.com","username":"whaom"},"change_message_id":"666e96f73a515bfb23c2a3e0936e5f7d01aa0596","unresolved":false,"context_lines":[{"line_number":1104,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."},{"line_number":1105,"context_line":""},{"line_number":1106,"context_line":"VLAN provisioning on switch(es) is dependent on ML2 driver functionality"},{"line_number":1107,"context_line":"being developed to support this feature."},{"line_number":1108,"context_line":""},{"line_number":1109,"context_line":""},{"line_number":1110,"context_line":"Testing"}],"source_content_type":"text/x-rst","patch_set":6,"id":"3a50d1a3_f3faaf74","line":1107,"in_reply_to":"9a41bdd9_007b20f6","updated":"2015-07-22 01:58:44.000000000","message":"ok, will check, thank you.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":6558,"name":"Sukhdev Kapur","email":"sukhdevkapur@gmail.com","username":"sukhdev-8"},"change_message_id":"12d12bb50db4d4749f0687c79a8ab023de7e77ff","unresolved":false,"context_lines":[{"line_number":1104,"context_line":"\u003chttps://blueprints.launchpad.net/ironic/+spec/network-provider\u003e`_."},{"line_number":1105,"context_line":""},{"line_number":1106,"context_line":"VLAN provisioning on switch(es) is dependent on ML2 driver functionality"},{"line_number":1107,"context_line":"being developed to support this feature."},{"line_number":1108,"context_line":""},{"line_number":1109,"context_line":""},{"line_number":1110,"context_line":"Testing"}],"source_content_type":"text/x-rst","patch_set":6,"id":"9a41bdd9_007b20f6","line":1107,"in_reply_to":"9a41bdd9_0674d3fa","updated":"2015-07-15 04:58:23.000000000","message":"VLAN ML2 type driver and mechanism drivers have been in existence since Havana Release. Please look for Neutron documents, you will find plenty of information.","commit_id":"2c5f405b4e071ff7f4320d46776ccdb7e725c287"},{"author":{"_account_id":10202,"name":"Wan-yen Hsu","email":"wanyenhsu@gmail.com"},"change_message_id":"c1867ed3c10d0da2ead0d59b42f28ad1f32020c9","unresolved":false,"context_lines":[{"line_number":76,"context_line":"disk installation only). The reason for this is because the nodes will not be"},{"line_number":77,"context_line":"able to PXE boot if they cannot reach the TFTP server. Local boot will"},{"line_number":78,"context_line":"need to be used for any node deployed outside of the provisioning network."},{"line_number":79,"context_line":"Instance netboot should not be permitted when this feature is used. This"},{"line_number":80,"context_line":"limitation could be addressed if routing was set up from every tenant network"},{"line_number":81,"context_line":"to the TFTP server or if Ironic would work with a TFTP server per tenant"},{"line_number":82,"context_line":"network, but these options are out of scope in this spec."}],"source_content_type":"text/x-rst","patch_set":7,"id":"3a50d1a3_e9b16290","line":79,"updated":"2015-07-27 23:47:43.000000000","message":"Some deploy drivers do not use tftp, e.g., virtual media drivers.  So, the spec should allow virtual media driver or other deploy method that does not use tftp to do netboot.  Am I missing something here?","commit_id":"b4db386b2fc4539815df0b86a715d2aaff544e03"},{"author":{"_account_id":18781,"name":"vinay muddu","email":"vinay50muddu@yahoo.com","username":"vmud213"},"change_message_id":"f3251e1bc185be986287d54682b215084eed662e","unresolved":false,"context_lines":[{"line_number":161,"context_line":""},{"line_number":162,"context_line":"     \"binding:profile\": {"},{"line_number":163,"context_line":""},{"line_number":164,"context_line":"         \"local_link_information\": ["},{"line_number":165,"context_line":"               {"},{"line_number":166,"context_line":"                  \"switch_id\": xxx,"},{"line_number":167,"context_line":""}],"source_content_type":"text/x-rst","patch_set":7,"id":"dab17558_df49cb32","line":164,"range":{"start_line":164,"start_character":10,"end_line":164,"end_character":32},"updated":"2016-05-11 05:19:54.000000000","message":"May be this is local_link_connection?","commit_id":"b4db386b2fc4539815df0b86a715d2aaff544e03"},{"author":{"_account_id":10343,"name":"Jim Rollenhagen","email":"jim@jimrollenhagen.com","username":"jimrollenhagen"},"change_message_id":"6d861e16a83a9a9370ec2fc0aab1b38e4eee1e0e","unresolved":false,"context_lines":[{"line_number":385,"context_line":"    \"updated_at\": \"2015-05-15T09:04:12.011844+00:00\","},{"line_number":386,"context_line":"    \"uuid\": \"1004e542-2f9f-4d9b-b8b9-5b719fa6613f\","},{"line_number":387,"context_line":"    \"local_link_connection\": {"},{"line_number":388,"context_line":"      \"swwitch_id\": \"0a:1b:2c:3d:4e:5f\","},{"line_number":389,"context_line":"      \"port_id\": \"Ethernet3/1\","},{"line_number":390,"context_line":"      \"switch_info\": \"switch1\","},{"line_number":391,"context_line":"    },"}],"source_content_type":"text/x-rst","patch_set":7,"id":"3a50d1a3_42796f2e","line":388,"updated":"2015-07-21 14:32:37.000000000","message":"Nit, if you push another patch: s/swwitch/switch/","commit_id":"b4db386b2fc4539815df0b86a715d2aaff544e03"}]}
