| *** mhen_ is now known as mhen | 01:54 | |
| aravindh5murugesan | Hello folks. I'm running into a problem when I set CPU models to custom and the specifying the models. The default config that we receive from our vendor sets the cpu-model to broadwell-notsx-IBS something. But our CPU model is saphirerapids and the oldest we need to support is Icelake. But When I set it to Icelake or SaphireRapids, I used the same name as listed in virsh cpu-models, but nova complains that CPU | 04:24 |
|---|---|---|
| aravindh5murugesan | can assist me here. 2025-07-17 20:44:36.072 ERROR oslo_service.service [req-fa1ce5c0-d687-4d09-ac10-7ca05eb98461 None None] Error starting thread.: nova.exception.InvalidCPUInfo: Configured CPU model: SapphireRapids-noTSX and CPU Flags [] ar not compatible with host CPU. Please correct your config and try again. Unacceptable CPU info: CPU doesn't have compatibility | 04:27 |
| aravindh5murugesan | CPU model in hypervisor - Vendor ID: GenuineIntel Model name: Intel(R) Xeon(R) Silver 4416+ | 04:28 |
| sean-k-mooney | aravindh5murugesan the problem you are having is that icelake has feature flags like tsk tha tintel disabled in microcode and does tno supprot on saphire rapids | 09:09 |
| sean-k-mooney | *like the flags for tsx that intel ... | 09:11 |
| opendevreview | Takashi Kajinami proposed openstack/nova master: Drop unused parameter type validation method https://review.opendev.org/c/openstack/nova/+/909189 | 10:53 |
| aravindh0murugesan | sean-k-mooney: In that case, would it possible to disable those unsupported flags and somehow make this work? | 12:03 |
| sean-k-mooney | yes there are 2 ways, libvirt provide some custom models for older cpu i.e. the *-notsx-IBS variants | 12:20 |
| sean-k-mooney | but nova also has a config option that allows you to add/remove cpu flags | 12:20 |
| aravindh0murugesan | I compared the icelake.xml and my lscpu flags and found these flags are missing in my host cpu - avx512vnni,mpx,pclmuldq,spec-ctrl,sse4.1,sse4.2,tsc-deadline, | 12:37 |
| aravindh0murugesan | so I just do a - on these flags in nova configs? | 12:37 |
| aravindh0murugesan | cpu_models = Cascadelake-Server-noTSX cpu-model-extra-flags = -avx512vnni,-mpx,-pclmuldq,-spec-ctrl,-sse4.1,-sse4.2,-tsc-deadline this does not work either | 12:38 |
| aravindh0murugesan | sean-k-mooney: Editing the XML to say removed='yes' kinda worked. But I'm not sure if this is the suggested way to do things. | 12:44 |
| aravindh0murugesan | My bad, changing this cpu-model-extra-flags to snake case fixed it. Thanks. | 12:58 |
| sean-k-mooney | ack so you were able to get this wokring | 13:23 |
| sean-k-mooney | and yes https://docs.openstack.org/nova/latest/configuration/config.html#libvirt.cpu_model_extra_flags is the cofnig opiton | 13:25 |
| sean-k-mooney | by the wya i have alwasy know cpu_model_extra_flags as SLUG case and cpu-model-extra-flags as snake case | 13:26 |
| opendevreview | Wesley Hershberger proposed openstack/nova stable/2024.2: Use dict object for request_specs_dict in the _list_view https://review.opendev.org/c/openstack/nova/+/955310 | 13:28 |
| opendevreview | Wesley Hershberger proposed openstack/nova stable/2024.1: Add ServersViewBuilderTestV296 unit test class https://review.opendev.org/c/openstack/nova/+/955304 | 13:28 |
| opendevreview | Wesley Hershberger proposed openstack/nova stable/2024.1: Use dict object for request_specs_dict in the _list_view https://review.opendev.org/c/openstack/nova/+/955305 | 13:28 |
| dansmith | gibi: so, with my patch to actually fix s-g, I'm hitting the db poison fixture on NoDBTestCase tests.. I was thinking that I must be working around one of the mocks for that, but I haven't found it yet | 14:31 |
| dansmith | I'm wondering if maybe before we were scheduling things to threads but not actually running them and these tests are actually missing db mocks that are now getting tickled | 14:32 |
| dansmith | have you run into that anywhere else? | 14:32 |
| gibi | I havent seen this type of issues yet with tasks using the futurists executors especially not in legacy mode spawning GreenThreads. | 15:08 |
| gibi | I can take a look at the actual failure on Monday | 15:09 |
| dansmith | gibi: wait, I think I'm just stupid | 15:10 |
| dansmith | I was raising the exception the task raised in my result() but I think our contract with the sg caller is to return the Exception if it raised | 15:10 |
| gibi | yes I think scatter gather expects the exception to be returned | 15:11 |
| gibi | https://github.com/openstack/nova/blob/770a6f1eabc6cfb77d3ac5afd91e6e8ce9f8b9d9/nova/context.py#L425 | 15:11 |
| gibi | https://github.com/openstack/nova/blob/770a6f1eabc6cfb77d3ac5afd91e6e8ce9f8b9d9/nova/context.py#L463 | 15:12 |
| gibi | so yes the exception caught on the task thread and returned az the result value of the task | 15:12 |
| dansmith | yup | 15:14 |
| opendevreview | Merged openstack/nova master: Make nova.utils.pass_context private https://review.opendev.org/c/openstack/nova/+/948188 | 15:23 |
| opendevreview | Dan Smith proposed openstack/nova master: Remove eventlet timer from multi_cell_list https://review.opendev.org/c/openstack/nova/+/954990 | 15:29 |
| opendevreview | Dan Smith proposed openstack/nova master: WIP: Parallelize s-g generators https://review.opendev.org/c/openstack/nova/+/955091 | 15:29 |
| dansmith | that ^ should fix the unit/fixture failure on the eventlet timer patch and get the WIP much closer to passing.. more work to be done there of course, but at least it's not 100% fail | 15:30 |
| *** haleyb is now known as haleyb|out | 18:33 | |
| -opendevstatus- NOTICE: The Gerrit service on review.opendev.org will be offline briefly for a configuration and version update, but should return to service momentarily | 20:07 | |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!