Description of problem: Nova placement has wrong information cpus In OSP 16.1, I have created 2 VM with cpu_policy set to dedicated. ~~~ (overcloud) [stack@undercloud-0 ~]$ openstack flavor show mid +----------------------------+-----------------------------------------------------------------------------------------------------------------------------------------+ | Field | Value | +----------------------------+-----------------------------------------------------------------------------------------------------------------------------------------+ | OS-FLV-DISABLED:disabled | False | | OS-FLV-EXT-DATA:ephemeral | 0 | | access_project_ids | None | | description | None | | disk | 20 | | extra_specs | {'hw:cpu_policy': 'dedicated', 'hw:mem_page_size': 'large', 'hw:emulator_threads_policy': 'isolate', 'hw:cpu_thread_policy': 'isolate'} | | id | e1b5c7dc-a441-4cb0-858d-0e64eeec6ee6 | | name | mid | | os-flavor-access:is_public | True | | properties | hw:cpu_policy='dedicated', hw:cpu_thread_policy='isolate', hw:emulator_threads_policy='isolate', hw:mem_page_size='large' | | ram | 2048 | | rxtx_factor | 1.0 | | swap | 0 | | vcpus | 2 | +----------------------------+-----------------------------------------------------------------------------------------------------------------------------------------+ ~~~ Both the VM are running on same ~~~ (overcloud) [stack@undercloud-0 ~]$ openstack server show 2-cpu-1 -c id -c name -c flavor -c OS-EXT-SRV-ATTR:host --fit-width +----------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | Field | Value | +----------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | OS-EXT-SRV-ATTR:host | nfv-compute-sriov-0.localdomain | | flavor | disk='20', ephemeral='0', extra_specs.hw:cpu_policy='dedicated', extra_specs.hw:cpu_thread_policy='isolate', extra_specs.hw:emulator_threads_policy='isolate', | | | extra_specs.hw:mem_page_size='large', original_name='mid', ram='2048', swap='0', vcpus='2' | | id | 059ec06f-c4f4-4ada-966b-409aa261eaa5 | | name | 2-cpu-1 | +----------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ (overcloud) [stack@undercloud-0 ~]$ openstack server show 2-cpu -c id -c name -c flavor -c OS-EXT-SRV-ATTR:host --fit-width +----------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | Field | Value | +----------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ | OS-EXT-SRV-ATTR:host | nfv-compute-sriov-0.localdomain | | flavor | disk='20', ephemeral='0', extra_specs.hw:cpu_policy='dedicated', extra_specs.hw:cpu_thread_policy='isolate', extra_specs.hw:emulator_threads_policy='isolate', | | | extra_specs.hw:mem_page_size='large', original_name='mid', ram='2048', swap='0', vcpus='2' | | id | bef7467d-500e-40b6-b60f-df7bcc32bdce | | name | 2-cpu | +----------------------+------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------+ (overcloud) [stack@undercloud-0 ~]$ ~~~ Both the VM are using cpu policy dedicated, b ut in nova placement usage i see VCPUS been use instead of PCPUS ~~~ (overcloud) [stack@undercloud-0 ~]$ openstack resource provider list +--------------------------------------+---------------------------------+------------+ | uuid | name | generation | +--------------------------------------+---------------------------------+------------+ | 77ed94b4-4e76-4552-a4fa-cbaa59984284 | nfv-compute-sriov-0.localdomain | 17 | | 96da5079-91e0-4857-a1d1-608b912e0df3 | nfv-compute-sriov-1.localdomain | 12 | +--------------------------------------+---------------------------------+------------+ (overcloud) [stack@undercloud-0 ~]$ openstack resource provider inventory list --os-placement-api-version 1.2 77ed94b4-4e76-4552-a4fa-cbaa59984284 +----------------+------------------+----------+----------+----------+-----------+-------+ | resource_class | allocation_ratio | min_unit | max_unit | reserved | step_size | total | +----------------+------------------+----------+----------+----------+-----------+-------+ | VCPU | 16.0 | 1 | 2 | 0 | 1 | 2 | | MEMORY_MB | 1.0 | 1 | 63776 | 4096 | 1 | 63776 | | DISK_GB | 1.0 | 1 | 558 | 0 | 1 | 558 | | PCPU | 1.0 | 1 | 29 | 0 | 1 | 29 | +----------------+------------------+----------+----------+----------+-----------+-------+ (overcloud) [stack@undercloud-0 ~]$ openstack resource provider usage show 77ed94b4-4e76-4552-a4fa-cbaa59984284 +----------------+-------+ | resource_class | usage | +----------------+-------+ | VCPU | 4 | <<<<< | MEMORY_MB | 4096 | | DISK_GB | 40 | | PCPU | 0 | +----------------+-------+ ~~~ Version-Release number of selected component (if applicable): Red Hat OpenStack Platform release 16.1.0 GA (Train) How reproducible: Steps to Reproduce: 1. 2. 3. Actual results: Expected results: I expected PCPU to be used instead of vcpu. Additional info:
Thanks for the report. This is a known issue and is targeted for 16.1.2. I'm closing this as a duplicate. *** This bug has been marked as a duplicate of bug 1862396 ***