Tuesday, 2014-03-18

*** anamalagon has quit IRC00:12
*** vkmc has quit IRC00:21
*** kgriffs is now known as kgriffs_afk00:23
*** rektide has joined #openstack-marconi00:29
*** jqiu has quit IRC00:41
*** prashanthr_1 has joined #openstack-marconi00:51
*** jraim has quit IRC00:53
*** mkoderer has quit IRC00:54
*** jraim has joined #openstack-marconi00:55
*** mkoderer has joined #openstack-marconi00:55
*** prashanthr_1 has left #openstack-marconi00:56
*** mpanetta has joined #openstack-marconi01:04
*** mkoderer has quit IRC01:12
*** mkoderer has joined #openstack-marconi01:14
*** rossk has quit IRC01:27
*** nosnos has joined #openstack-marconi01:31
*** jergerber has joined #openstack-marconi01:37
*** nosnos has quit IRC01:38
*** nosnos has joined #openstack-marconi01:38
*** mpanetta has quit IRC01:54
*** haomai___ has quit IRC02:28
*** haomaiwang has joined #openstack-marconi02:28
*** flwang has quit IRC02:33
prashanthr_Can i know which are the storage back ends under which marconi has been tested successfully ?03:12
*** flwang has joined #openstack-marconi03:16
*** jergerber has quit IRC03:21
*** nosnos has quit IRC03:58
*** ayoung is now known as ayoung__zzzZZZZ04:06
*** mwagner_lap is now known as mwagner_zzz04:25
*** chandan_kumar has joined #openstack-marconi04:31
*** nidhi has joined #openstack-marconi04:39
*** haomaiw__ has joined #openstack-marconi04:57
*** nosnos has joined #openstack-marconi04:59
*** haomaiwang has quit IRC05:01
nidhino longer affects: marconi/icehouse Changed in marconi: milestone: icehouse-3 → none05:26
nidhidoes this mean the issue has been solved or that it is not required now05:26
*** AAzza has joined #openstack-marconi06:12
*** nidhi has quit IRC06:28
*** balajiiyer has left #openstack-marconi06:40
*** saju_m has joined #openstack-marconi06:44
*** saju_m has quit IRC06:45
*** saju_m has joined #openstack-marconi06:46
*** AAzza has quit IRC06:49
*** AAzza has joined #openstack-marconi07:10
*** nidhi has joined #openstack-marconi07:21
*** haomaiw__ has quit IRC07:43
*** haomaiwang has joined #openstack-marconi07:43
*** haomaiw__ has joined #openstack-marconi07:48
*** haomaiwang has quit IRC07:51
*** nidhi has quit IRC07:55
*** ykaplan has joined #openstack-marconi08:21
*** njirap has joined #openstack-marconi08:21
*** saju_m has quit IRC08:50
*** njirap has quit IRC08:55
*** haomaiw__ has quit IRC08:55
*** njirap has joined #openstack-marconi08:56
*** saju_m has joined #openstack-marconi08:56
*** flwang has quit IRC09:08
*** haomaiwa_ has joined #openstack-marconi09:13
*** njirap has quit IRC09:18
*** njirap has joined #openstack-marconi09:20
*** renlt has joined #openstack-marconi09:29
*** renlt has quit IRC09:35
*** jamieh has joined #openstack-marconi09:48
*** jamieh is now known as Guest8345609:48
*** njirap has quit IRC09:49
*** ykaplan has quit IRC09:50
*** haomaiw__ has joined #openstack-marconi09:55
*** haomaiwa_ has quit IRC09:56
*** nosnos_ has joined #openstack-marconi10:02
*** nosnos has quit IRC10:05
*** rwsu has quit IRC11:13
*** rwsu has joined #openstack-marconi11:14
*** tedross has joined #openstack-marconi11:30
*** nosnos_ has quit IRC11:30
*** nosnos has joined #openstack-marconi11:31
*** mwagner_zzz has quit IRC11:35
*** westmaas has quit IRC11:36
*** westmaas has joined #openstack-marconi11:37
*** jcru has joined #openstack-marconi12:06
*** jcru has quit IRC12:07
*** jcru has joined #openstack-marconi12:07
*** westmaas has quit IRC12:10
*** westmaas has joined #openstack-marconi12:11
*** jcru has quit IRC12:12
*** AAzza has quit IRC12:22
*** sriram has joined #openstack-marconi12:32
*** sriram has joined #openstack-marconi12:32
*** prashanthr_ has quit IRC12:35
*** flaper87|afk is now known as flaper8712:35
flaper87guess who's back? Back on-line... Guess who's back? Back on-line12:36
*** ayoung__zzzZZZZ is now known as ayoung13:02
*** flaper87 is now known as flaper87|afk13:10
sriramwelcome back flaper87 :)13:10
*** alcabrera has joined #openstack-marconi13:11
*** njirap has joined #openstack-marconi13:11
alcabreraGood morning! :D13:11
*** flaper87|afk is now known as flaper8713:19
*** AAzza has joined #openstack-marconi13:23
maliniflaper87: https://review.openstack.org/#/c/81001/ can use your opinion13:27
malinithis is wrt the mongo version at the gate13:27
malinithere is some discussion going on in the dev ML wrt mongo versions in the gate13:32
*** mpanetta has joined #openstack-marconi13:33
*** whenry has quit IRC13:34
alcabrerathat's good to know13:34
*** kgriffs_afk is now known as kgriffs13:39
*** njirap has quit IRC13:39
alcabrerakgriffs: hey. :)13:41
kgriffsso, what's everyone up to today?13:41
mpanettaReading that link you sent me :)13:43
alcabrerareviewing GSoC & OPW proposals, thinking about Marconi Notifications, prepping for the Marconi meeting, and tinkering with genetic algorithms in Haskell. :D13:43
alcabrera**Marconi meeting(s)13:43
*** mwagner_lap has joined #openstack-marconi13:43
kgriffsmpanetta: was that the kernel tuning link?13:43
mpanettaGood stuff13:43
mpanettaThanks for sending it to me!13:43
kgriffsyw! Someone else mentioned it to me and I thought I'd pass it along.13:44
*** stritz has joined #openstack-marconi13:44
kgriffsfor everyone else, here's the link again: http://goo.gl/rmsxIs13:44
mpanettaOhh I see, they get the short version :P13:44
alcabreraooh, nice13:44
*** mpanetta has quit IRC13:45
kgriffsmpanetta: I already updated my gist according to that article - it was pretty close already. https://gist.github.com/kgriffs/402783513:45
*** amitgandhi has joined #openstack-marconi13:45
flaper87alcabrera: kgriffs heeeey13:45
*** amitgandhi has quit IRC13:45
flaper87guess who's back on-line ?13:45
*** mpanetta has joined #openstack-marconi13:45
alcabreraflaper87: yo! welcome back! :D13:45
kgriffsflaper87: yo, aren't you supposed to be missing in action (MIA)?13:45
mpanettaOnly issue, I wonder how many of these have to be tuned on the host as well as the guest...13:45
alcabrerare: perf tuning, I think this book will be an *amazing* read: http://www.amazon.com/Systems-Performance-Enterprise-Brendan-Gregg/dp/013339009813:45
kgriffsdarn, now I can't say all those things I had planned to say13:45
mpanettaGrr dumb connection13:45
*** amitgandhi has joined #openstack-marconi13:46
flaper87kgriffs: You've no idea how addicted to internet I am13:46
flaper87So, I'm at my grandma's place now (long story) but she didn't have internet 'til now13:46
kgriffsalcabrera: nice book13:46
flaper87I called my ISP and asked to activate internet connection here13:46
flaper87they, surprissingly enough, did it in NO TIME13:47
flaper87and I brought an old router here13:47
kgriffsalcabrera: I came across this blog recently that is also relevant: http://mechanical-sympathy.blogspot.com/13:47
flaper87BUT, I forgot the fucking power adapter of that router at home13:47
kgriffsnothing is every easy13:47
flaper87I went to get a new power adapter and they didn't have one in that store13:47
flaper87so, since I'm not addicted to internet </sarcasm> I bought a new router13:47
alcabrerahehe, flaper87! Going to great lengths to keep the world (and you) connected!13:48
flaper87Man, I've to attend to to some anonymous internet addicts group13:48
flaper87anyway, enough about me!13:48
alcabrerakgriffs: bookmarked. much reading. :)13:48
flaper87I should be around for our meeting but I'm not sure I'll make it to today's graduation review13:49
flaper87we'll see13:49
kgriffsflaper87: ok. I am putting together a wiki page to answer all the stuff from http://git.openstack.org/cgit/openstack/governance/tree/reference/incubation-integration-requirements13:49
kgriffsmalini: you online?13:49
malinikgriffs: yes for another 5 min13:50
kgriffsok, sorry to hear about your daughter - hope she feels better soon!13:50
kgriffsmalini: will you be in today's team mtg?13:51
malinikgriffs: I will miss the team meeting13:51
maliniBut will be online for the grad eval13:51
kgriffsmalini: ok, any outstanding items for tempest/devstack?13:51
*** njirap has joined #openstack-marconi13:52
*** flwang has joined #openstack-marconi13:52
malinikgriffs: https://bugs.launchpad.net/devstack/+bug/129213613:52
malinithe above might be due to mongo version at the gate13:53
malinithere is a discussion going on in ML abt mongo version at the gate13:53
maliniIt'll be awesome if someboy in marconi team can respond as well13:53
malinithe ceilometer team is driving the discussion there13:53
malinithe other is https://bugs.launchpad.net/devstack/+bug/129406813:54
kgriffsflaper87: can you respond to that email thread?13:54
malinikgriffs: I am adding a post test hook to find out what is going on there13:54
maliniSee https://review.openstack.org/#/c/81249/13:54
malinialcabrera: Can you please take over https://review.openstack.org/#/c/81249/13:54
flaper87kgriffs: malini sure thing?13:55
maliniI have some feedback, which shud be easy to address13:55
flaper87is that thread new?13:55
flaper87I haven't read it13:55
* flaper87 is still going through email13:55
* alcabrera clicks the link13:55
maliniflaper87: [openstack-dev] [Ceilometer][QA][Tempest][Infra] Ceilometertempest testing in gate13:55
malinialcabrera: I cannot get to tht patch in the next 2-3 hours13:55
flaper87funny that there's no mention of mongo in the title13:56
alcabreramalini: I'll fix that up, np13:56
malinialcabrera: thanks a lot!! Appreciate your help !!13:56
maliniGetting that patch merged will give us some more info on why sqla driver doesnt work in the gate13:56
*** nosnos has quit IRC13:56
maliniAs is marconi can run with only sqlite in the gate13:57
malinikgriffs: In short at ou current state, marconi can run with sqlite driver at the gate13:57
flwangmorning all13:57
malinibye for now..will be back later13:57
flaper87malini: wait, why just sqlite?13:58
flaper87Why not mysql?13:58
maliniflaper87: no idea ..https://review.openstack.org/81249 shud give more info once merged13:58
* flaper87 wonders what's going on with Marconi's storage layer lately13:58
alcabreraflwang: good morning. :)13:58
kgriffsflaper87: it would be awesome if you have some time to try mysql in devstack and help malini out13:59
alcabrera(package-install yaml-mode; now my emacs is a yaml editing monster)13:59
flaper87kgriffs: that's the point, I did and it worked :P13:59
* flaper87 confused13:59
flaper87yeah, as pretty much everything regarding devstack and the gate13:59
flaper87it always works locally13:59
kgriffswe need docker, pronto!14:00
flaper87I'll give it another try and see if something changed but, I'm sure as hell it worked14:00
kgriffsflaper87, alcabrera: I need some help figuring out the pypy issue that is blocking all our patches14:00
kgriffslooks like a library is incompatible14:00
flaper87kgriffs: ah yeah, that too, I was looking at it yday14:00
flaper87we should try syncing the requirements in marconi14:00
flaper87that same error happened on other modules a couple of weeks ago14:01
flaper87I thought we were safe14:01
flaper87that might have been exactly the issue14:01
flaper87didn't we synced the requirements lately?14:01
flaper87I remember seeing a patch for the requirements14:01
flaper87lets try reverting that patch, submit the revert for review and see if it succeedes14:01
alcabreradang, an incompatible library is blocking pypy? :/14:03
alcabrerathat's frustrating14:03
kgriffsso, it works fine if you run tox on your local box14:05
kgriffsbut the gate runs a different command line14:05
*** whenry has joined #openstack-marconi14:05
kgriffsflaper87: I just tried a sync and it looks like we are already up to date14:06
*** njirap has quit IRC14:07
kgriffsflaper87: do you want to try the revert?14:07
flaper87kgriffs: sure, lemme give that a try14:08
alcabreramccabe package is driving pypy crazy, as far as I can tell14:09
alcabrerabut it works on my machine... @_@14:10
*** Kuwagata has joined #openstack-marconi14:14
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Revert "Updated from global requirements"  https://review.openstack.org/8127014:19
flaper87alcabrera: kgriffs ^^^^^ lets see what happens14:19
flaper87alcabrera: kgriffs gtg guys, I'll be back later14:19
*** prashanthr_ has joined #openstack-marconi14:20
alcabreraflaper87: thanks! o/14:20
*** njirap has joined #openstack-marconi14:20
*** balajiiyer has joined #openstack-marconi14:20
flwangflaper87: cross fingers14:20
flaper87ok, that didn't work14:21
*** AAzza has quit IRC14:22
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Revert "Updated from global requirements"  https://review.openstack.org/8127114:22
flaper87kgriffs: alcabrera flwang ^^^^ second revert14:22
alcabrerahmmm, I'm dubious about that last one w/ six being reverted.14:23
flaper87the second one didn't work either https://jenkins05.openstack.org/job/gate-marconi-pypy/19/console14:24
flaper87we need to go deeper in the module failing and see what's happening there14:24
flaper87I gtg now guys, I'll be back in a bit14:24
*** flaper87 is now known as flaper87|afk14:25
*** jcru has joined #openstack-marconi14:25
*** jergerber has joined #openstack-marconi14:27
*** ykaplan has joined #openstack-marconi14:45
openstackgerritFei Long Wang proposed a change to openstack/marconi: Add requirements sphinx and oslo.sphinx  https://review.openstack.org/8128014:48
alcabrerakgriffs: will we be meeting today (in 3 mins)?14:57
*** cpallares has joined #openstack-marconi14:59
cpallareskgriffs: graduation meeting? For some reason I have that on my calendar.15:00
kgriffscpallares: team meeting now, graduation review is this afternoon at 3pm CT15:01
cpallaresah ok thanks kgriffs15:02
*** chandan_kumar has quit IRC15:02
*** vkmc has joined #openstack-marconi15:03
*** vkmc has quit IRC15:03
*** vkmc has joined #openstack-marconi15:03
*** chandan_kumar has joined #openstack-marconi15:04
*** AAzza has joined #openstack-marconi15:10
*** anamalagon has joined #openstack-marconi15:15
*** Guest83456 has quit IRC15:20
*** jamieh has joined #openstack-marconi15:20
*** jamieh is now known as Guest6230815:21
*** saju_m has quit IRC15:23
*** saju_m has joined #openstack-marconi15:40
*** reed has joined #openstack-marconi15:50
*** rupsky has joined #openstack-marconi15:56
*** saju_m has quit IRC15:57
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: fix(pypy-gate): short term setuptools workaround  https://review.openstack.org/8129315:57
*** rupsky has quit IRC15:58
*** tjanczuk has joined #openstack-marconi16:00
*** alcabrera is now known as alcabrera|afk16:03
flwangkgriffs: ping16:04
kgriffsflwang: pong16:04
flwangkgriffs: may I use your chart on the HongKong summit session for internal training?16:05
kgriffswhich chart?16:05
flwangkgriffs: delivered by you, flaper87 and Allan16:05
kgriffsah, from the slides16:06
kgriffssure, no problem16:06
kgriffsthe PDF should be up there16:06
flwangkgriffs: great, thanks16:06
kgriffsI may have the original graphics somewhere if you need them16:06
amettsametts wants royalties16:06
flwangkgriffs: btw, seems we were already using the 'topics' ATM16:06
flwangametts: oh, hi Allan :D16:07
flwangkgriffs: thanks, I think current pdf should be ok16:08
*** alcabrera|afk is now known as alcabrera16:10
*** stritz has quit IRC16:11
tjanczukfolks, I have a few follow up questions from openstack-meeting-alt16:13
*** balajiiyer has quit IRC16:13
*** balajiiyer has joined #openstack-marconi16:13
tjanczukkgriffs: you mentioned AMQP would be on the table in the Juno time frame and there are no current plans for STOMP or MQTT. In terms of AMQP, is it considered at the transport or storage layer? I saw some older notes indicating this would be used at the storage layer, not transport.16:14
kgriffsso, flaper87|afk has been looking at it for the storage layer16:15
kgriffsTBH, I don't know what the benefit would be for making it a transport - why not just use Qpid or Rabbit instead?16:15
tjanczukWhat was the perceived benefit of using it at the storage layer?16:16
kgriffstjanczuk: iirc, people wanted to reuse their existing AMQP server and not have to manage another box, but most people who are interested in using Marconi actually like it because it is easier to deploy and use than AMQP in some ways16:17
kgriffstjanczuk: flaper87 will probably have further thoughts on this - hopefully he will pop in later, but if not he should be around tomorrow16:18
kgriffsdon't get me wrong; if there is a strong use case for it, I'm happy to consider any of these ideas16:19
kgriffsjust remembered16:19
kgriffssomeone said that having AMQP support could help bridge to other systems. But I think the consensus was that building a standalone bridge service would be a better direction architecturally. We might need to revisit that discussion.16:20
tjanczukI am trying to understand the relationship between other messaging solutions (Rabbit. Qpid, ActiveMQ etc) and Marconi. What is the high level goal of Marconi: to provide some form of an abstraction over those (perhaps by bridging from HTTP protocol over) or an alternative solution that offers a form of differentiation?16:20
kgriffsbalajiiyer: can you let me know when you post to the ML?16:22
tjanczukThe reason I am confused is that notion of using AMQP at the storage layer. I understand plugging in Mongo or MySQL at that level, but AMQP implies a whole message broker semantics (at least 0.9)16:23
kgriffstjanczuk: yep, and you aren't alone. There are two camps of reactions to Marconi16:24
kgriffsfirst, "yay, this is great, it is a more natural fit for building cloud apps than AMQP and is easier to operate at scale with HA"16:24
*** ykaplan has quit IRC16:25
kgriffssecond camp: "yay, this will create a standard API over the myriad of existing message brokers out there."16:25
kgriffsoh, one more16:25
kgriffsthird camp: "why don't you guys just create a provisioning service for [fill in the blank]"?16:25
kgriffsIMO, marconi's mission is really to address the first one, and build on that to provide notifications-as-a-service16:26
kgriffsso, reasons people like marconi over AMQP include:16:27
kgriffs1. first-class, idiomatic HTTP API16:27
kgriffs2. hybrid feed+claim semantics16:28
kgriffs3. guaranteed once-and-only-once delivery and FIFO for a single producer to a single queue16:28
kgriffs(note that #3 will likely become optional, depending on queue "flavor")16:29
kgriffs4. multi-tenant API16:29
alcabreraisn't (3) really (3) and (4), since OaOO exists apart from FIFO?16:29
kgriffs5. easier HA than Rabbit16:29
kgriffsalcabrera: possibly.16:29
*** flaper87|afk is now known as flaper8716:30
kgriffs6. easier scale-out16:31
* flaper87 kinda back16:31
flaper87How was the meeting ?16:31
kgriffsso, I think that there would be room for a "AMQP" or whatever provisioning service in OpenStack16:32
flaper87any updates ? highlights ?16:32
* flaper87 gets the link16:32
tjanczukRe: 1. Is the HTTP API appealing to people because of the environment of the clients (e.g. web, mobile app), or because HTTP is just plain simpler to use than AMQP?16:32
kgriffstjanczuk: I've heard both reasons16:32
kgriffswe are also thinking about supporting json-p and/or CORS so web apps can hit the service directly16:33
*** rossk has joined #openstack-marconi16:33
flaper87The http API is definitely easier to use than AMQP but I'd say there are more things to it than just being easier16:33
flaper87for instance, you've to know "less" things about the messaging technology you're using and can focus on what you need from it16:34
kgriffsflaper87: went well. Pecan eval is on the wiki. alcabrera is helping with pypy thing. I am working on writing up a wiki page that gives answers to all the latest grad reqs.16:34
*** chandankumar_ has joined #openstack-marconi16:34
flaper87that is to say, that with marconi you won't need to worry about queues, exchanges etc. You just need to worry about sending a message and consuming it16:34
tjanczukRe 2: The "claim" semantics is somethigng that is quite ubiquitous in other solutions; the "feed" semantics is quite unique to Marconi IMO. Where there specific scenarios that were targeted with the feed semantics? Having such semantics at the transport level is quite restrictive in terms of what you can use at the storage level. Specifically, I am not sure how you could represent the feed semantics using AMQP at the stora16:34
flaper87kgriffs: +1 thanks man!16:35
flaper87alcabrera: any luck with  the pypy issue ?16:35
alcabreraflaper87: waiting on the gate. We should have a short-term workaround as provided by adrian_otto and the ML. :)16:35
alcabreraflaper87: https://review.openstack.org/#/c/81293/16:36
kgriffstjanczuk: there are certain messaging patterns that the feed semantics affords in an elegant way (i.e., pub sub)16:36
kgriffstjanczuk: and yes, it does restrict backends. That is what I alluded to when I said that we were considering allowing drivers to implement only a subset of the API16:36
*** chandan_kumar has quit IRC16:37
flaper87alcabrera: sooooooooooo, it's setuptools fault ?16:37
kgriffstjanczuk: another use case: having an auditor observe the stream of messages as they are produced and consumed by workers16:37
alcabreraflaper87: yup - 3.3 broke the world, again. :P16:37
*** oz_akan_ has joined #openstack-marconi16:37
tjanczukkgriffs: I see, I get it now. So if AMQP technology would be used as storage, feed semantics would not be supported?16:37
flaper87oh man16:37
kgriffstjanczuk: that seattle companies' service can be a real pain when trying to audit/debug what is going on16:38
kgriffs(gained from personal experience - in a previous company we used lots of their cloud sevices)16:38
flaper87tjanczuk: there are certainly some things about the AMQP storage that will be hidden and others won't exists.16:38
tjanczukkgriffs: there are at least 2 I can think of you are alluding to: Amazon and Microsoft ;)16:39
flaper87I think the AMQP driver will be implemented on top of amqp 1.016:39
flaper87at least that's what I've been lenaing towards in my last brainstorms16:39
flaper87something on top of qpid proton that can talk to rabbitmq, qpid and other things16:39
flaper87there are some bits that still worry me but we'll figure those out as we go16:40
*** balajiiyer has quit IRC16:40
flaper87alcabrera: pypy succeeded but the docs gate failed16:40
*** njirap has quit IRC16:40
alcabreraah, I see what I need to do16:41
alcabreraI need to layer my patch on top of flwang's16:41
tjanczukflapper87: interesting. Why are you leaning more towards 1.0 vs 0.9?16:41
alcabrerasince that one fixes the doc gate16:41
kgriffsflaper87, tjanczuk: this discussion makes me think we need a FAQ on our wiki.16:41
alcabreradouble breaking gates are troublesome, fwiw16:41
kgriffsalcabrera: but... the doc is non-voting, right?16:42
alcabrerahmmm, oh16:42
alcabrerawell, that simplifies this a little. :)16:42
kgriffsmaybe not. we'll see16:42
kgriffscome on tempest!16:43
kgriffsand, for the record16:43
flaper87kgriffs: alcabrera it's voting16:43
kgriffsob boy16:44
alcabrerathat would lead me to believe that both gates need to be fixed in the same patch16:44
kgriffsflwang: still around?16:44
kgriffswe need to merge these two patches16:44
*** balajiiyer has joined #openstack-marconi16:44
kgriffsalcabrera: you took the words right out of my keyboard!16:44
kgriffsalcabrera: make it so16:44
kgriffsalcabrera: And blowing into maximum warp speed, you appeared for an instant to be in two places at once.16:45
tjanczukkgriffs: Re 4: easier HA than Rabbit. Why would Marconi be easier to HA than Rabbit? I can see how the HTTP transport layer can be HAed easily given a stateless implementation, but the real HA heavy lifting happens at the storage level, especially when delivery guarantees are involved. How would Marconi storage layer be easier to HA than Rabbit?16:45
kgriffstjanczuk: specifically, mongodb is easier16:46
kgriffsit is basically HA out of the box16:46
kgriffsassuming you have a replica set and the app handles auto-reconnects16:46
alcabrerakgriffs: on it - going to fuse some patches~16:47
kgriffsalso, when things crash I've heard that it gets pretty ugly for Rabbit, at least16:47
kgriffs(things, meaning Rabbit)16:48
kgriffsbut that is anecdotal16:48
tjanczukI think there were some growing pains with Rabbit, but I am hearing (also anecdotal), these are largely remedied with the new active/active clustering.16:50
kgriffsvery possible16:51
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: fix(gate): one for pypy, one for docs  https://review.openstack.org/8129316:51
tjanczukOne more question that popped in my mind when looking at the HTTP APIs: was there a reason to restrict messages to JSON? Is there a plan to relax this?16:51
kgriffstjanczuk: meaning XML?16:52
kgriffsa couple things there16:52
tjanczukNo format in particular. Just not being restrictive. Leverage fully Content-Type.16:53
kgriffswe did JSON first because we were short on hands and didn't have time to do both and make XML look xml-y and JSON json-y16:53
* flaper87 is still planning to write another transport for marconi16:53
kgriffsI personally am open to other types. We have talked about XML and msgpack, for instance.16:53
flaper87at some point, when there will be time for that16:54
tjanczukThere is nothing in the current API that requires Marconi to understand the semantics of the message. Why prescribe the message format in such case?16:54
kgriffsXML is interesting in particular, because Nova is moving away from supporting it. May or may not be temporary.16:54
kgriffstjanczuk: because we have to implement it and maintain that implementation ad-naseum. Also, client libs need to have a standard set of types that work across clouds16:55
kgriffstjanczuk: technically, you could support any content-type you wanted in deployment X16:56
kgriffsyou could fork the code or write an API extension (we don't really have support for the latter yet, but will probably do it at some point)16:56
kgriffstjanczuk: at the last summit it was proposed that we allow arbitrary content-type payloads ala swift16:57
kgriffsis that what you had in mind?16:57
tjanczukI am not sure how swift implements it. I would expect the body of my HTTP message to be opaque to Marconi and therefore of any Content-Type the application developer desires. All the Marconi metadata surrounding that payload would be represented using HTTP headers and/or URL.17:00
kgriffstjanczuk: that is basically what swift does17:09
kgriffsmetadata in the headers can get ugly though, but that is a valid approach17:10
kgriffstjanczuk: we can consider doing this for the v2.0 api but you'll have to help me remember to discuss it. :)17:10
kgriffsmaybe a bp or something17:11
tjanczukSo if a change like that were to happen, how would it be handled with respect to existing implementation? Would be additive, a replacement, or an alternative? I suppose I am asking about API versioning story in Marconi.17:13
kgriffstjanczuk: tbh, we are still fleshing out that story. My personal preference would be to avoid deprecating previous API versions as long as possible.17:14
tjanczukGreat, thanks guys, I learned a lot. I am sure I will be back. What time zones are you in? Or rather, what time zones do you _work_ in? ;)17:15
kgriffsbut even with that in mind, there are plenty of questions, like, how do we handle someone accessing the same messages via v1 and v2?17:15
kgriffstjanczuk: great chat!17:16
kgriffsI'm in central time / US17:16
kgriffssome folks on eastern17:16
kgriffsflaper87 is in Italy17:16
kgriffsflwang is in China i believe17:16
flaper87tjanczuk: as for TZs I work in, erm, kinda -∞ UTC +∞17:18
tjanczukI am in PST. That is, I live in PST and there is some weak correlation to when I work.17:18
tjanczukAnyway, great chat and talk to you soon.17:19
*** tjanczuk has left #openstack-marconi17:21
kgriffsrock on17:22
*** stritz has joined #openstack-marconi17:28
alcabrerathis https://review.openstack.org/#/c/81293/ fixed both gates. Waiting on py26 and tempest - ETA 37 min according to Zuul17:29
alcabreraoh, there goes py2617:29
flaper87alcabrera: thanks for fixing that!17:29
flaper87It's all kgriffs fault17:30
* flaper87 ducks17:30
maliniback o/17:30
malinialcabrera: thanks for dealing with the infra patch17:30
alcabreramalini: welcome back!17:30
alcabreranp. :)17:30
alcabrerapy26 succeeded!17:33
alcabreraso now, even if temptest fails17:33
alcabrerathis patch will get +1 from jenkins17:33
alcabrerawaiting on tempest17:33
flaper87LOL @ "even if tempest fails"17:33
flaper87malini: I'd be really mad ^^^17:33
flaper87malini: go to alcabrera's desk and kick his legs17:33
*** Guest62308 has quit IRC17:34
kgriffsflaper87: that's ok, I'll take the blame *this* time.17:35
kgriffsflaper87: one of these days I'm so going to stop covering for you, tho! ;)17:35
flaper87is there anything I can help with for the graduation meeting ?17:35
flaper87kgriffs: yeah yeah, I know you wont17:36
* flaper87 gives kgriffs some money17:36
*** jamieh has joined #openstack-marconi17:37
* kgriffs whistles as he walks away17:37
flaper87kgriffs: https://review.openstack.org/#/c/81293/ <- A+17:38
*** jamieh is now known as Guest9931517:38
flaper87lets short-circuit the gate17:38
alcabrerashort-circuiting sounds nice, but how? :P17:39
mpanettaApanner in the works :P17:40
flaper87alcabrera: by approving a patch that is still gating, you basically saying to jenikins: "Suck it, I don't care what you say"17:44
flaper87but then, if the merge fails, you'll hear Jenkin's laugh right behind you17:44
balajiiyerflaper87: lol17:44
*** haomaiw__ has quit IRC17:45
flaper87you'll feel Jenkins breathing on your neck17:45
alcabreraah, I see17:45
alcabrerano, don't want that17:45
flaper87and you'll start trembling like a really tiny scared cat17:45
flaper87alcabrera: don't worry, kgriffs approved it so you're safe17:45
alcabreraand this is why we replace our build systems once every few years17:45
*** oz_akan_ has quit IRC17:45
kgriffsand this is why we replace our PTLs once every few years17:47
* kgriffs is going to need some serious therapy when this is over17:47
*** AAzza has quit IRC17:48
flaper87because jenkins scares the hell out of them17:49
*** jcru has quit IRC17:50
flaper87gtg guys! Have a great meeting! I'll try to make it17:51
alcabreraflaper87: o/17:52
*** flaper87 is now known as flaper87|afk17:53
kgriffsmalini: ping17:57
kgriffsalcabrera: ping17:58
alcabrerakgriffs: pong17:58
kgriffsalcabrera: have you done the devstack thing?17:58
kgriffsi mean, are you familiar with running marconi devstack?17:58
alcabrerakgriffs: not at all - no devstack experience on my end18:00
amettskgriffs: I though jay-atl has done it18:01
* ametts should check where his right hand is on the keyboard before he types18:01
alcabrerayeah, jay-atl and malini are our resident devstack experts18:01
kgriffsoh, I need some help answering that question18:02
amettsmalini should also be back online shortly18:03
openstackgerritA change was merged to openstack/marconi: fix(gate): one for pypy, one for docs  https://review.openstack.org/8129318:03
balajiiyerkgriffs: Reg. devstack, in the past month or so jay-atl and malini submitted couple of devstack patches to get Marconi working. Looking at how old that question is, I wonder whether those patches may have solved the issue.18:05
kgriffsbalajiiyer: I think so. I commented to that effect.18:06
kgriffsbut I would like others to chime in too18:06
malinicatching up18:08
kgriffswelcome back!18:10
maliniI think we should submit a patch to devstack to use sqlite as the backend till the mongo/mysql issues are resolved18:12
maliniThis will help folks run marconi on devstack18:12
kgriffsmalini: ok, let's do that18:13
malinikgriffs: I am submitting a patch for that now18:13
kgriffsmalini: do you have the bug handy?18:14
kgriffsI want to schedule it for rc118:14
kgriffsthought I had, but i can't find it18:15
maliniI think we have diff issues wrt mongo & mysql18:15
*** AAzza has joined #openstack-marconi18:15
malinimongo db issue might have something to do with the gate version as well18:16
*** prashanthr_ has quit IRC18:17
malinipatch to use sqlite as default backend https://review.openstack.org/#/c/81326/18:19
kgriffsmalini: the bugs are now assigned to you for rc1: https://launchpad.net/marconi/+milestone/icehouse-rc118:19
alcabreramalini: +118:19
kgriffsothers can help, but you being assigned mean you will report on the progress, is that OK?18:20
maliniI have some patches going to get the logs etc.18:20
maliniif they accept the sqlite backend patch, we can make the tempest job voting18:21
kgriffsjraim: fyi, pecan eval has been published18:23
jraimyep, reading it now18:24
*** Guest99315 is now known as jamie_h18:29
*** haomaiwang has joined #openstack-marconi18:40
alcabrerakgriffs: doc patches failed to merge again. Seems they really want a rebase18:40
kgriffsin oz around?18:46
alcabreraI think not. I believe he's out sick, kgriffs18:46
kgriffsalcabrera: can you rebase it then?18:46
kgriffsalso catherine's18:46
alcabrerasure thing18:46
kgriffsthey should still get credit automatically from a simple rebase18:47
kgriffsalcabrera, balajiiyer, malini, ametts, flaper87|afk: https://wiki.openstack.org/wiki/Marconi/Incubation/Graduation18:47
kgriffsmegan_w: ^^^18:47
kgriffsmeeting in just over an hour18:47
megan_wkgriffs: yep, on the calendar and ready to roll18:47
megan_wkgriffs: let me know if you want me to be prepared to answer anything specific...like what we're seeing from customers here at rack18:48
alcabrerarebase w/ conflicts18:48
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: adds docs directory with dev guide  https://review.openstack.org/7200118:49
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: Marconi Operations Document  https://review.openstack.org/7829418:49
alcabrerathat should do it18:49
kgriffsmegan_w: kk. We may need to leverage Cloud Queues in arguing for marconi's maturity18:51
megan_wsounds good.  let me pull out some of our usage numbers from larger customers.....18:52
*** flaper87|afk is now known as flaper8718:59
openstackgerritAlejandro Cabrera proposed a change to openstack/marconi: docs(readme): update for mysql, daemon server  https://review.openstack.org/8103319:00
*** Kuwagata has left #openstack-marconi19:03
*** saju_m has joined #openstack-marconi19:18
*** jay-atl has left #openstack-marconi19:25
malinikgriffs: We can make the tempest job voting with sqlite backend..https://wiki.openstack.org/wiki/Marconi/Incubation/Graduation  has a couple of references tht imply otherwise19:37
kgriffsmalini: maybe we should go ahead and do that, assuming the sqlite devstack patch is merged19:41
maliniyes..I'll do tht once the sqlite patch is merged19:41
kgriffsmalini: I updated the text for that req.19:42
kgriffslook OK?19:42
malinikgriffs: do you want to mention make it voting 'after https://review.openstack.org/#/c/81326/ is merged'19:46
amitgandhikgriffs: quick update on horizon - in case the TC bring it up ...19:47
amitgandhithe horizon work itself is done, currently its blocked due to some devstack issues, and there was an issue with the way marconi was communicating with keystone that horizon wasnt liking which flavio was looking into19:48
amettsWow, amitgandhi blamed devstack, keystone, horizon, and flavio -- all in once sentence. :)19:49
maliniI hope we'll get to the last of the devstack issues in the near future ^o)19:51
*** jamie_h has quit IRC19:51
* amitgandhi feels like hitting devstack with a big devstick19:51
* alcabrera finds that we're often devstuck19:52
malinidevstack needs more tests19:52
kgriffsamitgandhi: kk, thanks19:53
*** cpallares has quit IRC19:55
amitgandhiis the TC meeting on #openstack-meeting ?19:55
alcabrera4 minutes 'til meeting time19:56
amettsamitgandhi:  Yes, I think so.19:56
*** cpallares has joined #openstack-marconi19:57
amitgandhi30 seconds...19:59
kgriffsmalini: if they ask for specifics on that, we can let them know about  https://review.openstack.org/#/c/81326/19:59
malinisounds good20:00
amitgandhitheyre ready for us20:00
*** stritz has quit IRC20:02
*** AAzza has quit IRC20:13
alcabreraThat's all for the TC meeting for now.20:26
*** flaper87 is now known as flaper87|afk20:27
openstackgerritA change was merged to openstack/marconi: adds docs directory with dev guide  https://review.openstack.org/7200120:28
*** flaper87|afk is now known as flaper8720:29
*** jay-atl has joined #openstack-marconi20:35
openstackgerritA change was merged to openstack/marconi: Marconi Operations Document  https://review.openstack.org/7829420:38
*** flaper87 is now known as flaper87|afk20:39
*** ayoung is now known as ayoung_Dad_mode20:48
*** alcabrera has left #openstack-marconi20:51
*** sriram has quit IRC20:59
*** mpanetta has quit IRC21:02
*** jamieh has joined #openstack-marconi21:02
*** jamieh is now known as Guest274921:03
*** tedross has quit IRC21:06
*** Guest2749 has quit IRC21:07
*** kgriffs is now known as kgriffs_afk21:11
*** malini is now known as malini_afk21:14
*** balajiiyer has left #openstack-marconi21:16
*** ayoung_ has joined #openstack-marconi21:21
*** ayoung_ is now known as ayoung21:22
*** ayoung_Dad_mode has quit IRC21:24
*** oz_akan_ has joined #openstack-marconi21:29
*** cpallares has quit IRC21:41
*** oz_akan_ has quit IRC21:42
*** saju_m has quit IRC21:42
*** amitgandhi has quit IRC21:45
*** kgriffs_afk is now known as kgriffs21:55
*** reed has quit IRC21:58
*** malini_afk is now known as malini22:10
*** mwagner_lap has quit IRC22:21
*** jcru has joined #openstack-marconi22:44
*** jcru has quit IRC22:45

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!