Comment 0 for bug 1999814

Hello,

This is very similar to pad.lv/1852437 (and the related blueprint at https://blueprints.launchpad.net/nova/+spec/allow-disabling-cpu-flags), but there is a very different and important nuance.

A customer I'm working with has two classes of blades that they're trying to use. Their existing ones are Cascade Lake-based; they are presently using the Cascadelake-Server-noTSX CPU model via libvirt.cpu_model in nova.conf. Their new blades are Ice Lake-based, which is a newer processor, which typically would also be able to run based on the Cascade Lake feature set - except that these Ice Lake processors lack the MPX feature defined in the Cascadelake-Server-noTSX model.

The result of this is evident when I try to start nova on the new blades with the Ice Lake CPUs. Even if I specify the following in my nova.conf:

[libvirt]
cpu_mode = custom
cpu_model = Cascadelake-Server-noTSX
cpu_model_extra_flags = -mpx

That is not enough to allow Nova to start; it fails in the libvirt driver in the _check_cpu_compatibility function:

2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service Traceback (most recent call last):
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service File "/usr/lib/python3/dist-packages/nova/virt/libvirt/driver.py", line 771, in _check_cpu_compatibility
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service self._compare_cpu(cpu, self._get_cpu_info(), None)
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service File "/usr/lib/python3/dist-packages/nova/virt/libvirt/driver.py", line 8817, in _compare_cpu
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service raise exception.InvalidCPUInfo(reason=m % {'ret': ret, 'u': u})
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service nova.exception.InvalidCPUInfo: Unacceptable CPU info: CPU doesn't have compatibility.
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service 0
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service Refer to http://libvirt.org/html/libvirt-libvirt-host.html#virCPUCompareResult
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service During handling of the above exception, another exception occurred:
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service Traceback (most recent call last):
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service File "/usr/lib/python3/dist-packages/oslo_service/service.py", line 810, in run_service
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service service.start()
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service File "/usr/lib/python3/dist-packages/nova/service.py", line 173, in start
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service self.manager.init_host()
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service File "/usr/lib/python3/dist-packages/nova/compute/manager.py", line 1404, in init_host
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service self.driver.init_host(host=self.host)
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service File "/usr/lib/python3/dist-packages/nova/virt/libvirt/driver.py", line 743, in init_host
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service self._check_cpu_compatibility()
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service File "/usr/lib/python3/dist-packages/nova/virt/libvirt/driver.py", line 777, in _check_cpu_compatibility
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service raise exception.InvalidCPUInfo(msg)
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service nova.exception.InvalidCPUInfo: Configured CPU model: Cascadelake-Server-noTSX is not compatible with host CPU. Please correct your config and try again. Unacceptable CPU info: CPU doesn't have compatibility.
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service 0
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service Refer to http://libvirt.org/html/libvirt-libvirt-host.html#virCPUCompareResult
2022-12-15 17:20:59.562 1836708 ERROR oslo_service.service

If I make a custom libvirt CPU map file which removes the "<feature name='mpx'/>" feature and specify that as the cpu_model instead, I am able to make Nova start - so it does indeed seem to specifically be that single feature which is blocking me. However, editing the libvirt CPU mapping files is probably not the right way to fix this - hence why I'm filing this bug, for discussion of how to support cases like this.

Currently the only "proper" way I'm aware of to work around this right now is to fall back to a Broadwell-based configuration which lacks the "mpx" feature to use as a common baseline, but that's a much older configuration than Cascade Lake and would mean missing out on all the other features which are common in both Cascade Lake and Ice Lake. I would rather if there were a way to use the Cascade Lake settings but simply remove that "mpx" feature from use.

----

Steps to reproduce
==================

On an Ice Lake system lacking the MPX feature (e.g. /proc/cpuinfo reporting model of "Intel(R) Xeon(R) Gold 5318Y"), specify the following settings in nova.conf in libvirt settings:

[libvirt]
cpu_mode = custom
cpu_model = Cascadelake-Server-noTSX
cpu_model_extra_flags = -mpx

Then try to start nova.

Expected result
===============

Nova should start since Cascadelake-Server-noTSX is a subset of Icelake-Server-noTSX, thus allowing the use of Cascadelake-Server-noTSX as a common baseline model for both Cascade Lake and Ice Lake servers.

Actual result
=============

Nova refuses to start, claiming the specified CPU model is incompatible. The "cpu_model_extra_flags = -mpx" config option does not help.

Environment
===========

Nova/OpenStack version: OpenStack Ussuri running on Ubuntu Focal. Specifically, nova packages are at version 2:21.2.4-0ubuntu2.

Hypervisor: libvirt + KVM

Other relevant notes
====================

There are some other open related bugs. The removal of the MPX feature in some Ice Lake processors has manifested in other ways as well. These bugs are primarily in regards to the missing MPX feature breaking how Ice Lake processors are detected, so the nuance is somewhat different - however, they may be worth reviewing as well.

* https://gitlab.com/libvirt/libvirt/-/issues/304: bug regarding the Icelake CPU maps in libvirt not working to detect certain Ice Lakes, instead detecting them as Broadwell-noTSX-IBRS according to "virsh capabilities" due to lacking the MPX feature. (I've personally tested that removing the mpx feature from the associated CPU mapping files allows for detecting as Ice Lake, but that's not the correct way to fix this.)

There is also an interesting comment on this bug at https://gitlab.com/libvirt/libvirt/-/issues/304#note_1065798706. It basically implies that rather than looking at "virsh capabilities", "virsh domcapabilities" should be used instead as it seems to more correctly identify the CPU model even if there are disabled flags like MPX.

* https://bugs.launchpad.net/ubuntu/+source/libvirt/+bug/1978064: Launchpad-side bug regarding the above issue as encountered in Ubuntu.