| *** VW has joined #craton | 00:05 | |
| *** VW has quit IRC | 00:50 | |
| *** VW has joined #craton | 01:11 | |
| *** VW has quit IRC | 01:49 | |
| *** VW has joined #craton | 01:52 | |
| *** syed__ has quit IRC | 01:55 | |
| *** Mudpuppy has quit IRC | 02:17 | |
| *** jovon has quit IRC | 02:17 | |
| *** Mudpuppy has joined #craton | 02:47 | |
| *** Mudpuppy has quit IRC | 02:52 | |
| *** VW has quit IRC | 03:15 | |
| *** harlowja has quit IRC | 03:33 | |
| *** tojuvone has quit IRC | 03:44 | |
| *** valw has joined #craton | 03:49 | |
| *** valw_ has joined #craton | 04:02 | |
| *** valw has quit IRC | 04:04 | |
| *** VW has joined #craton | 04:12 | |
| *** tojuvone has joined #craton | 04:16 | |
| *** tojuvone_ has joined #craton | 04:16 | |
| *** tojuvone has quit IRC | 04:17 | |
| *** tojuvone_ has quit IRC | 04:17 | |
| *** harlowja has joined #craton | 04:38 | |
| *** VW has quit IRC | 04:40 | |
| *** valw_ has quit IRC | 05:17 | |
| *** valw has joined #craton | 05:18 | |
| *** VW has joined #craton | 05:33 | |
| *** VW has quit IRC | 05:38 | |
| *** harlowja has quit IRC | 05:51 | |
| *** harlowja has joined #craton | 05:52 | |
| *** VW has joined #craton | 07:35 | |
| *** VW has quit IRC | 07:40 | |
| *** valw has quit IRC | 07:47 | |
| *** valw has joined #craton | 07:56 | |
| *** harlowja has quit IRC | 08:03 | |
| *** VW has joined #craton | 09:38 | |
| *** VW has quit IRC | 09:43 | |
| *** VW has joined #craton | 10:40 | |
| *** VW has quit IRC | 10:46 | |
| *** Mudpuppy has joined #craton | 10:51 | |
| *** Mudpuppy has quit IRC | 10:56 | |
| -openstackstatus- NOTICE: Our OpenStack CI systems are stuck and no new jobs are submitted. Please do not recheck - and do not approve changes until this is fixed. | 11:59 | |
| *** Mudpuppy has joined #craton | 12:48 | |
| -openstackstatus- NOTICE: Our OpenStack CI system is coming back online again. Thanks for your patience. | 12:58 | |
| *** VW has joined #craton | 13:16 | |
| *** VW has quit IRC | 13:21 | |
| *** VW has joined #craton | 13:28 | |
| *** VW has quit IRC | 13:30 | |
| *** Mudpuppy has quit IRC | 13:43 | |
| *** valw has quit IRC | 14:12 | |
| *** syed__ has joined #craton | 14:23 | |
| *** VW has joined #craton | 14:33 | |
| *** VW has quit IRC | 14:37 | |
| *** Mudpuppy has joined #craton | 14:57 | |
| *** valw has joined #craton | 15:02 | |
| *** valw has quit IRC | 15:10 | |
| *** jovon has joined #craton | 15:10 | |
| *** valw has joined #craton | 15:23 | |
| *** valw has quit IRC | 15:35 | |
| *** valw has joined #craton | 15:37 | |
| sigmavirus | So, I haven't tested this at all, but https://review.openstack.org/396678 is the start to the caching work on cratonclient | 15:46 |
|---|---|---|
| sigmavirus | It *might* Just Work⢠| 15:46 |
| sigmavirus | ^ jimbaker sulo palendae odyssey4me | 15:46 |
| *** valw has quit IRC | 15:55 | |
| *** valw has joined #craton | 16:02 | |
| jovon | it was brought to my attention that flask-autodoc is a better solution to generate docs because it automatically creates documentation for your endpoints based on the routes, function arguments and docstrings. Should the manual efforts here: https://review.openstack.org/#/c/390386/ be ditched in favor of this method? | 16:02 |
| sigmavirus | jovon: so, all of the big tent projects document their APIs manually | 16:09 |
| sigmavirus | And none of them use flask | 16:09 |
| sigmavirus | I suspect we should move on with manually documenting our API because otherwise we're going to be too tightly tied to flask | 16:10 |
| jovon | I see your point. I'll continue on then. | 16:11 |
| *** VW has joined #craton | 16:11 | |
| jovon | thank you | 16:11 |
| git-harry | sigmavirus: I wouldn't describe it as being tied. All the docs would be in the method doc strings. Manual seems like wasted effort but if going that way it would probably make sense to use https://github.com/openstack/os-api-ref | 16:12 |
| sigmavirus | git-harry: os-api-ref was going to be a suggestion, but I think we've had jovon stuck with this work quite a while | 16:14 |
| jovon | i don't mind. i just want this to be done correctly and efficiently | 16:16 |
| git-harry | sigmavirus: if the rest of OpenStack isn't using flask, why was it chosen for this project? | 16:17 |
| sigmavirus | git-harry: there are non-big tent projects that use it and it's good for rapid prototyping | 16:18 |
| *** valw has quit IRC | 16:53 | |
| *** valw has joined #craton | 16:55 | |
| *** david-lyle has quit IRC | 16:58 | |
| *** tojuvone has joined #craton | 17:06 | |
| *** tojuvone_ has joined #craton | 17:06 | |
| *** tojuvone_ has quit IRC | 17:07 | |
| sulo | sigmavirus: can you give your thoughts on https://review.openstack.org/#/c/360605 as well | 17:26 |
| sulo | i want to make sure i capture eveyones view before making more changes | 17:26 |
| sigmavirus | sure, will be right back and then give thoughts | 17:27 |
| *** valw has quit IRC | 17:49 | |
| *** valw has joined #craton | 17:55 | |
| *** david-lyle has joined #craton | 17:58 | |
| *** valw has quit IRC | 17:59 | |
| *** david-lyle has quit IRC | 18:03 | |
| jimbaker | git-harry, flask is used iirc by designate | 18:07 |
| jimbaker | and is in global requirements | 18:08 |
| jimbaker | more importantly, we didn't see a need for building out code unnecessarily with async approaches. all interaction with the db is sync, which means that complex models in sqlalchemy work without bugs (very important!) | 18:09 |
| *** valw has joined #craton | 18:09 | |
| jimbaker | and taskflow can do things independently in a worker pool | 18:09 |
| jimbaker | sulo, re http://specs.openstack.org/openstack/api-wg/guidelines/pagination_filter_sort.html#filtering and https://review.openstack.org/#/c/360605/ | 18:11 |
| jimbaker | how about this approach | 18:11 |
| jimbaker | ?vars=x:in:baz, etc | 18:12 |
| jimbaker | this means that if the value we are matching against has an embedded colon, it needs to be urlencode | 18:13 |
| jimbaker | d | 18:13 |
| jimbaker | but that might be reasonable for supporting this for vars | 18:13 |
| *** VW has quit IRC | 18:13 | |
| jimbaker | also we can do the same operators for well-known reserved keys | 18:13 |
| *** valw has quit IRC | 18:13 | |
| jimbaker | anyway, i will discuss more on the change proposal | 18:14 |
| *** VW has joined #craton | 18:22 | |
| *** VW has quit IRC | 18:26 | |
| sigmavirus | git-harry: that was a great review btw | 18:27 |
| sulo | yes, +1 | 18:32 |
| sulo | one other feedback i got was is that other project like nova, are removing this type of filtering since it causes problems with the db and acts like a Ddos vector. So I am now thinking this would be more of a problem with us as the data grows. | 18:33 |
| sulo | so another view here is not to allow this type of query at all | 18:33 |
| sulo | so no filter when its un indexed | 18:34 |
| *** valw has joined #craton | 18:36 | |
| *** david-lyle has joined #craton | 18:36 | |
| sigmavirus | hm | 18:37 |
| sigmavirus | sulo: so maybe, but nova and craton have fundamentally different user-bases and expected sizes of users, no? | 18:38 |
| sigmavirus | I'm not saying we should do it, but I would expect that craton's not going to have that kind of trouble | 18:38 |
| sulo | perhaps, but the problem is still the same, more true for us than nova i would think | 18:38 |
| sulo | i expect the data part to be the most bulky info for each resource | 18:39 |
| sulo | but either ways .. i think we have a good path forward .. either not allow, but if allow .. do ?vars=x:y | 18:40 |
| sulo | we should probably not worry about operator filtering .. cuz that even crazier query | 18:40 |
| *** valw has quit IRC | 18:41 | |
| *** VW has joined #craton | 18:49 | |
| *** VW has quit IRC | 18:54 | |
| *** harlowja has joined #craton | 18:55 | |
| jimbaker | sulo, agreed with deferring on operator filtering for now | 18:57 |
| jimbaker | simplifies the change | 18:57 |
| jimbaker | re user base: yeah, i don't think we are ever going to be public facing with these APIs, but a badly implemented worker could still do a DOS :) | 18:58 |
| jimbaker | and sounds good about ?vars=x:y - this makes me happy | 18:58 |
| sulo | jimbaker: yeah, well i have quite a few experience in public cloud where its easy to kinda dos .. take one of the bigger regions | 18:59 |
| sulo | so i do think its kinda important to keep this in mind | 18:59 |
| *** VW has joined #craton | 19:00 | |
| sulo | i tihnk once we have the feature .. its harder to stop bad usage i guess | 19:00 |
| *** valw has joined #craton | 19:00 | |
| sulo | the good thing is this discussion is making us think about all this | 19:01 |
| syed__ | sigmavirus: a quick question. https://github.com/openstack/python-cratonclient/blob/master/cratonclient/v1/regions.py#L32 | 19:01 |
| syed__ | Is this populated from schema .. | 19:01 |
| sigmavirus | syed__: I would ask chrisspencer how he decided what should be in that (and the other similar dictionaries) | 19:02 |
| syed__ | sigmavirus: Cool, yeah will do that, thanks | 19:03 |
| sigmavirus | syed__: please do let me know when you've found out :) | 19:03 |
| syed__ | sigmavirus: sure | 19:03 |
| *** klindgren__ is now known as klindgren | 19:03 | |
| chrisspencer | syed__ sigmavirus those were created from the original class UML diagrams (not sure where they ended up after we moved to openstack infra) | 19:08 |
| syed__ | cool, thanks chrisspencer | 19:12 |
| jimbaker | git-harry, good catch on that review - nothing like demonstrating a problem! | 19:12 |
| syed__ | sigmavirus: should i just go and fix that in regions, cells, hosts etc | 19:12 |
| jimbaker | anyway, by pulling operator support out, we will have something more manageable for that review - and hopefully a better query mechanism in the future. i think we will want operators at some point | 19:14 |
| jimbaker | but one concern: we will want the operators to be json aware | 19:14 |
| *** valw has quit IRC | 19:14 | |
| jimbaker | some variant on jsonpath? probably a good idea | 19:14 |
| sigmavirus | jimbaker: jsonpatch you mean? | 19:18 |
| sigmavirus | syed__: go for it | 19:19 |
| syed__ | sigmavirus: cool, on it | 19:19 |
| *** valw has joined #craton | 19:22 | |
| *** valw has quit IRC | 19:26 | |
| *** valw has joined #craton | 19:36 | |
| *** valw has quit IRC | 19:59 | |
| *** valw has joined #craton | 20:04 | |
| sigmavirus | palendae: dogpile has file backends and that review includes one way of using a filebackend | 20:20 |
| palendae | Oh, I missed on_disk | 20:21 |
| palendae | Thanks | 20:21 |
| sigmavirus | palendae: happy to help :) | 20:26 |
| sigmavirus | just updated that patch too with tests | 20:26 |
| *** valw has quit IRC | 20:28 | |
| *** valw has joined #craton | 20:35 | |
| *** valw has quit IRC | 20:51 | |
| *** valw has joined #craton | 20:52 | |
| syed__ | sigmavirus: Just for my knowledge, https://github.com/openstack/python-cratonclient/blob/7562ab3e3a13f9e463123deeed05ab994bc12363/cratonclient/tests/unit/shell/v1/test_hosts_shell.py#L257 | 21:02 |
| syed__ | sigmavirus: What is access_secret_id for ? If we go by schema then there shouldn't be any such thing too. Just double checking | 21:02 |
| sigmavirus | syed__: I didn't write the shell. I wrote tests so refactoring the code would be doable | 21:05 |
| syed__ | sigmavirus: cool | 21:05 |
| sigmavirus | syed__: if --access-secret-id is no longer a thing tear out the stuff related to that if you must | 21:05 |
| sigmavirus | I'd ask jimbaker and sulo though whether there were plans to add that | 21:05 |
| syed__ | sigmavirus: okay, thank you | 21:06 |
| chrisspencer | syed__ sigmavirus before removing all of those, verify with jimbaker and sulo to make sure those items will not be part of implementation going forward. I had added those based off of their initial docs | 21:13 |
| syed__ | chrisspencer: cool, thanks | 21:13 |
| sigmavirus | syed__: if they're going to be there, we can comment them out for now | 21:15 |
| sigmavirus | Anyway, I'm off for the day. See y'all Monday | 21:15 |
| syed__ | sigmavirus: thanks, have a great weekend | 21:16 |
| *** valw has quit IRC | 21:29 | |
| *** Mudpuppy_ has joined #craton | 21:30 | |
| *** harlowja has quit IRC | 21:32 | |
| *** Mudpuppy has quit IRC | 21:33 | |
| *** Mudpuppy_ has quit IRC | 21:34 | |
| jimbaker | sigmavirus, no, jsonpath for query filters | 21:51 |
| jimbaker | syed__, access_secret is not yet a thing, but we will be supporting in the future | 21:52 |
| jimbaker | basically it's what we need for a worker to connect to a target device | 21:52 |
| syed__ | jimbaker: cool, will comment that out then so in case when we do need in future will pick up from there | 21:52 |
| jimbaker | syed__, it's fine to do that, i don't believe it's exposed by the REST API | 21:53 |
| syed__ | cool. | 21:53 |
| *** VW has quit IRC | 22:12 | |
| *** VW has joined #craton | 22:12 | |
| *** VW has quit IRC | 22:16 | |
| *** VW has joined #craton | 22:36 | |
| *** VW has quit IRC | 22:40 | |
| *** harlowja has joined #craton | 23:01 | |
| *** VW has joined #craton | 23:03 | |
| *** VW has quit IRC | 23:08 | |
| *** VW has joined #craton | 23:30 | |
| *** VW has quit IRC | 23:34 | |
| *** VW has joined #craton | 23:37 | |
| *** VW_ has joined #craton | 23:39 | |
| *** VW has quit IRC | 23:42 | |
| *** VW_ has quit IRC | 23:44 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!