| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2025-11-26 | |||
| 14:53:36 | noonedeadpunk | the reason why I'm raising this, is that we are actually having cross_az_attach=False here.... | |
| 14:53:46 | noonedeadpunk | And made it working *somehow* | |
| 14:53:52 | noonedeadpunk | In *some* cases... | |
| 14:53:52 | sean-k-mooney | it built on the false premiss that nova AZ names and AZ names in any other service shoudl align | |
| 14:54:48 | sean-k-mooney | so if you set cross_az_attach=False in the api then we will end up pinnign the vm to the az of the voluem if and only if you pass in a volume | |
| 14:54:53 | noonedeadpunk | So I'd really love to leave "status quo" for this specific spec wrt cross_az_attach | |
| 14:55:14 | sean-k-mooney | if on the other hand nova creates the voluem i dont think we pin it because we create teh volume after schdulign i think | |
| 14:55:32 | noonedeadpunk | Yes, I think it's true actually | |
| 14:55:33 | sean-k-mooney | noonedeadpunk: yes im not asking you to change the behvior fomr today | |
| 14:55:37 | sean-k-mooney | just document what it iss | |
| 14:55:40 | noonedeadpunk | ++ | |
| 14:55:53 | noonedeadpunk | ok, shoot... | |
| 14:56:15 | noonedeadpunk | so basically, it would make things worse, or indeed I'd need to address that in implementation.... | |
| 14:56:24 | noonedeadpunk | or well... | |
| 14:56:42 | sean-k-mooney | you can litrally punt an just say "this will not change the behvior of cross_az_attach" and if its disabel then it will fail if there is a confilct | |
| 14:57:22 | noonedeadpunk | ++ | |
| 14:57:59 | noonedeadpunk | I am literally failing now to set in my head cross_az_attach behavior in all scenarios.... | |
| 14:58:13 | noonedeadpunk | ok, thanks, will do that then | |
| 14:59:26 | sean-k-mooney | yep because its stupily complicated. the default is true which mean sthe api does nto condier the voluem az by default | |
| 14:59:40 | sean-k-mooney | if you only set it to false on the compute we only check it there | |
| 15:00:05 | sean-k-mooney | if you onely set it on the api we will validate teh az request in the build request vs the volume | |
| 15:00:22 | sean-k-mooney | and we will use the volume's az if there isnt a conflict | |
| 15:00:40 | sean-k-mooney | there is some other behvior thace chages as well | |
| 15:00:56 | noonedeadpunk | unless you create an instance with `--boot-fropm-volume` | |
| 15:01:17 | sean-k-mooney | yep in which case the voluem i think get create on the compute node :) | |
| 15:02:42 | opendevreview | Dmitriy Rabotyagov proposed openstack/nova-specs master: [spec] Add Cross-AZ scheduling blueprint https://review.opendev.org/c/openstack/nova-specs/+/900296 | |
| 15:12:27 | opendevreview | Dmitriy Rabotyagov proposed openstack/nova-specs master: [spec] Add Cross-AZ scheduling blueprint https://review.opendev.org/c/openstack/nova-specs/+/900296 | |
| 17:21:01 | opendevreview | Merged openstack/nova-specs master: Re-propose vTPM live migration https://review.opendev.org/c/openstack/nova-specs/+/961564 | |
| 17:26:46 | tkajinam | dansmith, hi ! It'd be nice if you can spare some time to check this discussion with Uggla https://review.opendev.org/c/openstack/nova-specs/+/966584/comment/24e4946c_0522d2cf/ ,regarding how we handle firmware used by existing vms during hard-reboot. | |
| 17:27:45 | dansmith | okay I'll try.. my list is stacked pretty deep and we're off the rest of the week here, as you probably know | |
| 17:29:25 | tkajinam | dansmith, thx and yeah I heard that | |
| 17:30:57 | tkajinam | I'll check with Uggla if I can update that part later (after merging the current version) or I can get spec freeze exception, if we need some more time/discussions | |
| 17:31:46 | Uggla | tkajinam, that's ok for me | |
| 18:47:05 | opendevreview | Will Szumski proposed openstack/nova master: Randomise best hosts https://review.opendev.org/c/openstack/nova/+/968544 | |
| 19:56:51 | opendevreview | Sean Mooney proposed openstack/nova master: [WIP] create ResourceSummaryNotification object https://review.opendev.org/c/openstack/nova/+/968558 | |
| #openstack-nova - 2025-11-27 | |||
| 02:17:10 | opendevreview | Seyeong Kim proposed openstack/nova master: libvirt: Support boot_index for multiple block devices https://review.opendev.org/c/openstack/nova/+/963665 | |
| 09:25:24 | opendevreview | Will Szumski proposed openstack/nova master: Randomise best hosts https://review.opendev.org/c/openstack/nova/+/968544 | |
| 10:38:02 | opendevreview | Johannes Kulik proposed openstack/nova-specs master: Spec for asynchronous volume-attach API https://review.opendev.org/c/openstack/nova-specs/+/958900 | |
| 11:15:57 | opendevreview | Johannes Kulik proposed openstack/nova-specs master: Spec for asynchronous volume-attach API https://review.opendev.org/c/openstack/nova-specs/+/958900 | |
| 11:33:12 | r-taketn | Hello everyone. I have uploaded the SEV/SEV-ES refactoring codes. Could you please review it and provide some feedback? | |
| 11:33:20 | r-taketn | #link https://blueprints.launchpad.net/nova/+spec/generalize-sev-code | |
| 11:37:23 | r-taketn | It was discussed at PTG (https://etherpad.opendev.org/p/nova-2026.1-ptg#L687) and decided to make this specless. And, I have run SEV/SEV-ES tests (both manually and using tempest) with uploaded codes. | |
| 11:49:14 | opendevreview | Dmitriy Rabotyagov proposed openstack/nova-specs master: [spec] Add Cross-AZ scheduling blueprint https://review.opendev.org/c/openstack/nova-specs/+/900296 | |
| 14:34:46 | masahito | hi nova team, please review the bug fix for 500 Internal Server error in the lock API https://review.opendev.org/c/openstack/nova/+/946223 . Thanks. | |
| 17:16:04 | noonedeadpunk | sean-k-mooney: sorry for bothering you again... I'm trying to process reviews on the spec, and got confused myself as well. Given that weights are run per hypervisor, then basically we can't use heapq basically? | |
| 17:17:54 | noonedeadpunk | there's no reason to build AZ to Host mapping, as we just need to get current hypervisor aggregates, and fetch AZs from it. And then fetch azs foreach request_spec.instance_group.members | |
| 17:18:44 | noonedeadpunk | and then like count how many instances are in the same az, and use that as a negative multiplier? | |
| 17:19:09 | sean-k-mooney | we can the heap woudl need to be be in the request spec object | |
| 17:19:27 | sean-k-mooney | althernitively you can impelent the wigher by overriding the follow fucntion | |
| 17:19:49 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/weights.py#L100-L120 | |
| 17:20:06 | sean-k-mooney | that will allow it to process all hosts at once and you jsut return the list of wights | |
| 17:20:56 | sean-k-mooney | noonedeadpunk so the point i was orgianly trying to make is you rare not ment to fetch data in the weighers | |
| 17:21:17 | sean-k-mooney | only use the data form the host state or request spec | |
| 17:22:38 | sean-k-mooney | we do not have any weigher that operate on aggreats today btu we do have filters that do | |
| 17:22:48 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/scheduler/filters/aggregate_instance_extra_specs.py#L50 | |
| 17:23:00 | noonedeadpunk | ok, I was just looking at existing weighters, but seems none of them is actually using weigh_objects | |
| 17:23:05 | sean-k-mooney | but even then you lotice that we never do lookups on object or db queries | |
| 17:23:28 | sean-k-mooney | noonedeadpunk: correct non of the intree ones weigh all the hosts at once | |
| 17:23:47 | sean-k-mooney | https://github.com/openstack/nova/blob/master/nova/scheduler/weights/affinity.py | |
| 17:24:04 | sean-k-mooney | is the best refence for what you currently want to do | |
| 17:24:12 | noonedeadpunk | yeah, was checking that, and it seems that it should have everything needed | |
| 17:24:23 | noonedeadpunk | except not doing it at once | |
| 17:25:07 | sean-k-mooney | so i ugess it does indrectly | |
| 17:25:16 | sean-k-mooney | actully no | |
| 17:25:38 | sean-k-mooney | it does not have the count of isntace in the curren group in the a given az | |
| 17:26:03 | sean-k-mooney | if you were to weigh all host you shoudl have the info you need however | |
| 17:26:07 | noonedeadpunk | but then if weigh_objects is used - does that still run for each hypervisor? | |
| 17:26:34 | sean-k-mooney | since you can gorup the host_state obejct by AZ then count the number of instnace form the isntnace grou pin each group | |
| 17:26:46 | sean-k-mooney | noonedeadpunk: no it will run once | |
| 17:27:14 | noonedeadpunk | so.... then also doing a query is not that bad? | |
| 17:27:31 | sean-k-mooney | you don tneed to do a query | |
| 17:27:37 | sean-k-mooney | the info shoudl already be there | |
| 17:27:55 | sean-k-mooney | if you look at the base impoealtion of weigh objects https://github.com/openstack/nova/blob/master/nova/weights.py#L100-L120 | |
| 17:28:08 | sean-k-mooney | that the forlop that runs the weigher over all host one by one | |
| 17:28:34 | sean-k-mooney | weighed_obj_list is the list of HostState objects | |
| 17:28:41 | noonedeadpunk | so weighed_obj_list is list of HostState? | |
| 17:28:42 | sean-k-mooney | weight_properties is the request spec | |
| 17:28:43 | noonedeadpunk | ok | |
| 17:28:45 | noonedeadpunk | yes | |
| 17:29:05 | sean-k-mooney | we never got a round to fixing the fact they are called diffent thing in diffeent palces | |
| 17:29:51 | noonedeadpunk | yeah, ok, makes sense | |
| 17:29:56 | noonedeadpunk | thanks for clarifying it | |
| 17:30:11 | noonedeadpunk | as I got a little bit confused | |
| 17:30:20 | sean-k-mooney | ya i proably could have explaid this better | |
| 17:30:31 | sean-k-mooney | it only clicked now that we shoudl ahve all the data we need already | |
| 17:30:51 | sean-k-mooney | but that to implement this weigher we shoudl jsut overried weigh_objects | |
| 17:31:03 | noonedeadpunk | yeah | |
| 17:31:26 | noonedeadpunk | I had no idea what these parameters are tbh | |
| 17:31:44 | noonedeadpunk | *arguments | |
| 17:31:57 | sean-k-mooney | ya so in newer filtere/weigher we name them what they are | |
| 17:32:08 | sean-k-mooney | but in th older once we never fixt them to make sense | |
| 17:32:27 | noonedeadpunk | ++ | |
| 17:33:15 | noonedeadpunk | now it looks *way* simpler then I originally anticipated | |
| 17:52:26 | opendevreview | Dmitriy Rabotyagov proposed openstack/nova-specs master: [spec] Add Cross-AZ scheduling blueprint https://review.opendev.org/c/openstack/nova-specs/+/900296 | |
| 18:09:22 | sean-k-mooney | noonedeadpunk: same | |
| 18:19:12 | noonedeadpunk | um? | |
| 18:19:37 | noonedeadpunk | sean-k-mooney: sorry, did not get what you meant | |
| 18:20:39 | sean-k-mooney | its simpler then i was expecting also | |
| 18:20:45 | noonedeadpunk | ah | |
| 18:20:46 | noonedeadpunk | :D | |