)]}'
{"doc/source/admin/cpu-models.rst":[{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"71583ce6259e216be344cb16b99f365b26105b05","unresolved":true,"context_lines":[{"line_number":31,"context_line":"of ``nova.conf`` define which type of CPU model is exposed to the hypervisor"},{"line_number":32,"context_line":"when using KVM: :oslo.config:option:`libvirt.cpu_mode` and"},{"line_number":33,"context_line":":oslo.config:option:`libvirt.cpu_models`."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":"The :oslo.config:option:`libvirt.cpu_mode` option can take one of the following"},{"line_number":36,"context_line":"values: ``none``, ``host-passthrough``, ``host-model``, and ``custom``."},{"line_number":37,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"293cab85_fe23c4b9","line":34,"updated":"2021-03-30 11:58:32.000000000","message":"i do find it confusing that we start the cpu mode seaction by talking baout cpu models.\n\nthis shoud really be in a preable or otherwise not under this heading.\ni would move everything from lines 23-34 to line 19","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"5827c9497f9c81df82eabb6c03672e47a3305abb","unresolved":false,"context_lines":[{"line_number":31,"context_line":"of ``nova.conf`` define which type of CPU model is exposed to the hypervisor"},{"line_number":32,"context_line":"when using KVM: :oslo.config:option:`libvirt.cpu_mode` and"},{"line_number":33,"context_line":":oslo.config:option:`libvirt.cpu_models`."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":"The :oslo.config:option:`libvirt.cpu_mode` option can take one of the following"},{"line_number":36,"context_line":"values: ``none``, ``host-passthrough``, ``host-model``, and ``custom``."},{"line_number":37,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"7e939483_d725843d","line":34,"in_reply_to":"293cab85_fe23c4b9","updated":"2021-03-31 10:32:13.000000000","message":"Reworked this in a follow-up. Hopefully it makes more sense now","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"35ed4d2d0610fdadf1c1bfa4477c61e6066603a7","unresolved":false,"context_lines":[{"line_number":31,"context_line":"of ``nova.conf`` define which type of CPU model is exposed to the hypervisor"},{"line_number":32,"context_line":"when using KVM: :oslo.config:option:`libvirt.cpu_mode` and"},{"line_number":33,"context_line":":oslo.config:option:`libvirt.cpu_models`."},{"line_number":34,"context_line":""},{"line_number":35,"context_line":"The :oslo.config:option:`libvirt.cpu_mode` option can take one of the following"},{"line_number":36,"context_line":"values: ``none``, ``host-passthrough``, ``host-model``, and ``custom``."},{"line_number":37,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"117635a0_1059dd74","line":34,"in_reply_to":"7e939483_d725843d","updated":"2021-07-20 20:19:22.000000000","message":"this is adressed in https://review.opendev.org/c/openstack/nova/+/784066","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":10135,"name":"Lee Yarwood","display_name":"Lee Yarwood","email":"lyarwood@redhat.com","username":"lyarwood"},"change_message_id":"45ef6e1bdbffd8e2b5109c0d2da4844694a31edb","unresolved":false,"context_lines":[{"line_number":44,"context_line":"~~~~~~~~~~"},{"line_number":45,"context_line":""},{"line_number":46,"context_line":"If :oslo.config:option:`cpu_mode\u003dhost-model \u003clibvirt.cpu_mode\u003e`, the CPU model"},{"line_number":47,"context_line":"in ``/usr/share/libvirt/cpu_map/*.xml`` that most closely matches the host and"},{"line_number":48,"context_line":"requests additional CPU flags to complete the match. This CPU model has a"},{"line_number":49,"context_line":"number of advantages:"},{"line_number":50,"context_line":""},{"line_number":51,"context_line":"* It provides almost all of the host CPU features to the guest, thus providing"}],"source_content_type":"text/x-rst","patch_set":1,"id":"6b5ea7af_9689e3e9","line":48,"range":{"start_line":47,"start_character":40,"end_line":48,"end_character":52},"updated":"2021-07-22 14:01:48.000000000","message":"I think you\u0027re missing an `is used` somewhere in this sentence right?\n\n*later* Oh right FUP, cool.","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"71583ce6259e216be344cb16b99f365b26105b05","unresolved":true,"context_lines":[{"line_number":62,"context_line":"largely identical. However, if your compute nodes span multiple processor"},{"line_number":63,"context_line":"generations, you may be better advised to select a ``custom`` CPU model."},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The ``host-model`` CPU model is the default for the KVM \u0026 QEMU hypervisors"},{"line_number":66,"context_line":"(:oslo.config:option:`libvirt.virt_type`\\ \u003d``kvm``/``qemu``)"},{"line_number":67,"context_line":""},{"line_number":68,"context_line":".. note::"}],"source_content_type":"text/x-rst","patch_set":1,"id":"b3120cae_f74edec2","line":65,"range":{"start_line":65,"start_character":36,"end_line":65,"end_character":43},"updated":"2021-03-30 11:58:32.000000000","message":"this is more or less true in common use but its not technically correct.\n\nthe default value for cpu_mode is None when unset, when using the libvirt virt dirver none get converted to host-model for x86 and most architectures but the behavior is also arch dependent\nhttps://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L4725-L4743\n\nfor aarch64 with the libvirt driver using kvm/qemu we default to host-passthough\n\nif you explicitly set the mode to the sting \"none\" we have a dirrernet behavior\nhttps://github.com/openstack/nova/blob/master/nova/virt/libvirt/driver.py#L4746-L4747\nwhich you have not documented. that should proably be added as a new section.\n\nthat will create a default initalisted guest cpu object which will delegate teh decision to libvirt\n\nhttps://github.com/openstack/nova/blob/16074287c0c60aa5b0a02ba8d03bf78a184a95fb/nova/virt/libvirt/config.py#L904-L906\n\n\nlibvirt default is currently also host-model but if they were ever to change that then cpu_mode\u003d\"none\" would pick up the new libvirt default so its not the same behavior as unset.","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"35ed4d2d0610fdadf1c1bfa4477c61e6066603a7","unresolved":false,"context_lines":[{"line_number":62,"context_line":"largely identical. However, if your compute nodes span multiple processor"},{"line_number":63,"context_line":"generations, you may be better advised to select a ``custom`` CPU model."},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The ``host-model`` CPU model is the default for the KVM \u0026 QEMU hypervisors"},{"line_number":66,"context_line":"(:oslo.config:option:`libvirt.virt_type`\\ \u003d``kvm``/``qemu``)"},{"line_number":67,"context_line":""},{"line_number":68,"context_line":".. note::"}],"source_content_type":"text/x-rst","patch_set":1,"id":"f1e5305c_e2ed3d18","line":65,"range":{"start_line":65,"start_character":36,"end_line":65,"end_character":43},"in_reply_to":"1a38920f_fac7be7f","updated":"2021-07-20 20:19:22.000000000","message":"ok this is adressed here https://review.opendev.org/c/openstack/nova/+/784066/2/doc/source/admin/cpu-models.rst#69","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"5827c9497f9c81df82eabb6c03672e47a3305abb","unresolved":false,"context_lines":[{"line_number":62,"context_line":"largely identical. However, if your compute nodes span multiple processor"},{"line_number":63,"context_line":"generations, you may be better advised to select a ``custom`` CPU model."},{"line_number":64,"context_line":""},{"line_number":65,"context_line":"The ``host-model`` CPU model is the default for the KVM \u0026 QEMU hypervisors"},{"line_number":66,"context_line":"(:oslo.config:option:`libvirt.virt_type`\\ \u003d``kvm``/``qemu``)"},{"line_number":67,"context_line":""},{"line_number":68,"context_line":".. note::"}],"source_content_type":"text/x-rst","patch_set":1,"id":"1a38920f_fac7be7f","line":65,"range":{"start_line":65,"start_character":36,"end_line":65,"end_character":43},"in_reply_to":"b3120cae_f74edec2","updated":"2021-03-31 10:32:13.000000000","message":"Done in the follow-up","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":9708,"name":"Balazs Gibizer","display_name":"gibi","email":"gibizer@gmail.com","username":"gibi"},"change_message_id":"d0abd90d8a2932886a44728ab748735d9b1b5f8d","unresolved":true,"context_lines":[{"line_number":72,"context_line":"   definition is transferred to the destination host as-is. This results in the"},{"line_number":73,"context_line":"   migrated guest on the destination seeing exactly the same CPU model as on"},{"line_number":74,"context_line":"   source even if the destination compute host is capable of providing more CPU"},{"line_number":75,"context_line":"   features. However, shutting down and restarting the guest on the may present"},{"line_number":76,"context_line":"   different hardware to the guest, as per the new capabilities of the"},{"line_number":77,"context_line":"   destination compute."},{"line_number":78,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"ebba428b_a2e7b41a","line":75,"range":{"start_line":75,"start_character":51,"end_line":75,"end_character":71},"updated":"2021-03-30 14:50:35.000000000","message":"I cannot parse this. Does the \"on the\" unnecessary here?","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":15334,"name":"Stephen Finucane","display_name":"stephenfin","email":"stephenfin@redhat.com","username":"sfinucan"},"change_message_id":"5827c9497f9c81df82eabb6c03672e47a3305abb","unresolved":false,"context_lines":[{"line_number":72,"context_line":"   definition is transferred to the destination host as-is. This results in the"},{"line_number":73,"context_line":"   migrated guest on the destination seeing exactly the same CPU model as on"},{"line_number":74,"context_line":"   source even if the destination compute host is capable of providing more CPU"},{"line_number":75,"context_line":"   features. However, shutting down and restarting the guest on the may present"},{"line_number":76,"context_line":"   different hardware to the guest, as per the new capabilities of the"},{"line_number":77,"context_line":"   destination compute."},{"line_number":78,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"c877892b_b0dcff26","line":75,"range":{"start_line":75,"start_character":51,"end_line":75,"end_character":71},"in_reply_to":"5e15c3d9_3e9d7ac4","updated":"2021-03-31 10:32:13.000000000","message":"Done","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"},{"author":{"_account_id":11604,"name":"sean mooney","email":"smooney@redhat.com","username":"sean-k-mooney"},"change_message_id":"f4ade390e00fe1f175972e70cb7ed0c9c21c703a","unresolved":true,"context_lines":[{"line_number":72,"context_line":"   definition is transferred to the destination host as-is. This results in the"},{"line_number":73,"context_line":"   migrated guest on the destination seeing exactly the same CPU model as on"},{"line_number":74,"context_line":"   source even if the destination compute host is capable of providing more CPU"},{"line_number":75,"context_line":"   features. However, shutting down and restarting the guest on the may present"},{"line_number":76,"context_line":"   different hardware to the guest, as per the new capabilities of the"},{"line_number":77,"context_line":"   destination compute."},{"line_number":78,"context_line":""}],"source_content_type":"text/x-rst","patch_set":1,"id":"5e15c3d9_3e9d7ac4","line":75,"range":{"start_line":75,"start_character":51,"end_line":75,"end_character":71},"in_reply_to":"ebba428b_a2e7b41a","updated":"2021-03-30 14:51:45.000000000","message":"on the shoudl be removed yes","commit_id":"be03ca7be793b5f67caed2dffbb73c1dce163497"}]}