)]}'
{"/COMMIT_MSG":[{"author":{"_account_id":8864,"name":"Artom Lifshitz","email":"notartom@gmail.com","username":"artom"},"change_message_id":"c3c12d9dda3f51199590fe95495080d583a95db1","unresolved":false,"context_lines":[{"line_number":6,"context_line":""},{"line_number":7,"context_line":"objects: adding serial, path, bus_address fields to bdm object"},{"line_number":8,"context_line":""},{"line_number":9,"context_line":"Adding fields in the title to the block device mapping object"},{"line_number":10,"context_line":"to be able to form disk metadata."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"path - Holds a fix disk path, when used in a conteiner"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"1a122d0e_1cee3eb4","line":9,"updated":"2016-05-03 14:15:49.000000000","message":"Nit: for clarity, I\u0027d repeat the fields here - ex: \"Adding fields serial, path, bus_address\"","commit_id":"10047db45c3b03eb1862b2ce64314ef30a0e5fb3"},{"author":{"_account_id":8864,"name":"Artom Lifshitz","email":"notartom@gmail.com","username":"artom"},"change_message_id":"c3c12d9dda3f51199590fe95495080d583a95db1","unresolved":false,"context_lines":[{"line_number":9,"context_line":"Adding fields in the title to the block device mapping object"},{"line_number":10,"context_line":"to be able to form disk metadata."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"path - Holds a fix disk path, when used in a conteiner"},{"line_number":13,"context_line":"serial - user set serial disk serial string"},{"line_number":14,"context_line":"bus_address - The format is based on the bus, the user requested to"},{"line_number":15,"context_line":"              this device to be put, can be any of the following:"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"1a122d0e_9c388e3f","line":12,"updated":"2016-05-03 14:15:49.000000000","message":"Spelling: \"container\"","commit_id":"10047db45c3b03eb1862b2ce64314ef30a0e5fb3"},{"author":{"_account_id":8864,"name":"Artom Lifshitz","email":"notartom@gmail.com","username":"artom"},"change_message_id":"c3c12d9dda3f51199590fe95495080d583a95db1","unresolved":false,"context_lines":[{"line_number":10,"context_line":"to be able to form disk metadata."},{"line_number":11,"context_line":""},{"line_number":12,"context_line":"path - Holds a fix disk path, when used in a conteiner"},{"line_number":13,"context_line":"serial - user set serial disk serial string"},{"line_number":14,"context_line":"bus_address - The format is based on the bus, the user requested to"},{"line_number":15,"context_line":"              this device to be put, can be any of the following:"},{"line_number":16,"context_line":"              PCI address, or SCSI address, of USB port, or IDE channel, etc."}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"1a122d0e_1c459eb2","line":13,"updated":"2016-05-03 14:15:49.000000000","message":"Nit: \"user-set\"","commit_id":"10047db45c3b03eb1862b2ce64314ef30a0e5fb3"},{"author":{"_account_id":8864,"name":"Artom Lifshitz","email":"notartom@gmail.com","username":"artom"},"change_message_id":"c3c12d9dda3f51199590fe95495080d583a95db1","unresolved":false,"context_lines":[{"line_number":13,"context_line":"serial - user set serial disk serial string"},{"line_number":14,"context_line":"bus_address - The format is based on the bus, the user requested to"},{"line_number":15,"context_line":"              this device to be put, can be any of the following:"},{"line_number":16,"context_line":"              PCI address, or SCSI address, of USB port, or IDE channel, etc."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Partially implements blueprint virt-device-role-tagging"},{"line_number":19,"context_line":"Co-authored-by: Artom Lifshitz \u003califshit@redhat.com\u003e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":2,"id":"1a122d0e_1c8f5ecd","line":16,"updated":"2016-05-03 14:15:49.000000000","message":"I\u0027d rephrase the whole paragraph to something like \"The block device\u0027s address as it appears to the guest. The format is based on the bus type, for example PCI address, SCSI address, etc.\"","commit_id":"10047db45c3b03eb1862b2ce64314ef30a0e5fb3"},{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"d6bbb572df0e8a9e58587c1a9f7844de91905059","unresolved":false,"context_lines":[{"line_number":13,"context_line":"serial - user-set serial disk serial string"},{"line_number":14,"context_line":"bus_address - The block device\u0027s address as it appears to the guest."},{"line_number":15,"context_line":"              The format is based on the bus, can be any of the following:"},{"line_number":16,"context_line":"              PCI address, or SCSI address, of USB port, or IDE channel, etc."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Partially implements blueprint virt-device-role-tagging"},{"line_number":19,"context_line":"Co-authored-by: Artom Lifshitz \u003califshit@redhat.com\u003e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"dab17558_5e46f921","line":16,"updated":"2016-05-13 10:16:25.000000000","message":"This really surprises me to see added.  This completely diverges from the spec design and the earlier implementations of this series.\n\nThe design calls for us to define an object model that formally represents device metadata information, that is directly serializable to JSON. The virt drivers would return an instance of this object model for use by the config driver generator and/or metadata service. As such there should be no need to persist this data against block devices and/or network interfaces in the database.\n\nI think this patch is really taking things in the wrong direction and not aligned with the approved spec.","commit_id":"6b7a476cf0a865a0809568b8c7e2615809e5925a"},{"author":{"_account_id":8802,"name":"Vladik Romanovsky","email":"vromanso@redhat.com","username":"vladikr"},"change_message_id":"4f85be30dd5f6e97ab38505462f8fe4cf5081e58","unresolved":false,"context_lines":[{"line_number":13,"context_line":"serial - user-set serial disk serial string"},{"line_number":14,"context_line":"bus_address - The block device\u0027s address as it appears to the guest."},{"line_number":15,"context_line":"              The format is based on the bus, can be any of the following:"},{"line_number":16,"context_line":"              PCI address, or SCSI address, of USB port, or IDE channel, etc."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Partially implements blueprint virt-device-role-tagging"},{"line_number":19,"context_line":"Co-authored-by: Artom Lifshitz \u003califshit@redhat.com\u003e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"dab17558_95070917","line":16,"in_reply_to":"dab17558_1e38f13c","updated":"2016-05-17 13:29:59.000000000","message":"I understand your concern, however, I don\u0027t see how to proceed here.\nDan Smith was opposing to the metadata object, because the API would have to make a blocking call to retrieve it. [1][2]\n\nWe have decided to store the necessary information the database/objects (bdm and virtual interface) for simplicity and easy access for it in the api, where we form the metadata object.\nI don\u0027t see how do I pave the way toward a scenario that you have mentioned, as we are not allowing it in the api..\n\nInstead of having these fields (bus_address, path..) directly in the BDM, would serializing the DiskMetadata and NetworkInterfaceMetadata into json and storing it in a \u0027metadata\u0027 field of the BDM and VirtualInterface objects will make things better?\n \n[1] https://review.openstack.org/#/c/304509/2/nova/compute/api.py@2825\n\n[2] https://review.openstack.org/#/c/304511/2/nova/api/metadata/base.py@200","commit_id":"6b7a476cf0a865a0809568b8c7e2615809e5925a"},{"author":{"_account_id":1779,"name":"Daniel Berrange","email":"berrange@redhat.com","username":"berrange"},"change_message_id":"373a1ec38be2c1deed50a47466b25037d1f28e69","unresolved":false,"context_lines":[{"line_number":13,"context_line":"serial - user-set serial disk serial string"},{"line_number":14,"context_line":"bus_address - The block device\u0027s address as it appears to the guest."},{"line_number":15,"context_line":"              The format is based on the bus, can be any of the following:"},{"line_number":16,"context_line":"              PCI address, or SCSI address, of USB port, or IDE channel, etc."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Partially implements blueprint virt-device-role-tagging"},{"line_number":19,"context_line":"Co-authored-by: Artom Lifshitz \u003califshit@redhat.com\u003e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"dab17558_1e38f13c","line":16,"in_reply_to":"dab17558_5e46f921","updated":"2016-05-13 10:22:19.000000000","message":"Further, by specifying addresses against the bdm, you\u0027re taking us straight towards a scenario where the user can specify addresses via the API when providing the BDM which is very much not what we want.\n\nI really don\u0027t think that device addresses should be exposed in the general nova database model at all.\n\nIn addition this is going to create some problems for cases where we need to be able to represent things in the device metadata which are not representable in the Nova bdm/virtual interface tables at all.\n\nFor example, in order to correctly handle SCSI addressing, I believe we\u0027re going to need to include details of the SCSI controller in the device metadata.\n\neg, a SCSI disk is associated with a SCSI controller, which is associated with a PCI address. Explicitly modelling the device metadata and having the virt driver return this object model allowed for this possibility. By just recording addresses against the BDM you\u0027ve made it impossible to express any kind of relationships between devices - they\u0027re just all isolated","commit_id":"6b7a476cf0a865a0809568b8c7e2615809e5925a"},{"author":{"_account_id":7730,"name":"Sahid Orentino Ferdjaoui","email":"sahid.ferdjaoui@industrialdiscipline.com","username":"sahid"},"change_message_id":"d55b5db1a58919de468034d1dfb5e45f6d436442","unresolved":false,"context_lines":[{"line_number":13,"context_line":"serial - user-set serial disk serial string"},{"line_number":14,"context_line":"bus_address - The block device\u0027s address as it appears to the guest."},{"line_number":15,"context_line":"              The format is based on the bus, can be any of the following:"},{"line_number":16,"context_line":"              PCI address, or SCSI address, of USB port, or IDE channel, etc."},{"line_number":17,"context_line":""},{"line_number":18,"context_line":"Partially implements blueprint virt-device-role-tagging"},{"line_number":19,"context_line":"Co-authored-by: Artom Lifshitz \u003califshit@redhat.com\u003e"}],"source_content_type":"text/x-gerrit-commit-message","patch_set":5,"id":"dab17558_b2b91918","line":16,"in_reply_to":"dab17558_5e46f921","updated":"2016-05-13 10:32:10.000000000","message":"I think there were comments from Dan Smith about to make it persistent to the database.\n\nFrom my thinking by that way each drivers don\u0027t have to build their own way to store this information (e.g: libvirt has domain XML to store metadata but we don\u0027t know about xen or vmware...), so if a new driver have to implement this feature it just have to correctly fill this object. - In the other way since now that is persistent to the database we should to add an extra effort to update this information when guests are moving between hosts.\n\nI do not have a strong opinion of what is the best and your knowledge would be probably really helpful.","commit_id":"6b7a476cf0a865a0809568b8c7e2615809e5925a"}],"nova/objects/block_device.py":[{"author":{"_account_id":8864,"name":"Artom Lifshitz","email":"notartom@gmail.com","username":"artom"},"change_message_id":"c3c12d9dda3f51199590fe95495080d583a95db1","unresolved":false,"context_lines":[{"line_number":319,"context_line":"    # Version 1.15: BlockDeviceMapping \u003c\u003d version 1.14"},{"line_number":320,"context_line":"    # Version 1.16: BlockDeviceMapping \u003c\u003d version 1.15"},{"line_number":321,"context_line":"    # Version 1.17: Add get_by_instance_uuids()"},{"line_number":322,"context_line":"    # Version 1.18: BlockDeviceMapping \u003c\u003d version 1.17"},{"line_number":323,"context_line":"    VERSION \u003d \u00271.18\u0027"},{"line_number":324,"context_line":""},{"line_number":325,"context_line":"    fields \u003d {"}],"source_content_type":"text/x-python","patch_set":2,"id":"1a122d0e_dcc5e63d","line":322,"updated":"2016-05-03 14:15:49.000000000","message":"Shouldn\u0027t this be \u003c\u003d version 1.18?","commit_id":"10047db45c3b03eb1862b2ce64314ef30a0e5fb3"},{"author":{"_account_id":8802,"name":"Vladik Romanovsky","email":"vromanso@redhat.com","username":"vladikr"},"change_message_id":"572a3b6590678b9a2037375afb2c47ab87524dff","unresolved":false,"context_lines":[{"line_number":319,"context_line":"    # Version 1.15: BlockDeviceMapping \u003c\u003d version 1.14"},{"line_number":320,"context_line":"    # Version 1.16: BlockDeviceMapping \u003c\u003d version 1.15"},{"line_number":321,"context_line":"    # Version 1.17: Add get_by_instance_uuids()"},{"line_number":322,"context_line":"    # Version 1.18: BlockDeviceMapping \u003c\u003d version 1.17"},{"line_number":323,"context_line":"    VERSION \u003d \u00271.18\u0027"},{"line_number":324,"context_line":""},{"line_number":325,"context_line":"    fields \u003d {"}],"source_content_type":"text/x-python","patch_set":2,"id":"1a122d0e_f7cf1d55","line":322,"in_reply_to":"1a122d0e_dcc5e63d","updated":"2016-05-03 14:25:32.000000000","message":"hmm.. probably. :)\nI\u0027m not sure if we even need to change this object\u0027s version?","commit_id":"10047db45c3b03eb1862b2ce64314ef30a0e5fb3"}]}