Tuesday, 2025-10-28

seyeongkimHello, could someone please review this patch? https://review.opendev.org/c/openstack/nova/+/963665 I would greatly appreciate any advices or ideas you might have. Thanks in advance.00:36
*** mhen_ is now known as mhen02:58
opendevreviewBalazs Gibizer proposed openstack/nova master: Rally job for eventlet-removal  https://review.opendev.org/c/openstack/nova/+/96013008:38
opendevreviewTakashi Kajinami proposed openstack/placement master: Remove remaining tox env for Python 3.9  https://review.opendev.org/c/openstack/placement/+/96507709:54
dmsimard[m]Uggla: Hi, any chance we could discuss about our prototype for monitoring-aware nova scheduling around the same time as resource provider weigher ? It feels related and we discussed about it with bauzas and sean-k-mooney at the summit.10:25
dmsimard[m]I can fill in some info in the pad as well.10:25
UgglaHi dmsimard[m], cool I was waiting input from you.10:26
Uggladmsimard[m] let's try to add this with resource provider weigher. sean-k-mooney do you think it is feasable ?10:28
dmsimard[m]If there's anything specific you'd like to know ahead of time I'll be happy to write it down, thanks for the help :)10:28
Uggladmsimard[m], bauzas said in the pad that he missed some materials. So if possible provide some info in the pad (at the bottom). Btw if you have a link to a spec that's probably the best option.10:30
dmsimard[m]There is no spec or blueprint yet. For better or for worse we kind of worked our way backwards such that we could find an approach that worked so we could discuss it and then know which direction to go with a spec.10:36
dmsimard[m]I'll add info to the pad within the next few hours.10:36
Uggladmsimard[m] thx. 👍10:38
sean-k-mooneyUggla: i think that shoudl be fine to combine although its a diffent aprpoch.11:02
sean-k-mooneyit might be wrorth havign  more tiem to be able to dicuss it11:03
dmsimard[m]Just following up, we added some info/context at the bottom of the pad13:23
lowercaseIs there a way to instruct a nova compute node to be instructed to ignore certain cpu features during live migration?13:29
lowercaseLive Migration failure: operation failed: guest CPU doesn't match specification: missing features: lbrv,tsc-scale,vmcb-clean,flushbyasid,pause-filter,pfthreshold: libvirt.libvirtError: operation failed: guest CPU doesn't match specification: missing features: lbrv,tsc-scale,vmcb-clean,flushbyasid,pause-filter,pfthreshold13:29
lowercaseI'm currently researching what those flags are, but they appear between 5.15 and 6.8 kernel upgrade13:39
cardoesean-k-mooney: how can I add that nova-spec to the agenda?13:44
sean-k-mooneyi think there shoudl still be time to cover more topics Uggla ^13:47
Ugglacardoe please write your topic at the end of the etherpad:  https://etherpad.opendev.org/p/nova-2026.1-ptg13:48
Ugglacardoe, then I'll try to find a slot. Please tell me expected duration.13:49
UgglaThe agenda is packed, but if the topic is not so long, I will be able to find a slot.13:49
lowercaseI believe I found the answer myself, https://docs.openstack.org/nova/2023.1/admin/cpu-models.html13:50
lowercasesorry for the bother13:50
bauzasUggla: do we meet back at 4pm UTC in the nova room ?15:43
gibibauzas: some of us are already there as the eventlet crossproject finished early15:46
gibicc Uggla :)15:46
Ugglabauzas that was the original plan15:46
Ugglayep I followed15:46
Ugglado you think we can restart a bit earlier ?15:46
gibiI think we can start around 5015:48
Ugglalet's say 55' ok so if people miss the notification they will not miss to much of the discussion15:49
opendevreviewGoutham Pacha Ravi proposed openstack/nova master: DNM: test dependency on devstack-plugin-ceph changes  https://review.opendev.org/c/openstack/nova/+/91895715:54
dansmithbauzas: lots of typing noise from you16:04
bauzasack16:05
dansmithbauzas: FYI, you sound different than normal and like maybe you're using your laptop mic?16:22
dansmithwe're hearing mouse clicks, keys, etc16:22
bauzasfucking Jitsi, it doesn't tell me which mic I use16:22
dansmithyeah, jitsi-- it's also not letting me unmute from my headset like better tools do :/16:22
bauzaswhat's weird is that I get the usual sound of the in-use mic when I move it in order to speak16:23
bauzasusually, the sounds I get is when my headset is in use16:23
dansmithand also sean-k-mooney was having one-way audio in two separate incidents today16:23
clarkbbauzas: it shows you the mic you are using if you clikc the little ^ on the mic icon16:24
bauzasclarkb: there was no checkmark when I was opening it16:24
bauzasbut once I click on one of the mics, then indeed I have the checkmark set against it16:24
bauzasI guess I probably wasn't forcing to use my headset mic and jitsi was using the laptop one as default 16:25
dansmithmine also did the "you have no mic checked" thing initially16:25
clarkbpavucontrol allows you to pick a default device and jitsi shows me "default" as an otpion. I wonder if it just falls back to that if you don't select something different16:26
sean-k-mooneydansmith: ack what broswer are you using. i swap to firefox and it seams to be workign currently16:26
dansmithchrome and macos16:27
sean-k-mooneyclarkb: at least form me i had 2 way audio with soem particapent and 1 way audio with others in teh same call16:27
sean-k-mooneyclarkb: so it felt more like a nat/peer to peer issue16:27
clarkbsean-k-mooney: weird, we should have the server set up to run all traffic through the server to avoid peer to peer issues. But could still have been nat/firewall issues potentially with the server as the remote endpoint. The other thing we see is that some people just don't toggle the browser perms and they get muted for input and output (the browsers should pop up asking you16:28
clarkbto allow this)16:28
clarkbI think one of the issues is that those of us that use jitsi regularly don't erally have issues with it so these are all unknown unknowns. using it more regularly would help shake out where the problems are but its a chicken and egg problem16:29
melwittit has been fine for me fwiw, I use brave and macos. I also don't use headset controls though16:36
sean-k-mooneyclarkb: honestly it seam to be working fine on firefox and apparenlty that is what i was usign last ptg16:36
clarkbsean-k-mooney: that is good to know (I use chrome and that works for me on linux with x windows)16:37
sean-k-mooneyclarkb: i normlaly use chrom for all webrtc stuff like google meet but it also could be on my side, i have intrusion detection enabeld on my ubiquit router16:37
sean-k-mooneyalthoug given firefox is fine its proably not that16:38
cardoeUggla: I've put https://review.opendev.org/c/openstack/nova-specs/+/471815 at the end of the agenda... still wondering how we can make this happen17:02
cardoeThere's code for the spec. There's tests for the spec. There's multiple downstream vendors that have been running the patches in production for over a year now.17:03
Ugglacardoe, I will find a slot tomorrow morning my time and I will keed you update.17:03
Uggla*keep17:03
cardoeUggla: happy to do it around the time of the ironic/nova cross project times since I'm representing Ironic for both cross project topics.17:05
Ugglayep I was thinking of it. Because the Ironic session is quite long. So maybe you'll have a timeslot.17:06
sean-k-mooneyUggla: we shoudl update the ptg bot to say we are don for today18:34
sean-k-mooneyclarkb: what is the channel again for the ptg is it opendev-events or something generic for that?18:35
Ugglasean-k-mooney, I was doing it when I see your msg poping18:35
Uggla:)18:35
sean-k-mooney:)18:35
clarkbsean-k-mooney: its #openinfra-events18:36
sean-k-mooneyah its openinfra18:37
sean-k-mooneyi tried opensdev and openstack18:37
BigToeTunaHi friends. I have a quick inquery, I have a use case which requires growing the instances directory mountpiont on the  hypervisor. However I can achieve this by having multiple instances_paths I.E /var/lib/nova/instances0, /var/lib/nova/instances1 and so on. Is there any way to specify multiple instance directories?19:38
dansmithnope20:01
BigToeTunaI figured, thanks!20:13

Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!