| openstackgerrit | Chris Dent proposed openstack/osc-placement master: Update tox and tests to work with modern setups https://review.openstack.org/639717 | 00:06 |
|---|---|---|
| openstackgerrit | Chris Dent proposed openstack/osc-placement master: Add support for 1.18 microversion https://review.openstack.org/639738 | 00:06 |
| *** mriedem is now known as mriedem_away | 00:26 | |
| *** ttsiouts has quit IRC | 01:17 | |
| *** mriedem_away has quit IRC | 02:40 | |
| *** bhagyashris has joined #openstack-placement | 03:58 | |
| openstackgerrit | Eric Fried proposed openstack/osc-placement master: Update tox and tests to work with modern setups https://review.openstack.org/639717 | 04:12 |
| openstackgerrit | Eric Fried proposed openstack/osc-placement master: Add support for 1.18 microversion https://review.openstack.org/639738 | 04:15 |
| *** tetsuro has joined #openstack-placement | 04:53 | |
| openstackgerrit | Merged openstack/osc-placement master: Update tox and tests to work with modern setups https://review.openstack.org/639717 | 05:26 |
| *** tetsuro has quit IRC | 06:20 | |
| *** e0ne has joined #openstack-placement | 07:28 | |
| *** tetsuro has joined #openstack-placement | 07:28 | |
| *** takashin has left #openstack-placement | 07:36 | |
| *** rubasov has quit IRC | 08:00 | |
| *** tetsuro has quit IRC | 08:01 | |
| *** tssurya has joined #openstack-placement | 08:25 | |
| *** mcgiggler has joined #openstack-placement | 08:39 | |
| *** ttsiouts has joined #openstack-placement | 08:56 | |
| *** rubasov has joined #openstack-placement | 09:00 | |
| *** bhagyashris has quit IRC | 09:49 | |
| *** rubasov has quit IRC | 10:07 | |
| *** rubasov has joined #openstack-placement | 10:09 | |
| *** cdent has joined #openstack-placement | 10:15 | |
| *** e0ne has quit IRC | 10:22 | |
| cdent | mornin | 10:23 |
| *** e0ne has joined #openstack-placement | 10:23 | |
| *** e0ne has quit IRC | 11:11 | |
| *** cdent has quit IRC | 11:12 | |
| *** e0ne has joined #openstack-placement | 11:14 | |
| *** ttsiouts has quit IRC | 11:18 | |
| *** cdent has joined #openstack-placement | 11:20 | |
| openstackgerrit | Tetsuro Nakamura proposed openstack/placement master: Refactor _get_trees_matching_all() https://review.openstack.org/639888 | 11:40 |
| openstackgerrit | Tetsuro Nakamura proposed openstack/placement master: Adds debug log in allocation candidates https://review.openstack.org/639889 | 11:40 |
| *** e0ne has quit IRC | 11:49 | |
| *** ttsiouts has joined #openstack-placement | 11:51 | |
| *** gibi is now known as giblet | 12:45 | |
| *** e0ne has joined #openstack-placement | 12:50 | |
| cdent | efried: that ↑ is on my list, at first glance there's a fair bit of difference between the other *LIst classes and the one introduced there, so probably fine | 13:06 |
| cdent | also, did you not go to bed, or stay up late, or get up early or some other kind of weird sleep | 13:07 |
| cdent | I put a special section in the pupdate just for mriedem | 13:19 |
| efried | cdent: Yes, the RPCandidates list thingy is definitely not like our ex-OVO lists. I think it adds value. I'm not sure it's perfect in its current form, but it's better than what was there, as described in the commit message (p[N] list comps everywhere). | 13:30 |
| efried | cdent: and, I've started doing more reviews on my phone, to try to keep up a bit better; so I go through a bunch of my openstack mail at odd times. | 13:31 |
| cdent | gerrit? on your phone? Or are you viewing with something else? | 13:32 |
| *** mriedem has joined #openstack-placement | 13:32 | |
| efried | yeah, just review.o.o on safari on iphone. | 13:33 |
| cdent | huh. that sounds painful. | 13:34 |
| efried | I don't try to do it for large or complicated reviews, because I make no attempt to go look at code repositories there. But for small changes or incremental patch set updates it's okay. | 13:34 |
| efried | Main objective is to delete emails. | 13:34 |
| efried | which makes my time in front of real computer more productive | 13:35 |
| sean-k-mooney | cdent: what did ye replace OVOs with? json schema? | 13:36 |
| efried | sean-k-mooney: PODs | 13:36 |
| cdent | sean-k-mooney: python. we don't need rpc in placement | 13:37 |
| sean-k-mooney | ah os no validation | 13:37 |
| efried | except the lists, which are lists. | 13:37 |
| sean-k-mooney | ovo are not just for rpc | 13:37 |
| cdent | the validation happens at the http api level, so yes, jsonschema, but that was already there | 13:37 |
| sean-k-mooney | but ok | 13:37 |
| efried | more ovorhead than we needed | 13:37 |
| efried | see what I did there | 13:37 |
| cdent | smh | 13:37 |
| sean-k-mooney | we have discussed the idea of dropping ovo in os-vif but we still wanted a schema validation layer which is why i asked | 13:38 |
| sean-k-mooney | for now we are sticking with ovos | 13:38 |
| cdent | sean-k-mooney: the reason we used ovo in the first place was because of a "maybe someday there will be an rpc or other kind of objecty interface", but enough time has passed to say "no" to that | 13:38 |
| cdent | I think if you have an object interface that needs validation within that interface, ovo can make some sense | 13:39 |
| sean-k-mooney | ya that was/is on the cards for os-vif but aparenetly peopel dont like the idea of ovo over the rest api | 13:39 |
| sean-k-mooney | which if that is not going to happen in the future for os-vif raise the question do we need ovo at all | 13:40 |
| sean-k-mooney | any thank back to reading your placement update | 13:40 |
| * cdent nods | 13:40 | |
| cdent | The performance characteristics for the change measured there is probably not normal for many uses of ovo | 13:41 |
| sean-k-mooney | * anyway thanks | 13:41 |
| sean-k-mooney | ya maybe | 13:41 |
| cdent | that's a response of /allocation_candidates from a set of 1000 resource providers | 13:41 |
| cdent | and with multiple objects created for each provider | 13:41 |
| sean-k-mooney | not needint to add backleveling code would be nice however | 13:41 |
| sean-k-mooney | ovo have a lot of developer over head too | 13:41 |
| efried | cdent: I don't think in_tree[N] was for bandwidth. It was for things like force-host and resize-to-same-host. | 13:43 |
| giblet | efried: in_tree in a_c can be used to attach ports with bandwidth to a running server | 13:43 |
| cdent | efried: please say so. It's hard to remember everything when there isn't a consistent thread of dialog (which is the real (unfulfilled) reason that I do the pupdates) | 13:44 |
| sean-k-mooney | giblet: it can be but the motivating usecase wes for fost-host | 13:44 |
| efried | giblet: Ah, yeah, that makes sense. | 13:45 |
| efried | and... | 13:45 |
| *** efried is now known as fried_rice | 13:45 | |
| sean-k-mooney | fried_rice: beyond traditon do you thing the friday nics actully work to reduce the amount of pings cores get on fridays so they can focus on getting things done | 13:46 |
| sean-k-mooney | fried_rice: i still like it but most people who are active on irc know the firday nics too | 13:47 |
| cdent | was that the reason? | 13:47 |
| cdent | I thought it was to create an exclusionary and elitist club? | 13:47 |
| cdent | A fraternity of sorts | 13:47 |
| giblet | casual friday maybe | 13:48 |
| sean-k-mooney | i think it stared when jaypipes or one of the cores said they were going into bunker mode to get something done and it just became a thing | 13:49 |
| sean-k-mooney | its been long enough that i dont remember the details | 13:49 |
| * cdent thinks maybe we should clone both tetsuro and giblet | 13:53 | |
| edleafe | It started as a joke about "casual Fridays". Since we're virtual, we had "casual nick Fridays" | 13:57 |
| edleafe | But since then it has morphed to more of an insider thing, which is why I stopped using mine. | 13:57 |
| * edleafe runs off to PT | 13:58 | |
| fried_rice | cdent: I did it to play along, didn't really give it any deep thought. | 14:04 |
| cdent | I'm 1/3 joking above, but I'm definitely aware of loads of people who have found it exclusionary | 14:04 |
| fried_rice | Having experienced it first as an outsider, I agree it was confusing. But it's not, like, secret or anything, and it doesn't take much to figure out, and (I think) it's "fun". | 14:05 |
| fried_rice | it's not exclusionary - anyone can have a friday nick. | 14:06 |
| cdent | our experience isn't the one that really matters here. it's the experience of people who have felt excluded by it | 14:06 |
| cdent | and I take people at their word when me tell me that | 14:06 |
| fried_rice | yeah, fair enough. | 14:06 |
| * cdent goes to check why that osc-placement thing failed | 14:08 | |
| *** Sundar has joined #openstack-placement | 14:08 | |
| fried_rice | cdent: ^ probably need a global trait cache :P | 14:08 |
| fried_rice | o/ Sundar | 14:08 |
| * cdent kills all caches | 14:09 | |
| Sundar | o/ Eric | 14:09 |
| Sundar | Hi cdent | 14:09 |
| cdent | I did try at one point to get rid of the string_to_id stuff but it was too embedded | 14:09 |
| cdent | hi Sundar | 14:09 |
| Sundar | Did i stumble into some involved talk on trait caches? | 14:10 |
| cdent | thankfully no | 14:12 |
| cdent | i was being taunted by an overcooked bowl of fried_rice | 14:12 |
| fried_rice | overcooked? overcooked?! | 14:13 |
| fried_rice | I guess it could be worse, I could be half-baked. | 14:13 |
| Sundar | cdent: Thanks very much for taking the time to review. | 14:15 |
| Sundar | cdent: Unfortunately -- for you -- I have more responses and questions. If you don't mind. | 14:16 |
| *** mcgiggler has quit IRC | 14:17 | |
| cdent | Sundar: no problem, can you remind me of the etherpad link please? | 14:18 |
| fried_rice | cdent: #link https://etherpad.openstack.org/p/nova-cyborg-api-objects | 14:20 |
| cdent | thanks | 14:20 |
| cdent | Sundar: I'll respond on the etherpad | 14:21 |
| Sundar | cdent: One question to be highlighted: the APIs return a links field, with self and other links. That is of the form /resource/uuid for a specific object. Could I return /resource?property=value ? | 14:30 |
| fried_rice | hm, it seems like it would be weird for a link to contain a querystring | 14:31 |
| cdent | it does seem weird, but strictly speaking it is not wrong. From the web's standpoint the URI is the entire string | 14:31 |
| fried_rice | is it still not wrong if that link returned multiple resources? | 14:32 |
| Sundar | fried_rice: We return the URI for a specific object in the links. | 14:32 |
| fried_rice | right, I get it, but to me ?property=value generically looks like a search which might have the potential to return multiple results. For a specific key I get that it would be unique. | 14:34 |
| Sundar | fried_rice: i..e if you ask GET /device_profiles?name=mydp, the response contains a URI for that. Whereas, if you ask GET /device_profiles, I believe there is a generic href but not for any specific object or a collection. | 14:34 |
| fried_rice | yeah, so as long as the API has a constraint that device_profile.name is globally unique, that ^ returns at most one result. | 14:34 |
| cdent | I think the question needs to be around wether the structure of the response is of one resource or a list of one resource | 14:37 |
| cdent | if it is a list that's probably not ideal | 14:37 |
| cdent | but, however | 14:37 |
| cdent | since it appears to be a filtering operation it should be a list | 14:37 |
| cdent | that's why I think it is important to have /device_profiles?filter and /device_profiles/{uuid} be different things with different structures | 14:38 |
| cdent | /device_profiles?{filter} should probalby always be a list | 14:38 |
| cdent | The key thing here is to have a consistent grammar | 14:38 |
| cdent | (which I think the document pretty much does, so mostly we are talking about the future, not about the now) | 14:39 |
| fried_rice | this is what I was getting at, yes | 14:41 |
| Sundar | cdent: IIUC, based on your comments in the doc and here, you seem to be kinda-sorta ok with the current API? If I add the proviso that we have versioning (and possibly microversioning), could we keep it as is *for now*? We could evolve it in the future based on operator feedback, but just trying to avoid having too many APIs to start with. | 14:43 |
| fried_rice | s/possibly// | 14:44 |
| fried_rice | IMO setting up for microversioning is a must | 14:44 |
| fried_rice | trying to use major versions is too much of a PITA, which is the whole reason microversions exist. | 14:45 |
| cdent | Sundar: yes, as I said at the start: no red flags, just some things to think about as things evolve. As far as I can tell you're not painting yourself into bad corners | 14:45 |
| Sundar | fired_rice: ha -- I expected that! I am open to it. Seems like a good idea. It's just me being paranoid about what we are going to get done. | 14:45 |
| Sundar | fried_rice: ^ | 14:46 |
| cdent | as fried_rice and I discussed yesterday, you can lay in a basic frame for future microversions without too much difficulty | 14:46 |
| cdent | however, one thing we didn't talk about is that I sometimes wish we had put off stamping the major version in placement until a bit later | 14:46 |
| cdent | placement doesn't really "work" until about 1.11 or so | 14:47 |
| fried_rice | in this case we're intending v2.0 of the cyborg API will "work" for the initial pass of the nova integration code. | 14:48 |
| cdent | I'm glad we're all nice and conscious of fragile notion of our work "working" | 14:49 |
| openstackgerrit | Chris Dent proposed openstack/osc-placement master: Add support for 1.18 microversion https://review.openstack.org/639738 | 14:50 |
| cdent | fried_rice, mriedem ↑ should fix the intermittent failure but I haven't tested it yet as I don't have access to a cloud from where I am right now | 14:50 |
| Sundar | cdent, fried_rice: Cyborg folks already released v1 APIs in Rocky. So, we have to start with v2. So, yes, v2.0 would be the first to work with Nova. What is the equivalent of "put off stamping the major version"? | 14:51 |
| fried_rice | sort of. | 14:52 |
| cdent | I should clarify: we built microversioning into placement into the very start, and versioned changes to the api even before anything was using placement. | 14:52 |
| cdent | What I'm saying is that I think that was overkill | 14:52 |
| cdent | What we should have done was started versioning at the point at which there was a first client | 14:53 |
| cdent | (declared that moment in time 1.0) | 14:53 |
| fried_rice | cdent: what you're saying is, we could have just merged API changes to placement without bumping microversions, calling the whole API a WIP, until we got to the point where it was usable, at which point we could have said "no longer a WIP" and called that v1.0. | 14:53 |
| cdent | yes | 14:53 |
| fried_rice | mm | 14:53 |
| cdent | that's only okay to do when there are no users | 14:54 |
| cdent | which was true initially | 14:54 |
| cdent | melwitt: this change to use storyboard will probaby need your stamp: https://review.openstack.org/#/c/639445/ | 14:57 |
| *** jaypipes is now known as leakypipes | 15:01 | |
| melwitt | cdent: ok, will look | 15:02 |
| cdent | thanks | 15:02 |
| *** Sundar is now known as DarthNau | 15:03 | |
| melwitt | does anyone know how to run osc-placement functional tests locally in a devstack? when I do a "source openrc admin admin" and then "tox -efunctional" I get weird errors like "'too few arguments' not in u"openstack: 'resource provider inventory set' is not an openstack command." | 15:31 |
| cdent | melwitt: are you up to date with the recently mered (yesterday) fixes? | 15:35 |
| cdent | there were lots of python3 problems that ought to be fixed | 15:35 |
| melwitt | maybe not | 15:35 |
| melwitt | ok, let me get them. thanks | 15:35 |
| *** e0ne has quit IRC | 15:38 | |
| cdent | melwitt: are you working in python3 or 2? If 2, then something else is likely wrong. But if 3, then yeah, it never worked (until yesterday`) | 15:39 |
| *** belmoreira has quit IRC | 15:39 | |
| *** e0ne has joined #openstack-placement | 15:39 | |
| melwitt | cdent: python3. working now, thank you | 15:42 |
| cdent | huzzah! | 15:42 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Factor listiness into an ObjectList base class https://review.openstack.org/637325 | 15:52 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Move _set_objects into ObjectList https://review.openstack.org/637328 | 15:52 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Move *List.__repr__ into ObjectList https://review.openstack.org/637332 | 15:52 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Clean up ObjectList._set_objects signature https://review.openstack.org/637335 | 15:52 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Move RC_CACHE in resource_class_cache https://review.openstack.org/640114 | 15:52 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Use native list for lists of Usage https://review.openstack.org/639391 | 15:52 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Move Allocation and AllocationList to own module https://review.openstack.org/640184 | 15:52 |
| openstackgerrit | Chris Dent proposed openstack/placement master: Make base test case file for object unit tests https://review.openstack.org/640406 | 15:52 |
| *** tssurya has quit IRC | 15:59 | |
| leakypipes | edleafe, sean-k-mooney, cdent: if one cannot have a little fun on Fridays, then what kind of world are we living in? :) | 16:01 |
| cdent | a dark and dismal one where brexit and trump exist? | 16:02 |
| sean-k-mooney | ouch | 16:03 |
| sean-k-mooney | cdent: your wit is sharp today | 16:04 |
| cdent | coffee is working? | 16:04 |
| sean-k-mooney | perhaps. politics is depressing currently? always? | 16:05 |
| cdent | I think it is because dr fried_rice sent me to get antibiotics and they are working | 16:05 |
| * fried_rice tries to decide whether he wants that credit/blame | 16:06 | |
| cdent | life is hard | 16:10 |
| *** e0ne has quit IRC | 16:14 | |
| *** e0ne has joined #openstack-placement | 16:15 | |
| *** e0ne has quit IRC | 16:17 | |
| fried_rice | cdent: https://review.openstack.org/#/c/640184/ -- I can take a whack at factoring increment_provider_generation and slotting it in under that patch if you like. | 16:24 |
| fried_rice | factoring out* | 16:24 |
| fried_rice | unless you're already embroiled | 16:24 |
| cdent | fried_rice: I'm not currently doing that, I'm doing the next step: killing AllocationList | 16:25 |
| cdent | the exist of which is part of why I'm not too fussed about that ^ one being super clean (although I haven't read your commets yet) | 16:26 |
| fried_rice | cdent: okay, then I need to avoid changing the commit order. I'll put it on top for now and we can slot it in later (or not) | 16:26 |
| cdent | cool | 16:26 |
| fried_rice | cdent: I did note the one non-clean place that I figured was going away in 'kill AllocationList' - I'm okay leaving that one alone. The others (indents broken by renaming module) I think should be fixed here. | 16:27 |
| cdent | I'm not clear how a cut and paste can break an indent? | 16:27 |
| fried_rice | because you renamed a module, and the new name had a different length | 16:28 |
| fried_rice | so param lists aligned on that opening paren were shifted. | 16:28 |
| fried_rice | ...which I suppose would be one reason to avoid that particular indent style in the first place, if one were inclined to be so anal at coding time. Cause it's not a pep violation until the refactor. (And at the moment it's not a pep violation at all because we're ignoring that rule.) | 16:29 |
| openstackgerrit | Merged openstack/osc-placement master: Add support for 1.18 microversion https://review.openstack.org/639738 | 16:30 |
| cdent | oh | 16:31 |
| fried_rice | leakypipes: cdent: uhm, how do you feel about _increment_provider_generation(ctx, rp) being ResourceProvider.increment_generation(self) (using self._context) ? | 16:32 |
| cdent | fried_rice: how would you feel if we stop doing that horrible indent style that I keep breaking? | 16:32 |
| cdent | fried_rice: I'd be somewhat meh about that because it is not _that_ resource providers generation we are changing | 16:33 |
| fried_rice | I would be fine with it, though it would be challenging to (remember to) enforce. I also suggested inline that we switch off some of those pep ignores. | 16:33 |
| cdent | is it? | 16:33 |
| fried_rice | yes, absolutely. | 16:33 |
| * cdent defers to leakypipes | 16:34 | |
| cdent | I don't have a strong opinion | 16:34 |
| fried_rice | it's not a classmethod, it's an instance method. | 16:34 |
| fried_rice | this_rp.increment_generation() | 16:34 |
| * fried_rice tries it | 16:34 | |
| cdent | yeah, if it is an instance that's being inc'd I'd be fine with it | 16:34 |
| cdent | I had thought we were perhaps not in an object context, just an id, but I guess to have generation at all, we must be | 16:35 |
| fried_rice | the only question is whether to use the context or pass from caller. | 16:36 |
| fried_rice | are they gonna be the same? | 16:36 |
| fried_rice | really just needs to be a context that has a session in it. | 16:37 |
| fried_rice | Which is what all of these contexts are for, right? | 16:37 |
| fried_rice | we already have this for consumer (except it calls the _increment_consumer_generation method - will collapse that too) | 16:39 |
| openstackgerrit | Eric Fried proposed openstack/placement master: ResourceProvider.increment_generation() https://review.openstack.org/640433 | 16:43 |
| fried_rice | cdent, leakypipes: ^ | 16:43 |
| cdent | it's the same conrtext | 16:44 |
| fried_rice | schweet | 16:46 |
| fried_rice | hm, why does _increment_consumer_generation have a @writer decorator on it, but _increment_provider_generation didn't? | 16:47 |
| *** DarthNau has quit IRC | 16:47 | |
| cdent | presumably everything that calls inc_gen already has a writer on it | 16:48 |
| cdent | and it is likely that everything that call inc consumer does too | 16:48 |
| cdent | but it was added anyway | 16:48 |
| cdent | we should probably make sure | 16:49 |
| cdent | but in most cases it is probably right to not waste the time | 16:49 |
| cdent | more rocks, more bugs | 16:49 |
| fried_rice | Does | 16:50 |
| fried_rice | @writer | 16:50 |
| fried_rice | ... | 16:50 |
| fried_rice | @writer | 16:50 |
| fried_rice | ... | 16:50 |
| fried_rice | do anything or is the second one a no-op? | 16:50 |
| fried_rice | ...as long as not .independent | 16:51 |
| *** Sundar has joined #openstack-placement | 16:52 | |
| fried_rice | hah, it's only called from one place. And it's right next to a rp.increment_generation. And that method is wrapped in a writer. This should be an easy refactor. | 16:52 |
| cdent | I used to think that @writer @writer opened another transactin within the existing one but last I heard that's not the case | 16:53 |
| cdent | I'm not 100% on that though | 16:54 |
| fried_rice | unrelated, tox -e py27 on the placement repo is now so quick it registers as 0.00000 seconds to stestr. | 16:55 |
| cdent | \o/ | 16:55 |
| fried_rice | The overall wallclock is still ~4.5s with the tox overhead. But that's acceptable. | 16:56 |
| cdent | I'm getting .2 seconds | 16:56 |
| * cdent time to get a better computer | 16:57 | |
| openstackgerrit | Eric Fried proposed openstack/placement master: Inline Consumer.increment_generation() https://review.openstack.org/640437 | 16:57 |
| fried_rice | cdent: ^ | 16:57 |
| cdent | seems fair | 17:00 |
| melwitt | in nova, we avoid nesting the database transaction contexts. you will have a problem if your outer method has @reader and the inner has @writer (we had a bug with that in the past by an accidental nest) | 17:01 |
| melwitt | so just be on the lookout for that | 17:02 |
| *** ttsiouts has quit IRC | 17:06 | |
| openstackgerrit | Chris Dent proposed openstack/placement master: WIP: Use native list of list of Allocation https://review.openstack.org/640450 | 17:20 |
| cdent | fried_rice: you're going to find that cringey. leakypipes you might like it | 17:21 |
| fried_rice | ack | 17:21 |
| cdent | it's not done because of the issues in the parent, but I need to go deal with dinner and related stuff | 17:21 |
| stephenfin | cdent, fried_rice: bauzas noted that there was an issue with placement in Queens where the request IDs didn't match what nova was using. Any chance you know what the bug report for that was? | 17:24 |
| cdent | stephenfin: do you mean that nova wasn't passing along the request id to placement, or that the form of the ids that placement had was wrong? | 17:25 |
| stephenfin | The former, I suspect | 17:25 |
| stephenfin | You couldn't take the request ID from nova logs and simply grep for that in placement logs, to correlate what placement was doing for that request | 17:26 |
| cdent | stephenfin: maybe this: https://bugs.launchpad.net/nova/+bug/1652099 | 17:26 |
| openstack | Launchpad bug 1734625 in OpenStack Compute (nova) queens "duplicate for #1652099 placement: Request IDs are not passed to placement service" [Medium,Fix committed] - Assigned to Takashi NATSUME (natsume-takashi) | 17:26 |
| stephenfin | cdent++ That looks like the one | 17:27 |
| cdent | and this: https://bugs.launchpad.net/nova/+bug/1734625 | 17:27 |
| openstack | Launchpad bug 1734625 in OpenStack Compute (nova) queens "placement: Request IDs are not passed to placement service" [Medium,Fix committed] - Assigned to Takashi NATSUME (natsume-takashi) | 17:27 |
| cdent | oh sorry, I guess the bot did that for me | 17:28 |
| stephenfin | clever bot | 17:28 |
| * cdent goes home | 17:28 | |
| * cdent waves | 17:28 | |
| stephenfin | o/ | 17:28 |
| *** cdent has quit IRC | 17:28 | |
| fried_rice | stephenfin: this and associated bug? I306ac6f5c6b67d77d91a7ba24d4d863ab3e1bf5c | 17:29 |
| fried_rice | https://review.openstack.org/#/c/531299/ | 17:29 |
| fried_rice | yeah, the bug cdent said | 17:29 |
| stephenfin | Yup. I think that's the one | 17:30 |
| *** mriedem is now known as mriedem_lunch | 17:56 | |
| *** edleafe_ has joined #openstack-placement | 18:09 | |
| *** irclogbot_2 has joined #openstack-placement | 18:10 | |
| *** cdent has joined #openstack-placement | 18:58 | |
| *** cdent has quit IRC | 18:58 | |
| *** e0ne has joined #openstack-placement | 19:12 | |
| *** purplerbot has quit IRC | 19:18 | |
| *** purplerbot has joined #openstack-placement | 19:18 | |
| *** mriedem_lunch is now known as mriedem | 19:28 | |
| *** e0ne has quit IRC | 19:31 | |
| *** edleafe_ has quit IRC | 19:36 | |
| openstackgerrit | Merged openstack/placement master: Use set instead of list https://review.openstack.org/639887 | 19:49 |
| *** irclogbot_2 has quit IRC | 19:49 | |
| *** irclogbot_2 has joined #openstack-placement | 20:01 | |
| *** e0ne has joined #openstack-placement | 20:05 | |
| *** Sundar has quit IRC | 20:13 | |
| openstackgerrit | Eric Fried proposed openstack/placement master: ResourceProvider.increment_generation() https://review.openstack.org/640433 | 21:35 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move Allocation and AllocationList to own module https://review.openstack.org/640184 | 21:35 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Inline Consumer.increment_generation() https://review.openstack.org/640437 | 21:35 |
| *** irclogbot_2 has quit IRC | 21:38 | |
| openstackgerrit | Eric Fried proposed openstack/placement master: ResourceProvider.increment_generation() https://review.openstack.org/640433 | 21:50 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move Allocation and AllocationList to own module https://review.openstack.org/640184 | 21:50 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Inline Consumer.increment_generation() https://review.openstack.org/640437 | 21:50 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move reshape() into placement.tasks.reshaper https://review.openstack.org/640540 | 21:50 |
| mriedem | fried_rice: can we mark this complete now? https://blueprints.launchpad.net/nova/+spec/alloc-candidates-in-tree | 21:59 |
| fried_rice | mriedem: Did the spec update merge? | 21:59 |
| mriedem | not yet | 21:59 |
| fried_rice | melwitt was talking about marking it complete anyway. | 21:59 |
| fried_rice | up to you. | 21:59 |
| fried_rice | The code is done | 21:59 |
| mriedem | done | 22:00 |
| fried_rice | \o/ | 22:00 |
| openstackgerrit | Eric Fried proposed openstack/placement master: WIP: Use native list of list of Allocation https://review.openstack.org/640450 | 22:02 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Inline Consumer.increment_generation() https://review.openstack.org/640437 | 22:02 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move Allocation and AllocationList to own module https://review.openstack.org/640184 | 22:20 |
| openstackgerrit | Eric Fried proposed openstack/placement master: WIP: Use native list of list of Allocation https://review.openstack.org/640450 | 22:20 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Inline Consumer.increment_generation() https://review.openstack.org/640437 | 22:20 |
| *** e0ne has quit IRC | 22:24 | |
| *** guilhermesp__ has joined #openstack-placement | 22:45 | |
| *** guilhermesp has quit IRC | 22:52 | |
| *** melwitt has quit IRC | 22:52 | |
| *** guilhermesp__ is now known as 18WAAFJGJ | 22:52 | |
| *** melwitt has joined #openstack-placement | 22:53 | |
| *** mriedem is now known as mriedem_afk | 23:08 | |
| openstackgerrit | Eric Fried proposed openstack/placement master: Remove pep8 whitespace ignores https://review.openstack.org/640554 | 23:37 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Nit collector https://review.openstack.org/640555 | 23:47 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Factor listiness into an ObjectList base class https://review.openstack.org/637325 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move _set_objects into ObjectList https://review.openstack.org/637328 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move *List.__repr__ into ObjectList https://review.openstack.org/637332 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Clean up ObjectList._set_objects signature https://review.openstack.org/637335 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move RC_CACHE in resource_class_cache https://review.openstack.org/640114 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Use native list for lists of Usage https://review.openstack.org/639391 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Make base test case file for object unit tests https://review.openstack.org/640406 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move reshape() into placement.tasks.reshaper https://review.openstack.org/640540 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: ResourceProvider.increment_generation() https://review.openstack.org/640433 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Move Allocation and AllocationList to own module https://review.openstack.org/640184 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: WIP: Use native list of list of Allocation https://review.openstack.org/640450 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Inline Consumer.increment_generation() https://review.openstack.org/640437 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Remove pep8 whitespace ignores https://review.openstack.org/640554 | 23:56 |
| openstackgerrit | Eric Fried proposed openstack/placement master: Nit collector https://review.openstack.org/640555 | 23:56 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!