vcpu_pin_set ignored in rocky on single socket

asked 2019-08-05 12:19:05 -0500

Brendan Kelly gravatar image

Hello, I am running devstack rocky. I am seeing problems with vcpu_pin_set not working.

I have single socket system with 8 cores and I have set the vcpu_pin_set to 3-7. I have hw:cpu_policy set to dedicated on the flavor. I consistently see the nova schedular try and pack the instance onto sibling cores. Also the available pCPU list shows 0-7 cores in the nova logs.

I have used vcpu pinning in the past on OSP 12 in a host NUMA environment.

cpuset in logs shows all cores 0-7 as available.

bkelly@devstack1:~$ lscpu

Architecture: x86_64
CPU op-mode(s): 32-bit, 64-bit
Byte Order: Little Endiant
CPU(s): 8t
On-line CPU(s) list: 0-7t
Thread(s) per core: 1t
Core(s) per socket: 8t
Socket(s): 1t
NUMA node(s): 1t
Vendor ID: GenuineIntelt
CPU family: 6t
Model: 86t
Model name: Intel(R) Xeon(R) CPU D-1548 @ 2.00GHzt
Stepping: 3t
CPU MHz: 1995.603t
CPU max MHz: 2000.0000t
CPU min MHz: 800.0000t
BogoMIPS: 3991.17t
Virtualization: VT-xt
L1d cache: 32Kt
L1i cache: 32Kt
L2 cache: 256K L3 cache: 12288K NUMA node0 CPU(s): 0-7t
Flags: fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfmperf pni pclmulqdq dtes64 monitor ds_cpl vmx smx est tm2 ssse3 sdbg fma cx16 xtpr pdcm pcid dca sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm abm 3dnowprefetch cpuid_fault epb cat_l3 cdp_l3 invpcid_single pti intel_ppin ssbd ibrs ibpb stibp tpr_shadow vnmi flexpriority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm cqm rdt_a rdseed adx smap intel_pt xsaveopt cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local dtherm arat pln pts flush_l1d

bkelly@devstack1:~$ cat /etc/nova/nova.conf

[DEFAULT] reserved_huge_pages = node:0,size:2048,count:64 vif_plugging_timeout = 300 vif_plugging_is_fatal = True use_neutron = True firewall_driver = nova.virt.firewall.NoopFirewallDriver compute_driver = libvirt.LibvirtDriver default_ephemeral_format = ext4 pointer_model = ps2mouse graceful_shutdown_timeout = 5 metadata_workers = 2 osapi_compute_workers = 2 transport_url = rabbit://stackrabbit:stackqueue@ logging_exception_prefix = ERROR %(name)s %(instance)s logging_default_format_string = %(color)s%(levelname)s %(name)s [-%(color)s] %(instance)s%(color)s%(message)s logging_context_format_string = %(color)s%(levelname)s %(name)s [%(global_request_id)s %(request_id)s %(project_name)s %(user_name)s%(color)s] %(instance)s%(color)s%(message)s logging_debug_format_suffix = {{(pid=%(process)d) %(funcName)s %(pathname)s:%(lineno)d}} instances_path = /opt/stack/data/nova/instances state_path = /opt/stack/data/nova enabled_apis = osapi_compute metadata_listen = osapi_compute_listen = instance_name_template = instance-%08x my_ip = default_floating_pool = public rootwrap_config = /etc/nova/rootwrap.conf allow_resize_to_same_host = True debug = True vcpu_pin_set="3,4,5,6,7" log_dir=/var/log/nova

[wsgi] api_paste_config = /etc/nova/api-paste.ini

[scheduler] workers = 2 driver = filter_scheduler

[filter_scheduler] track_instance_changes = False


==> nova-scheduler.log <== DEBUG nova.scheduler.filter_scheduler [None req-d7a5e837-8eda-4458-b061-cf11d5cdb8af admin admin] [instance: 9543d511-6d29-4cf4-861e-8d4f36850061] Selected host: (devstack1, devstack1 ... (more)

edit retag flag offensive close merge delete

1 answer

Sort by ยป oldest newest most voted

answered 2019-08-11 15:59:39 -0500

Brendan Kelly gravatar image

Hello, I did not realize devstack uses a file "nova-cpu.conf". I only had made changes to nova.conf. It is now working.

edit flag offensive delete link more

Get to know Ask OpenStack

Resources for moderators

Question Tools

1 follower


Asked: 2019-08-05 12:19:05 -0500

Seen: 46 times

Last updated: Aug 05 '19