Monday, 2013-09-23

*** notmyname_ has joined #openstack-marconi00:19
*** ekarlso has quit IRC00:23
*** notmyname has quit IRC00:23
*** wirehead_ has quit IRC00:23
*** notmyname_ is now known as notmyname00:24
*** ekarlso has joined #openstack-marconi00:29
*** jdprax has joined #openstack-marconi00:53
*** jdprax has quit IRC00:54
*** wirehead_ has joined #openstack-marconi02:17
*** flaper87|afk is now known as flaper8707:32
*** ykaplan has joined #openstack-marconi07:41
*** gordonsim has joined #openstack-marconi08:18
*** key3 has quit IRC09:13
*** ykaplan has quit IRC10:05
al-maisanGood afternoon!10:50
flaper87al-maisan: goooooood afternoon11:07
al-maisanflaper87: hello there! How are things?11:08
al-maisanWe have sunshine in CH :)11:09
flaper87al-maisan: all good, we've got a half-cloudy day :D11:09
flaper87meaning, the clouds are right above me, all other places are sunny11:10
al-maisanhe-he :)11:12
al-maisanreal clouds or "clouds"?11:12
flaper87nah, real clouds. It was cloudy yesterday and it still is today. Weather forecast says temperatures will be high this week, though.11:13
al-maisanso, how do we progress this week? I'll spend some time and look at various queuing APIs and the client discussion etherpad11:17
flaper87al-maisan: I suggest you to take a look to Proton11:18
al-maisanIs this the right etherpad: ?11:18
al-maisanOK, will do.11:18
flaper87I'll review patches in a bit11:18
flaper87I've been covering some internal things today11:18
al-maisansure .. same here :-P11:19
* al-maisan checks for a proton package on Fedora11:19
al-maisanhmm .. python-qpid-proton looks promising :)11:20
*** ykaplan has joined #openstack-marconi11:29
*** oz_akan_ has joined #openstack-marconi11:41
*** oz_akan_ has quit IRC11:43
*** oz_akan_ has joined #openstack-marconi11:44
*** tedross has joined #openstack-marconi11:59
*** oz_akan_ has quit IRC12:00
*** cppcabrera has joined #openstack-marconi12:04
cppcabreraGood morning! :D12:05
flaper87cppcabrera: good morning12:07
flaper87how was your weekend?12:07
flaper87al-maisan: reviewed your patch, I added some comments12:08
flaper87al-maisan: you may want to take a look here as well:
cppcabreraflaper87: My weekend was awesome. Lots of relaxing, lots of time with my wife, and I caught up on most of my bookmark'd article/video backlog. :D12:09
cppcabreraHow was yours?12:09
flaper87pretty good as well, spent some time with my family and coding. :D12:10
cppcabrerasweet. :)12:11
cppcabreraI read the live blog from CUFP 2013. The keynote was an especially good read. (#link:
al-maisanflaper87: thanks for the review!12:17
* flaper87 bookmarks that link12:19
flaper87al-maisan: np, I'll wait for your new PS to rebase mine12:20
*** jraim has joined #openstack-marconi12:28
*** oz_akan_ has joined #openstack-marconi12:57
*** oz_akan_ has quit IRC12:58
*** oz_akan_ has joined #openstack-marconi12:59
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: proxy: mirror structure of marconi queues + bootstrap
*** mpanetta has joined #openstack-marconi13:02
cppcabreraI'm going to write out a spec for marconi proxy. It's developed enough at this point that it has: storage interface, cache interface, transport interface, and a well-defined API.13:02
flaper87cppcabrera: +113:07
cppcabreraI think it also might make reviewing the thing a little easier. :P13:08
*** rustlebee is now known as russellb13:08
flaper87hehe, indeed13:09
cppcabreraI realized while thinking about it during the weekend that reviewing this series of patches must be especially annoying, since decisions made in early patches are deprecated/fixed by the time a new patch rolls out, so reading it in sequential order is misleading.13:09
flaper87I'm reviewing your patch now13:09
cppcabrerafor example, the patch that brings in oslo.cache uses static configuration (the code must be modified to swap out backends), but that's fixed by the latest patch where -proxy:transport mirrors -queues:transport.13:10
cppcabrerathanks, flaper87! :)13:10
mpanettaSomeone has been having fun with marconi lately I see :)13:12
ekarlsooslo.cache is what ?13:13
cppcabreraekarlso: oslo.cache is a caching interface supporting multiple backends. flaper87 is the primary author and can tell you all the things about it. :)13:14
cppcabreraIt's quite handy.13:14
flaper87ekarlso: under development cache abstraction for oslo
cppcabrerampanetta: haha, I'm always having fun with marconi. :D13:15
ekarlsook, so not a sep library ?13:15
cppcabreramy understanding is that it is a separate library/module - flaper87, is that correct?13:16
flaper87yeah, that's kinda corret. So, it'll stay in oslo-incubator as a separate package for a bit and then I'll pull it out into its own repo13:18
*** ykaplan has quit IRC13:18
*** ykaplan has joined #openstack-marconi13:18
*** amitgandhi has joined #openstack-marconi13:23
flaper87cppcabrera: why is it that you need to define a resource for ClaimCreate and one for Claims ?13:30
cppcabreraBecause they forward to different endpoints, flaper87. :)13:31
cppcabreraClaimCreate handles /v1/queues/{queue}/claims13:32
cppcabreraClaim handles /v1/queues/{queue}/claims/{claim_id}13:32
flaper87cppcabrera: aaaahhh, good point! It wasn't obvious to me at first glance13:32
mpanettaOhh has claiming messages been changed a bit?13:33
*** malini_afk is now known as malini13:33
cppcabreranah, mpanetta.13:33
flaper87mpanetta: what did you want to change about it?13:33
cppcabrerait's the same it's been for at least a month. :)13:33
mpanettaAh ok.  Last time I actually played with this was before it was moved from the staging git repo.13:34
cppcabreraflaper87: The marconi-proxy spec is born! #link
cppcabreraAt least the endpoints summary portion of that spec is filled out.13:34
mpanettaI don't remember the claimcreate api.  But I may not have seen it :)13:34
cppcabrerampanetta: I wouldn't be able to identify it as claimcreate, usually. I'm more used to seeing it as `POST /v1/queues/{queue}/claims`. :P13:36
flaper87al-maisan: btw, do you think you can get those fixes done by the end of the day? Kinda need them :D13:36
mpanettaAh, yes.13:36
flaper87no preasures, though. Just hurry up but, you know, no hurries. Just finish that very quickly but hey, I don't want to stress you13:36
flaper87al-maisan: ^13:36
flaper87cppcabrera: +1 for the spec13:37
flaper87you know, soemthing I love about the way we're developing marconi is that nothing gets released without being documented13:38
flaper87I want to keep it that way13:38
flaper87cppcabrera: you've been very helpful at documenting / reviewing things. Big thanks to you!13:38
cppcabreraI'm just happy to make things a little easier for everyone. Keeping all the details in mind if *hard*. I need to refresh on the proxy every morning to see what it's supposed to do. :D13:41
cppcabreraKeeping a doc like that around helps me, too. ;)13:41
cppcabrera** *hard*...13:41
flaper87I wonder we'll be able to sue marconiclient for request forwarding in marconi-proxy. We can implement very nice tricks there. Like, pick the fastest transport enabled and forward the request there13:45
flaper87I wonder whether*13:45
flaper87poor marconiclient, it's already being sued13:45
* mpanetta sues marconi for not proxying his request.13:45
flaper87the idea is to avoid duplicating code in the proxy + reuse connections and whatever improvements can be made there that are also useful for end-users13:46
cppcabreraI wonder as well how much of marconiclient we can reuse.13:49
cppcabreraI have to say, though, that the forwarding mixin you proposed early on flaper87 helped clean up the implementation of the proxy **a lot**. I'm very grateful for that suggestion. :D13:49
cppcabreraAlso, zyuan has been working on this ( that might clean things up even further in the proxy.13:49
*** ametts has joined #openstack-marconi13:50
flaper87cppcabrera: glad it helped :D13:50
flaper87cppcabrera: that looks very useful13:50
flaper87cppcabrera: reviewed13:59
cppcabreralots and lots of comments (starts reading)14:00
cppcabrerathanks, flaper87!14:00
*** whenry has quit IRC14:26
cppcabreraflaper87: responded to review. :)14:26
* flaper87 reading14:28
flaper87cppcabrera: awesome, I agree with the planned approach since this patch and the enxt one are pretty big14:30
flaper87cppcabrera: thanks for working on that14:30
flaper87I don't think we'll need to defer any other fix after this 2 patches land14:30
flaper87we'll be able to focus on a per-patch review + fixes14:31
cppcabrera+1 - things are getting more solid with the proxy with each passing day.14:31
flaper87I'd like to get those 2 in before reviewing the next set of patches14:32
flaper87gosh, I haven't written 1 line of code today14:32
cppcabreraworks for me, flaper87. :)14:32
cppcabreraThanks for understanding!14:32
cppcabreraI'll have the -proxy spec finished by the end of the day today. It's a good draft at this point, since I've had a bit of experience implementing this thing to go with all the discussions we had awhile ago.14:33
cppcabreraIt'll also help with the review of the remaining 2 patches.14:33
cppcabreralol, no code yet? Are meetings holding you up, flaper87? :P14:33
flaper87ah yeah, + internal stuff + reviews.14:34
*** kgriffs_afk is now known as kgriffs14:36
cppcabrerabusy mondays~14:37
cppcabrerakgriffs: o/14:37
flaper87kgriffs: yo, Happy monday :)14:43
cppcabreralol kgriffs14:48
kgriffsThe only thing that saved this day for me was a hot toaster strudel. I have another one in my backpack just in case things get *really* bad today.14:48
cppcabrera"Boy, I wish I had a fifty pound pan of lasagna" - that made me realize I'm quite hungry. >.>14:49
* kgriffs is so glad we have a toaster at work14:49
cppcabrerammm, strudel...14:49
amettskgriffs, flaper87:  Looks like one of our two Summit sessions got approved14:50
flaper87ametts: YEEEESS!!14:50
flaper87ametts: don't you read twitter????? :D14:50
flaper87ametts: that's definitely AWESOME14:50
flaper87we've got to rock in that session14:51
* ametts goes and looks14:51
cppcabreraI'm very happy about this session approval.14:51
* cppcabrera saw it yesterday14:51
cppcabreraLet me know if you guys need any help putting together content for the presentations14:51
* amitgandhi wonders how ametts misses twitter - was sure its his home page14:51
kgriffsstrudel muß sein14:52
amettsamitgancdi, flaper87: Twitter comes up every time I launch a new browser page, and I wrote a little python thing that gives me one twitter post as a system notification every few minutes.  I also see my most recent twitter on my phone.14:54
amettsThe catch is that I never go back and read history if I'm not there to notice something when it comes in.14:55
amettsI'm only "in the moment". :D14:55
flaper87HAHAHAH, realtime or nothing14:55
cppcabreraI check mine in the morning and at lunch, and probably once more at the end of the day. :P14:55
amettsflaper87:  Yeah, I'm backlogged on work, backlogged on email, and backlogged on projects around the house.  I don't need to be backlogged on Twitter as well.14:56
ametts(this is also why I've never had a TiVo)14:56
amettsI never want to be backed up on "things I have to watch".14:56
amettsAlthough I do have a "things to read list that's a mile long" -- I'm addicted to OReilly "Deals of the Day"14:57
cppcabreraheh, that's what bookmarks are for, ametts! ;)14:57
* cppcabrera has a lengthy To Watch bookmarks list14:57
cppcabreraOohh, O'Reilly Deal of the Day...14:57
cppcabreraThose get me, too.14:57
cppcabreraI've ended up buying almost all the books from their catalogue that I want, at this point. :P14:57
flaper87hahahahhaha, cppcabrera your bookmarks seem expensive14:59
cppcabrerathey cost more in time than in $$$14:59
cppcabreraI finally got around to watching "Escaping the Ivory Tower: The Haskell Journey" and "The Most Important Minute of Your Life" this weekend. I highly recommend both talks!15:00
flaper87cppcabrera: man, pocketed both of them15:06
flaper87thanks :)15:06
cppcabreranp. I've usually got links in the pursuit of knowledge, and I'm happy to share! :D15:07
kgriffshas anybody followed up with the openstack/requirements team to find out whether we can still get something like mmh3 in since it only affects Marconi?15:16
*** reed has joined #openstack-marconi15:18
flaper87kgriffs: not me, let me ping -infra guys15:18
kgriffsgroovy, thanks. The multi-db patch needs some kind of hashing lib. I am looking at XX as well15:19
cppcabrerakgriffs: I haven't followed up, either.15:21
cppcabreraI'll do so after this round of fixes to the latest proxy patch.15:21
cppcabrera(almost done with some local patching)15:21
kgriffsat first glance, i think mmh3 will be easier to update to support cffi15:22
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: proxy: mirror structure of marconi queues + bootstrap
kgriffsmtg in 1515:48
flaper87mmh, damnit, not sure I'll be able to stick around the whole meeting15:48
cppcabreraI'll be there.15:49
kgriffsflaper87: ok, let's audit the API first thing then15:50
amettsGreat.  Now I have "I'll be there" from the Jackson Five stuck in my head....15:50
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: proxy: mirror structure of marconi queues + bootstrap
cppcabrera"Is this really what you meant to do? Type 'yes' to confirm: yeees" -> #fail :P15:51
cppcabreralol ametts15:51
cppcabrerakgriffs: ^^16:05
*** ykaplan has quit IRC16:09
flaper87mmh, I think kgriffs dissapeared :D16:09
flaper87I got to go guys, I'm sorry. It's the second meeting I miss in a row. Won't happen next week16:09
cppcabreraflaper87: take care! ping me later and I'll update you on any parts you'd like to ask about. :)16:10
flaper87cppcabrera: awesome, thanks man! ttyl16:11
*** flaper87 is now known as flaper87|afk16:11
*** cthulhup_ has joined #openstack-marconi16:16
al-maisanflaper87|afk: yes, should be done tonight16:23
*** reed has quit IRC16:37
*** reed has joined #openstack-marconi16:38
*** reed has quit IRC16:41
*** reed has joined #openstack-marconi16:42
*** hub_cap has joined #openstack-marconi16:50
kgriffsbinascii.crc32 is pretty fast17:07
kgriffsnot as fast as mmh3, but close17:07
*** aababilov has joined #openstack-marconi17:07
hub_capdouble hi!17:08
hub_capwe come bearing questions17:08
cppcabreraaababilov: It's been awhile. :D17:08
hub_capso im looking to integrate aababilov's apiclient code17:08
cppcabreraI've had your email sitting in my inbox for a bit regarding the apiclient code.17:08
cppcabreraMy apologies for not responding sooner. :x17:09
aababilovI am a volunteer, so, I write the code when I have free time :)17:09
hub_capso lets chat about it!17:09
kgriffszyuan, cppcabrera:
kgriffsbbl (lunch)17:09
malinicppcabrera: do you limit access on who can create partitions ? We dont want every user to be able to create a partition17:10
cppcabrerayes, let's. So, re: the apiclient patch and why it was reverted awhile ago.17:10
* hub_cap listens17:10
kgriffsI am hopeful that binascii.crc32 should be fine and we don't need pylru17:10
cppcabreramalini: it's up to rbac to limit access to admin users only17:10
kgriffsit would be great if someone could benchmark that17:10
hub_capcppcabrera: fwiw, im looking to add it to troveclient17:10
aababilovwhat;s "bbl"?17:11
hub_capbe back later17:11
*** kgriffs is now known as kgriffs_afk17:11
cppcabrerathe patch was reverted because we needed to make faster progress on the implementation of our client and there were issues with it that we did not/could not address at the time.17:11
cppcabreraThere was a list of those issues maintained at some point17:12
cppcabreraI lost that link. >.>17:12
aababilovbut marconiclient is almost empty now17:12
aababilovit does not send HTTP requests at all17:12
cppcabrerayup - though we have several patches sitting in our review queue17:12
aababilovwhy haven't you added me as a reviewer?17:13
aababilovmaybe I could help to adopt them17:13
cppcabreraI was still new to OS dev at the time - that was my bad.17:13
aababilovyou know, that was quite disappointing to see my code removed :(17:14
hub_capaababilov: ill help readd in trove :P17:14
cppcabreraaababilov: I'd be interested in taking a second look at the apiclient code at this time, now that I'm more familiar with both OS and the marconi project17:15
* al-maisan starts fixing the branch with the transport layer request spec17:16
cppcabreraI'll also add you in to pending patches as a reviewer17:16
hub_capagain fwiw, i like the code... im just sad that the auth stuff doesnt exist anywhere17:16
cppcabrera*pending marconiclient patches17:16
cppcabrerahub_cap: we've got this going for auth atm:
hub_capcppcabrera: so that looks VERY similar to what i need17:17
hub_caphow come this isint 1) in keystoneclient or 2) in oslo like the apiclient stuff?17:17
cppcabreraflaper87|afk would be able to answer that much better than I.17:18
hub_capheh lucky guy17:18
cppcabrerahe's at meetings for a good part of the day, from what I've heard. :P17:18
cppcabreraMy best guess at why that patch isn't part of oslo just yet is that it's being tested here to see if this'll work17:19
cppcabrerathat's me guessing - I gave 0 design input on that particular patch. :)17:19
aababilovauth plugins are not in oslo just because Adam Young asked me to put them into keystoneclient17:20
aababilovthen he asked me to put the whole apiclient into keystoneclient17:20
aababilovand finally Jamie Lennox started writing an alternative auth plugin implementation that is still on review17:21
cppcabreralots of auth plugin impls going on in parallel!17:21
cppcabreraI wasn't aware.17:21
aababilovdo you know how to influent keystoneclient maintainers? ;)17:21
cppcabrerahaha, I have no sway in keystone land. I've spoken with ayoung once or twice and that's the extent of my involvement with keystone. :)17:22
maliniaababilov: chocolates work in most situations ;)17:22
cppcabreramalini: +117:22
cppcabrerapoptarts work **really** well with the marconi team17:23
aababilovwhen I will work in Zurich, I will send him chocolate :)17:23
hub_capaababilov: i can help17:23
hub_capi know peoples17:23
hub_capdo you have reviews or do we need to jump into keystone chat room and start talking w/ them?17:23
aababilovI have an abandoned review:
cppcabreraaababilov: It's good hearing from you again. I'm going to go grab lunch and then return shortly.17:25
aababilovI relied on Jamie and suspended my work on it17:25
aababilovjamie is jamielennox on irc17:25
aababilovhe is in Australia17:25
aababilovit's night there :)17:26
hub_capyes it is!17:27
hub_capwe might have to wait till late afternoon UTC-8 time17:27
hub_capwhichi will be night time for you aababilov ;)17:27
aababilovI usually write jamie about 08:00 UTC+3 (my morning)17:27
aababilovI can try to catch him tomorrow17:28
hub_capaababilov: let me do this... let me get all the info i can from people today17:28
hub_capand i will send out to the ML and try to get consensus17:28
hub_capso i can keep making progress17:28
hub_capmordred wants me to finish this and he has offered review / smacking people help17:28
aababilovhere are jamie's commits:,n,z17:30
hub_capugh these are 2+wks old17:32
aababilovthat's why I begin to worry17:32
hub_capyes ill bring it up on the mailing list. i need to make progress and i dont have 2 more wks to wait17:32
aababilovok! I will wait for mail tomorrow and I will go have a rest today17:35
aababilovsee you!17:35
*** aababilov has left #openstack-marconi17:35
*** kgriffs_afk is now known as kgriffs17:41
*** kgriffs is now known as kgriffs_afk17:55
*** ayoung is now known as ayoung-vrooom18:05
openstackgerritMuharem Hrnjadovic proposed a change to openstack/python-marconiclient: This adds a transport layer request spec class
al-maisanflaper87|afk: please see ^18:07
cppcabreraal-maisan: I read your patchset for the third time now to get a feel for how this will work.18:27
cppcabreraMy understanding is that this will be coupled with some form of httpclient to issue the actual requests and will serve as the low level layer for all transport implementations.18:27
cppcabreraFor example, building 'client.queues' on top of 'Request.list_queues'.18:28
cppcabrerawhere client.queues() -> lists queues from API using a Python generator18:28
cppcabreraAnd the Request portion provides a lot of the validation machinery.18:28
cppcabreraAm I understanding the purpose of your patchset correctly?18:28
* cppcabrera wants to review this patch with understanding of what it's trying to solve18:29
al-maisancppcabrera: yes, the request spec captures what needs done and is passed to the transport layer which turns it into an actual (http, zmq, etc.) request18:35
al-maisanit's a bit of an API toward the transport layer18:36
al-maisanwrt the validation: it's supposed to be specific to a Marconi API version18:36
cppcabreramakes sense that the validation would be version specific. I can also see another layer of validation being added that would be transport-specific, but that's down the line. This spec aspect is a first line of defense, perse.18:37
al-maisangoing forward we might have version specific code like here:
cppcabreraooohh, link18:38
cppcabreraThanks for sharing!18:38
cppcabreraI'll give your patchset a review after I check out that link18:38
al-maisanand '' will then probably live underneath a 'v1' directory18:38
cppcabreraBy the end of the day, really. :)18:38
al-maisantake your time18:38
al-maisanthanks for looking at it!18:38
cppcabreranp. Thanks for helping marconiclient grow! :D18:40
*** oz_akan_ has quit IRC18:43
*** oz_akan_ has joined #openstack-marconi18:44
*** kgriffs_afk is now known as kgriffs18:46
*** oz_akan_ has quit IRC18:48
*** cthulhup_ has quit IRC18:49
*** oz_akan_ has joined #openstack-marconi18:55
al-maisannot at all sure whether flaper87|afk will be happy ;-) let's see what he says..18:59
* al-maisan signs off18:59
hub_capcppcabrera: hey just to make sure, the apiclient stuff didnt get removed cuz it was broken, right?19:37
hub_capas in, buggy...19:37
cppcabrerathat's correct, hub_cap. We didn't find bugs in the run-time, though we didn't test. We had design concerns.19:39
hub_capok thx cppcabrera. ill let the design concerns leak into oslo :)19:39
kgriffscppcabrera: got a minute?19:42
kgriffsI thought I would just comment on these patches as I go through them rather than do it in gerrit, since i guess you are just making notes and will update in the next batch of patches...19:43
cppcabrera+1 kgriffs19:44
*** malini is now known as malini_afk19:44
kgriffsfirst up:19:44
kgriffswill these results need to be pageable?19:45
cppcabreraI'm not sure. One project could have many queues, and it might be useful for an admin to be able to paginate over the results rather than having them all dumped in a single shot.19:46
kgriffsis there a doc somewhere showing the API design?19:47
cppcabreraIf pagination is more than a nice-to-have, this'll bring markers + limits into the discussion19:47
cppcabrera#link (proxy API spec)19:47
cppcabreraThis page is not yet linked to by anything because it;ll likely evolve after a review of the proxy itself.19:48
*** ayoung-vrooom is now known as ayoung19:50
* kgriffs is looking19:52
kgriffs"Paging is not supported at this time."19:52
kgriffsok, so I think you should start a list of "planned enhancements" or something19:52
kgriffspaging is definitely something you will want since a single project can have thousands of queues19:52
cppcabreraI may just tack this list onto the marconi-proxy-issues etherpad for now19:53
kgriffsgenerally speaking, we should take some time over the next few team mtgs to review this admin API19:53
cppcabreraI'd appreciate that!19:54
cppcabreraIt's been fun crafting an API, but nothing beats team input for making it even better19:54
kgriffsyep, it's always good to seek feedback early, often19:55
kgriffs(added to next week's agenda)19:55
cppcabreraquick side note - I need to add a context manager to timeutils to handle time overrides19:57
cppcabrerajust noticed this while fixing an age bug in marconi queues storage19:57
cppcabrerametadata - that's got the worst hack of the system, IMO19:58
kgriffsso, weren't we going to query partitions and merge the listings, so you didn't have to dupe data in the catalog?19:59
cppcabrerathat's a planned enhancement, kgriffs. It's also a pretty high priority one, so I can start working on that one tomorrow20:00
kgriffsok, just checking20:00
cppcabrerathat enhancement eliminates the metadata hack, for one. :D20:00
kgriffsthat docstring at the top has "Name" and "Host" capitalized?20:07
kgriffsactually, looks like other caps snuck in there as well20:08
kgriffscppcabrera: ^^^20:08
cppcabrerathose caps are intentional20:09
cppcabreraI'm attempting to connet the rationale of the one-letter key to the word20:09
cppcabreraassociations. :)20:10
kgriffsmaybe a slightly different syntax would make it more clear20:13
kgriffs'p': [p]roject name :: str20:13
kgriffsTBH, i don't think it is worth worrying about. People are smart. They will figure it out20:14
kgriffsthe underscores just look strange to me, I guess20:14
* kgriffs is picking nits20:14
cppcabrerait's cool. I'll take all the ideas.20:14
cppcabreraI'm actually pretty fond of the [p] syntax, so I'll keep that in mind for the future.20:15
cppcabreraP was the first thing that came to mind.20:15
kgriffsaren't you basically taking everything except '_id' here?20:15
cppcabreraIs that overspecifying things?20:18
cppcabreraI think I got a bit zealous with fields while writing this. :P20:18
openstackgerritA change was merged to openstack/marconi: proxy: memory storage driver + tests + fixes
kgriffswhat do you think about only specifying {'_id': 0}20:20
kgriffssame goes for L6620:23
cppcabreraI think it's a great idea - keeps the impl. flexible20:23
cppcabreraIn case schema changes20:23
kgriffsok, moving on...20:25
kgriffsL94: multi is False by default iirc20:27
kgriffs"The default value for multi is currentlyFalse, but this might eventually change to True"20:28
kgriffsah, gtk20:28
kgriffslet's keep that in there.20:28
cppcabreracool. I had to read pymongo docs a lot.20:29
cppcabreraI essentially learned my mongo off of this proxy dev.20:29
kgriffsthe name is a a little awkward...20:42
kgriffshow about "PartitioningController"20:43
kgriffsor just "PartitionController"20:43
* kgriffs picking more nits20:43
kgriffsL 4620:44
kgriffsI mean20:44
kgriffsthere again, you should only need '_id': 020:44
cppcabreraI'm inclined to stick with PartitionsController since it handles many. :x20:46
cppcabreraYes, the '_id' fields problem is globally noted.20:46
kgriffsmoving on20:46
cppcabreraI'll come around and fix the fields issue in a single patch targeting just that.20:46
cppcabreraIt's rather prevalent.20:46
kgriffsFWIW, I think "normalize" may not be the best name for that function, but I know it is used elsewhere already, so until we change it globally this is ok20:48
kgriffsany other patches you want me to ninjapprove?20:49
cppcabreraThe associated unit test patches that zyuan has been working on.20:49
cppcabreraI think I have them listed in the issues pad...20:50
* cppcabrera checks20:50
cppcabrerayup, they're there20:50
cppcabreraThanks for looking through this, kgriffs!20:50
* kgriffs doesn't want to miss his flight20:51
kgriffsmarconi service returns 20420:54
kgriffswhy the change to 200 here?20:54
kgriffszyuan: ^^^20:54
kgriffsalso, it needs on_head20:54
cppcabreranoted re: on_head20:56
cppcabreraMany resources are missing this.20:57
kgriffswell, health needs on_head since LB's like to use it20:57
openstackgerritA change was merged to openstack/marconi: proxy: mongodb storage driver
kgriffswhat's the story on this?20:59
cppcabreraI haven't looked too thoroughly into that, but zyuan mentioned headers...21:02
cppcabreraI may be able to find the falcon comments21:02
cppcabrerakgriffs: ^^21:02
cppcabreraI'm not sure if capitalize is necessary, but that's the story.21:03
kgriffsso there are two issues at play21:03
kgriffsone is that one in falcon21:03
kgriffswhen you set header and want to override an earlier one21:03
kgriffsthere is a nuance there with WSGI and/or HTTP that is on the tip of my tongue...21:04
*** malini_afk is now known as malini21:04
cppcabreraHTTP allows duplicate headers, WSGI does not, IIRC21:05
cppcabreraI'm heading out for the night. Let me know if you find any more issues, kgriffs! I'll fix them up tomorrow. So far, the majority of the issues have been on the easy side to handle. ;)21:07
* cppcabrera hopes he hasn't jinxed it21:07
cppcabreraGood night~21:07
*** cppcabrera has left #openstack-marconi21:08
*** oz_akan_ has quit IRC21:11
*** iampims has joined #openstack-marconi21:11
*** oz_akan_ has joined #openstack-marconi21:12
*** oz_akan_ has quit IRC21:16
*** gordonsim has quit IRC21:48
*** amitgandhi has quit IRC21:56
*** amitgandhi has joined #openstack-marconi21:56
*** hub_cap has left #openstack-marconi21:58
*** oz_akan_ has joined #openstack-marconi22:22
*** amitgandhi has quit IRC22:22
*** mpanetta has quit IRC22:23
*** tedross has quit IRC22:26
*** oz_akan_ has quit IRC22:28
*** openstack has joined #openstack-marconi22:31
*** cthulhup has joined #openstack-marconi22:50
*** cthulhup has quit IRC22:54
openstackgerritKurt Griffiths proposed a change to openstack/marconi: fix(mongodb): Marker generation creates a bottleneck under heavy load
*** kgriffs is now known as kgriffs_afk23:15
*** malini is now known as malini_afk23:50

Generated by 2.14.0 by Marius Gedminas - find it at!