| openstackgerrit | Merged openstack/nova master: Add retrying to requirements.txt https://review.openstack.org/568692 | 00:08 |
|---|---|---|
| openstackgerrit | Merged openstack/nova master: Fix HTTP500 error of changes-since on v2.0 API https://review.openstack.org/567916 | 00:31 |
| mriedem_afk | in the same 1 hour span, i've said both "i'm not feeding the entire neighborhood!" and "i've gotta pay for that water!" | 00:38 |
| mriedem_afk | i feel pretty good about that | 00:38 |
| openstackgerrit | Naichuan Sun proposed openstack/nova master: xenapi(N-R-P): support compute node resource provider update https://review.openstack.org/521041 | 01:04 |
| *** edmondsw has joined #openstack-placement | 01:18 | |
| *** edmondsw has quit IRC | 01:22 | |
| *** mriedem_afk has quit IRC | 01:27 | |
| openstackgerrit | Tsuyoshi Nagata proposed openstack/nova master: nova improvement of maximum attach volumes more than 26 vols https://review.openstack.org/567472 | 01:58 |
| *** gjayavelu has quit IRC | 02:20 | |
| openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Support nested alloc cands with sharing providers https://review.openstack.org/567508 | 02:48 |
| openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Return all resources in provider_summaries https://review.openstack.org/558045 | 02:48 |
| openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Return all nested providers in tree https://review.openstack.org/559480 | 02:48 |
| openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Add microversion for nested allocation candidate https://review.openstack.org/565487 | 02:48 |
| openstackgerrit | Vu Cong Tuan proposed openstack/nova stable/queens: Update auth_url in install docs https://review.openstack.org/568730 | 02:55 |
| openstackgerrit | jichenjc proposed openstack/nova master: WIP: Remove support for /os-fixed-ips REST API https://review.openstack.org/568516 | 03:17 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Deprecate running API services under eventlet https://review.openstack.org/549510 | 03:48 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: DNM: Test nova-api eventlet deprecation https://review.openstack.org/549770 | 03:49 |
| openstackgerrit | Tsuyoshi Nagata proposed openstack/nova master: nova improvement of maximum attach volumes more than 26 vols https://review.openstack.org/567472 | 03:56 |
| openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Add tests for _get_trees_matching_all() function https://review.openstack.org/568737 | 04:12 |
| *** tetsuro has joined #openstack-placement | 04:13 | |
| tetsuro | ^ @jaypipes: this patch of https://review.openstack.org/#/c/568737/ answers your comment in https://review.openstack.org/#/c/567150/11 | 04:16 |
| openstackgerrit | Tetsuro Nakamura proposed openstack/nova master: Add tests for _get_trees_matching_all() function https://review.openstack.org/568737 | 04:19 |
| *** edmondsw has joined #openstack-placement | 04:32 | |
| *** tetsuro has left #openstack-placement | 04:52 | |
| *** tetsuro_ has joined #openstack-placement | 04:54 | |
| *** tetsuro_ has quit IRC | 04:55 | |
| openstackgerrit | Tsuyoshi Nagata proposed openstack/nova master: nova improvement of maximum attach volumes more than 26 vols https://review.openstack.org/567472 | 04:56 |
| *** diga has joined #openstack-placement | 05:25 | |
| *** gjayavelu has joined #openstack-placement | 05:25 | |
| *** diga has quit IRC | 05:32 | |
| openstackgerrit | Tsuyoshi Nagata proposed openstack/nova master: nova improvement of maximum attach volumes more than 26 vols https://review.openstack.org/567472 | 05:44 |
| openstackgerrit | Naichuan Sun proposed openstack/nova master: XenAPI: update the document related to vdi streaming https://review.openstack.org/568444 | 05:46 |
| openstackgerrit | Vishakha Agarwal proposed openstack/nova master: Re-using the code of os brick cinder https://review.openstack.org/565703 | 05:55 |
| openstackgerrit | Naichuan Sun proposed openstack/nova master: xenapi(N-R-P): support compute node resource provider update https://review.openstack.org/521041 | 06:13 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Remove mox in libvirt/test_driver.py (2) https://review.openstack.org/568757 | 06:23 |
| *** edmondsw has quit IRC | 06:49 | |
| openstackgerrit | Merged openstack/nova-specs master: Complex (Anti)-Affinity Policies https://review.openstack.org/546925 | 06:56 |
| *** avolkov has joined #openstack-placement | 07:35 | |
| openstackgerrit | Nguyen Hai proposed openstack/nova-specs master: Follow the new PTI for document build https://review.openstack.org/551802 | 07:39 |
| *** jaypipes has quit IRC | 07:50 | |
| *** jaypipes has joined #openstack-placement | 07:51 | |
| openstackgerrit | Chen Hanxiao proposed openstack/nova master: libvirt: guest: introduce blockStats instead of domain.blockStats https://review.openstack.org/526833 | 08:01 |
| openstackgerrit | Naichuan Sun proposed openstack/nova master: XenAPI: update the document related to vdi streaming https://review.openstack.org/568444 | 08:01 |
| openstackgerrit | Tsuyoshi Nagata proposed openstack/nova master: nova improvement of maximum attach volumes more than 26 vols https://review.openstack.org/567472 | 08:03 |
| *** gjayavelu has quit IRC | 08:14 | |
| *** tssurya has joined #openstack-placement | 08:15 | |
| *** edmondsw has joined #openstack-placement | 08:20 | |
| *** edmondsw has quit IRC | 08:25 | |
| gibi | efried: I think we have a problem with the flavor extra_spec in the granular patch: https://review.openstack.org/#/c/515811/18/releasenotes/notes/granular-extra-specs-50b26b8f63717942.yaml@9 | 08:33 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova-specs master: Allow having placement inventories with reserved value equal to total https://review.openstack.org/568613 | 08:39 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova stable/queens: placement: Fix HTTP error generation https://review.openstack.org/568784 | 08:45 |
| openstackgerrit | Surya Seetharaman proposed openstack/nova stable/queens: POC: Filter/ignore NotSupportedWarning ['use_tpool'] in nova-manage https://review.openstack.org/568785 | 08:51 |
| *** cdent has joined #openstack-placement | 08:57 | |
| *** diga has joined #openstack-placement | 08:58 | |
| *** gjayavelu has joined #openstack-placement | 09:00 | |
| *** diga has quit IRC | 09:35 | |
| openstackgerrit | Yikun Jiang (Kero) proposed openstack/nova master: WIP: Microversion 2.63 - Use new format policy in server group https://review.openstack.org/567534 | 09:56 |
| *** edmondsw has joined #openstack-placement | 10:08 | |
| *** edmondsw has quit IRC | 10:13 | |
| *** e0ne has joined #openstack-placement | 10:39 | |
| *** cdent has quit IRC | 10:51 | |
| *** cdent has joined #openstack-placement | 11:33 | |
| *** edmondsw has joined #openstack-placement | 11:57 | |
| *** edmondsw has quit IRC | 12:01 | |
| *** mriedem has joined #openstack-placement | 12:05 | |
| *** edmondsw has joined #openstack-placement | 12:07 | |
| *** edmondsw has quit IRC | 12:07 | |
| efried | gibi: Good catch, will fix right away. | 12:11 |
| openstackgerrit | Eric Fried proposed openstack/nova master: Granular requests to get_allocation_candidates https://review.openstack.org/515811 | 12:14 |
| efried | gibi: Done ^ | 12:14 |
| gibi | efried: thanks, I'm +2 | 12:17 |
| * cdent looks at that again since the fix | 12:18 | |
| *** edmondsw has joined #openstack-placement | 12:36 | |
| cdent | I'm sorry jaypipes I think you're taking me on bad faith and over the line. | 12:37 |
| * cdent takes a walk | 12:39 | |
| edleafe | cdent: Now you know the main reason I volunteered to do more internal work at IBM and less in Nova / Placement. | 12:41 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Implement granular policy rules for placement https://review.openstack.org/524425 | 12:54 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add granular policy rules for /resource_classes* https://review.openstack.org/565578 | 12:54 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add granular policy rules for resource providers inventories https://review.openstack.org/568666 | 12:54 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add granular policy rules for usages https://review.openstack.org/568706 | 12:54 |
| openstackgerrit | Eric Fried proposed openstack/nova master: Use placement.inventory.inuse in report client https://review.openstack.org/568639 | 13:01 |
| jaypipes | edleafe: I'm the reason huh? | 13:02 |
| openstackgerrit | Eric Fried proposed openstack/nova master: Use GET.get instead of GET.getall in alloc-cands https://review.openstack.org/568713 | 13:05 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Add requested_resources field to RequestSpec https://review.openstack.org/567267 | 13:06 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Transfer port.resource_request to the scheduler https://review.openstack.org/567268 | 13:06 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Make placement.lib.RequestGroup a versioned object https://review.openstack.org/568840 | 13:06 |
| edleafe | jaypipes: A big part, yes | 13:16 |
| jaypipes | edleafe: awesome. | 13:16 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova-specs master: Allow having placement inventories with reserved value equal to total https://review.openstack.org/568613 | 13:18 |
| bauzas | cdent: FWIW, just saw that your forum session on splitting placement is conflicting with my own vGPU presentation :/ | 13:21 |
| bauzas | exact same time | 13:21 |
| bauzas | shit | 13:21 |
| efried | boo | 13:22 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Deduplicate config/policy reference docs from main index https://review.openstack.org/564846 | 13:23 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Implement granular policy rules for placement https://review.openstack.org/524425 | 13:23 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add granular policy rules for /resource_classes* https://review.openstack.org/565578 | 13:23 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add granular policy rules for resource providers inventories https://review.openstack.org/568666 | 13:23 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add granular policy rules for usages https://review.openstack.org/568706 | 13:23 |
| * bauzas needs to look at the Summit agenda, but no time yet :( | 13:24 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Cleanup ugly stub in TestLocalDeleteAllocations https://review.openstack.org/561997 | 13:24 |
| mriedem | bauzas: there is always time when you're on the crapper | 13:25 |
| * bauzas google translates that expression | 13:26 | |
| bauzas | hah | 13:26 |
| bauzas | I'm a remote person, so that problem is now solved since 4+ years | 13:26 |
| bauzas | managing access to a specific door without being seen by your colleagues is no longer a problem for me | 13:27 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Make placement.lib.RequestGroup a versioned object https://review.openstack.org/568840 | 13:27 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Add requested_resources field to RequestSpec https://review.openstack.org/567267 | 13:27 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Transfer port.resource_request to the scheduler https://review.openstack.org/567268 | 13:27 |
| mriedem | bauzas: i think you're missing the point | 13:27 |
| mriedem | you can multitask | 13:28 |
| bauzas | crapper == toilets ? | 13:28 |
| efried | gibi: Nooooo, not a RequestGroup ovo! | 13:28 |
| bauzas | that's where I hit the language barrier :) | 13:28 |
| sean-k-mooney | bauzas: yes | 13:28 |
| vdrok | :D | 13:28 |
| efried | gibi: I'm sure there's a good reason for it, sigh, just having PTSD from the non-OVO OVOs in placement.objects.resource_provider | 13:29 |
| bauzas | mriedem: well, I can surely multitask, but like I always said, it's done by concurrency, not real parallelism | 13:29 |
| vdrok | mriedem: morning, could you reopen this one? https://bugs.launchpad.net/nova/+bug/1751472 | 13:29 |
| openstack | Launchpad bug 1751472 in OpenStack Compute (nova) queens "InventoryInUse exception is periodically logged as ERROR" [Medium,Fix committed] - Assigned to Sylvain Bauza (sylvain-bauza) | 13:29 |
| bauzas | I only have one BPU | 13:29 |
| sean-k-mooney | there was one session on an alternitve to the nova schduler that i spoted that i wanted to flag to you guys too | 13:29 |
| bauzas | BPU = Brain Processing Unit | 13:29 |
| gibi | efried: what causing the pain regarding OVOs? | 13:29 |
| mriedem | vdrok: re-open? it's marked as fixed. if there is a new bug, it should be a new bug | 13:29 |
| mriedem | sean-k-mooney: yes the nfv one | 13:30 |
| mriedem | random "we made everything better w/o ever talking to anyone" sessions are always a blast | 13:30 |
| vdrok | mriedem: ah, hrm, ok. I already forgot how launchpad works :( | 13:30 |
| gibi | efried: for me the reason to make it an OVO is to be able to send through RPC as part of the RequestSpec object | 13:30 |
| efried | gibi: Because when I tried adding fields that were useful, but only internally, there was pushback that we shouldn't do that to OVOs. But the ones in question were only OVOs because that was a convenient way to get features, not because they needed to be sent across the wire (they don't). | 13:31 |
| sean-k-mooney | mriedem: proably i wanted to flag it to jaypipes in partcalar | 13:31 |
| mriedem | https://www.openstack.org/summit/vancouver-2018/summit-schedule/events/21708/revisiting-scalability-and-applicability-of-openstack-placement | 13:31 |
| efried | gibi: Yeah, in that case I guess it makes sense. | 13:31 |
| mriedem | sean-k-mooney: i already showed it to him, and jay won't be at the summit anyway | 13:31 |
| edleafe | gibi: are you talking about the nova side of things, or placement? | 13:31 |
| sean-k-mooney | mriedem: from the abstract i dont even know if it would use placement | 13:31 |
| gibi | edleafe: nova side | 13:32 |
| efried | sean-k-mooney: No, they're not talking about *our* placement. | 13:32 |
| edleafe | gibi: that's why it's a bad idea to use the same object in both services. Completely different needs | 13:32 |
| efried | sean-k-mooney: But it should be noted that our placement was architected to solve a lot of the problems they're complaining about. | 13:32 |
| gibi | efried: and alternative would be to keep the RequestGroup in the placement lib as a plain object and add an OVO to the nova side only and convert back and forth | 13:32 |
| jaypipes | sean-k-mooney: it wasn't an alternative to placement... | 13:32 |
| gibi | edleafe: ^ | 13:32 |
| edleafe | sean-k-mooney: right - they mean s/placement/scheduler | 13:32 |
| efried | gibi: Well, the "conversion" from nova to placement happens via a querystring, not via rpc serialization. | 13:33 |
| efried | gibi: And we certainly do not need it to be ovo on the placement side. Because again, it's only used internally. | 13:33 |
| mriedem | "This session will survey placement in OpenStack, explain how placement is done within the Nova OpenStack project, and highlight its limitations. We will show that the current scheduler does not scale to large clouds, and that the bottleneck is caused by attaining a complete view of the available resources. Instead, we will show an adaptive state caching technique that mitigates the bottleneck while preserving the | 13:33 |
| mriedem | ement quality." | 13:33 |
| gibi | efried: by convert I mean convert inside nova | 13:33 |
| mriedem | so they are using placement + cachingscheduler... | 13:33 |
| sean-k-mooney | yes i know the plan to replace the schduler not placement but i did not see any mention of if there schduler would interact with placement at all | 13:33 |
| edleafe | gibi: the fact that they represent the same information doesn't mean that they have to be the same object | 13:33 |
| efried | mriedem: They're talking about placement, not Placement. | 13:33 |
| mriedem | ok, so just cachingscheduler then | 13:34 |
| gibi | edleafe: fair point. but today we use the same object as well | 13:34 |
| mriedem | or whatever the intel guy did in austin | 13:34 |
| gibi | edleafe: actually this is the only object in the placement.lib today | 13:34 |
| mriedem | https://www.openstack.org/videos/austin-2016/dive-into-nova-scheduler-performance-where-is-the-bottleneck | 13:34 |
| edleafe | gibi: yeah, and I don't like having it there at all. It's an internal implementation to placement | 13:35 |
| sean-k-mooney | i guess we will find out i just was surprsed to see it as a session since i did not see any blueprints or irc discussion on the topic | 13:35 |
| efried | edleafe: It's internal to nova and internal to placement. It just happens to be identical on both sides. | 13:35 |
| edleafe | gibi: nova should have its own internal implementation. This way nova can modify it for its needs | 13:35 |
| edleafe | efried: not identical if it needs to be sent over RPC | 13:35 |
| efried | right, not identical _anymore_ I suppose. | 13:36 |
| gibi | efried, edleafe: I'm happy to use an OVO on Nova side and keep the plain object on placement side | 13:36 |
| mriedem | sean-k-mooney: not really surprised, it has something to do with nfv | 13:36 |
| edleafe | efried: the "anymore" is why they should be two separate objects in separate code bases | 13:36 |
| mriedem | so the track chairs were probably just like, "WOW AWESOME!!!" | 13:36 |
| efried | Which actually would make me a little bit happy, because I could put my to_querystring method in the nova-side one rather than inlined in the ResourceRequest method. | 13:36 |
| gibi | efried, edleafe: thanks for the quick feedback, I will do the separation | 13:37 |
| efried | so gibi, edleafe: I'm on board with the idea of splitting it into separate objects at this point. | 13:37 |
| efried | yeah, that. | 13:37 |
| sean-k-mooney | this is the session by the way https://www.openstack.org/summit/vancouver-2018/summit-schedule/events/21708/revisiting-scalability-and-applicability-of-openstack-placement | 13:38 |
| mriedem | the thing i linked 7 minutes ago.... :) | 13:38 |
| sean-k-mooney | mriedem: oh sorry ya i just say the intel one you linked after | 13:39 |
| sean-k-mooney | efried: on ovo, using it for rest api or even jsut internal uses has some beniftes even from a documentaiton point of view. | 13:41 |
| sean-k-mooney | efried: we dont jsut need to confine ovo to rpc calls | 13:42 |
| efried | sean-k-mooney: I agree about the benefits, totally. | 13:42 |
| efried | sean-k-mooney: The problem wound up being that once a thing was an OVO, even if it wasn't an RPC'd OVO, it started being assumed to have certain characteristics and restrictions that it shouldn't have to have. | 13:42 |
| sean-k-mooney | efried: such as versioning ? | 13:43 |
| sean-k-mooney | or the persecption its a stable interface | 13:43 |
| efried | sean-k-mooney: In this case it was addition of fields that were convenient for internal use, but not part of the "data" per se. | 13:44 |
| sean-k-mooney | efried: well nova internallly adds field to ovo for caching at runtime that are striped out when serialised | 13:44 |
| sean-k-mooney | its a little freaky the first time you see it but ovo dont prevent that | 13:45 |
| efried | sean-k-mooney: Yeah, nothing about ovo prevents doing this stuff. And I got to do it in the end. But there was some contention along the way. Specifically https://review.openstack.org/#/c/564351/ | 13:49 |
| bauzas | oh man | 13:56 |
| bauzas | https://www.openstack.org/summit/vancouver-2018/summit-schedule/events/21708/revisiting-scalability-and-applicability-of-openstack-placement is yet again a new story like https://storage.googleapis.com/pub-tools-public-publication-data/pdf/41684.pdf | 13:57 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova-specs master: Allow having placement inventories with reserved value equal to total https://review.openstack.org/568613 | 14:00 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova-specs master: Allow having placement inventories with reserved value equal to total https://review.openstack.org/568613 | 14:09 |
| openstackgerrit | Merged openstack/nova master: Move image conversion to privsep. https://review.openstack.org/554437 | 14:45 |
| openstackgerrit | Merged openstack/nova master: We don't need utils.trycmd any more. https://review.openstack.org/554439 | 14:45 |
| openstackgerrit | Merged openstack/nova-specs master: Allow having placement inventories with reserved value equal to total https://review.openstack.org/568613 | 14:52 |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs master: Rename spec to match blueprint https://review.openstack.org/568870 | 14:56 |
| openstackgerrit | Stephen Finucane proposed openstack/nova master: Simplify instance name generation https://review.openstack.org/516573 | 15:05 |
| openstackgerrit | Eric Fried proposed openstack/nova master: Debug logs for allocation_candidates filters https://review.openstack.org/568712 | 15:15 |
| efried | melwitt, jaypipes: Thanks for the review there, fixed up all prettier now ^ | 15:16 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Remove deprecated monkey_patch config options https://review.openstack.org/568880 | 15:29 |
| melwitt | efried: I kinda hate to say, I think that method should have had a unit test based on how it wasn't doing what you intended in PS1 | 15:34 |
| cdent | efried: yeah, I was going to ask the same thing: does a test matter? | 15:35 |
| efried | melwitt: It was something I wouldn't have caught with my eyeballs even writing the test. A boolean was ending up False (because missing) instead of <?>. | 15:40 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Remove deprecated monkey_patch config options https://review.openstack.org/568880 | 15:40 |
| efried | Do we have unit tests for the other __repr__s? | 15:41 |
| melwitt | efried: oh, I was thinking one of the test cases would verify a '<?>' showing up in the log message and would fail because of the bug that was there | 15:42 |
| melwitt | we do, I've written one before for something. lemme see | 15:42 |
| efried | melwitt: No biggie, I can do it if you think it's warranted. I just wasn't too worried about it since it's essentially a helper for a debug log. | 15:43 |
| mriedem | cdent: not sure i'm following your comment on https://review.openstack.org/#/c/568706/ but the /usages stuff does require a project_id, even if it's not *your* project_id, like if you're an admin, | 15:44 |
| mriedem | i kind of put my brain in a loop thinking about the implications of how to do that with our policy check, | 15:44 |
| mriedem | but i think we could (maybe should?) pass the requested project_id as the 'target' to the policy check rather than the context (which is the default) | 15:44 |
| cdent | mriedem: I meant that at some point we might want to check that the project_id in the query string matches the project_id of the the user-in-context | 15:44 |
| mriedem | we might be saying the same thing | 15:45 |
| mriedem | context.can() takes a target resource which if not set, is the current context | 15:45 |
| mriedem | so if i have admin A, project B and project C, right now only the admin user can see usages for either project B or C | 15:45 |
| mriedem | but if we changed that policy to be admin_or_owner, then the owner part gets weird | 15:45 |
| mriedem | and i think at that point we say the query param is the target rather than the current context | 15:46 |
| mriedem | so then project B can't see project C's usage | 15:46 |
| mriedem | *i think* | 15:46 |
| openstackgerrit | Merged openstack/nova master: PowerVM Driver: Snapshot https://review.openstack.org/543023 | 15:47 |
| mriedem | for most other things in placement, "owner" doesn't mean anything b/c most resources dont have project/user | 15:47 |
| mriedem | melwitt: i've marked https://blueprints.launchpad.net/nova/+spec/powervm-snapshot complete if you're keeping tally | 15:47 |
| melwitt | efried: yeah, I was just thinking it would have resulted in the wrong thing (wrong part of UUID, weird output of "False" when it was suppose to be "<?>" when someone goes to use the messages while they're debugging and maybe cause some confusion/wasted time. I dunno, maybe I'm overthinking it | 15:48 |
| melwitt | mriedem: thank you, I am keeping tally (though admittedly with gaps in the dates because sometimes I forget to check) | 15:49 |
| efried | melwitt: I await your suggestion. IMO it's not important enough to spend the time on, unless you say it is :) | 15:49 |
| openstackgerrit | Zack Cornelius proposed openstack/nova master: Implement file backed memory for instances in libvirt https://review.openstack.org/567876 | 15:50 |
| melwitt | efried: it's not now, after the fact :P it would have been useful on PS1 where it would have caught the bugs. hence my hesitation. I'd say next time let's just test it | 15:51 |
| openstackgerrit | Zack Cornelius proposed openstack/nova-specs master: Libvirt file backed memory https://review.openstack.org/563704 | 15:51 |
| efried | melwitt: Roger wilco. (What I was saying earlier is I probably would have written the test wrong too - but yeah.) | 15:52 |
| melwitt | yeah, I got what you mean but I didn't understand how the test wouldn't catch it if you had a test case that expected a "<?>" in the message because it would have been missing, unless I've totally misunderstood what would happen there | 15:53 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Add request_spec.RequestGroup versioned object https://review.openstack.org/568840 | 15:56 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Add requested_resources field to RequestSpec https://review.openstack.org/567267 | 15:56 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Transfer port.resource_request to the scheduler https://review.openstack.org/567268 | 15:56 |
| gibi | efried, edleafe: ^^ did the RequestGroup split as we discussed | 15:59 |
| efried | gibi: ack | 16:01 |
| efried | melwitt: No, you're correct; I'm saying I totally would have made the mistake of putting False into the test instead of <?>, just knowing how my brain works and the process I went through when I was working on this. | 16:01 |
| melwitt | oh, I see | 16:02 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova master: Placement: allow to set reserved value equal to total for inventory https://review.openstack.org/564838 | 16:06 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova master: Placement: allow to set reserved value equal to total for inventory https://review.openstack.org/564838 | 16:08 |
| *** efried is now known as efried_rollin | 16:10 | |
| *** edmondsw has quit IRC | 16:15 | |
| *** edmondsw_ has joined #openstack-placement | 16:18 | |
| openstackgerrit | Stephen Finucane proposed openstack/nova-specs master: Add 'numa-aware-vswitches' spec https://review.openstack.org/541290 | 16:22 |
| *** edmondsw_ has quit IRC | 16:23 | |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: DiskAdapter parent class https://review.openstack.org/549053 | 16:24 |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: Localdisk https://review.openstack.org/549300 | 16:24 |
| mriedem | cdent: check my note to self in https://review.openstack.org/#/c/568706/4/nova/api/openstack/placement/handlers/usage.py without reading the irc log linked in and see if that makes sense to you | 16:32 |
| cdent | roget | 16:32 |
| cdent | or roger even | 16:32 |
| cdent | mriedem: a) yes that makes good sense, b) I agree a TODO would make sense too. | 16:34 |
| mriedem | ok will throw that up quick | 16:35 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add granular policy rules for usages https://review.openstack.org/568706 | 16:39 |
| *** tssurya has quit IRC | 16:42 | |
| *** gjayavelu has quit IRC | 17:16 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Use RequestSpec.user_id in scheduler.utils.claim_resources https://review.openstack.org/568917 | 18:02 |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: DiskAdapter parent class https://review.openstack.org/549053 | 18:14 |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: Localdisk https://review.openstack.org/549300 | 18:14 |
| *** gjayavelu has joined #openstack-placement | 18:20 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Update nova-status and docs for required placement 1.24 https://review.openstack.org/568923 | 18:23 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Update nova-status and docs for required placement 1.24 https://review.openstack.org/568923 | 18:25 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Update placement upgrade docs for nova-api dependency on placement https://review.openstack.org/568925 | 18:31 |
| openstackgerrit | Merged openstack/nova master: Debug logs for allocation_candidates filters https://review.openstack.org/568712 | 18:54 |
| *** edmondsw has joined #openstack-placement | 18:57 | |
| *** e0ne has quit IRC | 19:05 | |
| *** e0ne has joined #openstack-placement | 19:05 | |
| *** mriedem1 has joined #openstack-placement | 19:09 | |
| *** mriedem has quit IRC | 19:11 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova-specs master: Add additional information for z/VM spec. https://review.openstack.org/562154 | 19:14 |
| *** mriedem1 is now known as mriedem | 19:15 | |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: DiskAdapter parent class https://review.openstack.org/549053 | 19:39 |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: Localdisk https://review.openstack.org/549300 | 19:39 |
| *** efried_rollin is now known as efried | 19:42 | |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: DiskAdapter parent class https://review.openstack.org/549053 | 19:58 |
| openstackgerrit | Eric Berglund proposed openstack/nova master: PowerVM Driver: Localdisk https://review.openstack.org/549300 | 19:58 |
| *** gjayavel_ has joined #openstack-placement | 20:01 | |
| *** gjayavelu has quit IRC | 20:03 | |
| efried | mriedem: https://review.openstack.org/#/c/568923/ -- does this need to be done any time something in report client uses a newer placement microversion? | 20:14 |
| efried | mriedem: i.e. do I need to do it here too: https://review.openstack.org/#/c/515811/ | 20:15 |
| cdent | efried: if you are requiring a newer version (you are), then "yes" | 20:16 |
| efried | cdent: Hm, okay. Any suggestions for a way to enforce that via tests? | 20:18 |
| *** e0ne has quit IRC | 20:18 | |
| cdent | efried: I'll think on it, but at this late hour I'm ill equipped to think of anything | 20:19 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: WIP: Don't count volume over quota when disabling compute https://review.openstack.org/568953 | 20:21 |
| mriedem | efried: it likely becomes less and less important now that placement is required and we expect/require deployers to be updating placement before nova, | 20:23 |
| mriedem | but for a time, as you recall, we tried to straddle microversions and fallback to earlier behavior if placement was behind | 20:24 |
| efried | right | 20:24 |
| mriedem | which i think long-term is still good practice since we try to do that for things like nova<>cinder and nova<>neutron | 20:24 |
| mriedem | and we / i try to tell people that you don't need to upgrade all services in your cloud at the same time every 6 months | 20:25 |
| mriedem | there was an ops ML thread related to this at one point, about adding a CI job that would be like grenade but keep at least nova behind a version, something like that | 20:25 |
| mriedem | since people upgrade nova last it sounds like | 20:25 |
| efried | mriedem: I'm thinking there may be a way to intercept the request primitives in our placement mocks in test_report (unit) and test_report_client (functional) to keep track of the microversion it sees; then import nova.cmd.status (where we move that min required microversion to a const at the top) and blow up as soon as it comes across one that's higher. | 20:26 |
| efried | cdent: ^ | 20:26 |
| efried | I didn't explain that particularly well, but you get the idea. | 20:27 |
| efried | do you think it's worth doing? | 20:27 |
| cdent | do you mean a sort of test at the end that says "the highest microversion I saw when I was testing the report client was X, now I'm checking to see if we use that in the status check"? | 20:30 |
| cdent | In parallel testing env that might be tricky to make sure it was accurate | 20:30 |
| efried | cdent: That was my original thought, but I don't think we need to do it that way. I think we can import nova.cmd.status.MIN_PLACEMENT_MICROVERSION and just compare it every time we see a version go by in a placement request in any test. We only need to blow up if we see one that's *higher* than MIN_PLACEMENT_MICROVERSION. Which of course relies on there being a unit *or* func test for every report client path that | 20:32 |
| efried | anyway. | 20:32 |
| cdent | oh, I see what you mean | 20:33 |
| cdent | so a sort of proxy client that looks at the request headers in every request | 20:33 |
| efried | yeah | 20:33 |
| efried | before passing them through | 20:33 |
| cdent | that ought to be doable, unclear if it is worth the effort | 20:34 |
| efried | agree on both counts, looking for feedback from mriedem on the latter. | 20:34 |
| mriedem | i certainly don't want to have to review that | 20:44 |
| mriedem | so a big ol meh from me | 20:44 |
| efried | okey dokey. It certainly wouldn't be simple to implement on the UT side as things stand, so no hard feelings from me to let it drop. | 20:48 |
| efried | on the func side it wouldn't be so bad, I don't think. | 20:49 |
| efried | and having that would be better than not having it, I suppose. | 20:49 |
| cdent | yeah, I had a quick look at the unit side and while it wouldn't be impossible it would be annoying | 20:55 |
| * cdent goes to bed | 20:55 | |
| cdent | 'night all | 20:56 |
| *** cdent has quit IRC | 20:56 | |
| *** sean-k-mooney has quit IRC | 20:58 | |
| *** avolkov has quit IRC | 21:13 | |
| *** sean-k-mooney has joined #openstack-placement | 21:25 | |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: Resource class set (v1.7) https://review.openstack.org/514644 | 21:39 |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: Usages per project and user (v1.8, v1.9) https://review.openstack.org/514646 | 21:39 |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: CLI allocation candidates (v1.10) https://review.openstack.org/514647 | 21:39 |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: New dict format of allocations (v1.11, v1.12) https://review.openstack.org/542819 | 21:39 |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: Transactionally update allocations (v1.13) https://review.openstack.org/546674 | 21:39 |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: Add nested resource providers (v1.14) https://review.openstack.org/546675 | 21:39 |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: Limit allocation candidates (v1.15, v1.16) https://review.openstack.org/548043 | 21:39 |
| openstackgerrit | Matt Riedemann proposed openstack/osc-placement master: Allocation candidates parameter: required (v1.17) https://review.openstack.org/548326 | 21:39 |
| mriedem | 1.7 should be ready to go ^ | 21:40 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Drop API compat handling for old compute error cases https://review.openstack.org/546256 | 21:41 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova master: ironic: Report resources as reserved when needed https://review.openstack.org/517921 | 21:41 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova master: ironic: Report resources as reserved when needed https://review.openstack.org/517921 | 21:41 |
| *** mriedem is now known as mriedem_afk | 21:41 | |
| openstackgerrit | Eric Fried proposed openstack/nova master: Update nova-status and docs for required placement 1.24 https://review.openstack.org/568923 | 21:54 |
| openstackgerrit | Eric Fried proposed openstack/nova master: Enforce placement minimum in nova.cmd.status https://review.openstack.org/568965 | 21:54 |
| efried | mriedem_afk: cdent: ^ | 21:54 |
| efried | That turned out to be not too bad at all. | 21:55 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova-specs master: Abort Cold Migration https://review.openstack.org/334732 | 21:57 |
| openstackgerrit | Merged openstack/nova master: Remove deprecated monkey_patch config options https://review.openstack.org/568880 | 22:38 |
| *** edmondsw has quit IRC | 22:50 | |
| *** edmondsw has joined #openstack-placement | 23:11 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Fix auth_url example in hypervisor-hyper-v.rst https://review.openstack.org/568983 | 23:40 |
| mriedem_afk | efried: nifty | 23:45 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!