)]}'
{"/PATCHSET_LEVEL":[{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"c4a54f4743a164a80f0b33452af82ecbecface8e","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":1,"id":"be6e782a_ffe6a851","updated":"2023-08-28 16:12:19.000000000","message":"almost all of the comments were telling what I understood, thanks for replies.\nI should have resolved ones which were not questions.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":16207,"name":"ribaudr","display_name":"uggla","email":"rene.ribaud@gmail.com","username":"uggla","status":"Red Hat"},"change_message_id":"645e5328a782f23ec4c2a413b3578853e6e1753f","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":2,"id":"b8430537_3471d367","updated":"2023-09-26 16:36:29.000000000","message":"\"Here is a first fairly quick review of the specification. Overall, the change seems okay and necessary to me. I\u0027ve marked it with a -1 to indicate that I\u0027ve made some comments to further improve the clarity of the specification and highlight the work regarding detaching a device.","commit_id":"5e31194e334e4f595e5bc92471c0784b710e38a7"},{"author":{"_account_id":16207,"name":"ribaudr","display_name":"uggla","email":"rene.ribaud@gmail.com","username":"uggla","status":"Red Hat"},"change_message_id":"9d0f878690ac498a16835e022fb9474f0bfe6db5","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"93bc88a9_045b25b3","updated":"2023-10-03 12:48:14.000000000","message":"-1 because I think we should discuss the removal or not of legacy XML files without aliases.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c678153993fc1709929329ccd3633f0c3a2a4952","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"88f96a33_4c4e9b8f","updated":"2023-09-29 15:47:24.000000000","message":"I do feel this spec is very important for very specific device types in libvirt. Accordingly this spec should be clear on which exact devices we plan to change and be clear on the fact we wouldn\u0027t target the others (as this is not needed yet)","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"6f2c6f8c7cda5c327646aa001cf35908d65e113a","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"5c2bd277_b22b4fe0","updated":"2023-10-27 13:46:28.000000000","message":"Looks good to me. I hold my +2 for Uggla\u0027s points to be settled.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"a81f6b0ed7e158c9d31a57448d8f3422189c7e24","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"9e87b21b_6b903425","updated":"2023-10-06 14:08:37.000000000","message":"Moving my opinion to +1 as I no longer see and holding concerns but I need to do a second pass before giving my go.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"19b7619f08aa8a0fba453ee644dee79e052d191b","unresolved":false,"context_lines":[],"source_content_type":"","patch_set":3,"id":"7d729656_64aecd90","updated":"2023-11-07 14:25:54.000000000","message":"OK, I think all concerns have been answered, nothing shall be holding this spec now.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"}],"specs/2024.1/approved/libvirt-dev-alias.rst":[{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"2f913d0be86e748ca45f088b79d97739258b5c46","unresolved":true,"context_lines":[{"line_number":20,"context_line":"Problem description"},{"line_number":21,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Nova currently looks up guest devices in XML for attach/detach and other"},{"line_number":24,"context_line":"modifications using a variety of methods. For example, disk devices use"},{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"}],"source_content_type":"text/x-rst","patch_set":1,"id":"829ff570_2da9df1e","line":25,"range":{"start_line":23,"start_character":24,"end_line":25,"end_character":47},"updated":"2023-08-28 13:44:09.000000000","message":"```\n  \u003cdevices\u003e\n   \u003cdisk block\u003e\n     \u003cserial\u003evol-id\u003c/serial\u003e\n   \u003c/disk\u003e\n  \u003c/devices\u003e```","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"47faa7c3e6901cd22475d94796bbcfe1cc04320d","unresolved":false,"context_lines":[{"line_number":20,"context_line":"Problem description"},{"line_number":21,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Nova currently looks up guest devices in XML for attach/detach and other"},{"line_number":24,"context_line":"modifications using a variety of methods. For example, disk devices use"},{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1646756f_3d64bc7e","line":25,"range":{"start_line":23,"start_character":24,"end_line":25,"end_character":47},"in_reply_to":"250e81ab_e3d3fc48","updated":"2023-08-28 17:28:48.000000000","message":"Okay got it, thanks.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"aea51aa6ec100c21fb57c6ddc47b4510876a8aa4","unresolved":true,"context_lines":[{"line_number":20,"context_line":"Problem description"},{"line_number":21,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Nova currently looks up guest devices in XML for attach/detach and other"},{"line_number":24,"context_line":"modifications using a variety of methods. For example, disk devices use"},{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"}],"source_content_type":"text/x-rst","patch_set":1,"id":"d1c4cda1_15d9ab3f","line":25,"range":{"start_line":23,"start_character":24,"end_line":25,"end_character":47},"in_reply_to":"829ff570_2da9df1e","updated":"2023-08-28 14:29:52.000000000","message":"Are you just providing an example or are you suggesting a change?","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"c4a54f4743a164a80f0b33452af82ecbecface8e","unresolved":false,"context_lines":[{"line_number":20,"context_line":"Problem description"},{"line_number":21,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":22,"context_line":""},{"line_number":23,"context_line":"Nova currently looks up guest devices in XML for attach/detach and other"},{"line_number":24,"context_line":"modifications using a variety of methods. For example, disk devices use"},{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"}],"source_content_type":"text/x-rst","patch_set":1,"id":"250e81ab_e3d3fc48","line":25,"range":{"start_line":23,"start_character":24,"end_line":25,"end_character":47},"in_reply_to":"d1c4cda1_15d9ab3f","updated":"2023-08-28 16:12:19.000000000","message":"providing example on what I understood, might be helpful for other reviewers, and if I am wrong, can get corrected.\n\nno change.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"2f913d0be86e748ca45f088b79d97739258b5c46","unresolved":true,"context_lines":[{"line_number":23,"context_line":"Nova currently looks up guest devices in XML for attach/detach and other"},{"line_number":24,"context_line":"modifications using a variety of methods. For example, disk devices use"},{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."}],"source_content_type":"text/x-rst","patch_set":1,"id":"396e06bd_35c8cf7c","line":26,"range":{"start_line":26,"start_character":8,"end_line":26,"end_character":60},"updated":"2023-08-28 13:44:09.000000000","message":"for disk type file\n```\n  \u003cdevices\u003e\n   \u003cdisk file\u003e\n     \u003cbackingstore\u003e\n       \u003csrc file\u003d\u0027...\u0027 /\u003e\n     \u003c/backingstore\u003e\n   \u003c/disk\u003e\n  \u003c/devices\u003e\n```\nthats what it mean right ?","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"c4a54f4743a164a80f0b33452af82ecbecface8e","unresolved":false,"context_lines":[{"line_number":23,"context_line":"Nova currently looks up guest devices in XML for attach/detach and other"},{"line_number":24,"context_line":"modifications using a variety of methods. For example, disk devices use"},{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."}],"source_content_type":"text/x-rst","patch_set":1,"id":"e7d1e6a1_10622a66","line":26,"range":{"start_line":26,"start_character":8,"end_line":26,"end_character":60},"in_reply_to":"1c19df85_14b827b7","updated":"2023-08-28 16:12:19.000000000","message":"Ack","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"aea51aa6ec100c21fb57c6ddc47b4510876a8aa4","unresolved":true,"context_lines":[{"line_number":23,"context_line":"Nova currently looks up guest devices in XML for attach/detach and other"},{"line_number":24,"context_line":"modifications using a variety of methods. For example, disk devices use"},{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."}],"source_content_type":"text/x-rst","patch_set":1,"id":"1c19df85_14b827b7","line":26,"range":{"start_line":26,"start_character":8,"end_line":26,"end_character":60},"in_reply_to":"396e06bd_35c8cf7c","updated":"2023-08-28 14:29:52.000000000","message":"No, I think you can set serial on file-backed disks. We stopped being able to set it on type\u003dlun disk devices a while ago. QEMU apparently never actually supported it and recently(ish) libvirt stopped letting you even try:\n\n```\nIf present, this specify serial number of virtual hard drive. For example, it may look like \u003cserial\u003eWD-WMAP9A966149\u003c/serial\u003e. Not supported for scsi-block devices, that is those using disk type \u0027block\u0027 using device \u0027lun\u0027 on bus \u0027scsi\u0027. Since 0.7.1\n```","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"2f913d0be86e748ca45f088b79d97739258b5c46","unresolved":true,"context_lines":[{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Use Cases"},{"line_number":32,"context_line":"---------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"da6c9f8f_9f8a3898","line":29,"range":{"start_line":28,"start_character":48,"end_line":29,"end_character":40},"updated":"2023-08-28 13:44:09.000000000","message":"for interfaces apart from mac-addr there is another unique property target-dev\n```\u003ctarget dev\u003d\u0027tap82b1b6cc-10\u0027/\u003e```","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"c4a54f4743a164a80f0b33452af82ecbecface8e","unresolved":false,"context_lines":[{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Use Cases"},{"line_number":32,"context_line":"---------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"70d45df6_c0112609","line":29,"range":{"start_line":28,"start_character":48,"end_line":29,"end_character":40},"in_reply_to":"bf5d0f5f_f917b410","updated":"2023-08-28 16:12:19.000000000","message":"ACk, \njust I wanted to point out target dev as a unique idetifier for interface.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"aea51aa6ec100c21fb57c6ddc47b4510876a8aa4","unresolved":true,"context_lines":[{"line_number":25,"context_line":"the \"serial\" property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Use Cases"},{"line_number":32,"context_line":"---------"}],"source_content_type":"text/x-rst","patch_set":1,"id":"bf5d0f5f_f917b410","line":29,"range":{"start_line":28,"start_character":48,"end_line":29,"end_character":40},"in_reply_to":"da6c9f8f_9f8a3898","updated":"2023-08-28 14:29:52.000000000","message":"Are you suggesting some change to what is written here? Not all interface types have a target dev (AFAIK) and I don\u0027t think we even have control over all the ones that do. Either way, I\u0027m not sure why it matters - the point here is to move to some consistent mechanism that lets us easily correlate virtual guest devices by the ID that *we* use to identify them.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"2f913d0be86e748ca45f088b79d97739258b5c46","unresolved":true,"context_lines":[{"line_number":41,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism for"},{"line_number":44,"context_line":"identifying devices of all types. For devices like volumes and network interfaces,"},{"line_number":45,"context_line":"the volume or port UUID should be used. For other devices, some other stable"},{"line_number":46,"context_line":"identifier that correlates to something in Nova or another service\u0027s database"},{"line_number":47,"context_line":"is required."}],"source_content_type":"text/x-rst","patch_set":1,"id":"08bf78be_23e8bc58","line":44,"range":{"start_line":44,"start_character":32,"end_line":44,"end_character":33},"updated":"2023-08-28 13:44:09.000000000","message":"right now by-default, alias-name is common property for all devices","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"c4a54f4743a164a80f0b33452af82ecbecface8e","unresolved":false,"context_lines":[{"line_number":41,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism for"},{"line_number":44,"context_line":"identifying devices of all types. For devices like volumes and network interfaces,"},{"line_number":45,"context_line":"the volume or port UUID should be used. For other devices, some other stable"},{"line_number":46,"context_line":"identifier that correlates to something in Nova or another service\u0027s database"},{"line_number":47,"context_line":"is required."}],"source_content_type":"text/x-rst","patch_set":1,"id":"ec68ce0e_59de6057","line":44,"range":{"start_line":44,"start_character":32,"end_line":44,"end_character":33},"in_reply_to":"07f152fe_214fb0b5","updated":"2023-08-28 16:12:19.000000000","message":"Ack","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"aea51aa6ec100c21fb57c6ddc47b4510876a8aa4","unresolved":true,"context_lines":[{"line_number":41,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism for"},{"line_number":44,"context_line":"identifying devices of all types. For devices like volumes and network interfaces,"},{"line_number":45,"context_line":"the volume or port UUID should be used. For other devices, some other stable"},{"line_number":46,"context_line":"identifier that correlates to something in Nova or another service\u0027s database"},{"line_number":47,"context_line":"is required."}],"source_content_type":"text/x-rst","patch_set":1,"id":"07f152fe_214fb0b5","line":44,"range":{"start_line":44,"start_character":32,"end_line":44,"end_character":33},"in_reply_to":"08bf78be_23e8bc58","updated":"2023-08-28 14:29:52.000000000","message":"I don\u0027t know what you mean by this. Are you referring to the auto-generated alias that libvirt adds when we don\u0027t specify it? The point here is not just to get aliases assigned to devices, but to set the alias to something we can use to identify which device matches which openstack element (i.e. which of four interfaces matches a specific neutron port).","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"2f913d0be86e748ca45f088b79d97739258b5c46","unresolved":true,"context_lines":[{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism for"},{"line_number":44,"context_line":"identifying devices of all types. For devices like volumes and network interfaces,"},{"line_number":45,"context_line":"the volume or port UUID should be used. For other devices, some other stable"},{"line_number":46,"context_line":"identifier that correlates to something in Nova or another service\u0027s database"},{"line_number":47,"context_line":"is required."},{"line_number":48,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"61c13cf7_c55989c9","line":45,"range":{"start_line":45,"start_character":14,"end_line":45,"end_character":38},"updated":"2023-08-28 13:44:09.000000000","message":"ack","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"c4a54f4743a164a80f0b33452af82ecbecface8e","unresolved":false,"context_lines":[{"line_number":42,"context_line":""},{"line_number":43,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism for"},{"line_number":44,"context_line":"identifying devices of all types. For devices like volumes and network interfaces,"},{"line_number":45,"context_line":"the volume or port UUID should be used. For other devices, some other stable"},{"line_number":46,"context_line":"identifier that correlates to something in Nova or another service\u0027s database"},{"line_number":47,"context_line":"is required."},{"line_number":48,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"c05d1567_297af4db","line":45,"range":{"start_line":45,"start_character":14,"end_line":45,"end_character":38},"in_reply_to":"61c13cf7_c55989c9","updated":"2023-08-28 16:12:19.000000000","message":"Done","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":34860,"name":"Amit Uniyal","email":"auniyal@redhat.com","username":"auniyal"},"change_message_id":"2f913d0be86e748ca45f088b79d97739258b5c46","unresolved":false,"context_lines":[{"line_number":45,"context_line":"the volume or port UUID should be used. For other devices, some other stable"},{"line_number":46,"context_line":"identifier that correlates to something in Nova or another service\u0027s database"},{"line_number":47,"context_line":"is required."},{"line_number":48,"context_line":""},{"line_number":49,"context_line":"Alternatives"},{"line_number":50,"context_line":"------------"},{"line_number":51,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"98d0ff55_2ee8fe63","line":48,"updated":"2023-08-28 13:44:09.000000000","message":"there are different type of devices (disk (file, block), interfaces, ), but as they all do not have same properties, to identify them Nova have to identify them using their specific idetifier/property. \n\nbut libvirt now provides an \"alias\" property which is common in all devices. Nova should use alias instead.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":6962,"name":"Kashyap Chamarthy","email":"kchamart@redhat.com","username":"kashyapc"},"change_message_id":"12d64aa7a3c8170f6cc4684553461c571ef36a08","unresolved":true,"context_lines":[{"line_number":130,"context_line":"Work Items"},{"line_number":131,"context_line":"----------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"- Enable setting and parsing the device alias on disk, interface, and pci devices"},{"line_number":134,"context_line":"- Actually set those device aliases in the various parts of the driver that create"},{"line_number":135,"context_line":"  those configs"},{"line_number":136,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1dda0384_2d1ec716","line":133,"updated":"2023-08-28 08:22:37.000000000","message":"Asking for clarity: since you explicitly list \"disk, interface, and PCI devices\",  is it safe to assume that we rule out the rest of the different kinds of devices on which the alias can be set?  E.g. the following devices are also alias-able (based on a quick local test):\n\n- RNG\n- Memory balloon\n- Mouse\n- Keyboard\n- Host/guest private communication \u0027channel\u0027 devices (\u003cchannel type\u003d\u0027unix\u0027\u003e followed by the specific channel  \u003ctarget type \u003d guestfwd|virtio|...\u003e)\n- Serial console devices\n\nThough, I doubt people want to set aliases for these devices.\n\nShould we make that assumption explicit, and declare something like: \"Initial alias support is only for disk, interface, and PCI devices\"?","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"aea51aa6ec100c21fb57c6ddc47b4510876a8aa4","unresolved":true,"context_lines":[{"line_number":130,"context_line":"Work Items"},{"line_number":131,"context_line":"----------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"- Enable setting and parsing the device alias on disk, interface, and pci devices"},{"line_number":134,"context_line":"- Actually set those device aliases in the various parts of the driver that create"},{"line_number":135,"context_line":"  those configs"},{"line_number":136,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"}],"source_content_type":"text/x-rst","patch_set":1,"id":"3b733565_e5a09797","line":133,"in_reply_to":"1dda0384_2d1ec716","updated":"2023-08-28 14:29:52.000000000","message":"Well, I think we only need to do it for devices we need to attach/detach or otherwise identify multiple instances. I\u0027m not aware of any place where we need to know which mouse is which in the XML. So perhaps  we can come up with some better description of the class of devices we need to do this for.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":6962,"name":"Kashyap Chamarthy","email":"kchamart@redhat.com","username":"kashyapc"},"change_message_id":"c6e05cc1481f7eaf543f0300426182b5ea5d3063","unresolved":true,"context_lines":[{"line_number":130,"context_line":"Work Items"},{"line_number":131,"context_line":"----------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"- Enable setting and parsing the device alias on disk, interface, and pci devices"},{"line_number":134,"context_line":"- Actually set those device aliases in the various parts of the driver that create"},{"line_number":135,"context_line":"  those configs"},{"line_number":136,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"}],"source_content_type":"text/x-rst","patch_set":1,"id":"a688d291_447e62e5","line":133,"in_reply_to":"3b733565_e5a09797","updated":"2023-08-29 09:20:52.000000000","message":"Heh, yeah, we definitely don\u0027t go \"mouse-hunting\" like that.  What you say here might be enough.  Something like:\n\n\"Only devices that can be attached/detached to an instance can have an alias: disks, network interfaces, and PCI devices.\"\n\nAlthough, the term \"PCI devices\" needs to be locked down a bit more.  But this is a minor point.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"8762f6ff18ddf10c453da66100ad919f2c29c54c","unresolved":false,"context_lines":[{"line_number":130,"context_line":"Work Items"},{"line_number":131,"context_line":"----------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"- Enable setting and parsing the device alias on disk, interface, and pci devices"},{"line_number":134,"context_line":"- Actually set those device aliases in the various parts of the driver that create"},{"line_number":135,"context_line":"  those configs"},{"line_number":136,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"}],"source_content_type":"text/x-rst","patch_set":1,"id":"e871132d_f3cd9f1d","line":133,"in_reply_to":"a688d291_447e62e5","updated":"2023-09-25 14:34:26.000000000","message":"This is the work items section, so listing the things I think need to be done to be considered finished here. But I\u0027ll adjust text earlier in the document to be explicit about the scope.","commit_id":"ef125a1aa55cea70e3479d68d2ea4059e536651a"},{"author":{"_account_id":16207,"name":"ribaudr","display_name":"uggla","email":"rene.ribaud@gmail.com","username":"uggla","status":"Red Hat"},"change_message_id":"645e5328a782f23ec4c2a413b3578853e6e1753f","unresolved":true,"context_lines":[{"line_number":45,"context_line":"like volumes and network interfaces, the volume or port UUID should be used."},{"line_number":46,"context_line":"For other devices, some other stable identifier that correlates to something"},{"line_number":47,"context_line":"in Nova or another service\u0027s database is required."},{"line_number":48,"context_line":""},{"line_number":49,"context_line":"Alternatives"},{"line_number":50,"context_line":"------------"},{"line_number":51,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"9c73fd9f_0316ab69","line":48,"updated":"2023-09-26 16:36:29.000000000","message":"\"Hello Dan, I\u0027d like to suggest a few points that, in my opinion, could enhance this section.\n\n1- Including a sample of libvirt XML for an alias or providing a link to the libvirt documentation on device aliases would be beneficial.\n\n2- It\u0027s important to reiterate the constraints related to aliases: an alias must begin with the prefix \u0027ua-\u0027 and may only contain the permitted characters [a-zA-Z0-9_-]. This means that a volume or port alias will take the form of \u0027ua-UUID\u0027.\n\n3- Also good to mention the benefits of aliases for detaching a device using the new method (virDomainDetachDeviceAlias).","commit_id":"5e31194e334e4f595e5bc92471c0784b710e38a7"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"60b9227a289debae5f7f4a25aafda875f18a3a3b","unresolved":true,"context_lines":[{"line_number":45,"context_line":"like volumes and network interfaces, the volume or port UUID should be used."},{"line_number":46,"context_line":"For other devices, some other stable identifier that correlates to something"},{"line_number":47,"context_line":"in Nova or another service\u0027s database is required."},{"line_number":48,"context_line":""},{"line_number":49,"context_line":"Alternatives"},{"line_number":50,"context_line":"------------"},{"line_number":51,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"ae63570d_5340d4e6","line":48,"in_reply_to":"9c73fd9f_0316ab69","updated":"2023-09-27 14:38:34.000000000","message":"\u003e 1- Including a sample of libvirt XML for an alias or providing a link to the libvirt documentation on device aliases would be beneficial.\n\nAck.\n\n\u003e 2- It\u0027s important to reiterate the constraints related to aliases: an alias must begin with the prefix \u0027ua-\u0027 and may only contain the permitted characters [a-zA-Z0-9_-]. This means that a volume or port alias will take the form of \u0027ua-UUID\u0027.\n\nI think this is really an implementation detail that is not that important to the overall design of the feature. It\u0027s something that is currently required by libvirt, but could change or be expanded in the future. I\u0027ll make a note here like \"according to libvirt\u0027s format requirements\" or something.\n\n\u003e 3- Also good to mention the benefits of aliases for detaching a device using the new method (virDomainDetachDeviceAlias).\n\nI make reference to this starting on L91, so I think I\u0027m good. I initially thought moving to this would be an obvious win, but after working on the implementation I\u0027m not sure we will really go this route, at least initially. So yes, it\u0027s a potential benefit, but to use it we\u0027d need to fully restructure some of our device handling code to make use of it. That may or may not be worth it, but if so, it\u0027ll have to come after we\u0027ve had a migration period where we can fully rely on the XML being alias-based.","commit_id":"5e31194e334e4f595e5bc92471c0784b710e38a7"},{"author":{"_account_id":16207,"name":"ribaudr","display_name":"uggla","email":"rene.ribaud@gmail.com","username":"uggla","status":"Red Hat"},"change_message_id":"9d0f878690ac498a16835e022fb9474f0bfe6db5","unresolved":false,"context_lines":[{"line_number":45,"context_line":"like volumes and network interfaces, the volume or port UUID should be used."},{"line_number":46,"context_line":"For other devices, some other stable identifier that correlates to something"},{"line_number":47,"context_line":"in Nova or another service\u0027s database is required."},{"line_number":48,"context_line":""},{"line_number":49,"context_line":"Alternatives"},{"line_number":50,"context_line":"------------"},{"line_number":51,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"18558aad_9a3a9a46","line":48,"in_reply_to":"ae63570d_5340d4e6","updated":"2023-10-03 12:48:14.000000000","message":"Ack","commit_id":"5e31194e334e4f595e5bc92471c0784b710e38a7"},{"author":{"_account_id":16207,"name":"ribaudr","display_name":"uggla","email":"rene.ribaud@gmail.com","username":"uggla","status":"Red Hat"},"change_message_id":"645e5328a782f23ec4c2a413b3578853e6e1753f","unresolved":true,"context_lines":[{"line_number":139,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"},{"line_number":140,"context_line":"  alias and fall back to the old way"},{"line_number":141,"context_line":"- Migrate existing instance XMLs on startup when device aliases are missing"},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Dependencies"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":145,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"feb0c3a2_e9ebde18","line":142,"updated":"2023-09-26 16:36:29.000000000","message":"You should also add the modifications around detaching a device using the new virDomainDetachDeviceAlias method.","commit_id":"5e31194e334e4f595e5bc92471c0784b710e38a7"},{"author":{"_account_id":16207,"name":"ribaudr","display_name":"uggla","email":"rene.ribaud@gmail.com","username":"uggla","status":"Red Hat"},"change_message_id":"9d0f878690ac498a16835e022fb9474f0bfe6db5","unresolved":false,"context_lines":[{"line_number":139,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"},{"line_number":140,"context_line":"  alias and fall back to the old way"},{"line_number":141,"context_line":"- Migrate existing instance XMLs on startup when device aliases are missing"},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Dependencies"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":145,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"10937331_35544728","line":142,"in_reply_to":"303e0a22_d74b84ae","updated":"2023-10-03 12:48:14.000000000","message":"I realized this after looking at your initial code in WIP.\nI also noticed virDomainDetachDeviceAlias is an async method, I think it will bring complexity, and it is probably not useful in our case. SO I agree you it is not worth to use this method.","commit_id":"5e31194e334e4f595e5bc92471c0784b710e38a7"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"60b9227a289debae5f7f4a25aafda875f18a3a3b","unresolved":true,"context_lines":[{"line_number":139,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"},{"line_number":140,"context_line":"  alias and fall back to the old way"},{"line_number":141,"context_line":"- Migrate existing instance XMLs on startup when device aliases are missing"},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Dependencies"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"},{"line_number":145,"context_line":""}],"source_content_type":"text/x-rst","patch_set":2,"id":"303e0a22_d74b84ae","line":142,"in_reply_to":"feb0c3a2_e9ebde18","updated":"2023-09-27 14:38:34.000000000","message":"As noted above, I\u0027m not planning to make that change because it will require a substantial rewrite of a bunch of other code. It seems like it would be nicer to use that method, but it\u0027s really not necessary to get the gain here. By using the alias to look up devices and then the usual detach-via-xml route, we can retain a lot of the existing code path but without the ambiguity.","commit_id":"5e31194e334e4f595e5bc92471c0784b710e38a7"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c678153993fc1709929329ccd3633f0c3a2a4952","unresolved":true,"context_lines":[{"line_number":11,"context_line":"https://blueprints.launchpad.net/nova/+spec/libvirt-dev-alias"},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Currently we identify devices in Libvirt guest XML by a variety of methods,"},{"line_number":14,"context_line":"which differs based on the device type (at least). Libvirt now provides a"},{"line_number":15,"context_line":"device alias mechanism by which we can tie virtual guest devices to an"},{"line_number":16,"context_line":"identifier we can use to look them up in a stable and generic way. Nova"},{"line_number":17,"context_line":"should move to using that, which will increase consistency, decrease some"}],"source_content_type":"text/x-rst","patch_set":3,"id":"e690629c_ba110a87","line":14,"range":{"start_line":14,"start_character":51,"end_line":14,"end_character":62},"updated":"2023-09-29 15:47:24.000000000","message":"femtonit: you shall explicitly say here or below which libvirt version supports the aliases (which is 3.9.0), and you could just say \u0027meh, as our min is 7.0.0\".","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"0a28e25ce898525044225af66869adbd4165ca80","unresolved":false,"context_lines":[{"line_number":11,"context_line":"https://blueprints.launchpad.net/nova/+spec/libvirt-dev-alias"},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Currently we identify devices in Libvirt guest XML by a variety of methods,"},{"line_number":14,"context_line":"which differs based on the device type (at least). Libvirt now provides a"},{"line_number":15,"context_line":"device alias mechanism by which we can tie virtual guest devices to an"},{"line_number":16,"context_line":"identifier we can use to look them up in a stable and generic way. Nova"},{"line_number":17,"context_line":"should move to using that, which will increase consistency, decrease some"}],"source_content_type":"text/x-rst","patch_set":3,"id":"675852bb_10aedf9a","line":14,"range":{"start_line":14,"start_character":51,"end_line":14,"end_character":62},"in_reply_to":"47399eb8_12bfe9fd","updated":"2023-10-02 14:05:44.000000000","message":"OK, I just wanted to be explicit (operators should wonder when it was supporting but meh).","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"06cd026a25c9a4cf101f2ede1118b1004b52a5fa","unresolved":true,"context_lines":[{"line_number":11,"context_line":"https://blueprints.launchpad.net/nova/+spec/libvirt-dev-alias"},{"line_number":12,"context_line":""},{"line_number":13,"context_line":"Currently we identify devices in Libvirt guest XML by a variety of methods,"},{"line_number":14,"context_line":"which differs based on the device type (at least). Libvirt now provides a"},{"line_number":15,"context_line":"device alias mechanism by which we can tie virtual guest devices to an"},{"line_number":16,"context_line":"identifier we can use to look them up in a stable and generic way. Nova"},{"line_number":17,"context_line":"should move to using that, which will increase consistency, decrease some"}],"source_content_type":"text/x-rst","patch_set":3,"id":"47399eb8_12bfe9fd","line":14,"range":{"start_line":14,"start_character":51,"end_line":14,"end_character":62},"in_reply_to":"e690629c_ba110a87","updated":"2023-09-29 16:34:09.000000000","message":"It was so long ago it seems pointless to mention this and not say anything about the other features we use since 0.9.7 :)","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c678153993fc1709929329ccd3633f0c3a2a4952","unresolved":false,"context_lines":[{"line_number":25,"context_line":"the ``serial`` property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Example volume attachment::"},{"line_number":32,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"5e9d04d7_9f47a25c","line":29,"range":{"start_line":28,"start_character":66,"end_line":29,"end_character":41},"updated":"2023-09-29 15:47:24.000000000","message":"++","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"761955c7e0cabb546d36b38da799f85f44d77f61","unresolved":false,"context_lines":[{"line_number":25,"context_line":"the ``serial`` property to identify them uniquely. However, libvirt and"},{"line_number":26,"context_line":"qemu do not support setting this property on all disk device types, which"},{"line_number":27,"context_line":"means Nova cannot use that to look up disk devices in a generic way. Further,"},{"line_number":28,"context_line":"if we have multiple network interfaces with the same MAC address, using that"},{"line_number":29,"context_line":"as a unique identifier is not sufficient."},{"line_number":30,"context_line":""},{"line_number":31,"context_line":"Example volume attachment::"},{"line_number":32,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"16e0d2d2_9f367a6b","line":29,"range":{"start_line":28,"start_character":66,"end_line":29,"end_character":41},"in_reply_to":"5e9d04d7_9f47a25c","updated":"2023-11-08 11:41:48.000000000","message":"ya this is a know limitation. \ni cant recall if we have a fallback today or not for that edgecase\n\nwe also have to do it slightly diffently for sriov\n\nso it woudl be nice to consolidate that.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c678153993fc1709929329ccd3633f0c3a2a4952","unresolved":true,"context_lines":[{"line_number":55,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism"},{"line_number":56,"context_line":"[1]_ for identifying all types of devices that are attach- or"},{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"}],"source_content_type":"text/x-rst","patch_set":3,"id":"f9fdc88f_ed32c825","line":58,"range":{"start_line":58,"start_character":37,"end_line":58,"end_character":54},"updated":"2023-09-29 15:47:24.000000000","message":"I know it will be hard to be exhaustive, but just about mdevs, I think we could also use their uuids, despite we don\u0027t persist them in Nova (only in Placement).","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"a81f6b0ed7e158c9d31a57448d8f3422189c7e24","unresolved":false,"context_lines":[{"line_number":55,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism"},{"line_number":56,"context_line":"[1]_ for identifying all types of devices that are attach- or"},{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"}],"source_content_type":"text/x-rst","patch_set":3,"id":"e2376a24_d4ad7be2","line":58,"range":{"start_line":58,"start_character":37,"end_line":58,"end_character":54},"in_reply_to":"03932678_e6a9bbd7","updated":"2023-10-06 14:08:37.000000000","message":"OK, after some discussion on IRC, it appears to me I misunderstood Dan\u0027s point : he was asking whether we do VGPU/MDEV attach/detach on the API, which isn\u0027t the case. We just bind vGPUs to instances when they are created or if they\u0027re migrated, but we don\u0027t really support something like \"I want to attach *this* VGPU to that instance\".\n\nAccordingly, my concern on mdevs seems useless as not in scope.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"0a28e25ce898525044225af66869adbd4165ca80","unresolved":true,"context_lines":[{"line_number":55,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism"},{"line_number":56,"context_line":"[1]_ for identifying all types of devices that are attach- or"},{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"}],"source_content_type":"text/x-rst","patch_set":3,"id":"03932678_e6a9bbd7","line":58,"range":{"start_line":58,"start_character":37,"end_line":58,"end_character":54},"in_reply_to":"a17beb06_7ed90761","updated":"2023-10-02 14:05:44.000000000","message":"Sure, if you\u0027re asking where we define them in the XML, it\u0027s here :\nhttps://github.com/openstack/nova/blob/87d4807848bb9546c1fca972da2eb2eda13eb08d/nova/virt/libvirt/driver.py#L7262\n\nIf you want to know how we verify whether an instance has a mdev when rebooting, it\u0027s there where we check this :\nhttps://github.com/openstack/nova/blob/87d4807848bb9546c1fca972da2eb2eda13eb08d/nova/virt/libvirt/driver.py#L3946-L3978\n\nIf you want to know how we allocate a mdev for a new instance, that\u0027s there : \nhttps://github.com/openstack/nova/blob/87d4807848bb9546c1fca972da2eb2eda13eb08d/nova/virt/libvirt/driver.py#L8484","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"06cd026a25c9a4cf101f2ede1118b1004b52a5fa","unresolved":true,"context_lines":[{"line_number":55,"context_line":"Nova\u0027s libvirt driver should move to using the device alias mechanism"},{"line_number":56,"context_line":"[1]_ for identifying all types of devices that are attach- or"},{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"}],"source_content_type":"text/x-rst","patch_set":3,"id":"a17beb06_7ed90761","line":58,"range":{"start_line":58,"start_character":37,"end_line":58,"end_character":54},"in_reply_to":"f9fdc88f_ed32c825","updated":"2023-09-29 16:34:09.000000000","message":"Can you point me to the attach/detach code for those?","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c678153993fc1709929329ccd3633f0c3a2a4952","unresolved":true,"context_lines":[{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"},{"line_number":62,"context_line":"devices that use a UUID as the primary identifier, we should be able"},{"line_number":63,"context_line":"to embed that within the alias."},{"line_number":64,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"638470c1_b639493f","line":61,"range":{"start_line":60,"start_character":32,"end_line":61,"end_character":48},"updated":"2023-09-29 15:47:24.000000000","message":"The docs says \"The identifier has to have \"ua-\" prefix and must be unique within the domain. Additionally, the identifier must consist only of the following characters: [a-zA-Z0-9_-]. Since 3.9.0\"\n\nTechnically, we wouldn\u0027t track the \u0027object uuid\u0027 (say the volume or port) but rather a defined pattern which is \u0027ua-%\u0027 % uuid.\n\nJust mention this in the spec, as you already gave an example below correctly.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"0a28e25ce898525044225af66869adbd4165ca80","unresolved":true,"context_lines":[{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"},{"line_number":62,"context_line":"devices that use a UUID as the primary identifier, we should be able"},{"line_number":63,"context_line":"to embed that within the alias."},{"line_number":64,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"97eb5850_c0664e7e","line":61,"range":{"start_line":60,"start_character":32,"end_line":61,"end_character":48},"in_reply_to":"504d51b4_b920b839","updated":"2023-10-02 14:05:44.000000000","message":"Well, meh, you\u0027re right, it\u0027s an implementation question... unless we have questions about what could happen if a new libvirt release would modify the prefix.\n\nThat said, we could just verify this when checking the existing devices and returning an exception saying \"heh, it looks to me that libvirt changed it\"","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"06cd026a25c9a4cf101f2ede1118b1004b52a5fa","unresolved":true,"context_lines":[{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"},{"line_number":62,"context_line":"devices that use a UUID as the primary identifier, we should be able"},{"line_number":63,"context_line":"to embed that within the alias."},{"line_number":64,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"504d51b4_b920b839","line":61,"range":{"start_line":60,"start_character":32,"end_line":61,"end_character":48},"in_reply_to":"638470c1_b639493f","updated":"2023-09-29 16:34:09.000000000","message":"Yeah, as noted in the previous revision comments (see below), I think this is an implementation detail, and could change over time in libvirt, so I just don\u0027t think it\u0027s necessary to replicate the current set of libvirt restrictions here. Do you think that\u0027s necessary?","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"19b7619f08aa8a0fba453ee644dee79e052d191b","unresolved":false,"context_lines":[{"line_number":57,"context_line":"detach-able. For devices like volumes and network interfaces, the"},{"line_number":58,"context_line":"volume or port UUID should be used.  For other devices, some other"},{"line_number":59,"context_line":"stable identifier that correlates to something in Nova or another"},{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"},{"line_number":62,"context_line":"devices that use a UUID as the primary identifier, we should be able"},{"line_number":63,"context_line":"to embed that within the alias."},{"line_number":64,"context_line":""}],"source_content_type":"text/x-rst","patch_set":3,"id":"ed57076b_b432f748","line":61,"range":{"start_line":60,"start_character":32,"end_line":61,"end_character":48},"in_reply_to":"97eb5850_c0664e7e","updated":"2023-11-07 14:25:54.000000000","message":"dan\u0027s point wfm.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"761955c7e0cabb546d36b38da799f85f44d77f61","unresolved":false,"context_lines":[{"line_number":60,"context_line":"service\u0027s database is required. Libvirt has specific requirements for"},{"line_number":61,"context_line":"the format of the alias, which must be followed. However, for most"},{"line_number":62,"context_line":"devices that use a UUID as the primary identifier, we should be able"},{"line_number":63,"context_line":"to embed that within the alias."},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"This is what the above disk example would look like with a"},{"line_number":66,"context_line":"nova-specified alias::"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7a320bd2_4e904544","line":63,"updated":"2023-11-08 11:41:48.000000000","message":"+1 i think we shoudl generall prefer usign the uuids when we have them.\nthat makes it easy to corralate and its useful for debugging too\n\ni.e. its very easy to look at the xml and say or \"it this port\" if we embed the port uuid as the alias.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c678153993fc1709929329ccd3633f0c3a2a4952","unresolved":true,"context_lines":[{"line_number":130,"context_line":"Developer impact"},{"line_number":131,"context_line":"----------------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"The libvirt driver will ultimately be simpler after this change."},{"line_number":134,"context_line":""},{"line_number":135,"context_line":"Upgrade impact"},{"line_number":136,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"7f925384_8250bb69","line":133,"updated":"2023-09-29 15:47:24.000000000","message":"given the work items below, you don\u0027t plan to modify each single device we already manage thru libvirt, right? (including mdevs).\nIf so, I\u0027d suggess you to write something here saying that developers would be incentived to use aliases for other resources not planned in that spec.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"a81f6b0ed7e158c9d31a57448d8f3422189c7e24","unresolved":false,"context_lines":[{"line_number":130,"context_line":"Developer impact"},{"line_number":131,"context_line":"----------------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"The libvirt driver will ultimately be simpler after this change."},{"line_number":134,"context_line":""},{"line_number":135,"context_line":"Upgrade impact"},{"line_number":136,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"e8475246_a3b92ea7","line":133,"in_reply_to":"477b92b2_cf031efb","updated":"2023-10-06 14:08:37.000000000","message":"OK, after some digging, I think this is indeed only related to volumes and ports (including VDPA ones). I don\u0027t see any other resources that require such change.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"06cd026a25c9a4cf101f2ede1118b1004b52a5fa","unresolved":true,"context_lines":[{"line_number":130,"context_line":"Developer impact"},{"line_number":131,"context_line":"----------------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"The libvirt driver will ultimately be simpler after this change."},{"line_number":134,"context_line":""},{"line_number":135,"context_line":"Upgrade impact"},{"line_number":136,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"c88c8ee5_ea8f5392","line":133,"in_reply_to":"7f925384_8250bb69","updated":"2023-09-29 16:34:09.000000000","message":"I think we should do it for any devices we support attach/detach for.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"0a28e25ce898525044225af66869adbd4165ca80","unresolved":true,"context_lines":[{"line_number":130,"context_line":"Developer impact"},{"line_number":131,"context_line":"----------------"},{"line_number":132,"context_line":""},{"line_number":133,"context_line":"The libvirt driver will ultimately be simpler after this change."},{"line_number":134,"context_line":""},{"line_number":135,"context_line":"Upgrade impact"},{"line_number":136,"context_line":"--------------"}],"source_content_type":"text/x-rst","patch_set":3,"id":"477b92b2_cf031efb","line":133,"in_reply_to":"c88c8ee5_ea8f5392","updated":"2023-10-02 14:05:44.000000000","message":"Hah ! if you want to do so, then we need to verify about other devices that we already support, and not only mdevs.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":16207,"name":"ribaudr","display_name":"uggla","email":"rene.ribaud@gmail.com","username":"uggla","status":"Red Hat"},"change_message_id":"9d0f878690ac498a16835e022fb9474f0bfe6db5","unresolved":true,"context_lines":[{"line_number":137,"context_line":""},{"line_number":138,"context_line":"The only upgrade impact comes from migrating existing instance XML documents"},{"line_number":139,"context_line":"to specify the device alias. Because we may be migrating instances to/from"},{"line_number":140,"context_line":"older nodes, we should retain compatibility with alias-less XMLs for some time"},{"line_number":141,"context_line":"to come."},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Implementation"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"0a1d5447_cc779571","line":141,"range":{"start_line":140,"start_character":65,"end_line":141,"end_character":7},"updated":"2023-10-03 12:48:14.000000000","message":"Discussing this point with Artom as well. I wonder how we could get rid of the compatibility code used by alias-less XMLs.\nOf course if we are waiting for a long time before removing compatibility, there are few chances we will break instances, but that\u0027s not 100% sure.\nOne idea on the top of my head:\n- we could add a check before upgrading to ensure all instances are not using old style XML without aliases.\nHowever I\u0027m not sure that\u0027s a good idea as the check will have to verify all instances, on all computes that could be a pain on large clusters.\n\n\nSo I think that\u0027s a point that needs to be discussed and maybe put some timelines for deprecation.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"761955c7e0cabb546d36b38da799f85f44d77f61","unresolved":false,"context_lines":[{"line_number":137,"context_line":""},{"line_number":138,"context_line":"The only upgrade impact comes from migrating existing instance XML documents"},{"line_number":139,"context_line":"to specify the device alias. Because we may be migrating instances to/from"},{"line_number":140,"context_line":"older nodes, we should retain compatibility with alias-less XMLs for some time"},{"line_number":141,"context_line":"to come."},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Implementation"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"375bca19_88314168","line":141,"range":{"start_line":140,"start_character":65,"end_line":141,"end_character":7},"in_reply_to":"083f9d48_a30076af","updated":"2023-11-08 11:41:48.000000000","message":"so we do need the non aliased version to be supproted for at least one release and likely 1 slurp at a minim.\n\nwehn we do live migration we shoudl end up rewriting the non alias form into the alias form\n\nand any hard reboot, cold migration , unshleve or evacuate willl also do it\n\nso  the old way is needed to support for vms that are not recreated acrros upgrades for a long period of time.\n\nwe likely shoudl do thils like we do for os-vif\n\nwe should have a common function to try  and find or detach a device by\nthe alias and return if that succeed otherwise fall back to the existing logic\n\nthen we can remove the fallback at some future date.\n\nso maybe in 2025.2 or 2026.1\nin general i think we are aligned here so ill mark this as resolved.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"50b3567057a2e84ef1c91ae8797240d7148377b2","unresolved":true,"context_lines":[{"line_number":137,"context_line":""},{"line_number":138,"context_line":"The only upgrade impact comes from migrating existing instance XML documents"},{"line_number":139,"context_line":"to specify the device alias. Because we may be migrating instances to/from"},{"line_number":140,"context_line":"older nodes, we should retain compatibility with alias-less XMLs for some time"},{"line_number":141,"context_line":"to come."},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Implementation"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"dcd2d333_684e460d","line":141,"range":{"start_line":140,"start_character":65,"end_line":141,"end_character":7},"in_reply_to":"0a1d5447_cc779571","updated":"2023-10-04 16:10:13.000000000","message":"\u003e Discussing this point with Artom as well. I wonder how we could get rid of the compatibility code used by alias-less XMLs.\n\nI don\u0027t think we need to prioritize removing support for non-alias-based operations any time soon, which is why I said what I did here in the spec. We can default to using the alias if we find it, else continue to fall back.\n\n\u003e Of course if we are waiting for a long time before removing \ncompatibility, there are few chances we will break instances, but that\u0027s not 100% sure.\n\nI\u0027m not sure what you mean by this.\n\n\u003e One idea on the top of my head:\n\u003e - we could add a check before upgrading to ensure all instances are not using old style XML without aliases.\n\nYes, I have that as the last work item for this effort. See L170.\n\n\u003e However I\u0027m not sure that\u0027s a good idea as the check will have to verify all instances, on all computes that could be a pain on large clusters.\n\nIt would be per-host during startup when we run `driver.init_host()` like we do for other upgrade-related hygiene stuff. It should scale horizontally and not be a problem even for large clusters.\n \n\u003e So I think that\u0027s a point that needs to be discussed and maybe put some timelines for deprecation.\n\nI don\u0027t think we need to address it right away. We should use aliases when we can, see about migrating instances on startup (per the work items) and we can discuss when we think it\u0027s appropriate to drop the old code later. I don\u0027t think we need to commit to any specific timeline right now.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"54ca2e7fffc467d270eee3bbe72abdab4813fb96","unresolved":false,"context_lines":[{"line_number":137,"context_line":""},{"line_number":138,"context_line":"The only upgrade impact comes from migrating existing instance XML documents"},{"line_number":139,"context_line":"to specify the device alias. Because we may be migrating instances to/from"},{"line_number":140,"context_line":"older nodes, we should retain compatibility with alias-less XMLs for some time"},{"line_number":141,"context_line":"to come."},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Implementation"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"c5d7ab45_96f3fe23","line":141,"range":{"start_line":140,"start_character":65,"end_line":141,"end_character":7},"in_reply_to":"375bca19_88314168","updated":"2023-11-08 14:39:50.000000000","message":"The thing is, some clouds don\u0027t move their instances around during every upgrade, so if you pick the longest OS lifespan they could survive without a reboot, you could have some *very* long-lived instanced that never get their XML rewritten. Since this is not expensive to keep, that\u0027s why I think it\u0027s worth waiting a good long time to purge.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"19b7619f08aa8a0fba453ee644dee79e052d191b","unresolved":true,"context_lines":[{"line_number":137,"context_line":""},{"line_number":138,"context_line":"The only upgrade impact comes from migrating existing instance XML documents"},{"line_number":139,"context_line":"to specify the device alias. Because we may be migrating instances to/from"},{"line_number":140,"context_line":"older nodes, we should retain compatibility with alias-less XMLs for some time"},{"line_number":141,"context_line":"to come."},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Implementation"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"dd0441ee_47a4a82b","line":141,"range":{"start_line":140,"start_character":65,"end_line":141,"end_character":7},"in_reply_to":"dcd2d333_684e460d","updated":"2023-11-07 14:25:54.000000000","message":"I personnally vote in favor of a statu-quo period (supporting non-aliased devices) for a certain time being (say 2 releases)","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":4393,"name":"Dan Smith","email":"dms@danplanet.com","username":"danms"},"change_message_id":"d34fc0032273c9cf796a8b1a4dcb2c45eacfdb89","unresolved":true,"context_lines":[{"line_number":137,"context_line":""},{"line_number":138,"context_line":"The only upgrade impact comes from migrating existing instance XML documents"},{"line_number":139,"context_line":"to specify the device alias. Because we may be migrating instances to/from"},{"line_number":140,"context_line":"older nodes, we should retain compatibility with alias-less XMLs for some time"},{"line_number":141,"context_line":"to come."},{"line_number":142,"context_line":""},{"line_number":143,"context_line":"Implementation"},{"line_number":144,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"083f9d48_a30076af","line":141,"range":{"start_line":140,"start_character":65,"end_line":141,"end_character":7},"in_reply_to":"dd0441ee_47a4a82b","updated":"2023-11-07 14:41:22.000000000","message":"It needs to be longer than that I suspect, and because of the way the we actually do detach, there\u0027s really very little to remove of the \"old way\" such that I\u0027m not sure it\u0027s a big rush to remove it.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"19b7619f08aa8a0fba453ee644dee79e052d191b","unresolved":false,"context_lines":[{"line_number":164,"context_line":"- Enable setting and parsing the device alias on disk, interface, and pci"},{"line_number":165,"context_line":"  devices"},{"line_number":166,"context_line":"- Actually set those device aliases in the various parts of the driver that"},{"line_number":167,"context_line":"  create those configs"},{"line_number":168,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"},{"line_number":169,"context_line":"  alias and fall back to the old way"},{"line_number":170,"context_line":"- Migrate existing instance XMLs on startup when device aliases are missing"}],"source_content_type":"text/x-rst","patch_set":3,"id":"33af2128_39b03b2c","line":167,"updated":"2023-11-07 14:25:54.000000000","message":"good luck :-)","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"c678153993fc1709929329ccd3633f0c3a2a4952","unresolved":true,"context_lines":[{"line_number":166,"context_line":"- Actually set those device aliases in the various parts of the driver that"},{"line_number":167,"context_line":"  create those configs"},{"line_number":168,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"},{"line_number":169,"context_line":"  alias and fall back to the old way"},{"line_number":170,"context_line":"- Migrate existing instance XMLs on startup when device aliases are missing"},{"line_number":171,"context_line":""},{"line_number":172,"context_line":"Dependencies"}],"source_content_type":"text/x-rst","patch_set":3,"id":"0207f6bd_25ba9c93","line":169,"updated":"2023-09-29 15:47:24.000000000","message":"++ that\u0027s the key thing in this spec, that\u0027s why I think we should only focus this spec on devices that we query thru their existing identifiers.","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"a81f6b0ed7e158c9d31a57448d8f3422189c7e24","unresolved":false,"context_lines":[{"line_number":166,"context_line":"- Actually set those device aliases in the various parts of the driver that"},{"line_number":167,"context_line":"  create those configs"},{"line_number":168,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"},{"line_number":169,"context_line":"  alias and fall back to the old way"},{"line_number":170,"context_line":"- Migrate existing instance XMLs on startup when device aliases are missing"},{"line_number":171,"context_line":""},{"line_number":172,"context_line":"Dependencies"}],"source_content_type":"text/x-rst","patch_set":3,"id":"161d267a_f9447b05","line":169,"in_reply_to":"0207f6bd_25ba9c93","updated":"2023-10-06 14:08:37.000000000","message":"Done","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"761955c7e0cabb546d36b38da799f85f44d77f61","unresolved":false,"context_lines":[{"line_number":166,"context_line":"- Actually set those device aliases in the various parts of the driver that"},{"line_number":167,"context_line":"  create those configs"},{"line_number":168,"context_line":"- Make the code that looks up devices by device-specific identifiers prefer the"},{"line_number":169,"context_line":"  alias and fall back to the old way"},{"line_number":170,"context_line":"- Migrate existing instance XMLs on startup when device aliases are missing"},{"line_number":171,"context_line":""},{"line_number":172,"context_line":"Dependencies"}],"source_content_type":"text/x-rst","patch_set":3,"id":"1a157711_7ae73f47","line":169,"in_reply_to":"161d267a_f9447b05","updated":"2023-11-08 11:41:48.000000000","message":"+1 this is inline with the converstaion we where having regarding upgrades","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"},{"author":{"_account_id":7166,"name":"Sylvain Bauza","email":"sbauza@redhat.com","username":"sbauza"},"change_message_id":"19b7619f08aa8a0fba453ee644dee79e052d191b","unresolved":false,"context_lines":[{"line_number":181,"context_line":"and functional coverage that will be added. Potentially enabling (and using)"},{"line_number":182,"context_line":"the LUN passthrough attachment mechanism would be beneficial, but that is"},{"line_number":183,"context_line":"somewhat beyond the scope of this effort which is just changing the enumeration"},{"line_number":184,"context_line":"behavior."},{"line_number":185,"context_line":""},{"line_number":186,"context_line":"Documentation Impact"},{"line_number":187,"context_line":"\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d\u003d"}],"source_content_type":"text/x-rst","patch_set":3,"id":"f9cdf2e0_c2406f46","line":184,"updated":"2023-11-07 14:25:54.000000000","message":"++","commit_id":"33c1b141fe181086b30d80294e296d588f66465b"}]}