| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Remove mox in libvirt/test_driver.py (5) https://review.openstack.org/570842 | 00:00 |
|---|---|---|
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Remove mox in libvirt/test_driver.py (6) https://review.openstack.org/571330 | 00:02 |
| openstackgerrit | Merged openstack/nova master: z/VM Driver: add power actions https://review.openstack.org/543340 | 00:35 |
| openstackgerrit | Merged openstack/nova master: z/VM Driver: add get console output https://review.openstack.org/543344 | 00:46 |
| openstackgerrit | Merged openstack/nova master: Adjust log style and remove ocata support https://review.openstack.org/576462 | 00:46 |
| openstackgerrit | Merged openstack/nova master: fix cellv2 delete_host https://review.openstack.org/582171 | 00:46 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Remove mox in libvirt/test_driver.py (7) https://review.openstack.org/571992 | 00:54 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Remove mox in libvirt/test_driver.py (8) https://review.openstack.org/571993 | 00:55 |
| *** mriedem has quit IRC | 01:36 | |
| *** lei-zh has joined #openstack-placement | 02:01 | |
| *** takashin has joined #openstack-placement | 02:04 | |
| *** tetsuro has joined #openstack-placement | 02:12 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Transform volume.usage notification https://review.openstack.org/580345 | 02:15 |
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Transform libvirt.error notification https://review.openstack.org/484851 | 02:36 |
| *** lei-zh has quit IRC | 02:57 | |
| *** lei-zh has joined #openstack-placement | 02:57 | |
| openstackgerrit | Takashi NATSUME proposed openstack/nova master: Add the guideline to write API reference https://review.openstack.org/569058 | 03:01 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Remove "DEPRECATED" tag from Obsolete APIs https://review.openstack.org/583092 | 03:20 |
| *** lei-zh has quit IRC | 03:23 | |
| *** lei-zh has joined #openstack-placement | 03:23 | |
| openstackgerrit | MultipleCrashes proposed openstack/nova master: Retry decorator fix for instances which go into ERROR state during bulk delete https://review.openstack.org/563418 | 04:40 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Remove "DEPRECATED" tag from Obsolete APIs https://review.openstack.org/583092 | 04:40 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Remove "DEPRECATED" tag from Obsolete APIs https://review.openstack.org/583092 | 04:42 |
| *** lei-zh has quit IRC | 04:48 | |
| *** lei-zh has joined #openstack-placement | 05:06 | |
| *** e0ne has joined #openstack-placement | 05:18 | |
| *** tetsuro has quit IRC | 05:19 | |
| *** e0ne has quit IRC | 05:20 | |
| *** tetsuro has joined #openstack-placement | 05:22 | |
| *** lei-zh has quit IRC | 05:33 | |
| *** lei-zh has joined #openstack-placement | 05:35 | |
| *** lei-zh1 has joined #openstack-placement | 05:57 | |
| *** lei-zh has quit IRC | 05:58 | |
| *** alex_xu has quit IRC | 05:59 | |
| *** alex_xu has joined #openstack-placement | 06:02 | |
| *** lei-zh- has joined #openstack-placement | 06:05 | |
| *** lei-zh1 has quit IRC | 06:06 | |
| openstackgerrit | Merged openstack/nova master: Address nits in server group policy series https://review.openstack.org/581616 | 06:23 |
| openstackgerrit | jichenjc proposed openstack/nova master: Enhance doc to guide user to use nova user https://review.openstack.org/583115 | 06:31 |
| *** alex_xu has quit IRC | 06:36 | |
| *** alex_xu has joined #openstack-placement | 06:36 | |
| *** lei-zh- has quit IRC | 06:38 | |
| *** lei-zh has joined #openstack-placement | 06:42 | |
| *** lei-zh has quit IRC | 06:53 | |
| *** lei-zh has joined #openstack-placement | 06:53 | |
| *** ttsiouts has joined #openstack-placement | 06:55 | |
| *** lei-zh has quit IRC | 07:01 | |
| *** lei-zh has joined #openstack-placement | 07:01 | |
| *** peereb has joined #openstack-placement | 07:04 | |
| *** ttsiouts has quit IRC | 07:20 | |
| *** ttsiouts has joined #openstack-placement | 07:59 | |
| *** e0ne has joined #openstack-placement | 08:05 | |
| *** ttsiouts has quit IRC | 08:16 | |
| *** ttsiouts has joined #openstack-placement | 08:19 | |
| *** takashin has left #openstack-placement | 08:31 | |
| *** tssurya has joined #openstack-placement | 08:35 | |
| *** alex_xu has quit IRC | 08:45 | |
| *** alex_xu has joined #openstack-placement | 08:46 | |
| *** cdent has joined #openstack-placement | 08:47 | |
| *** ttsiouts has quit IRC | 08:56 | |
| *** ttsiouts has joined #openstack-placement | 08:57 | |
| *** sean-k-mooney has joined #openstack-placement | 09:20 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Merge server create for scheduler hint extension https://review.openstack.org/579067 | 09:40 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Merge server create for multiple-create extension https://review.openstack.org/580017 | 09:41 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Merge server create for scheduler hint extension https://review.openstack.org/579067 | 09:48 |
| openstackgerrit | Stephen Finucane proposed openstack/nova master: scheduler: Start utilizing RequestSpec.network_metadata https://review.openstack.org/564452 | 09:51 |
| openstackgerrit | Stephen Finucane proposed openstack/nova master: Assorted cleanups from numa-aware-vswitches series https://review.openstack.org/582651 | 09:51 |
| *** lei-zh has quit IRC | 10:08 | |
| *** ttsiouts has quit IRC | 10:16 | |
| cdent | bbs | 10:21 |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Merge server create for multiple-create extension https://review.openstack.org/580017 | 10:53 |
| *** tetsuro has quit IRC | 11:01 | |
| *** ttsiouts has joined #openstack-placement | 11:16 | |
| openstackgerrit | Ghanshyam Mann proposed openstack/nova master: Merge server create for keypair extension https://review.openstack.org/583203 | 11:19 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Escalate UUID validation warning to error in test https://review.openstack.org/540386 | 11:26 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Fix a newly introduced UUID warning in the unit test https://review.openstack.org/583205 | 11:26 |
| openstackgerrit | Zhenyu Zheng proposed openstack/nova master: Address nit in 79dac41fee178dabb547f4d7bc10609630767131 https://review.openstack.org/583212 | 11:40 |
| *** edmondsw has joined #openstack-placement | 12:06 | |
| *** ttsiouts has quit IRC | 12:09 | |
| *** ttsiouts has joined #openstack-placement | 12:20 | |
| cdent | jaypipes, efried found the bug I was experiencing, left a comment about it on https://review.openstack.org/#/c/582383/ | 12:30 |
| jaypipes | looking | 12:47 |
| jaypipes | cdent: see line 4010 on my original submission: https://review.openstack.org/#/c/582383/1/nova/api/openstack/placement/objects/resource_provider.py | 12:49 |
| efried | Good one Chris. | 12:50 |
| efried | jaypipes: 4010? | 12:50 |
| jaypipes | efried: note that was addressed in my original | 12:51 |
| jaypipes | cdent, efried: sorry, 4101 | 12:51 |
| jaypipes | dyslexia sets in | 12:51 |
| efried | that makes more sense. | 12:51 |
| jaypipes | efried: you'd mentioned that wouldn't be necessary if we did the two-step in your revision. | 12:52 |
| jaypipes | efried: but maybe it's still needed? | 12:52 |
| efried | jaypipes: I didn't say that, but yes, I agree something like it is still needed. | 12:52 |
| efried | cdent, jaypipes: So we could stuff that same loop into the new function. But it's not quite that simple - we also need to *check* the rp generations in the allocations at some point before we just overwrite them. | 12:53 |
| efried | So like maybe we start off before we do anything else and make sure they match each other | 12:53 |
| efried | i.e. for each rp found in both the inventories side and the allocations side, make sure the generations are the same; if not, that's like a 400 I think. | 12:54 |
| efried | Once that check is out of the way, then yes, we can make sure the same rp objs are used in both the alloc and inv sides. | 12:54 |
| efried | and in fact, all of that should arguably be done from the handler. | 12:55 |
| efried | Because it's the handler that's responsible for converting the incoming dicts to objects. | 12:55 |
| jaypipes | efried: why do we need to check them? | 12:55 |
| efried | jaypipes: To make sure they're the same. If I pass an RP with gen 1 on the allocs side and the same RP with gen 2 on the inv side, I done f'ed up. | 12:55 |
| jaypipes | efried: just use the resource provider objects that were used in the first step. if they then come back as concurrent update detected, that's a *good* thing, right? | 12:56 |
| *** peereb has quit IRC | 12:56 | |
| jaypipes | efried: it means we prevented overwriting when someone made a change we didn't know about. | 12:56 |
| efried | jaypipes: In that scenario it would be a bad thing if they didn't raise concurrent update; but that doesn't mean it's always a good thing if they do. | 12:57 |
| jaypipes | efried: if we use the actual objects that were used in the first part of the operation (like my code did originally), we can trust nobody else messed with those object generations, right? and that means if we submit the final replace_all() and get back a concurrent update detected, that means the generation increment failed which is the correct protection we wanted (a *different* thread updated those records from our own). | 12:58 |
| jaypipes | efried: ok, walk me through what you're thinking w.r.t. the "good thing". | 12:58 |
| efried | jaypipes: I think we're saying the same thing. Reusing objects. Walking... | 12:58 |
| jaypipes | ah, ok | 12:58 |
| efried | 1) The handler first makes sure that, in the *dict* representations, the caller passed the same generations for a given RP that shows up in both the inv and alloc sides. This is a sanity check. The payload is simply invalid, bogus, garbled, nonsense if this constraint is not met. | 12:59 |
| efried | 2) The handler keeps a cache of rp objs it sees as it populates the inv object, and then reuses those as it populates the allocs object. | 12:59 |
| efried | 3) The db side method remains unchanged. | 13:00 |
| jaypipes | efried: isn't that basically what cdent's PSOT /reshaper code is already doing? https://review.openstack.org/#/c/576927/8/nova/api/openstack/placement/handlers/reshaper.py | 13:02 |
| jaypipes | efried: except instead of "keeping a cache of rps", it's just sending the Inventory objects that contain the providers that it already generation-checked. | 13:02 |
| jaypipes | efried: obviously, cdent's code doesn't "reuse those as it populates the allocs object", but that's basically what my original line 4101 code did in the reshape() function, so we could just re-use that, right? and keep that logic out of hte handler. | 13:05 |
| jaypipes | which I'm sure cdent would like ;) | 13:05 |
| jaypipes | (keeping code out of the handler) | 13:05 |
| efried | stand by, let me make sure I'm going to mean what I'm about to say... | 13:06 |
| cdent | cdent loves it when code is not in the handler | 13:07 |
| * cdent is catching up, was hanging laundry | 13:07 | |
| jaypipes | cdent: like a good Brit. well done. | 13:08 |
| cdent | also like a good brit, it should start raining on it any minute | 13:08 |
| jaypipes | heh | 13:09 |
| *** jaypipes is now known as was | 13:09 | |
| *** was is now known as jaypipes | 13:10 | |
| *** e0ne has quit IRC | 13:11 | |
| cdent | presumably as a first step, changing the test to not reuse objects is one way to "look like the real world" and then fixing things so the tests passes moves us along | 13:12 |
| efried | Putting aside the question of where the code will live (handler or db side) for now: | 13:13 |
| efried | 1) We still need the sanity check to make sure the RP gens on the inventory side are consistent with those on the allocs side. That is not happening anywhere now, in old code or new. | 13:13 |
| efried | 2) We can either a) reuse rp objects built while creating the InventoryList as we populate the AllocationList; or b) replace the rp objects in the AllocationList after the fact like 4101. | 13:13 |
| efried | For 2a we would need to change some method signatures (like passing a dict of already-retrieved rp objects into create_allocation_list). | 13:13 |
| efried | 2b seems a bit ugly and inefficient, because we already went through the work of creating rp objects that we're now going to throw away. | 13:13 |
| efried | In any case, the goal being that: | 13:13 |
| efried | 3) By the time the existing db-side code runs, the inv and alloc objects are sharing rp objects. | 13:13 |
| cdent | I would _very_ much prefer that the sharing only be inside the obj.rp.reshape method and that we not require it by the caller | 13:16 |
| jaypipes | cdent: ++ | 13:16 |
| efried | As for the question of handler or db side, today it seems like the handler side is responsible for converting payload dicts into objects; and the db side is responsible for using those objects to effect db changes. So if we go with 2a, it makes sense for it to live on the handler side fo sho, because it's part of *constructing* the objects... | 13:16 |
| jaypipes | efried: let us free cdent of this worry. | 13:16 |
| efried | ...so since both of you agree the sharing needs to be enforced on the db side, I guess that means 2b. | 13:16 |
| jaypipes | efried: I'm not entirely sure I agree it's important to check the alloc gens vs. the inv gens. | 13:17 |
| efried | ...which I still think *could* have been done on the handler side; but I get y'all's point, and it's a good reason to do it in the db side. | 13:17 |
| efried | jaypipes: Really? Then which would you go with? | 13:17 |
| cdent | efried: the allocations don't come with gens when sent as data. they only get them by loading them | 13:18 |
| cdent | so if we treat the inv gens as authoritative, then we get the protection you want | 13:18 |
| jaypipes | efried: I would go with 2b. I'm just saying I don't think we need a separate step that looks at the data sent in the allocations block's generations and makes sure those rp gens are identical to the rp gens used in the inventories block. | 13:18 |
| jaypipes | cdent: precisely. | 13:18 |
| efried | oh? hold... | 13:18 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Transform missing delete notifications https://review.openstack.org/410297 | 13:18 |
| openstackgerrit | Balazs Gibizer proposed openstack/nova master: Send soft_delete from context manager https://review.openstack.org/476459 | 13:18 |
| efried | The allocs do come with rp gens. | 13:19 |
| efried | oh, shit | 13:19 |
| efried | I hope the docs are wrong for POST /allocations | 13:20 |
| jaypipes | they are not. | 13:20 |
| efried | GET /allocations/{c} has the rp gen. | 13:20 |
| jaypipes | yes. | 13:20 |
| efried | And PUT /allocations/{c} includes it, but (at least according to the docs) ignores it. If we didn't change that in 1.28 to *not* ignore, we were supposed to. | 13:21 |
| cdent | no we were not | 13:21 |
| cdent | rp generations are not something we want to watch during allocations | 13:21 |
| cdent | that we don't is why we created consumer generations | 13:22 |
| cdent | we discussed this in dublin and it was quite clear that rp gens would not help (and would in fact hinder) | 13:22 |
| efried | okay okay, I'm remembering now - and was indeed confusing rp gens with consumer gens. | 13:23 |
| cdent | when we write an allocation we are _not_ validating that our view of inventory hasn't change | 13:23 |
| jaypipes | right. | 13:23 |
| cdent | we don't want to do that because allocations happen often | 13:23 |
| jaypipes | right, and all we want to ensure is that capacity isn't exceeded, not that nothing has changed about the provider at all. | 13:23 |
| efried | then... why do we increment provider gens when we change allocs? | 13:23 |
| jaypipes | efried: to indicate there has been a change. | 13:24 |
| efried | you just said two opposite things. | 13:24 |
| cdent | no, we didn't | 13:24 |
| jaypipes | no I didn't :) | 13:24 |
| cdent | brb | 13:24 |
| cdent | the difference is in who the observer is | 13:26 |
| efried | Here's my confusion: it seems like we should either be using rp gens when we do alloc operations, or not. It sounds like we're incrementing, but not validating, which seems inconsistent and wrong. | 13:26 |
| cdent | a thing making allocations is not the same observer as the thing changing rps and inventories | 13:27 |
| cdent | plus, we need them for the compare and swap | 13:27 |
| cdent | gens provide two purposes | 13:27 |
| cdent | one for the remote callers consistency of view etc | 13:27 |
| cdent | the other for (to quote jay) http://www.joinfu.com/2015/01/understanding-reservations-concurrency-locking-in-nova/ | 13:28 |
| cdent | (please correct me if I just got all that wrong jaypipes ) | 13:28 |
| openstackgerrit | Vladyslav Drok proposed openstack/nova master: ironic: Report resources as reserved when needed https://review.openstack.org/517921 | 13:28 |
| openstackgerrit | Merged openstack/nova master: objects: Add NUMACell.network_metadata https://review.openstack.org/564439 | 13:30 |
| jaypipes | cdent: no, that's correct. | 13:31 |
| jaypipes | efried: shall I hack on this patch and add back the line 4101 thing? and add tests for this as cdent has asked for? | 13:32 |
| efried | jaypipes: Sounds like a plan, man. | 13:33 |
| jaypipes | efried: k, will do. | 13:33 |
| cdent | yay, a plan | 13:34 |
| jaypipes | cdent: I am *only* pulling the db trx patch, not anything above it. ok with you? | 13:34 |
| jaypipes | (i.e. I'm not going to rebase the entire series0 | 13:34 |
| cdent | yes | 13:34 |
| jaypipes | k | 13:34 |
| jaypipes | just checking. | 13:34 |
| cdent | yeah, wise, it's bound to cause some confusion | 13:35 |
| jaypipes | cdent: FYI, I have a patch series that breaks up resource_provider.py into multiple files. waiting to push it until we're done with this surgery, but you should like it when it finally does get pushed. | 13:36 |
| cdent | huzzah | 13:36 |
| *** mriedem has joined #openstack-placement | 13:51 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Add unshelve instance error info to fault table https://review.openstack.org/579747 | 13:51 |
| openstackgerrit | sahid proposed openstack/nova master: libvirt: remove unused attribute driver for LibvirtConfigNodeDevice https://review.openstack.org/583246 | 13:59 |
| openstackgerrit | sahid proposed openstack/nova master: libvirt: add format_dom for ConfigNodeDevicePciSubFunctionCap https://review.openstack.org/583247 | 13:59 |
| openstackgerrit | sahid proposed openstack/nova master: libvirt: add missing type for LibvirtConfigNodeDevicePciCap https://review.openstack.org/583248 | 13:59 |
| openstackgerrit | sahid proposed openstack/nova master: libvirt: add format_dom() for config LibvirtConfigNodeDevicePciCap https://review.openstack.org/583249 | 13:59 |
| openstackgerrit | sahid proposed openstack/nova master: libvirt: add format_dom() method for LibvirtConfigNodeDevice https://review.openstack.org/583250 | 13:59 |
| openstackgerrit | sahid proposed openstack/nova master: libvirt: add generater of fake pci devices capabilities https://review.openstack.org/583251 | 13:59 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Call generate_image_url only for legacy notification https://review.openstack.org/564528 | 14:01 |
| openstackgerrit | Merged openstack/nova master: Replace support matrix ext with common library https://review.openstack.org/481304 | 14:25 |
| openstackgerrit | Jay Pipes proposed openstack/nova master: perform reshaper operations in single transaction https://review.openstack.org/582383 | 14:27 |
| jaypipes | cdent, efried: ^ | 14:28 |
| * cdent looks | 14:28 | |
| openstackgerrit | Chris Dent proposed openstack/nova master: WIP: [placement] Add /reshaper handler for POST https://review.openstack.org/576927 | 14:30 |
| cdent | jaypipes: fixes the issue I was seeing | 14:31 |
| cdent | I have successfully reshaped. | 14:32 |
| cdent | I'll carry on, and add some more tests a bit later in the day | 14:32 |
| jaypipes | cdent: cool. obviously, I can't do gabbit tests for reshape in that patch (since the handler code builds on top of that patch) | 14:33 |
| * cdent nods | 14:33 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Fix "XLibvirt KVM (ppc64)" typo in feature support matrix docs https://review.openstack.org/583267 | 14:37 |
| gibi | cdent, efried: FYI I slowly but making progress locally to change the report client to use 1.28 for allocations. I hope I will have something to push this week | 15:02 |
| *** mriedem has left #openstack-placement | 15:04 | |
| *** mriedem has joined #openstack-placement | 15:05 | |
| jaypipes | is it Friday yet? | 15:06 |
| gibi | jaypipes: that would be soo nice, but unfortunately not. | 15:07 |
| openstackgerrit | Eric Fried proposed openstack/nova master: Fix "XLibvirt KVM (ppc64)" typo in feature support matrix docs https://review.openstack.org/583267 | 15:08 |
| cdent | gibi++ | 15:09 |
| *** ttsiouts has quit IRC | 15:09 | |
| efried | gibi: Sweet dude. | 15:11 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Transform metrics.update notification https://review.openstack.org/580567 | 15:24 |
| openstackgerrit | Stephen Finucane proposed openstack/nova master: WIP: Add functional tests for numa-aware-vswitches https://review.openstack.org/583288 | 15:28 |
| *** ttsiouts has joined #openstack-placement | 15:48 | |
| *** ttsiouts has quit IRC | 15:48 | |
| *** ttsiouts has joined #openstack-placement | 15:49 | |
| *** ttsiouts has quit IRC | 15:54 | |
| jaypipes | efried: I can haz factor out those common LOC in a followup patch plz? | 16:05 |
| efried | jaypipes: sho. Are we looking to merge this right away, or wait until more of the series has matured so we can flush out more bugs? | 16:05 |
| jaypipes | efried: up to you and cdent. | 16:06 |
| cdent | jaypipes: fup fine with me | 16:11 |
| cdent | I reckon we should hold off merging the bottom until we've had a successful reshape from the resource client? | 16:11 |
| jaypipes | cdent: that's fine with me, sure | 16:13 |
| cdent | s/resource/report/ | 16:14 |
| cdent | all words are the same in my mind | 16:14 |
| efried | So at this point, jaypipes if you're done messing with the bottom, have you had a chance to look at the client patch yet? | 16:14 |
| jaypipes | efried: I have not, no. Is there a patch up for that? | 16:14 |
| efried | It's still needing tests, but I want to make sure the broad strokes of the algo are good with y'all before sinking a bunch of time into that. | 16:14 |
| efried | jaypipes: https://review.openstack.org/#/c/576236/ | 16:15 |
| * cdent enqueue | 16:15 | |
| *** tssurya has quit IRC | 16:19 | |
| *** e0ne has joined #openstack-placement | 16:31 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Remove support for /os-floating-ip-dns REST API https://review.openstack.org/582943 | 16:35 |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Remove "DEPRECATED" tag from Obsolete APIs https://review.openstack.org/583092 | 16:35 |
| openstackgerrit | Merged openstack/nova master: Fix a newly introduced UUID warning in the unit test https://review.openstack.org/583205 | 17:14 |
| *** e0ne has quit IRC | 17:15 | |
| *** e0ne has joined #openstack-placement | 17:21 | |
| *** diga has joined #openstack-placement | 17:25 | |
| diga | Hello guys | 17:37 |
| openstackgerrit | Merged openstack/nova master: Escalate UUID validation warning to error in test https://review.openstack.org/540386 | 17:38 |
| efried | diga: Howdy. What's up? | 17:39 |
| diga | cdent: efried : want to start work on placement, can you give me to get started ? | 17:39 |
| diga | efried: Hey | 17:39 |
| diga | efried: Do you have some work item for me ? | 17:40 |
| * efried looks for bugs... | 17:40 | |
| diga | okay | 17:41 |
| efried | diga: Here's the open placement bugs: https://bugs.launchpad.net/nova/+bugs?field.tag=placement&orderby=-id | 17:42 |
| efried | diga: Make sure you pick one that doesn't already have an owner. | 17:42 |
| diga | efried: sure | 17:42 |
| efried | diga: Let me know if you want help/guidance. | 17:42 |
| diga | efried: yup | 17:43 |
| efried | diga: You can also review open patches. | 17:43 |
| diga | efried: okay | 17:43 |
| diga | efried: can you post link here ? | 17:44 |
| diga | will review it | 17:44 |
| efried | diga: The easiest way to find reviews is to follow the placement update that gets sent out to the dev mailing list weekly. | 17:44 |
| diga | yeah | 17:45 |
| efried | It may get a bit stale over the course of the week, but it ought to be a good starting point. | 17:45 |
| efried | diga: Here's the latest one: http://lists.openstack.org/pipermail/openstack-dev/2018-July/132252.html | 17:45 |
| diga | I have that mail, will pick it up from there | 17:45 |
| diga | efried: okay | 17:45 |
| efried | Great. We really appreciate the help! | 17:45 |
| diga | efried: :) I will try my best friend | 17:48 |
| *** diga has quit IRC | 17:52 | |
| openstackgerrit | Eric Fried proposed openstack/nova master: WIP: Compute: Handle reshaped provider trees https://review.openstack.org/576236 | 18:00 |
| efried | cdent: Wanted to rebase to get test results, but it looks like the microversion patch is in merge conflict. May I fix it? | 18:01 |
| efried | jaypipes: ...by rebasing the whole series? | 18:01 |
| jaypipes | efried: please go ahead | 18:03 |
| jaypipes | efried: if cool with cdent | 18:03 |
| efried | ready to push on your nod cdent | 18:05 |
| *** e0ne has quit IRC | 18:15 | |
| cdent | go for it efried, was dining | 18:32 |
| openstackgerrit | Eric Fried proposed openstack/nova master: perform reshaper operations in single transaction https://review.openstack.org/582383 | 18:32 |
| openstackgerrit | Eric Fried proposed openstack/nova master: WIP: [placement] Add /reshaper handler for POST https://review.openstack.org/576927 | 18:32 |
| openstackgerrit | Eric Fried proposed openstack/nova master: WIP: Compute: Handle reshaped provider trees https://review.openstack.org/576236 | 18:32 |
| efried | done ^ | 18:32 |
| efried | didn't want to eff with any local edits you had cdent | 18:32 |
| cdent | efried: I think I'm hands off until tomorrow morning, so if you need to do it again, feel free | 18:33 |
| efried | rgr | 18:33 |
| *** cdent has quit IRC | 18:42 | |
| openstackgerrit | sean mooney proposed openstack/nova master: update tox venv env to install all requirements https://review.openstack.org/583336 | 18:48 |
| *** e0ne has joined #openstack-placement | 18:49 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: WIP: Add regression test for bug 1781710 https://review.openstack.org/583339 | 18:53 |
| openstack | bug 1781710 in OpenStack Compute (nova) "ServersOnMultiNodesTest.test_create_server_with_scheduler_hint_group_anti_affinity failing with "Servers are on the same host"" [High,Fix released] https://launchpad.net/bugs/1781710 - Assigned to Matt Riedemann (mriedem) | 18:53 |
| *** efried has quit IRC | 19:14 | |
| *** efried has joined #openstack-placement | 19:14 | |
| openstackgerrit | Merged openstack/nova master: Merge server create for scheduler hint extension https://review.openstack.org/579067 | 19:35 |
| openstackgerrit | Merged openstack/nova master: Fix "XLibvirt KVM (ppc64)" typo in feature support matrix docs https://review.openstack.org/583267 | 19:35 |
| *** e0ne has quit IRC | 19:41 | |
| openstackgerrit | sean mooney proposed openstack/nova master: random hacking for bug 1781710 https://review.openstack.org/583347 | 19:55 |
| openstack | bug 1781710 in OpenStack Compute (nova) "ServersOnMultiNodesTest.test_create_server_with_scheduler_hint_group_anti_affinity failing with "Servers are on the same host"" [High,Fix released] https://launchpad.net/bugs/1781710 - Assigned to Matt Riedemann (mriedem) | 19:55 |
| openstackgerrit | Merged openstack/nova master: Add unshelve instance error info to fault table https://review.openstack.org/579747 | 19:59 |
| *** tssurya has joined #openstack-placement | 20:04 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Additional/alternative fix for bug 1781710 https://review.openstack.org/583351 | 20:20 |
| openstack | bug 1781710 in OpenStack Compute (nova) "ServersOnMultiNodesTest.test_create_server_with_scheduler_hint_group_anti_affinity failing with "Servers are on the same host"" [High,Fix released] https://launchpad.net/bugs/1781710 - Assigned to Matt Riedemann (mriedem) | 20:20 |
| *** tssurya has quit IRC | 20:23 | |
| efried | jaypipes: When we do yaml-based generic device passthrough, I anticipate there being a config file at the conductor (which can specify device types at the level of vendor/product ID) whose entries can be superseded by a file per compute (which can specify the same, but can also get down to specific devices by address etc. if needed). | 20:56 |
| efried | jaypipes: Are you okay if we strawman that conf file name as /etc/nova/passthrough.yaml ? | 20:56 |
| efried | The one on the compute can/will be virt type-specific, but no reason for it to have a different name, is there? | 20:57 |
| jaypipes | efried: /etc/nova/devices.yaml or /etc/nova/inventory.yaml would be my preference. | 21:26 |
| jaypipes | /etc/nova/inventory.yaml being what I had originally proposed I believe... | 21:26 |
| efried | jaypipes: inventory.yaml is pretty generic; but maybe that's your intent. | 21:26 |
| jaypipes | efried: ack, yeah it was my intent to be generic. | 21:26 |
| efried | i.e. can potentially encompass things like VCPU... | 21:26 |
| efried | which may be how NUMA topology gets specified, or something, dunno. | 21:27 |
| efried | anyway, dig. | 21:27 |
| openstackgerrit | Merged openstack/nova master: Call generate_image_url only for legacy notification https://review.openstack.org/564528 | 21:29 |
| efried | jaypipes: Have you thought about how to reconcile generic content vs. virt-specific content in such a file? | 21:30 |
| *** edmondsw has quit IRC | 21:36 | |
| *** edmondsw_ has joined #openstack-placement | 21:45 | |
| *** edmondsw_ has quit IRC | 21:49 | |
| openstackgerrit | Matt Riedemann proposed openstack/nova master: Update RequestSpec.instance_uuid during scheduling https://review.openstack.org/583347 | 21:58 |
| jaypipes | efried: to me it's all about the node's inventory and traits... | 22:13 |
| jaypipes | efried: could have all that information in a single YAML file IMHO. | 22:13 |
| jaypipes | efried: but I need to run off for the day now... | 22:13 |
| efried | ight, later. | 22:13 |
| jaypipes | talk to you tomorrow. | 22:13 |
| jaypipes | :) | 22:13 |
| openstackgerrit | Eric Fried proposed openstack/nova master: WIP: Compute: Handle reshaped provider trees https://review.openstack.org/576236 | 22:38 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!