Monday, 2026-03-02

*** rosmaita1 is now known as rosmaita13:20
*** vhari_ is now known as vhari14:14
sean-k-mooneydo we have a better solution then rejecting a host in a pre playbook to ensure a job runs on a vm with 8vcpus16:35
opendevreviewMerged openstack/project-config master: Stop publishing collection to PyPi  https://review.opendev.org/c/openstack/project-config/+/97808416:35
sean-k-mooneynova and cyborg have feature that reallly do need more then 8vcpu to be able to test16:35
sean-k-mooneyi think the nested virt lables happen to be 8vpcu by aciddnet but that not actully the reqiurement16:36
clarkbsean-k-mooney: we are already rejecting nodes with only 1 vcpu16:37
sean-k-mooneyright but the job can fail with the retry limit16:37
sean-k-mooneyso that not really relibale16:37
clarkbthat is the xen bug which can be worked around with a kernel boot parameters but it will impact every context switch the cpus do for all jobs on that image (ubuntu noble)16:38
clarkbthen the new rax flex region instances only have 4 vcpus but that has been deemed fine as they are zoomy16:38
sean-k-mooneyright but that not actully fine for some nova feature16:38
sean-k-mooneylike cpu pinning16:38
sean-k-mooneyor some feature we want to test in cybrog16:38
fungithe nested virt label avoids rackspace legacy, and therefore the xen bug, but yeah doesn't avoid 4cpu nodes in flex16:38
clarkbright the only way to properly fix that is to either accept a performance decrease across the board on that image with nox2apic. Stop using the cloud region entirely (its like half our quota). Or stop using newer kernels16:39
sean-k-mooneythat will use kernel moduel where we we need ot allcoate a core to the kernel module16:39
sean-k-mooneyclarkb: so the usecause im talkign about is unrelated to the xen issue16:39
clarkbwe;ve never promised 8 vcpus https://docs.opendev.org/opendev/infra-manual/latest/testing.html we call out different cpu copunts and an explicit difference between nodes16:40
sean-k-mooneyim well aware16:40
fungiif this is something that we want to solve, then a custom label limited to flavors with 8 cpus or more would be the most efficient solution16:40
sean-k-mooneyim stating that i would like a way to request >6 vcpu or spcificly 816:40
clarkbright if the issue isn't the xen issue but rax flex vs everything else then we'd need different labels16:41
fungirunning jobs on the 16gb or 32gb labels might work around it, but will consume more memory quota obviously16:41
clarkbthat only map to clouds where there are 8vcpus. There isn't a way to use 4vcpu? allocating one cpu still leaves you 316:41
sean-k-mooneyright so im wondierign if it would be ok to createa single new lable ubuntu-noble-8vgpu-8G 16:41
sean-k-mooneyand use that just in the job or jobs that woudl need it16:42
sean-k-mooneyclarkb: we need 2 core for the nvme emulator and we need to use isocpus 16:42
sean-k-mooneymeain runing devstack with effeictly 2 core for eveything else16:43
sean-k-mooneyim currently wondering if we coudl make this a 2 node job with a deicated compute16:43
sean-k-mooneythat might work16:43
sean-k-mooneyfor https://review.opendev.org/c/openstack/cyborg/+/977083/9#message-13a90c9ce5b990729858bc6a80f53a254589a84116:43
clarkbyes 2 node should double the total cpu count if you can slice things effectively across the nodes16:44
sean-k-mooneyfor feature like cpu pinnign haveing 4 cores total is hard16:44
sean-k-mooneywell a compute only has nova/neutorn libvirt and ovs/ovn16:45
sean-k-mooneyso that proably woudl be workable16:45
sean-k-mooneywe coudl perhaps test cpu pinning in nova with 4 cores if concrrancy was 116:45
sean-k-mooneyso that coudl be an option as well but we are really jumping through hoops to make it workable16:46
sean-k-mooneyfewer faster cores is proably fine for most services but it ligitmatly inviladge or limits testing for some projects16:47
sean-k-mooneyits why we historically have not tested these feature in the first party ci in the past16:47
opendevreviewMichal Nasiadka proposed openstack/project-config master: propose-updates: Add test-requirements target  https://review.opendev.org/c/openstack/project-config/+/97856616:52
sean-k-mooneythe nvmevirt repos say "t is highly recommended to use the isolcpus Linux command-line configuration to avoid schedulers putting tasks on the CPUs that NVMeVirt uses:"16:54
sean-k-mooneybut i wonder if we coudl get away with not doing that16:54
sean-k-mooneyas long as its fucntional enough for ci im not sure we care about the performance16:55
sean-k-mooneysicne most test wont use the virtualised nvme device16:55
sean-k-mooneyonly the one that is testing that feature16:55
sean-k-mooneywe migt get away with oversubcibign those cores16:55
*** mtreinish_ is now known as mtreinish17:30
*** mtreinish_ is now known as mtreinish17:53

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!