Tuesday, 2014-04-29

*** kgriffs is now known as kgriffs|afk00:04
*** pquerna has quit IRC00:11
*** pquerna has joined #openstack-marconi00:13
*** pquerna has joined #openstack-marconi00:13
*** tjanczuk has joined #openstack-marconi00:34
*** tjanczuk has quit IRC00:38
*** shakamunyi has quit IRC02:11
*** ayoung has quit IRC02:17
*** balajiiyer has joined #openstack-marconi02:44
*** rossk has quit IRC02:45
*** balajiiyer has left #openstack-marconi02:46
*** balajiiyer has joined #openstack-marconi03:06
*** haoma____ has quit IRC03:07
*** haomaiwang has joined #openstack-marconi03:09
*** balajiiyer has quit IRC03:11
*** chandan_kumar has joined #openstack-marconi04:05
*** haomaiwang has quit IRC04:08
*** haomaiwang has joined #openstack-marconi04:10
*** haomai___ has joined #openstack-marconi04:22
*** haomaiwang has quit IRC04:25
*** tjanczuk has joined #openstack-marconi04:36
*** tjanczuk has quit IRC04:41
*** tjanczuk has joined #openstack-marconi05:17
*** chandan_kumar has quit IRC06:01
*** chandan_kumar has joined #openstack-marconi06:16
*** chandan_kumar has quit IRC06:22
*** haomai___ has quit IRC06:36
*** haomaiwang has joined #openstack-marconi06:36
*** flaper87|afk is now known as flaper8706:39
*** chandan_kumar has joined #openstack-marconi06:40
*** haomaiw__ has joined #openstack-marconi07:02
*** haomaiwang has quit IRC07:05
*** reed has joined #openstack-marconi07:28
*** ykaplan has joined #openstack-marconi08:22
*** jamie_h has joined #openstack-marconi08:23
*** chandan_kumar has quit IRC09:14
*** tjanczuk has quit IRC09:14
*** chandan_kumar has joined #openstack-marconi09:15
*** haomaiw__ has quit IRC09:34
*** haomaiwa_ has joined #openstack-marconi09:35
*** haomaiwa_ has quit IRC09:55
*** haomaiwang has joined #openstack-marconi09:56
*** fifieldt_ has joined #openstack-marconi11:16
*** fifieldt has quit IRC11:17
*** tjanczuk has joined #openstack-marconi11:42
*** haomaiwang has quit IRC11:42
*** haomaiwang has joined #openstack-marconi11:43
*** tjanczuk has quit IRC11:47
*** jchai has joined #openstack-marconi12:06
*** ykaplan has quit IRC12:35
*** mpanetta has joined #openstack-marconi12:44
*** sriram has joined #openstack-marconi12:46
*** sriram has quit IRC12:46
*** sriram has joined #openstack-marconi12:46
*** ykaplan has joined #openstack-marconi12:51
*** jamie_h has quit IRC13:01
*** jamie_h has joined #openstack-marconi13:01
*** fifieldt_ is now known as fifieldt13:03
*** flaper87 is now known as flaper87|afk13:20
*** mwagner_lap has quit IRC13:22
*** flaper87|afk is now known as flaper8713:25
*** jmckind has joined #openstack-marconi13:26
*** Obulpathi has joined #openstack-marconi13:27
*** Obulpathi has quit IRC13:28
*** Obulpathi has joined #openstack-marconi13:28
*** ykaplan has quit IRC13:31
*** prashanthr_ has joined #openstack-marconi13:40
*** ykaplan has joined #openstack-marconi13:46
*** amitgandhi has joined #openstack-marconi13:49
*** flaper87 is now known as flaper87|afk13:53
*** malini_afk is now known as malini13:54
*** balajiiyer has joined #openstack-marconi13:55
*** flaper87|afk is now known as flaper8713:56
*** cpallares has joined #openstack-marconi13:57
*** whenry has joined #openstack-marconi14:04
*** alcabrera|afk is now known as alcabrera14:09
*** ykaplan has quit IRC14:10
alcabreraGood morning, all. :)14:10
cpallaresalcabrera: good morning!14:24
*** mwagner_lap has joined #openstack-marconi14:31
alcabreracpallares: hey! How are you? :)14:34
*** kgriffs|afk is now known as kgriffs14:35
flaper87goooooooooooooooooooooood morning14:35
malinihelllooooooo !!!14:35
sriramhhhheeeelllllooooo :D14:36
alcabrerahey, all. :)14:37
*** alcabrera is now known as alcabrera|afk14:41
*** tjanczuk has joined #openstack-marconi14:46
*** ykaplan has joined #openstack-marconi14:49
*** alcabrera|afk is now known as alcabrera14:50
*** tjanczuk has quit IRC14:51
kgriffsok everybody14:53
kgriffsfirst one to decipher this gets a virtual prize14:54
kgriffsWbva gur GRNZ ZRRGVAT va #bcrafgnpx-zrrgvat-nyg14:54
*** jchai is now known as jchai_afk14:54
Alex_GaynorJoin the TEAM MEETING in #openstack-meeting-alt :-)14:55
alcabreraJoin *** **** MEETING in #openstack-meeting-alt14:55
alcabreranicely done, Alex_Gaynor. :)14:56
* kgriffs gives Alex_Gaynor https://gist.github.com/anonymous/7dbe487df24082c9c76d14:56
alcabrerasweet chocolate14:56
* alcabrera heads over to #openstack-meeting-alt14:57
*** whenry has quit IRC14:58
*** ykaplan has quit IRC14:58
*** flwang has joined #openstack-marconi15:02
*** jchai_afk is now known as jchai15:02
*** jamie_h has quit IRC15:17
*** jamie_h has joined #openstack-marconi15:18
*** shakamunyi has joined #openstack-marconi15:21
*** haomaiw__ has joined #openstack-marconi15:26
openstackgerritIhar Hrachyshka proposed a change to openstack/marconi: Synced jsonutils from oslo-incubator  https://review.openstack.org/9107015:27
*** haomaiwang has quit IRC15:28
*** ykaplan has joined #openstack-marconi15:31
*** megan_w|afk is now known as megan_w15:34
*** sriram has quit IRC15:46
*** sriram has joined #openstack-marconi15:48
*** balajiiyer has quit IRC15:48
*** amitgandhi has quit IRC15:48
*** megan_w is now known as megan_w|afk15:49
*** balajiiyer has joined #openstack-marconi15:50
openstackgerritBalaji Iyer proposed a change to openstack/marconi: Removes host header from functional tests  https://review.openstack.org/9108315:51
*** Obulpathi has quit IRC15:57
*** ykaplan has quit IRC16:00
flaper87sriram: have you done any progress on the lazy queue stuff?16:00
sriramflaper87: yeah I had discussed it with malini and alcabrera16:01
sriramand got lazy queue creation working.16:02
flaper87sriram: cool, I was curious to know how is that comming along16:02
sriramI was working on the status codes that need to be returned.16:02
flaper87sriram: looking forward to your patch16:02
sriramflaper87: Its coming along well :)16:02
flaper87kgriffs: malini flwang alcabrera * : What do you guys think about the msgpack proposal?16:03
sriramits insanely faster.16:04
flwangflaper87: link?16:04
flaper87I'd like to move that forward16:04
flaper87and do the switch asap16:05
alcabrerathat'll be fun16:05
flaper87flwang: https://blueprints.launchpad.net/marconi/+spec/api-v1.1-msgpack16:05
flaper87kgriffs: have you done any work on that front ?16:05
*** Obulpathi has joined #openstack-marconi16:05
kgriffsso, I sent out some feelers to the SDK devs at rackspace16:05
kgriffsthey do a lot of meetups and stuff, so they have a good feel for the developer community at large16:07
kgriffsand they were pretty enthusiastic about the idea.16:07
openstackgerritBalaji Iyer proposed a change to openstack/marconi: Removes host header from functional tests  https://review.openstack.org/8939316:07
kgriffsin fact, one person said it would be cool to see other OpenStack APIs support application/x-msgpack as well16:07
kgriffsso maybe we should grab some folks and discuss this at the summit16:08
kgriffsas far as actual implementation work, no I haven't started that16:08
flaper87so, from a marocni's POV, I see some other benefits. For instance, it'd allow people to send binary data16:08
kgriffsI did create a v1.1 blueprint for it though16:08
kgriffsflaper87: +116:09
flaper87I put some thoughts on this earlier today and I think it'll be a good thing for marconi16:09
flaper87I also thought about the "supporting both" thing we discussed and I'm not convinced about that16:09
kgriffsthe SDK guys seem to prefer supporting both for now16:10
kgriffsI think dropping JSON may be an option for v2.016:10
flaper87but we should probably start adding support for msgpack and then figuring out the other bit16:10
flaper87kgriffs: exactly16:10
kgriffsI think we need to contribute to the broader ecosystem to make dropping JSON practical, e.g., contributing to requests/httpie/postman16:11
flaper87kgriffs: agreed16:11
kgriffsFor those of you following along at home, the blueprint is listed here with the other v1.1 ones: https://launchpad.net/marconi/+milestone/juno-116:11
kgriffsspecifically: https://blueprints.launchpad.net/marconi/+spec/api-v1.1-msgpack16:12
kgriffsflaper87: you are teh approver. :D16:12
kgriffsflaper87: have you floated the idea with your peeps at RH?16:12
kgriffsi mean, the general idea of supporting msgpack in openstack APIs16:12
flaper87yeah, the feeling is pretty much the same we had. I haven't asked whether chagning other modules throughout openstack would make sense or not16:13
flaper87but I will16:13
flaper87 /we had / we have16:13
kgriffscool, I just want to get some thoughts from a few different angles16:13
flaper87one more question: https://blueprints.launchpad.net/marconi/+spec/api-v1.1-plain-message-option16:14
flaper87I created this one today16:14
flaper87if you have some spare time today, I'd love to hear what you guys think about this proposal16:14
flaper87I see it implemented in both: get and pop16:15
kgriffshmm, so you would get a list of bodies?16:15
*** ayoung has joined #openstack-marconi16:15
kgriffscan you mock up what that would look like and link from the bp?16:15
flaper87but yeah, you'd get *just* the body16:16
flaper87I'm not sure it makes sense in the listing endpoint16:16
flaper87but it does for the message GET and POP actions16:16
kgriffsremember that pop can return >1 message16:16
kgriffsi think it will help me to see some different options mocked up in a gist16:18
flaper87sure, I'll add some examples16:18
flaper87I'd like to avoid doing this in the client library16:18
*** alcabrera is now known as alcabrera|afk16:19
kgriffsflaper87: btw, once we have a rough consensus around msgpack, could you mark that bp as approved16:19
*** chandan_kumar has quit IRC16:19
flaper87kgriffs: I already did :s16:20
kgriffsi think everyone is feeling pretty good about it16:20
kgriffsI will unassign myself in case someone would like to pick it up; I may not get to it for a little while16:21
flaper87in that case, I think I'll get it16:21
flaper87I would like to move that one forward, so we can get support for that in the client library16:21
flaper87and experiment with msgpack16:22
flaper87I'd like the benchmarks to be done with both, msgpack and json16:22
kgriffsflaper87: in other news, I think I'd like to work on this during juno: https://blueprints.launchpad.net/marconi/+spec/perf-testing16:23
flaper87kgriffs: +1, I put some thoughts today on what would make the results of those tests be good :P16:24
flaper87in a real world scenario16:24
flaper87kgriffs: what needs to be done there?16:24
flaper87there = perf-testing16:24
flaper87what's missing?16:24
flaper87Do we have a rough plan?16:24
kgriffsI put down some requirements on a wiki page16:25
kgriffsthe bulk of the work will be writing a tool that can benchmark end-to-end scenarios16:25
kgriffssince it has to trace message lifetime across >1 client16:26
kgriffsI was thinking to use Graphite to store and present the data16:26
*** tjanczuk has joined #openstack-marconi16:27
kgriffsand sensu to get box metrics16:29
kgriffsflaper87, malini: the big question is whether to try and do this with rally16:30
maliniwe could do a POC with rally, to check if it'll meet our requirements16:30
*** tjanczuk has quit IRC16:33
kgriffsok, I started a "tools" section at the bottom of the page: https://wiki.openstack.org/wiki/Marconi/bp/havana/perf-testing#Marconi:_Automated_Performance_Testing16:34
kgriffsflaper87: Do you remember why we removed simplejson?16:40
* kgriffs goes to log at git log16:40
flaper87kgriffs: there wasn't a huge performance gain16:40
flaper87plus it was an extra dependency16:41
flaper87and json ~= simplejson16:41
flaper87that's what I remember16:41
kgriffscommit 69559de3421377e9a7c40b9604c5b944a1f3801316:41
kgriffsYeah, seems like simplejson had subtle difference in unicode handling or something16:42
kgriffsflaper87: do you think we should add mspack retroactively to v1.0 api?16:45
flaper87kgriffs: I was planning to do so :/16:45
kgriffsI don't think it would hurt16:45
flaper87it should be pretty transparent from a user's point of view16:45
flaper87plus the client would do something like if !msgpack use json16:46
kgriffsbtw: https://bugs.launchpad.net/marconi/+bug/131412916:46
kgriffsI'm triaging this for marconi16:46
*** amitgandhi has joined #openstack-marconi16:46
flaper87kgriffs: there's a patch for that already16:47
flaper87I saw it poping up in gerrit16:48
flaper87but now I can't access gerrit anymore16:48
maliniI have a hard time getting used to the new look16:48
kgriffsyeah, it is different16:49
kgriffsso yes, jsonutils now uses it16:49
flaper87I haven't even seen it16:50
flaper87it just won't load16:50
flaper87I think I've some dns issues16:50
kgriffsjsonutils.to_primitive looks like an attempt to replace pickle16:50
flaper87review.openstack.org.   182     IN      A
flaper87I hate that function16:51
kgriffsrelevant: http://pyvideo.org/video/2566/pickles-are-for-delis-not-software16:51
flaper87and what I hate even more is that to_primitive makes sense but from_primitive doesn't16:51
flaper87I'd like to be able to sue gerrit16:52
kgriffsdo we want to switch to using jsontuils.loads/dumps/etc. ?16:52
kgriffsi don't see any value-add for marconi unless we want to sort of get simplejson, except simplejson isn't a drop-in replacement (has subtle differences)16:53
flaper87I don't see it either16:53
kgriffsbetter to just say "use msgpack if you want fast"16:53
flaper87kgriffs: could you dig review.openstack.org ?16:54
kgriffs(and we don't need pickle-like behavior)16:54
kgriffssimplejson must already be required16:56
kgriffsby what?16:56
flaper87kgriffs: could you $ dig review.openstack.org ? pls :D16:56
kgriffsoh, I can see it16:57
kgriffsdo you want the trace or something?16:57
flaper87just the DNS record16:57
flaper87same result here16:57
flaper87but I can't ping that IP16:57
kgriffsthat's because you are obviously an evil terrorist and we are blocking your internets16:58
kgriffsflaper87: have you done a traceroute?16:58
*** prashanthr_ has quit IRC16:58
kgriffsand have you complained loudly in #openstack-infra ?16:59
flaper87doing it now, I bet my ISP is blocking me16:59
kgriffsyou need to get yourself some Tor16:59
*** balajiiyer has quit IRC17:02
sriramflaper87: kgriffs: re lazy queue, querying messages of a non-existent queue-> should it gives us a 204 or 200 OK with empty list of messages?17:04
*** flaper87 is now known as flaper87|afk17:05
*** amitgandhi has quit IRC17:05
sriramalso if its the latter, the wiki defines a marker, should that be removed?17:05
*** amitgandhi has joined #openstack-marconi17:05
kgriffssriram: so, first off, we need to clone the "error codes" wiki page from v1.0 if that isn't already done17:05
kgriffsand get that up to date17:05
kgriffswrt the error code, that is a good question17:06
kgriffsif we return 204 it means the client may want to know a queue doesn't exist17:06
kgriffsbut if we auto-create queues, a client probably doesn't care whether the queue exists yet17:06
kgriffsalso, this sets people up for thinking more about topics rather than "physical" queues17:07
kgriffsso, I vote for 20017:07
kgriffsand if they give a marker, it will just be ignored if the queue doesn't exist17:07
*** flaper87|afk is now known as flaper8717:08
kgriffsin fact, you may not need to even check for the queue existing as a separate operation in the storage driver, if you the query is clever enough to just return an empty set for nonexistent queue17:08
openstackgerritA change was merged to openstack/marconi: Synced jsonutils from oslo-incubator  https://review.openstack.org/9107017:09
kgriffssriram: make sense?17:09
sriramyes makes sense, I was thinking of doing the queue check at the api level itself.17:09
sriraminfact that's what I have done currently17:10
kgriffsmalini, alcabrera|afk, flaper87, sriram: need another round of reviews on this: https://review.openstack.org/#/c/76093/17:10
kgriffssriram: rock on17:10
* kgriffs likes telling people to do things they are already doing17:10
sriramthanks kgriffs, I shall get back with more such questions :)17:10
* kgriffs grows pointy hair17:10
*** jchai is now known as jchai_afk17:11
* kgriffs is becoming more of a manager-type every day17:11
*** mkoderer has quit IRC17:12
*** amitgandhi has quit IRC17:18
*** amitgandhi has joined #openstack-marconi17:18
*** chandan_kumar has joined #openstack-marconi17:20
*** anamalagon has quit IRC17:23
*** amitgandhi has quit IRC17:24
*** amitgandhi has joined #openstack-marconi17:25
*** openstackgerrit has quit IRC17:32
*** openstackgerrit has joined #openstack-marconi17:33
*** rossk has joined #openstack-marconi17:34
*** balajiiyer has joined #openstack-marconi17:35
*** mpanetta_ has joined #openstack-marconi17:35
*** Obulpath_ has joined #openstack-marconi17:35
*** sriram1 has joined #openstack-marconi17:36
*** sriram has quit IRC17:38
*** Obulpathi has quit IRC17:39
*** mpanetta has quit IRC17:39
*** balajiiyer1 has joined #openstack-marconi17:44
*** balajiiyer has quit IRC17:44
*** jchai_afk is now known as jchai17:46
*** chandan_kumar has quit IRC17:49
*** balajiiyer1 has quit IRC17:53
*** balajiiyer has joined #openstack-marconi17:53
*** mpanetta_ has quit IRC18:06
*** mpanetta has joined #openstack-marconi18:07
malinikgriffs: do you think we should make  'message_ids=all' to be the default value to use when we want messages to be deleted ?18:08
maliniI have this thing at the back of my mind, that somebody is going to accidentally miss the params & end up losing everything in the queue18:08
malini& we don't have the option of asking 'Do you really want to do this?'18:09
kgriffshmm, yeah, that has been brought up before and TBH I forgot to work that out :p18:09
amitgandhimalini: +1.  ALL should not be the default18:09
*** sriram1 has quit IRC18:09
*** sriram has joined #openstack-marconi18:10
kgriffsthe way it is currently implemented is neat and clean, but not exactly safe.18:11
maliniamitgandhi, kgriffs: I mean have the user explicitly state 'DELETE /v1.1/queues/fizbit/messages?ids=18:11
kgriffsyep, got it18:11
* kgriffs is thinking18:11
kgriffsI think since several people have been nervous about this, it is reasonable to error on the side of safety18:13
kgriffsI'm just trying to decide if ids=all is the best way to go18:15
kgriffsi think so. I don't want to introduce yet another XOR param. :p18:15
mpanettaAdd enough XORs and you may just make yourself an adder :P18:16
mpanettaOr at least a half adder...18:16
*** mpanetta has quit IRC18:17
*** mpanetta has joined #openstack-marconi18:17
*** mpanetta_ has joined #openstack-marconi18:18
*** balajiiyer has quit IRC18:19
maliniI dont like the jenkins vote from old patches getting carried over :(18:20
*** balajiiyer has joined #openstack-marconi18:20
malinimakes it look like flavio still doesnt like this patch https://review.openstack.org/#/c/89393/, but still got merged18:20
kgriffsyeah, that is sort of lame18:20
kgriffsI guess we will need to be more careful18:21
maliniwonder if a +2 from an old patch, will get carried over to a new updated patch18:21
malinitht would be bad18:21
*** mpanetta has quit IRC18:22
*** whenry has joined #openstack-marconi18:28
*** whenry has quit IRC18:38
*** mpanetta_ is now known as mpanetta18:41
*** alcabrera|afk is now known as alcabrera18:51
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: [WIP]-Implement POP in v1.1 API  https://review.openstack.org/9020218:57
malinigrr..there is no longer a work in progres button18:57
*** jamie_h has quit IRC18:59
sriram"when listing messages in a queue or claiming messages, allow the queue to not exist (conceptually, queues "always exist" and if we don't have a record for the queue, we just treat it as being "empty")"19:15
sriramkgriffs: ^19:15
sriramI understand listing messages19:15
srirambut how does claiming fit in?19:16
alcabreraseems like that would be an empty claim, as well19:16
sriramour current behavior is a 20419:18
sriramI think we should maintain it.19:18
alcabreraI agree.19:18
sriramthanks alcabrera :) just wanted to make sure.19:18
*** jmckind has quit IRC19:20
*** Obulpath_ has quit IRC19:26
*** Obulpathi has joined #openstack-marconi19:26
* flaper87 can't use gerrit19:35
flaper87damn, what the heck is going on here19:35
mpanettaIt must be solar flares19:35
flaper87mpanetta: :(19:36
flaper87no gerrit, no +2's19:36
maliniflaper87: you cant access it at all?19:36
flaper87nope :/ something is blocking gerrits IP for me19:37
flaper87not sure if it's my ISP or my router19:37
flaper87but the router seems fine19:37
flaper87it happened to me once19:37
flaper87I... dunno what I did to make it work19:37
mpanettaReverse the polarity of the neutron flow.  That always worked for the 3rd Doctor.19:38
* mpanetta watches as peoples minds go EHH?! over that one :P19:38
flaper87holy crap, you lost me there19:38
*** ayoung has quit IRC19:39
mpanettaWell, neutrons have no polarity to reverse... heh19:39
maliniits all alcabrera's fault..he started talking physics19:40
mpanettaflaper87: https://www.youtube.com/watch?v=yLD03bU5R5M19:42
*** malini is now known as malini_afk19:42
alcabreraoh, I see I'm breaking things19:47
mpanettaYou broke something?19:51
mpanettaalcabrera: Reverse the polarity, that will fix it :PO19:54
openstackgerritA change was merged to openstack/marconi: Removes host header from functional tests  https://review.openstack.org/8939319:58
*** fifieldt has quit IRC20:08
*** fifieldt has joined #openstack-marconi20:08
*** Obulpathi has quit IRC20:10
*** amitgandhi has quit IRC20:12
*** Obulpathi has joined #openstack-marconi20:22
*** sriram has quit IRC20:28
*** alcabrera is now known as alcabrera|afk20:36
*** ayoung has joined #openstack-marconi20:37
*** whenry has joined #openstack-marconi20:40
*** whenry has quit IRC21:09
*** jchai has quit IRC21:09
*** mpanetta has quit IRC21:31
*** Obulpathi has quit IRC21:32
*** mwagner_lap has quit IRC21:39
*** sriram has joined #openstack-marconi22:13
*** sriram has quit IRC22:20
*** balajiiyer1 has joined #openstack-marconi22:25
*** balajiiyer1 has quit IRC22:26
*** balajiiyer has quit IRC22:26
*** cpallares has quit IRC22:45
*** tmu has quit IRC22:56
*** tmu has joined #openstack-marconi22:57
*** rossk has quit IRC22:57
*** flaper87 is now known as flaper87|afk22:57
*** mwagner_lap has joined #openstack-marconi23:01
*** rossk has joined #openstack-marconi23:03
*** shakamunyi has quit IRC23:14
*** shakamunyi has joined #openstack-marconi23:19
*** kgriffs is now known as kgriffs|afk23:57

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