Earlier  
Later
New Search
Posted Nick Remark
#openstack-nova - 2020-03-26
15:32:34 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Update and correct typing information https://review.opendev.org/714694
15:32:35 openstackgerrit Stephen Finucane proposed openstack/nova master: objects: Replace 'cpu_pinning_requested' helper https://review.opendev.org/714697
15:32:35 openstackgerrit Stephen Finucane proposed openstack/nova master: tests: Split instance NUMA object tests https://review.opendev.org/714696
15:32:36 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Remove handling of pre-Train compute nodes https://review.opendev.org/714699
15:32:36 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Don't consider overhead CPUs for unpinned instances https://review.opendev.org/714698
15:32:37 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Tweak the 'cpu_realtime_mask' handling slightly https://review.opendev.org/461456
15:32:37 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Add validation for 'cpu_realtime_mask' https://review.opendev.org/468203
15:32:38 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Invert order of NUMA topology generation https://review.opendev.org/714701
15:32:38 openstackgerrit Stephen Finucane proposed openstack/nova master: hardware: Rework 'get_realtime_constraint' https://review.opendev.org/714700
15:33:23 gibi gmann: do I understand correctly that you also got convinced that the PUT solution is OK
15:33:26 gibi ?
15:36:55 gmann gibi: yeah. I am ok to add in existing PUT as making other PUT or action API is a big change from implementation and users perspective. having api-ref clear about this PUT = swap + updating resources based on how you use this API.
15:37:18 gibi gmann: ack, thanks
15:37:53 gmann though it makes this API more complex but its tradeoff as dansmith mentioned that any other way is not less complex for client.
15:40:04 sean-k-mooney gmann: i think calling it a swap + updating resources is still the wrong way to think about it
15:40:19 gmann gibi: on the instance fault things: i will wai for dansmith to come back and discuss if somehow we can minimize the info leak to non-admin. other part like policy things is ok for me.
15:40:25 sean-k-mooney gmann: it is just declaritivly updatin the resouces and a sidefect of that can be to trigger swap
15:40:43 sean-k-mooney that is how we should explain it to endusers
15:40:55 gibi gmann: re instance fault: thanks
15:41:08 johnthetubaguy gmann: my preference is to not document the crazy cinder API, and move that to some developer docs or something?
15:41:36 johnthetubaguy should stop some confusion for the 99% of API users
15:41:37 sean-k-mooney johnthetubaguy: do end users ever call that directly
15:41:40 gmann sean-k-mooney: humm, having the swap as side-efect is my concern. that is big change to enduser machine so it has to be very explicit operation from usage doc side.
15:41:47 johnthetubaguy sean-k-mooney: they do, but they never should
15:41:58 johnthetubaguy its for cinder only, else odd things happen
15:42:07 sean-k-mooney gmann: not really that is how i alwasy tought of it already
15:42:42 sean-k-mooney johnthetubaguy: ok so this is just for the callback form cinder
15:42:51 johnthetubaguy that is my understanding, yes
15:42:56 johnthetubaguy hence the big red warning
15:43:04 gmann johnthetubaguy: can we make that internal than ? at least from doc side. i mean api-ref not at all talk about swap operation for users as currently we have ?
15:43:09 sean-k-mooney johnthetubaguy: and yes im pretty sure we have had downstream bugs filed as a result of users calling it directly
15:43:20 johnthetubaguy gmann: +1 that
15:43:24 gmann sean-k-mooney: yeah we had few in past
15:43:35 johnthetubaguy sean-k-mooney: yeah, what they wanted to do was call the Cinder swap API, but they got confused
15:43:39 gibi for me the difference is like change my house (swap) or repaint the kitchen in my house (update)
15:44:36 sean-k-mooney actully i think our customer wanted to bypass a check in the ciner side a and force it so that is why they used nova
15:45:35 johnthetubaguy gibi: or you swap out your oven to a hot one, or just change tell it to self destruct when your pizza is done. its the same thing you are changing.
15:45:45 openstackgerrit Stephen Finucane proposed openstack/nova master: api: Add framework for extra spec validation https://review.opendev.org/704643
15:45:46 openstackgerrit Stephen Finucane proposed openstack/nova master: docs: Add documentation for flavor extra specs https://review.opendev.org/710037
15:45:46 sean-k-mooney anyway when we have had issue related to swap volume we havne pretty much always fixed it once for them and told them not to do that again
15:45:46 openstackgerrit Stephen Finucane proposed openstack/nova master: api: Add microversion 2.84, extra spec validation https://review.opendev.org/708436
15:45:55 stephenfin gibi: I'd to rebase that ^ after losing the race for 2.83. Could you hit https://review.opendev.org/#/c/708436/9 again?
15:46:18 johnthetubaguy sean-k-mooney: yeah, sounds like they were doing something bad. something we shouldn't tell them to do in the docs
15:46:19 stephenfin and johnthetubaguy, any chance you'll stick that on your list? iirc, that was also on your list of gripes ^
15:46:21 gibi johnthetubaguy: yeah like that. replace the whole v.s. tweak a prt of the whole
15:46:45 gibi stephenfin: sure
15:46:56 johnthetubaguy stephenfin: well played, and yes on both counts
15:47:05 gibi stephenfin: I think I caused you to lose :)
15:47:34 stephenfin well, thankfully it's just lyarwood and I duking it out for 2.84 now, afaict :)
15:48:05 gibi :)
15:50:38 sean-k-mooney johnthetubaguy: have you had any futher issues with the libosinfo feature where we try to set some default based on the disto name/version
15:51:13 johnthetubaguy sean-k-mooney: mostly given up on it I think, not had time to dig
15:51:19 sean-k-mooney johnthetubaguy: we backported the fix for the last one downstream and fond that new libosinfo broke use in another way
15:51:54 sean-k-mooney johnthetubaguy: ok so i would like to remove it in victoria since it was broken from the start given it ignores architeture and machine type
15:52:13 sean-k-mooney would that cause issues for you if we did that
15:52:30 johnthetubaguy sean-k-mooney: it would make my life easier I think
15:52:54 sean-k-mooney we likely would fix the issue we hit so we can backport then kill it with fire in the next patch
15:52:55 johnthetubaguy sounds like a good idea, and just replace it with some decent documentation
15:53:29 sean-k-mooney ya we have explict image properties to set everyting it can set anyway
15:53:43 gibi stephenfin: done.
15:53:48 stephenfin thanks
15:54:00 sean-k-mooney and i alreay recommend people to use those anyway since libosinfo change its default out of our contol anyway
15:54:34 gibi nova meeting starts in 5 and a half minutes on #openstack-meeting-3
16:10:54 lyarwood stephenfin: oOOoOoo shiny.. WAIT A MINUTE!
16:11:19 lyarwood stephenfin: feel free to land on 2.84 btw, I need to rework something in my series today anyway
16:11:33 lyarwood stephenfin: if it's just the two of us I'll land on 2.85
16:12:09 openstackgerrit Stephen Finucane proposed openstack/nova master: Use neutronclient's port binding APIs https://review.opendev.org/706295
16:19:00 artom stephenfin, for the record, I hate https://review.opendev.org/#/c/631053/ and everything it stands for
16:19:17 artom I'll still review it because I started, but I will forever hold this grudge in my hear.
16:19:20 artom *heart.
16:20:14 stephenfin 🕺🕺🕺
16:20:14 stephenfin 🕺🕺🕺
16:21:15 artom ... man dancing?
16:25:32 openstackgerrit Huaqiang Wang proposed openstack/nova master: libvirt: support to create instance with dedicated and shared CPUs https://review.opendev.org/714655
16:25:32 openstackgerrit Huaqiang Wang proposed openstack/nova master: Take the instance dedicated CPU list from 'cpu_pinning' if possible https://review.opendev.org/713352
16:25:33 openstackgerrit Huaqiang Wang proposed openstack/nova master: Remove 'InstanceNUMACell.cpu_pinning_requested' field https://review.opendev.org/714656
16:25:33 openstackgerrit Huaqiang Wang proposed openstack/nova master: 'cpu_pinning_requested' property could be derived directly from cpu_policy https://review.opendev.org/713353
16:25:34 openstackgerrit Huaqiang Wang proposed openstack/nova master: Introduce 'MIXED' CPU allocation policy for instance https://review.opendev.org/713354
16:25:34 openstackgerrit Huaqiang Wang proposed openstack/nova master: Refactor the code in checking available host CPUs https://review.opendev.org/714657
16:25:35 openstackgerrit Huaqiang Wang proposed openstack/nova master: metadata: export the vCPU IDs that are pinning on the host CPUs https://review.opendev.org/688936
16:25:35 openstackgerrit Huaqiang Wang proposed openstack/nova master: Introduce the interface of creating 'MIXED' policy instance through 'PCPU' and 'VCPU' https://review.opendev.org/713355
16:25:36 openstackgerrit Huaqiang Wang proposed openstack/nova master: [WIP] An alternative way for keeping instance dedicated CPUs https://review.opendev.org/714658
16:26:09 gmann johnthetubaguy: you already have noticed this. I added no-deprecation tests for admin-api also which does not make a difference than enable-scope tests but still to check if nothing weird happens when remove deprecation. - https://review.opendev.org/#/c/715085/2/nova/tests/unit/policies/test_instance_usage_audit_log.py@99
16:26:24 gmann hope that is fine, in some cases it is just removing the deprecatiob from base rule and checks
16:26:47 johnthetubaguy gmann: yeah, seemed good to check anyways
16:26:55 gmann ok
16:27:16 johnthetubaguy its a bit like nova-next tests
16:27:22 huaqiang stephenfin: I updated the series for 'mixed policy instance', but only the last patch I changed this time
16:27:22 gmann yeah
16:27:48 huaqiang https://review.opendev.org/714658 [WIP] An alternative way for keeping instance dedicated CPUs
16:28:14 huaqiang this 'pcpuset approach' one,
16:28:59 huaqiang If you have time please have a look for this patch,
16:29:27 huaqiang I just want your comments for if it is ready to drop the 'cpu_pinning' approach now
16:30:00 huaqiang it the answer is comfirmative, I'll drop it and submit new series.
16:32:27 huaqiang this patch is still a 'WIP' patch, I still have several 'TODO's, I'll address tomorrow.
16:32:30 openstackgerrit John Garbutt proposed openstack/nova master: Add legacy limits and usage to unified limits https://review.opendev.org/713498
16:32:30 openstackgerrit John Garbutt proposed openstack/nova master: Enforce resource limits using oslo.limit https://review.opendev.org/615180
16:32:31 openstackgerrit John Garbutt proposed openstack/nova master: Update quota apis with keystone limits and usage https://review.opendev.org/713499
16:48:32 openstackgerrit John Garbutt proposed openstack/nova master: Tell oslo.limit how to count nova resources https://review.opendev.org/713301
16:49:08 openstackgerrit John Garbutt proposed openstack/nova master: Enforce resource limits using oslo.limit https://review.opendev.org/615180

Earlier   Later