Wednesday, 2013-11-13

openstackgerritAngus Salkeld proposed a change to openstack/python-marconiclient: Fix some links in the README
*** amitgandhi has joined #openstack-marconi00:04
*** amitgandhi has quit IRC00:07
*** amitgandhi has joined #openstack-marconi00:07
*** wirehead_ has quit IRC00:09
openstackgerritA change was merged to openstack/python-marconiclient: Fix some links in the README
*** jamieh__ has quit IRC00:20
*** tedross has quit IRC00:20
*** wirehead_ has joined #openstack-marconi00:21
*** jcru has quit IRC00:26
*** adrian_otto has joined #openstack-marconi00:47
*** adrian_otto has quit IRC01:06
*** nosnos has joined #openstack-marconi01:13
*** nosnos has quit IRC01:13
*** nosnos has joined #openstack-marconi01:14
*** ayoung has joined #openstack-marconi01:20
*** amitgandhi has quit IRC01:32
*** oz_akan_ has quit IRC01:39
*** oz_akan_ has joined #openstack-marconi01:40
*** oz_akan_ has quit IRC01:44
*** cpallares has quit IRC02:21
*** oz_akan_ has joined #openstack-marconi02:22
*** nosnos_ has joined #openstack-marconi02:27
*** haomaiwang has joined #openstack-marconi02:30
*** nosnos has quit IRC02:30
*** amitgandhi has joined #openstack-marconi02:33
*** amitgandhi has quit IRC02:39
*** amitgandhi has joined #openstack-marconi02:43
*** amitgandhi has quit IRC02:47
*** fifieldt has joined #openstack-marconi02:59
*** amitgandhi has joined #openstack-marconi03:36
*** amitgandhi has quit IRC03:40
*** oz_akan_ has quit IRC03:53
*** amitgandhi has joined #openstack-marconi04:37
*** amitgandhi has quit IRC04:42
*** nosnos_ has quit IRC04:44
*** nosnos has joined #openstack-marconi04:44
*** amitgandhi has joined #openstack-marconi04:47
*** amitgandhi has quit IRC04:51
*** adrian_otto has joined #openstack-marconi05:17
*** adrian_otto has quit IRC05:37
*** amitgandhi has joined #openstack-marconi05:37
*** amitgandhi has quit IRC05:42
*** amitgandhi has joined #openstack-marconi06:38
*** adrian_otto has joined #openstack-marconi06:39
*** amitgandhi has quit IRC06:42
*** flaper87|afk is now known as flaper8707:08
*** nosnos_ has joined #openstack-marconi07:11
*** nosnos has quit IRC07:14
*** adrian_otto has quit IRC07:30
*** nosnos_ has quit IRC07:31
*** nosnos has joined #openstack-marconi07:32
*** amitgandhi has joined #openstack-marconi07:38
*** amitgandhi has quit IRC07:43
*** Krast has joined #openstack-marconi08:32
*** amitgandhi has joined #openstack-marconi08:39
*** amitgandhi has quit IRC08:43
*** amitgandhi has joined #openstack-marconi08:49
*** amitgandhi has quit IRC08:54
*** yassine has joined #openstack-marconi09:14
*** yassine_ has joined #openstack-marconi09:18
*** jamieh__ has joined #openstack-marconi09:20
*** nosnos_ has joined #openstack-marconi09:33
*** nosnos has quit IRC09:37
*** amitgandhi has joined #openstack-marconi09:40
*** amitgandhi has quit IRC09:44
*** Krast has quit IRC10:07
*** Krast has joined #openstack-marconi10:07
*** Krast has quit IRC10:07
*** amitgandhi has joined #openstack-marconi10:41
*** amitgandhi has quit IRC10:45
*** amitgandhi has joined #openstack-marconi10:51
*** amitgandhi has quit IRC10:55
*** amitgandhi has joined #openstack-marconi11:41
*** amitgandhi has quit IRC11:45
*** nosnos_ has quit IRC11:52
*** nosnos has joined #openstack-marconi11:53
*** nosnos has quit IRC11:53
*** nosnos has joined #openstack-marconi11:54
*** nosnos has quit IRC11:58
*** nosnos has joined #openstack-marconi11:59
*** nosnos has quit IRC12:04
*** fifieldt has quit IRC12:25
*** yassine_ has quit IRC12:33
*** yassine has quit IRC12:33
*** yassine has joined #openstack-marconi12:33
*** amitgandhi has joined #openstack-marconi12:42
*** amitgandhi has quit IRC12:47
*** amitgandhi has joined #openstack-marconi12:52
*** amitgandhi has quit IRC12:56
*** tedross has joined #openstack-marconi12:57
*** mcpierce has left #openstack-marconi13:47
*** amitgandhi has joined #openstack-marconi14:06
*** jcru has joined #openstack-marconi14:06
*** amitgandhi has quit IRC14:07
*** amitgandhi has joined #openstack-marconi14:07
*** mpanetta has joined #openstack-marconi14:16
*** whenry has quit IRC14:18
*** alcabrera has joined #openstack-marconi14:19
alcabreraGood morning!14:26
*** whenry has joined #openstack-marconi14:26
mpanettaMorning :)14:27
*** oz_akan_ has joined #openstack-marconi14:27
mpanettaHow is your cat doing alcabrera?14:33
alcabreraShe's okay-ish. She's a little sniffly this morning, but was happy to eat.14:35
mpanettaThat is good to hear :)14:35
flaper87alcabrera: mpanetta goood morning!!!!!14:36
*** ykaplan has joined #openstack-marconi14:37
alcabreraflaper87: hey! :D14:37
* flaper87 likes how goood with 3 Os looks14:37
mpanettaflaper87: Hello!14:37
* flaper87 back as a full-day developer14:37
mpanettaWelcome back! :)14:37
*** malini_afk is now known as malini14:46
*** jamieh__ has quit IRC15:13
*** flwang has joined #openstack-marconi15:17
flaper87flwang: welcome! :)15:18
flaper87alcabrera: mpanetta come on guys, lets say Hi to flwang :D15:18
flwangflaper87: haha, yep15:18
alcabreraflwang: hey! :)15:18
flaper87flwang: this is a low-hanging-fruit you could start working on:
flwanghey guys, just joined a session in the summit, so I'm interested in it15:18
flwanghope I can contribute a little bit :)15:19
flaper87please, feel free to yell if something doesn't make sense15:19
* flaper87 pints out that flwang is also a Glance-core fellow15:19
maliniwelcome flwang15:19
alcabrerayeah, ask away. Glad to have you joining us!15:20
flwangflaper87: seems Zhihao Yuan is working on that bug?15:20
flaper87zyuan: are you? ^15:20
flaper87it's assinged to him but no patch has been submitted, yet!15:21
* flaper87 waits for a patch in 3, 2, 1....15:21
flaper87aaaawwww :(15:21
flaper87didn't work15:21
flwangflaper87: got it, I think I need a little bit time to dive into the code and then ask some stupid questions :)15:22
flaper87flwang: there are no stupid questions! You may also want to watch the talk we gave at the summit15:23
flaper87ah damn15:23
flaper87youtube <-> china15:23
flwangflaper87: you got it15:23
flaper87flwang: not sure how I can help there but I know you know what to do ;)15:23
mpanettaHi flwang!15:25
flwangflaper87: you know, cross the wall is the basic skill for Chinese developer
flwangmapanetta: hey there15:25
*** jergerber has joined #openstack-marconi15:26
flwangflaper87: let me ask the first stupid question, what's the mission of Marconi? provide a message service for cloud user, instead of Openstack itself, is it?15:27
zyuanflaper87: i'm not working on that, i'm not sure how much it helps15:27
mpanettaflwang: It is an extension to openstack, a new api.15:28
flwangmapanetta: got, so what's the extensions for? I mean who is the consumer?15:29
flwangI assume it's cloud tenant user, instead of openstack itself, right?15:30
alcabreraflwang: the consumer would typically be a cloud user.15:31
flwangalcabrera: got it, thx15:31
alcabreraI can see Marconi being used as a messaging bus for local use, as well, since it is pretty simple to get up and running. :)15:31
openstackgerritA change was merged to openstack/python-marconiclient: Add list of required fields to the API definition
alcabreraIt's one of the friendliest messaging systems I've seen so far, but then again, I'm both biased and not widely versed in messaging systems. :P15:31
flwangalcabrera: thanks, so seems like it's an implement of SQS/SNS in OpenStack15:33
alcabreraexactly. :)15:33
flwangalcabrera: cool15:34
flaper87wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwhhhhhhhhhhhaaaaaaaaaaaaattttttttttttt??? a patch merged in the client?15:34
alcabreraThe SQS part is mostly done (for a v1.0 release). The SNS portion is still in the design phases.15:34
* flaper87 cries15:34
flaper87kgriffs_afk: THANK YOUUU!!15:34
flaper87kgriffs_afk: now, next one :D15:34
flwangbtw, I realllllllllllllly like the pronunciation of Marconi15:35
alcabreraflaper87: congrats!15:36
alcabreraflwang: it *does* roll off the tongue nicely. :)15:36
flwangalcabrera: yep :)15:37
flwangvery cool15:37
mpanettaOf course, it is italian :P15:37
flaper87mpanetta: hahahah15:38
flwangflaper87: I want an Italian name :P15:38
flaper87flwang: I can give you one and start calling you like that! :D15:40
* flaper87 is trying to keep religion out of this conversation15:40
flwangflaper87: waiting and may use it as my nickname :)15:40
mpanettaBut they all end in vowels :P hehe15:41
* mpanetta is being silly15:41
flwangdon't tell me all of you guys are Italian :)15:41
mpanettaI'm not, only by name.15:42
mpanettaI've visited my relatives in Italy though, does that count? :P15:42
flwanghehe, hope I have chance to join the summit in Paris next year, then we can go to Italy, I'm a fans of Juventus club15:43
maliniI am so excited abt the next summit in Atlanta !15:44
malinihome sweet home !!15:44
malininot that I wud mind going to Paris ;)15:45
mpanettaWhen is that one malini?15:45
mpanettaI want to go to Paris!15:45
malininext April ? flaper87 might know better15:45
mpanettamalini: I mean the one here in atl15:45
flaper87April -> Atlanta, November -> Paris15:45
mpanettaOh 2 in one year?15:46
flaper87mpanetta: every 6 months15:46
mpanettaGood, that means I get to attend one ;)15:46
maliniwe wud be graduating in the ATL summit15:46
malinitht got me motivated to start tinkering with tempest15:46
*** ykaplan has quit IRC15:47
*** kgriffs_afk is now known as kgriffs15:47
* alcabrera still needs to learn him some Heat for great good15:47
mpanettaheat? tempest?15:47
mpanettaWhat are these?15:47
malinimpanetta: tempest is the openstack integration test suite15:48
flaper87btw, did I mentioned that Heat, Horizon and Ceilometer are looking forward to use Marconi ?15:48
flaper87kgriffs: 'sup buddy? How are you doing?15:49
flaper87good morning15:49
mpanettamalini: Oh?15:49
flaper87malini: how's the work on tempest integration going?15:49
mpanettaI will have to investigate15:49
mpanettaSo many new things to investigate15:49
* mpanetta is reminded of Cat on Red Dwarf and his investigations15:50
maliniflaper87: I have just started on it..nothing substantial so far. But I hope to have something solid by end of this month15:50
alcabreraahh, the great anvestor Cat that started it all...15:50
kgriffsspeaking of graduation15:50
flaper87malini: AWESOME!15:51
flaper87looking forward to your patches15:51
kgriffsJudging by the last TC list digest, I don't think they necessarily have us on the radar for Icehouse graduation15:51
flwangflaper87: i'm interested in that you mentioned Heat, Horizon and CM will use Marconi, any more details?15:52
kgriffs"OpenStack is currently 9 integrated projects in Havana, 10 once we get to Icehouse (+Trove)."15:52
mpanettaalcabrera: Red Dwarf fan?15:52
alcabrerampanetta: only slightly. It was fun to watch, but I remember very little. I like to think of myself as "Red Dwarf aware". :)15:52
flwangkgriffs: will Marconi graduate in Icehouse and to be in integrated state?15:52
mpanettaalcabrera: Ahh!  Cool :)15:53
kgriffsI would like it to be in Icehouse, but we need to prove to the TC that we are ready in plenty of time, like 8+ weeks before hand15:53
flwangkgriffs: got, thanks15:53
kgriffsflaper87: you were working on devstack-gate, right?15:54
flaper87kgriffs: TBH, I'd like it to be in Icehouse as well but, I'm also fine with us taking our time to make things right15:54
flaper87kgriffs: I am15:54
flaper87I think we've enough time to make it15:54
kgriffswe also need "useful tests in the integrated gate"15:54
malinikgriffs: are these all the stuff we need to graduate ?15:54
flaper87I just don't want us to 'freak out' if we don't15:54
malinikgriffs: "useful tests" as in tempest tests, rt?15:54
kgriffsflaper87: right. We will have a solid, stable release in the Icehouse timeframe whether or not we are officially integrated15:55
kgriffsmalini: good question, idk15:55
kgriffsI was going to ask if anyone knew what Sean was talking about15:55
maliniSean Dague?15:55
kgriffsAll integrated projects should be in the integrated gate, as this is15:55
kgriffsthe only way we provably know that they can all work together, at the15:55
kgriffssame level of requirements, in a consistent way.15:55
flaper87kgriffs: correct, that's how I see it15:55
kgriffsmalini: right15:56
kgriffssounds like Tempest to me15:56
flaper87kgriffs: that's mostly related to having support for devstack, tempest and making marconi tests run in the same environment with all other services running15:56
maliniI'll ping him to figure out what exactly is needed15:56
kgriffsmalini: thanks!15:56
flwangflaper87: i'm interested in that you mentioned Heat, Horizon and CM will use Marconi, any more details?15:57
kgriffsfwiw, here is the complete email:15:57
flaper87flwang: The three of them want to use the notification part of MArconi15:57
kgriffsmalini: question re. Tsung reports15:58
flaper87Heat to figure when service's bootstrap has been completed, AFAIU. Ceilometer to send alarms and notify others15:58
flwangso what's the relationship between marconi's notification and the oslo.messaging?15:58
kgriffsI was looking at this:
flaper87and same for Horizon15:58
*** jergerber has quit IRC15:58
kgriffsmalini: what is a "page"?15:58
kgriffsLooking at transactions, for example, I see ~3000 "pages" per sec15:58
malinikgriffs:a page is a group of request not separated by a thinktime15:59
kgriffswhat is a think time?16:00
maliniit's a delay you add between requests16:00
malinifor us the page response time does not really give any useful insight16:00
*** jamieh__ has joined #openstack-marconi16:00
kgriffsso, looking at this report, we can currently sustain only 1000 tps?16:00
kgriffsis that our max for one shard?16:01
malinikgriffs: requests is what we are interested in this report16:01
maliniwe can sustain ~4000 rps with 3 shards16:01
kgriffsoh, so ignore that transactions graph, and look at "Requests"?16:02
malinikgriffs: yes16:02
kgriffsaccording to this, we sit at around 300016:02
kgriffsthis graph was for a single shard iirc16:03
*** alcabrera is now known as alcabrera|afk16:03
*** jergerber has joined #openstack-marconi16:03
kgriffsmalini: once we have these tests running automatically (nightly or triggered for each patch) it would be cool to plot our performance over time16:04
kgriffsso we can see trends16:04
kgriffslike, "look, we are 1.5x faster than we were a month ago"16:04
malinikgriffs: tht wud be cool.we also need to include more API's than what we have here16:04
malini& we can also try a heavier load than the one here16:05
malinilast week was weird..we found out that one of the reasons we hit a plateau was because we didn't have a heavy enough load :D16:06
kgriffsIn other news, has anyone come across any leaked information about SQS architecture?16:08
kgriffsRight now it is just a black box as far as I know.16:09
mpanettamalini: Should I add tsung-8 to the xml files?16:10
malinimpanetta:'ll be good to add tht16:13
malinikgriffs: is tht geek gossip ;)16:14
kgriffsyes. I'm looking for gossip.16:21
kgriffsflaper87: I16:29
kgriffsam reviewing the client patches16:29
kgriffsbtw, can you also look at the sharding patches?16:30
*** adrian_otto has joined #openstack-marconi16:33
* flaper87 back16:34
flaper87kgriffs: yeah, I started doing that today, I got distracted by some internal stuff but I'll get back to that asap16:35
kgriffsflaper87: RabbitMQ can do something like 40K messages/sec, right? I haven't benchmarked it myself, but that is what I am seeing reported in blogs.16:36
kgriffsflaper87: kewl, thanks. BTW, I am planning on taking notes from the summit and converting to blueprints and stuff as appropriate over the next couple of weeks16:37
flaper87kgriffs: roughly yes!16:37
kgriffsalso cleaning up the wiki16:37
kgriffsflaper87: OK, I just wanted to make sure I wasn't crazy.16:37
*** amitgandhi has quit IRC16:37
flaper87kgriffs: +1, I can take core of the notificaion and AMQP sessions16:37
kgriffsflaper87: that would be great, thx16:38
flaper87kgriffs: from the first session, I think we agreed one putting more thoughts on the deserialization / content-type thing, and the qack endpoint16:38
flaper87The former kinda made sense but we didn't want it16:39
flaper87the later, instead, is something we definitely need16:39
flaper87alcabrera|afk: proposed it as a topic for the next meeting16:39
kgriffsby "qack" you are thinking basically "pop"?16:39
* flaper87 had a call with amit, alcabrera and zyuan on monday about 'What happened in the summit'16:39
flaper87kgriffs: yeah16:40
flaper87I just wanted to create some confussions16:40
kgriffsI was worthless monday16:40
kgriffsgot home Monday morning at 2am after 32 hours in transit16:40
kgriffs(due to delayed flight in HK which caused a missed connection in SFO)16:40
kgriffsso, sorry I wasn't around!16:41
flaper87no worries, hopefully I didn't forget anything. I used the ehterpads so, they seem to have everything we talked about16:41
kgriffscool beans16:42
kgriffsflaper87: I wonder how fast we can get Marconi given a redis backend and a TCP transport16:43
kgriffsI don't want to chase RabbitMQ necessarily16:44
kgriffsat some point it may make sense to start a broker provisioning service ala Trove under the Marconi program umbrella16:45
kgriffslike, spin me up an HA, single-tenant AMQP broker and give me nice monitoring and admin features around it16:46
flaper87kgriffs: agreed16:46
kgriffsit would obviously cost more in terms of compute resources, but I can see some people willing to pay for it16:46
flaper87so, re speed, I think we can improve it a lot by using a tcp transport and some other backend16:47
kgriffstbh, I think the main value proposition for the TCP transport in queues is providing push to notification consumers16:47
flaper87I still think we could improve mongodb's one, though16:47
kgriffsSo, maybe we target something like 10-20K tps16:48
*** alcabrera|afk is now known as alcabrera16:48
* alcabrera catches up16:48
kgriffsand for people who want higher throuput, they would use a RabbitMQ or MQ series or whatever directly16:48
kgriffsflaper87: definitely16:48
flaper87kgriffs: also, at some point, we could have something like 'direct-access' to the backend16:48
*** amitgandhi has joined #openstack-marconi16:48
kgriffsMongo once can be improved16:48
kgriffsesp. if we make FIFO and high durability optional16:49
flaper87that's something I've been thinking about but I'm still unsure as to how we could achieve that16:49
kgriffsif we didn't have to generate the marker we would be really fast16:49
kgriffsplus we have some indexing work to look at16:49
flaper87kgriffs: indeed!16:49
kgriffsflaper87: idk about direct access to the backend. I think that would be a separate service like I said before, since it would imply single-tenancy anyway, wouldn't it?16:50
flaper87kgriffs: well, you know, since we'll have a discoverable API and we'll also ship partially implemented APIs, making FIFO optional in the mongodb driver sounds reasonable16:50
kgriffsactually, I could see that being a queue "flavor"16:50
kgriffsyou could say "i don't need strict ordering"16:50
kgriffsand then we put that on a shard where the driver is configured to just use timestamp-based markers16:51
flaper87kgriffs: exactly, I think we should start writing some notes wrt queue's flavors16:51
kgriffsnot just strict ordering, but you might also get duplicate messages - client would have to watch out for that and skip things it has seen already16:51
kgriffsflaper87: +116:51
alcabrera+1 for queue flavors and optional FIFO16:51
kgriffsi think flavors is going to turn out to be VERY useful16:51
alcabreraI'm a big fan of the queue flavors idea16:51
alcabreraDeclarative queueing16:52
kgriffscan someone register bp and start a wiki page?16:52
flaper87kgriffs: o/16:52
alcabreraGoes *really* well with heterogeneous backends + sharding16:52
kgriffsi can't remember if we have one already16:52
flaper87not for flavors16:52
flaper87not that I know16:52
alcabreraI don't think we have anything for queue flavors yet16:52
*** yassine has quit IRC16:52
kgriffslet's check and create one if not there already, and seed a wiki page16:52
kgriffsspeaking of bps16:53
kgriffswe have a backlog to triage, also bugs16:53
alcabreraoh yes...16:53
kgriffsI was thinking we could spend 30 mins each day at 1600 UTC going over those16:53
alcabreraA lot of BPs, a lot of bugs.16:53
alcabrera1600 UTC works for me.16:54
kgriffsflaper87: ^^^16:55
flaper87kgriffs: +116:55
flaper87sounds like a plan16:55
alcabrerakgriffs: I'm giving a tech talk today we might be able to link to the falconframework page - "The Definitive Introduction to Falcon"16:55
alcabreraflaper87: +116:55
kgriffsalcabrera: oh, nice16:55
kgriffsI need to do a Falcon sprint in the near future16:56
kgriffsalcabrera: let me send you some slides in case you want to plagerize16:56
alcabrerakgriffs: thanks!16:56
alcabreraI'm tempted to do a docstring sprint for falcon.16:56
alcabreraGet us on RTD and such.16:56
alcabrera(esp. since RTD looks *really* snazzy now)16:57
*** jamieh__ has quit IRC16:57
kgriffsalcabrera: yes!!!!!16:59
kgriffsalong with that, I am tempted to create a google comment style ---> rtd/sphinx tool16:59
kgriffsalcabrera: Keep an eye on the nav arrows. Sometimes you have to go "down" to get to the next slide.17:00
alcabreraI missed that.17:00
kgriffsflaper87: re WSME vs. JSON schema17:00
kgriffspersonally, I think WSME's fundamental philosophy of ORM-based, static definitions if flawed17:01
kgriffsGiven that nova and glance will be using JSON schema17:01
kgriffsand it is looking like we will as well17:01
kgriffsI think we should do a survey of Python serialization and validation frameworks17:02
flaper87jsonschema all the way down!17:02
kgriffsserialization is an interesting one17:02
kgriffsI am suspicious that an ORM can generate truly XML-y XML and JSON-y JSON17:02
flaper87damn, 2 secs. brb!17:02
flaper87I don't think we need WSME17:03
flaper87jsonschema will work better for us and the cross-transport support17:03
flaper87WSME is *too* HTTP / Rest oriented, AFAIU17:03
kgriffsflaper87: actually, it is ironic17:03
kgriffsI don't think it is actually RESTful17:03
kgriffsREST media types tend to be more flexible, and WSME seems to be more static17:04
kgriffsI suspect that is due to it's SOAP heritage17:04
*** jergerber has quit IRC17:14
kgriffsanybody play around with this?17:17
*** jamieh__ has joined #openstack-marconi17:18
alcabreraI've never heard of it. :/17:20
kgriffsI just saw it proposed as an alternative to httplib for swift client17:20
alcabreraand it doesn't have a pypi registry.17:20
*** jamieh__ has quit IRC17:22
alcabreraah, I see.17:22
alcabreraI wonder if they've considered requests?17:22
ekarlsowhat about the oslo apiclient stuff ?17:22
kgriffsI recall something someone said at the summit in passing about requests17:22
kgriffslike, it didn't work in once case - wish I could remember what that was17:23
kgriffsI feel like we should just contribute patches to request17:23
kgriffsif it isn't quite where it needs to be for something17:24
kgriffs(i mean)17:24
ekarlsoI just used the apiclient stuff for python-libraclient17:24
ekarlsoworks awesomely which is based on request17:24
*** jergerber has joined #openstack-marconi17:25
*** jamieh__ has joined #openstack-marconi17:26
*** cindy has joined #openstack-marconi17:32
*** reed has joined #openstack-marconi17:34
*** cindy is now known as cpallares17:37
*** adrian_otto has left #openstack-marconi18:09
*** gleicon___ has quit IRC18:40
*** alcabrera is now known as alcabrera|afk18:45
*** alcabrera|afk is now known as alcabrera19:03
*** jamieh__ has quit IRC19:25
*** jamieh__ has joined #openstack-marconi19:28
amitgandhikgriffs: you around?19:37
amitgandhi<--- meeting19:38
kgriffsyes, sorry, didn't realize i wasn't connected to the other IRC server19:38
*** alcabrera is now known as alcabrera|afk20:02
*** mpanetta_ has joined #openstack-marconi20:05
*** alcabrera has joined #openstack-marconi20:06
*** mpanetta has quit IRC20:08
*** alcabrera|afk has quit IRC20:09
*** kgriffs is now known as kgriffs_afk20:12
flaper87kgriffs_afk: when you're back, could you ? :D20:17
*** kgriffs_afk is now known as kgriffs20:22
*** malini is now known as malini_afk20:22
*** mpanetta_ has quit IRC20:45
*** mpanetta has joined #openstack-marconi20:46
*** acabrera has joined #openstack-marconi20:48
*** alcabrera has quit IRC20:49
*** acabrera is now known as alcabrera20:57
flaper87alcabrera: just wanted to give you another 'Way to go' for the work you've been doing on the partition bp21:05
alcabreraflaper87: thanks! :D21:26
* flaper87 is writing a blog post after a long time of pure silence21:37
cpallareshi flaper87!21:38
flaper87cpallares: THERE YOU ARE!!!!21:38
cpallaresflaper87: how was HK? :D21:38
flaper87and I meant to yell21:39
flaper87cpallares: it was GREAT!!!!21:39
flaper87cpallares: so, how are you doing?21:39
cpallaresflaper87: I haven't seen you around! I'm doing good :)21:39
flaper87cpallares: yeah, Last week was pretty damn busy.21:40
flaper87cpallares: any new contribution to Marconi?21:40
flaper87you better say yes21:41
cpallaresflaper87: Haha I'm still waiting on my other one to get reviewed!21:41
* alcabrera goes to check review queue - it's been awhile.21:42
alcabreracpallares: hey! :D21:42
cpallaresalcabrera: hey :)21:42
flaper87OpenStack Queuing and Notification Service || Smile :D || Meetings every Monday 16:00 UTC || Wiki: ttps:// || Paste: || Send messages and make some noise :D21:44
*** flaper87 changes topic to "OpenStack Queuing and Notification Service || Smile :D || Meetings every Monday 16:00 UTC || Wiki: ttps:// || Paste: || Send messages and make some noise :D"21:44
alcabrera"Send messages and make some noise :D"21:44
alcabrerawill do21:44
* cpallares makes noise21:45
flaper87gotta keep this channel alive! There are 41 nicks in here!21:45
flaper8736 of those are bots21:45
flaper87nah, kidding!21:45
* alcabrera suspects amitgandhi is a bot21:45
flaper87all real people, silent people21:45
flaper87amitgandhi: you a bot?21:45
* flaper87 stares21:46
mpanettayes he is :P21:46
amitgandhiawaiting command...21:46
kgriffsMy friends call me R. Daneel.21:46
amitgandhi!amitgandhi help21:46
openstackamitgandhi: Error: "amitgandhi" is not a valid command.21:47
*** kgriffs is now known as rdaneel21:47
flaper87amitgandhi: a bot would've insulted me21:47
rdaneelyou are all part of my master plan to save humanity21:47
mpanettaamitgandhi: just got owned by openstack21:47
mpanettaI need to read that book...21:47
amitgandhiflaper87: bots need some gelato21:47
amitgandhimpanetta: lol21:48
mpanettaMmm gelato21:48
flaper87I sent some. Didn't rdaneel give it to you?21:48
mpanettaProbably not, he eats everything21:50
alcabrerardaneel, mpanetta: makes sense, given a loose interpretation of the 0th law of robotics. You've gotta prevent them from eating themselves to death. ;)21:50
rdaneelanybody know if there's an executor for eventlet?21:52
*** rdaneel is now known as kgriffs21:53
alcabrerakgriffs: I've seen no such compatibility layer, though it makes me think of something like this:
flaper87given the fact that eventlet does not support python 3.3 and hell it won't support python 3.2, I doubt there's support of it21:55
flaper87however, it may make sense for python 2.721:55
kgriffsi heard rumors that eventlet work for py3 was planned21:56
kgriffsdid I hear wrong?21:56
flaper87kgriffs: owwww reallly???21:56
flaper87I've heard all kind of things but that21:56
flaper87kgriffs: maybe gevent ?21:56
alcabrerakgriffs, flaper87: (eventlet on py3k)21:57
alcabrerathat's the latest21:57
kgriffslooks like a lot of talk but no action21:59
kgriffsflaper87: I was just thinking, the client lib could return a future instead of taking a callback. or something21:59
kgriffswe could then implement an executor for eventlet or gevent22:00
kgriffsand then require py 3.4 with tulip or whatever22:00
kgriffsfor running on Py322:00
kgriffsbasically, using the Executor abstraction as our portability layer22:00
flaper87kgriffs: yup, that's kinda the idea, I think we'll still need the callback, though.22:03
flaper87I thought about always returning a Future but I don't want people not interested in async results to get a `future` back22:04
flaper87I thought about having async=True|False and return a Future there22:04
flaper87but, I don't want to break the functions consistency, I'd like them to always return the result or None. So, I thought about having a callback and using an Actor like execution model22:06
flaper87I've to admit, that future library looks sexy22:08
flaper87and kinda makes sense22:08
flaper87but I was looking more into Pykka22:10
flaper87anywho, I think we should discuss it further, maybe in one of our meetings.22:11
alcabreraflaper87: I remembered something about actors, and then I found this post -
alcabreraJust a thought, and I agree - we need more discussion on the client concurrency primitives. :)22:27
alcabreraConsistency + 122:27
flaper87alcabrera: wow, thanks for sharing. I'll read it tomorrow, I don't think my brain can take that post right now22:28
flaper87alcabrera: I'm not an Actors fan, TBH22:28
flaper87I read the original paper - did you know the whole model cam out in 1985~ ? - and some of the papers that improved the original one22:29
alcabreraMy brain can't take it either, atm. :D22:29
flaper87and still, there are some things I don't like much22:29
alcabreraflaper87: it feels like all the concurrency models have been around at least 10+ years22:29
flaper87however, it seems to make sense for something like marconilib that needs the executor part to be forward-backward compatible22:30
flaper87and it doesn't have 'strong' concurrent requirements22:30
alcabrerayeah, it's pretty much just 'send' and 'retrieve result'22:31
alcabreraLike futures.22:31
flaper87alcabrera: yeah, it can return futures22:32
*** oz_akan_ has quit IRC22:33
alcabreraMore thoughts next time! :D22:33
alcabreraI'm out for the night.22:33
*** kgriffs is now known as kgriffs_afk22:33
alcabreraTake care. :)22:33
*** alcabrera has quit IRC22:33
*** oz_akan__ has joined #openstack-marconi22:34
*** ryanpetrello has joined #openstack-marconi22:37
*** oz_akan__ has quit IRC22:39
*** flaper87 is now known as flaper87|afk22:47
*** kgriffs_afk is now known as kgriffs22:49
*** whenry has quit IRC22:50
*** ryanpetrello has left #openstack-marconi22:52
*** jamieh__ has quit IRC23:00
*** mpanetta has quit IRC23:08
*** tedross has quit IRC23:08
*** jergerber has quit IRC23:18
*** adrian_otto has joined #openstack-marconi23:29
*** adrian_otto has quit IRC23:33
*** amitgandhi has quit IRC23:34
*** amitgandhi has joined #openstack-marconi23:35
*** amitgandhi has quit IRC23:39
*** kgriffs is now known as kgriffs_afk23:50

Generated by 2.14.0 by Marius Gedminas - find it at!