Thursday, 2014-05-22

*** openstackgerrit has quit IRC00:19
*** openstackgerrit has joined #openstack-marconi00:21
*** openstackgerrit has quit IRC00:34
*** openstackgerrit has joined #openstack-marconi00:35
*** openstackgerrit has quit IRC00:49
*** openstackgerrit has joined #openstack-marconi00:51
*** vkmc has joined #openstack-marconi00:54
*** vkmc has quit IRC00:54
*** haomaiwang has joined #openstack-marconi01:30
*** haomaiwang has quit IRC01:34
*** prashanthr_ has joined #openstack-marconi01:59
*** peopleme1ge has joined #openstack-marconi02:56
*** nosnos_ has joined #openstack-marconi02:57
*** rossk_ has joined #openstack-marconi02:57
*** ametts_ has joined #openstack-marconi02:57
*** peoplemerge has quit IRC03:03
*** pquerna has quit IRC03:03
*** nosnos has quit IRC03:03
*** russell_h has quit IRC03:03
*** ametts has quit IRC03:03
*** rossk has quit IRC03:03
*** russell_h has joined #openstack-marconi03:03
*** russell_h has quit IRC03:03
*** russell_h has joined #openstack-marconi03:04
*** russell_h has quit IRC03:05
*** russell_h has joined #openstack-marconi03:05
*** pquerna has joined #openstack-marconi03:09
*** pquerna has joined #openstack-marconi03:09
*** malini_afk is now known as malini03:15
*** boris-42 has quit IRC03:23
*** boris-42 has joined #openstack-marconi03:29
*** malini is now known as malini_afk03:39
*** nosnos_ has quit IRC03:52
*** haomaiwang has joined #openstack-marconi04:35
*** nosnos has joined #openstack-marconi04:35
*** haomaiwang has quit IRC04:40
*** prashanthr_ has quit IRC04:46
*** prashanthr_ has joined #openstack-marconi04:47
*** cpallares has quit IRC04:51
*** reed has quit IRC04:54
*** igor_ has joined #openstack-marconi04:59
*** igor__ has quit IRC05:02
*** mkoderer has joined #openstack-marconi05:54
*** prashanthr_ has quit IRC06:02
*** prashanthr_ has joined #openstack-marconi06:04
*** AAzza has joined #openstack-marconi06:34
*** shivamshukla has joined #openstack-marconi07:01
*** rossk_ has quit IRC07:12
*** flwang has quit IRC07:31
*** flwang has joined #openstack-marconi08:16
*** jamie_h has joined #openstack-marconi08:26
*** ykaplan has joined #openstack-marconi08:26
*** haomaiwa_ has joined #openstack-marconi08:42
*** flwang has left #openstack-marconi08:43
*** haomaiwa_ has quit IRC08:47
*** ashishsnigam has joined #openstack-marconi08:51
*** prashanthr_ has quit IRC09:37
*** AAzza has quit IRC09:38
*** haomaiwang has joined #openstack-marconi10:45
*** haomaiwang has quit IRC11:01
*** flaper87|afk is now known as flaper8711:05
*** flaper87 is now known as flaper87|afk11:27
*** ashishsnigam has quit IRC11:30
*** tedross has joined #openstack-marconi11:31
shivamshuklaHi All11:47
*** Gammo has quit IRC11:49
*** AAzza has joined #openstack-marconi11:49
*** mwagner_lap has quit IRC11:54
*** AAzza has left #openstack-marconi11:57
*** oz_akan_ has joined #openstack-marconi11:59
*** _rohitm has joined #openstack-marconi12:05
_rohitmhi all12:05
shivamshuklaHi _rohitm12:05
_rohitmhey man..wassup12:06
shivamshuklai m kinda new to marconi can you help me build it12:06
_rohitmoh....i am also new and wanted some pointers12:07
shivamshuklafound the starting point here
shivamshuklaand here12:09
*** _rohitm has quit IRC12:14
*** oz_akan_ has quit IRC12:17
*** oz_akan_ has joined #openstack-marconi12:17
*** flaper87|afk is now known as flaper8712:18
flaper87WHAT'S UP PEOPLE!?!?!??!?!!12:18
flaper87shivamshukla: you12:18
flaper87shivamshukla: you're very welcome to this channel12:18
flaper87hope you'll find marconi amazingly fun to work with12:18
*** oz_akan_ has quit IRC12:23
*** malini_afk is now known as malini12:26
*** nosnos has quit IRC12:31
flaper87kgriffs: yo, it's too early for you to be around. What happened?12:34
flaper87malini: hey :)12:34
flaper87malini: hope you're doing great :)12:35
maliniyes!!! except check-tempest-dsvm-marconi is failing in one of the jobs, & it was just made voting y'day :-#12:36
flaper87holy crap12:36
flaper87I actually +1'd that patch12:37
flaper87is it because we reverted the metadata thing?12:37
flaper87I guess no. tempest should hit v1.012:37
malini'/opt/stack/new/devstack/extras.d/ line 17: create_marconi_accounts: command not found'12:38
malinitht is weird..12:38
malinisince we didnt change anything around tht12:38
flaper87that is weird indeed12:38
malinithe pnly recent change is
flaper87also, the gate passed when the patch was approved12:38
flaper87was another patch approved after ours ?12:38
malininothing in marconi12:39
malinilet me check all of devstack12:39
flaper87I meant in devstack12:39
maliniI meant lib/marconi12:39
malininothing tht shud impact us12:39
flaper87malini: did you try running it in a brand new box?12:41
maliniwas just about to do tht12:42
*** sriram has joined #openstack-marconi12:54
*** sriram has quit IRC12:55
*** sriram has joined #openstack-marconi12:55
*** abettadapur has joined #openstack-marconi12:58
*** jchai has joined #openstack-marconi13:00
*** prashanthr_ has joined #openstack-marconi13:19
*** mpanetta has joined #openstack-marconi13:27
*** alcabrera|afk is now known as alcabrera13:47
alcabreragood morning! :)13:47
flaper87alcabrera: gooooood morning13:47
flaper87I see that atlanta just woke up13:48
flaper87I've an idea. Lets shake Texas and scare the hell out of the folks based there13:48
flaper87please, please, please, please13:48
flaper87alcabrera: you're suppose to help me with that13:48
flaper87I can't do everything13:49
flaper87come one13:49
alcabrerascare them to what ends? >.>13:49
* alcabrera does not sit well with this idea13:49
*** balajiiyer has joined #openstack-marconi13:56
flaper87alcabrera: :(13:56
flaper87alcabrera: just to wake them up13:56
*** balajiiyer has quit IRC13:56
*** balajiiyer has joined #openstack-marconi13:56
alcabreraI don't do well when I'm woken up by surprise. :P13:57
prashanthr_alcabrera and flaper87 : Good morningggg :)13:58
flaper87alcabrera: that's why you should always wake up before everyone else :P13:58
flaper87prashanthr_: hey :)13:58
* mpanetta is still sleeping :P13:59
sriramwake up mpanetta! :P14:00
mpanettaNo :P14:00
alcabreraprashanthr_: hey! :)14:01
* flaper87 shakes mpanetta14:01
flaper87mpanetta: scared? Woken up?14:01
* mpanetta flops like a ragdoll :P14:02
*** haomaiwang has joined #openstack-marconi14:02
*** Obulpathi has joined #openstack-marconi14:03
*** Obulpathi has quit IRC14:05
*** Obulpathi has joined #openstack-marconi14:05
*** jamie_h has quit IRC14:11
*** jamie_h has joined #openstack-marconi14:12
*** cpallares has joined #openstack-marconi14:13
*** AAzza has joined #openstack-marconi14:14
*** oz_akan_ has joined #openstack-marconi14:19
*** amalagon has quit IRC14:19
*** whenry has joined #openstack-marconi14:24
mpanettaYeah, I am definitely still asleep...  I just went to look for my bananna, that I already ate...14:31
alcabreraoh no, banana's gone. :(14:31
*** haomaiwang has quit IRC14:34
*** jchai is now known as jchai_afk14:40
alcabreraprashanthr_: how goes the redis effort? :)14:42
prashanthr_alcabrera: Redis is going awesome :) completed the basic version of queue management14:42
prashanthr_now able to create , delete queues14:43
flaper87prashanthr_: do you know if the redis client library is Py3K compatible?14:43
prashanthr_and also manage queue metadata :)14:43
flaper87prashanthr_: you're fast14:43
flaper87is the redis work based on alcabrera's previous work?14:43
alcabreraflaper87: redis-client is py3 compatible, according to
flaper87oh cool14:44
* flaper87 is too lazy to search on pypi14:44
prashanthr_flaper87 : :) ha ha it's still the basic version.14:44
flaper87prashanthr_: may I ask how you created the package? Did you copy mongodb's ?14:45
flaper87peopleme1ge: there?14:45
prashanthr_flaper87: Nope i just copied the base classes initially.14:46
prashanthr_now implementing the methods one-by-one14:46
prashanthr_alcabrera: I had some questions. do we have a meeting today ?14:48
alcabreraprashanthr_: yup!14:48
alcabrerawe'll get started in 10m14:48
prashanthr_alcabrera: Awesome :) sure :)14:48
alcabreraAAzza: thanks for filing the py3 bug about the requirements files. :)14:48
*** oz_akan_ has quit IRC14:50
*** oz_akan_ has joined #openstack-marconi14:51
alcabrerabefore I get distracted15:01
alcabreraprashanthr_, AAzza: do you all have 20-30m to meet and talk about how project work is going?15:02
prashanthr_alcabrera: Sure.15:02
AAzzayeap, of course15:03
alcabreralet's get started15:03
alcabreraAAzza: could you share what the latest is on marconi-py3?15:03
AAzzaalcabrera: not so much actually for now, first have configured testing environment for myself with 4 versions of python)15:05
alcabreranice! which versions, out of curiosity?15:05
AAzzaalcabrera: py26, py27, py33 and py3415:05
alcabreracould you also install pypy-2.3, AAzza?15:06
AAzzaalcabrera: yes, could try, good idea forgot about it15:06
alcabrerawe want to keep our pypy support top notch15:06
alcabreraAAzza: I also noticed you submitted this patch:
*** igor_ has quit IRC15:07
AAzzaalcabrera: yes, patch with requirements15:07
*** igor has joined #openstack-marconi15:07
AAzzaalcabrera: if it will be merged, then we can see that my patches actually reduce number of failing tests15:08
alcabreraI'll work with kgriffs to see that move forward15:08
AAzzaalcabrera: plan to start bugs shown by tests. will do investiagations for now15:08
AAzzaalcabrera: i suppose have a problem that unit tests are not running under py33, not failing just not running. kinda strange15:09
alcabreraoh yeah15:09
alcabreraI remember that15:09
alcabrerathey took *really* long to run on my machine15:10
AAzzaalcabrera: not that problem15:10
alcabreraAAzza: I'll look into it today, and get back to you with how long they take to run from a fresh environment.15:10
AAzzaalcabrera: it works for me now, but only functional15:10
AAzzaalcabrera: looks like unit tests are just skiped) but really not sure, need to dig into it15:11
alcabreralet's investigate that15:11
alcabreracan you sync up with me tomorrow for about 5m around the same time AAzza?15:11
*** igor has quit IRC15:12
AAzzaalcabrera: tomorrow, about same time? yes sure15:12
alcabrerathank you. :)15:12
alcabreraAAzza: anything else you'd like to discuss before we hear prashanthr_'s updates?15:12
AAzzaalcabrera: hmmm, prorably tomorrow, after investigation :) so can switch to prashanthr_15:13
alcabreraprashanthr_: how goes marconi-redis? :)15:14
prashanthr_alcabrera and AAzza: py3k compatibility is a constant exercise i guess :)15:14
alcabreraoh yes, definitely!15:14
prashanthr_so many new stuffs in py315:14
prashanthr_marconi-redis is now accepting new requests for queues15:15
prashanthr_i have tested with wsgi v1.015:15
alcabrerawsgi 1.0?15:16
alcabreracould you elaborate, prashanthr_?15:16
prashanthr_today i implemented the methods for QueueController for redis15:16
prashanthr_so now i am able to create / delete queues15:17
prashanthr_and also set and get metadata related to queues15:17
prashanthr_then i tested it with marconi API v1.015:17
prashanthr_it works fine15:17
alcabreraahhh, I see15:17
alcabrerawsgi 1.0 -> marconi API 1.015:18
alcabreraI got confused, since the WSGI specification in the python ecosystem is pep 333 and pep 333315:18
alcabrerait sounds like things are going well!15:18
prashanthr_alcabrera : yes. yeah me also was inititally confused because they are in wsgi/v1.0 folders :(15:18
alcabreraI'm taking a quick look at your repo at the moment15:18
prashanthr_sure. I have some doubts will ask them after that.15:19
AAzzaprashanthr_: just curious, how do you represent/store queue in redis?)15:19
alcabreraI need to check with flaper87 on this, but -- will we be bringing in redis support as a core storage driver for marconi?15:19
prashanthr_AAzza: i represent the queue information as a Redis Hash15:20
flaper87alcabrera: yup15:20
flaper87as-in, it'll land in the code base and we'll guarantee maintenance for it15:20
alcabreraalright, your current approach and repo setup looks good prashanthr_ :)15:20
alcabreraso - another thought just occurred to me15:20
alcabreraflaper87: should we be using the gerrit patchset approach to developing redis storage, rather than GH commits?15:21
prashanthr_alcabrera: thank you :).15:21
flaper87oh yeah, that'd be the recommended way to do it15:21
alcabreraso we need to make a change there15:21
flaper87that'll invite the community to chime in and it'll give the driver more visibility15:22
alcabreraprashanthr_: are you familiar with the openstack gerrit patch workflow? :)15:22
flaper87I'd recommend using the same approach we used for sqlalchemy15:22
alcabreraflaper87: agreed15:22
prashanthr_alcabrera: Yes. I have submitted 2 patches already.15:22
flaper871. setup the driver base code and test cases (no logic)15:22
flaper872. Start adding some logic for queues15:22
alcabreraprashanthr_: cool! Could you submit a redis storage starter patch sometime soon?15:22
flaper873. Add logic required for messages (some parts need the claim code to be ready, those can be put on-hold)15:23
flaper874. add logic for claims15:23
flaper875. Add the remaining parts in the message controller15:23
flaper876. Add stats in the queue controller15:23
flaper877. add support for redis pools15:23
flaper87pools = partitions = sharding (this is all marconi related)15:23
prashanthr_Sure :) Will start with it tomorrow. so how should i go about creating the bugs for the workflow ?15:24
prashanthr_Will it be on similar lines as the points mentioned by flaper87 ?15:24
alcabreraprashanthr_: I'd like for you to submit patches referencing a redis-storage blueprint15:24
alcabreraI need to find that and make sure you're assigned to it15:24
alcabreraand flaper87's implementation strategy sounds perfect15:25
alcabreraah, yes15:25
alcabrerathis thing: prashanthr_15:25
prashanthr_alcabrera: Ohh i did not know that i could submit patches related to bp's. Yeah kgriffs assigned it to me.15:25
alcabrerawhich you're already assigned to. :D15:25
alcabreraso to address a blueprint in a patch15:25
alcabrerainclude the following comment in your patch15:25
alcabrera#Partially-Implements bp/redis-storage-driver15:26
alcabreraor if it finishees support15:26
alcabrera#Implements bp/redis-storage-driver15:26
alcabreraflaper87: does that look correct?15:26
alcabrera(by comment, I mean: part of your git commit message)15:26
prashanthr_alcabrera: Sure. I will do that.15:27
alcabreraprashanthr_: (Ctrl+F Partially implements)15:27
alcabreraI got the syntax wrong. >.>15:27
alcabreraprashanthr_: thank you!15:27
flaper87Partially-implements blueprint: redis-storage-driver15:27
prashanthr_alcabrera: I have some questions. Can I ask them now ?15:28
flaper87the last patch should have a Implements blueprint: redis-storage-driver15:28
*** igor_ has joined #openstack-marconi15:28
alcabreraprashanthr_: yup, go ahead. :)15:28
prashanthr_the marker for listing of queues has to be the name of the last queue which has been returned to the client.15:29
prashanthr_Am i right ?15:29
* alcabrera thinks15:30
alcabreraso they can then pass in that marker on future queries15:30
alcabrerato paginate from the right place15:30
prashanthr_alcabrera: I have a small glitch related to Redis here. I am unable to use any redis datastructures to implement this.15:30
prashanthr_I can do it with lists but the operation will be very slow15:31
alcabreraIIRC, I had to use zsets, prashanthr_15:31
alcabrerathey provide lexicographical ordering15:31
alcabreraprashanthr_: (example)15:31
prashanthr_alcabrera: Sure. I will check it.15:32
alcabrerayeah, that tripped me up to when I tinkered with a redis driver. :)15:32
alcabreraI hope my efforts will help you!15:32
*** igor_ has quit IRC15:32
alcabrera*to -> too15:32
prashanthr_alcabrera: this is an invaluable reference for me :)15:32
prashanthr_ahh there is msgpack here as well :)15:33
alcabrerayup! :D15:33
alcabrerado you have other concerns, prashanthr_?15:33
prashanthr_alcabrera: I do have 2 more questions. Am i taking too much time ?15:34
alcabreranot at all -- go ahead. :)15:34
prashanthr_alcabrera: this is a stupid question sorry. Actually when seeing the license for the files there were 3-4 types of them15:34
prashanthr_from Rackspace/RedHat/Apache etc.15:34
prashanthr_what should i use ?15:35
alcabrerano question is a stupid question. :)15:35
alcabreraas for the license header15:35
alcabreraI'm not entirely sure15:35
alcabreraI'm inclined to believe that Openstack Foundation is probably the way to go15:35
alcabreraflaper87: what do you think?15:35
maliniprashanthr_: I have seen somebody use their personal name in one of the projects15:35
malinilet me chk if I can use tht, since you are not affiliated to any company15:36
alcabreramalini: that might work, too!15:36
alcabreramalini: thank you for chiming in. :)15:36
* flaper87 reads backlog15:36
flaper87the important piece is the license not the copyright15:36
flaper87so, as long as the license is there, you can omit the copyright line15:36
flaper87if you want to add it, you can add your name there15:37
flaper87or the foundation15:37
prashanthr_flaper87 and malini : thank you :)15:37
flaper87and by important I meant, requied15:37
flaper87required :)15:37
alcabreraI'd like to emphasize that there's no stupid questions, prashanthr_, AAzza. Ask anything, and we'll take it seriously. :D15:37
prashanthr_alcabrera: ha ha sure :)15:37
maliniprashanthr_:  See example here
flaper87I'd like to emphasize that I'm full of bad jokes15:38
prashanthr_flaper87 :)15:38
*** igor_ has joined #openstack-marconi15:38
* alcabrera verifies that flaper87 is full of bad jokes15:38
alcabrerayou said you had one more question, prashanthr_?15:38
*** jamie_h has quit IRC15:38
prashanthr_yes last question : Some of my code kinda of violates DRY in the utils part15:38
prashanthr_is that f9 ?15:39
alcabreraf9 -> fine?15:39
prashanthr_like scoped_project_queue and the reconnection decorators15:39
malinif9 sounds cool :)15:39
prashanthr_alcabrera: ha ha yeah. watsapp lingo :) got into me now.15:39
alcabreraheh. I've never seen it, but it makes sense. :)15:39
alcabrerabut yes15:40
alcabrerawrt DRYness15:40
*** igor__ has joined #openstack-marconi15:40
alcabreraI think for starters, we can try to be DRY as we go15:40
alcabreraand if it's something that's far too duplicated, it'll be brought up in code review15:40
alcabreracase by case15:40
prashanthr_alcabrera: Sure.  I have made a small note in the code as well.15:41
alcabreraworks for me. Just submit a patch and we'll give it a look over. :)15:41
*** flaper87 is now known as flaper87|afk15:41
alcabrerawith that --15:41
alcabreraany closing thoughts/questions/opinions AAzza, prashanthr_?15:42
*** cpallares has quit IRC15:42
prashanthr_alcabrera: Sure. will do it.15:42
AAzzaalcabrera: no, everything is clear;) except 5 version of python on one system is not what you want from life)15:43
prashanthr_alcabrera: None from my side :). crystal clear for today :)15:43
prashanthr_AAzza : ha ha :)15:43
*** igor_ has quit IRC15:43
*** jamie_h has joined #openstack-marconi15:43
alcabreraAAzza: haha, so true15:43
alcabrerawith that, we close our first GSoC/OPW meeting15:43
alcabrerathanks all15:43
prashanthr_awesome :) thank you !!15:44
*** igor__ has quit IRC15:44
*** abettadapur has quit IRC15:46
*** vkmc has joined #openstack-marconi15:47
*** vkmc has quit IRC15:47
*** vkmc has joined #openstack-marconi15:47
*** reed has joined #openstack-marconi15:50
alcabreraAAzza: it took 5m16.805s for all my unit tests to fail when running 'tox -e py33' w/ memcached removed from the reqs fail15:51
alcabrerareqs file15:51
vkmchowdie! :)15:52
alcabreravkmc: heyyyy!15:52
alcabrerahow are you? :)15:52
vkmchi alcabrera! fine, and you?15:58
vkmccatching up with emails and stuff, just came from college15:58
alcabreradoing great, thanks!15:59
alcabreravkmc: If you've some spare time, I'd love to catch up with you regarding amqp. :)15:59
*** abettadapur has joined #openstack-marconi16:00
AAzzaalcabrera: you see if you filter tests with tox -e py33 -- tests.functional, that tests are running and failing, but with  tox -- e py33 -- tests.unit no tests are running and task finishes sucessfully :-( the same commands for py27 works well16:01
*** _rohitm has joined #openstack-marconi16:02
*** jchai_afk is now known as jchai16:04
alcabreraAAzza: oh yeah! tests.unit does skip things, and I'm not sure why. L/16:05
AAzzaalcabrera: is searching now...16:08
vkmcalcabrera, sure, I started reading some of the documentation but I still have some details to check out16:11
alcabreraI did some reading, too, vkmc.16:12
alcabrerathe connection protocol looks a bit complex at the low level side. It mimicks TCP, as far as I can tell. :/16:12
alcabreraI think we can sidestep most of that, though16:13
alcabrerasince I'm sure a python-amqp client exists16:13
vkmcalcabrera, yeah it looks like TCP16:13
vkmcfor the v1.0 version16:14
alcabreraI'm still unsure on whether we should work on a 1.0 or a 0.9.1 amqp storage layer. :/16:15
alcabrerawe'll need to figure this out at the next marconi meeting16:16
alcabreravkmc: I'm adding an agenda item as "AMQP in Marconi: What Version(s) to Support"16:17
alcabrerado you mind if I list you as a co-contributor to this topic?16:18
vkmcplease do! :)16:18
alcabreracool! thanks. :)16:19
*** mwagner_lap has joined #openstack-marconi16:19
alcabreranext meeting is going to be busy16:19
vkmcflaper87 suggested me to focus on 1.0 only... thought it would be great if we can discuss if when we know the features better16:22
alcabrerathe main python client only supports amqp 0.9.116:23
*** prashanthr_ has quit IRC16:23
alcabrerathis worries me a little16:23
alcabrerathough I could be missing something16:23
alcabreravkmc, flaper87|afk: let me know if you find a python client that supports amqp 1.016:24
vkmcI'm on it!16:24
vkmc(can't guarantee I'll find it though)16:24
alcabrerano worries if you can't. :)16:27
alcabrerait just means we have more information16:27
*** jergerber has joined #openstack-marconi16:39
*** igor__ has joined #openstack-marconi16:40
AAzzaalcabrera: it looks like the problem is with testr itself, bug with filtering on py3, so actually when run all tests, all tests are run, just can't filter them for now:-(16:44
alcabrerasadness. :(16:44
alcabrerathanks for finding this, AAzza!16:44
*** igor__ has quit IRC16:45
*** alcabrera is now known as alcabrera|afk17:00
*** balajiiyer has quit IRC17:07
*** ykaplan has quit IRC17:10
*** alcabrera|afk is now known as alcabrera17:36
*** _rohitm has quit IRC17:37
*** balajiiyer has joined #openstack-marconi17:41
*** igor_ has joined #openstack-marconi17:42
*** igor_ has quit IRC17:46
peopleme1geflaper87|afk: I'm back17:52
*** peopleme1ge is now known as peoplemerge17:58
*** vkmc has quit IRC18:01
*** jamie_h has quit IRC18:12
*** vkmc has joined #openstack-marconi18:15
*** Obulpath_ has joined #openstack-marconi18:16
*** Obulpathi has quit IRC18:20
*** vkmc has quit IRC18:23
srirammalini: we need this prioritized as well.18:32
*** Obulpath_ has quit IRC18:36
*** Obulpathi has joined #openstack-marconi18:37
*** igor_ has joined #openstack-marconi18:42
*** vkmc has joined #openstack-marconi18:45
*** igor_ has quit IRC18:47
*** vkmc has quit IRC19:23
*** vkmc has joined #openstack-marconi19:24
*** vkmc has quit IRC19:24
*** vkmc has joined #openstack-marconi19:24
*** flaper87|afk is now known as flaper8719:42
*** igor_ has joined #openstack-marconi19:43
*** igor_ has quit IRC19:47
* flaper87 is back and says hi19:57
* flaper87 reads backlog19:58
flaper87I think we've pretty much decided to go with 1.019:58
flaper87we can bring it up again in our next meeting but unless there's something critical that will change our minds, I think 1.0 is the way we'll go19:58
vkmchi flaper87 o/19:59
flaper87vkmc: hey19:59
flaper87vkmc: Have you started reading about amqp 1.0 ?19:59
vkmcI'm on it yeah, still have test it20:00
flaper87as for the library, we'll use qpid-proton20:00
flaper87it's a messaging library based on the amqp 1.0 standard20:00
vkmcI was checking on pika, but it's AMQP 0.920:01
alcabrera -- this one flaper87?20:01
flaper87nope, 2 secs20:01
flaper87The qpid family may be confusing20:02
alcabreraoh, wait20:02
flaper87qpid-python should be the qpid library (0.9)20:02
flaper87there's qpid-dispatch which is a message router20:02
alcabrerait's not on pypi yet it seems. >.>20:02
flaper87then qpid-proton20:02
flaper87that one20:02
flaper87it's weird, it should be there20:02
flaper87I'll check on the guys behind it20:03
tjanczuk_So... AMQP... ;)20:03
tjanczuk_What is the rationale again for 1.0 vs 0.9?20:03
alcabreraI'm out for the day, all. :)20:05
alcabreratake care. o/20:05
flaper87cross-broker, single implementation, message oriented, support for peer-to-peer, easier to scale, support for federation and easier to put behind HA20:05
flaper87alcabrera: thanks20:05
flaper87take care20:05
*** alcabrera is now known as alcabrera|afk20:06
vkmcttfn alcabrera, enjoy your evening o/20:07
tjanczuk_Given that this is storage layer and the protocol is not exposed, folks really don't care if this is AMQP or FooBar, right? I wonder the focus of this decision should be on the actual product rather than protocol.20:07
flaper87that's a good concern. I think it should be both, TBH. What Marconi will support won't only help marconi but the overall community.20:10
flaper87I think the benefits that AMQP 1.0 brings as a standard will help marconi as well20:11
flaper87some of those are the ones I mentioned before20:11
tjanczuk_Looking at this from the product perspective, Rabbit has a lot going for it: it is very popular, well supported, large community, active development, already used in OpenStack (so folks deploying OpenStack need one less area of expertise).20:13
tjanczuk_And frankly I think it is less important if the intergation is over AMQP 0.9, MQTT, or STOMP. OK, maybe not STOMP ;)20:13
tjanczuk_Cutting to the chase: I would love to see a Rabbit-based storage layer in Marconi. I can write it. If I do, can this become part of the project?20:21
*** igor_ has joined #openstack-marconi20:25
*** igor__ has joined #openstack-marconi20:27
flaper87Again, I think that's a valid concern but I'm less worried about that20:29
flaper87I spoke with some rabbitmq guys and the only reason they haven't started dedicating *LOT* of time to it is because there wasn't a huge demand20:29
*** igor_ has quit IRC20:30
flaper87I think OpenStack is huge enough and is worth the amqp 1.0 support20:30
flaper87if what rabbitmq needs is a reason to support it, I'll be more than happy to give them that motivation20:30
flaper87I'll hopefully have more info about this soon20:31
*** igor__ has quit IRC20:32
*** mwagner_lap has quit IRC20:33
*** sriram has quit IRC20:33
*** balajiiyer has quit IRC20:33
malinifyi…the broken tempest job was due to my devstack patch20:34
maliniI have submitted to fix this
flaper87malini: :( what happened ?20:35
maliniBut I am waiting for the experimental job to get back the results, to make sure tht fixes it20:35
flaper87oh :S20:35
maliniflaper87: I am embarassed to admit it.But bad syntax in the shell script20:35
flaper87but, how did it pass?20:35
tjanczuk_flapper87: Waiting for AMQP 1.0 to get popular enough for Rabbit to decide to support it, even if this happens, does not sound like timely way to enable people to use Rabbit with Marconi.20:36
malinidevstack doesnt have gate jobs to check if the script it got past20:36
tjanczuk_So again: if I write it, can this become part of the project?20:36
*** abettadapur has quit IRC20:36
*** balajiiyer has joined #openstack-marconi20:36
malinitjanczuk_: why can't it become part of the project, if you write it?20:37
flaper87To be clear. They support it as in they respond to bug reports20:37
flaper87I've asked for the list of supported features20:38
flaper87(and missing ones)20:38
flaper87I'll hopefully have it in the next few days20:38
flaper87tjanczuk_: as for 0.9 in Marconi's code base, I wouldn't be fully against it but I'm not convinced. For starters, I'd hate to have support for rabbit and not other brokers20:39
flaper87and I do care about that, regardless rabbit being the most deployed broker20:39
tjanczuk_flapper87: There are no notable deployments of rabbit with amqp 1.0 I am aware of. Everyone I know uses it with the built-in 0.9. Including OpenStack.20:39
tjanczuk_Is this necessarily an XOR? Maybe we can have 1.0 and 0.9?20:40
flaper87tjanczuk_: It's not an XOR. As of now, I think 1.0 will happen regardless. 0.9 will be an addition20:40
flaper87You could start working on it as an external plugin and we can discuss adding it to the codebase later20:41
flaper87I think that's a fair approach20:41
flaper87Installing marconi w/ the rabbit driver would be a matter of doing: $ pip install marconi-rabbit-driver20:42
tjanczuk_Is the redis driver going to be internal or external?20:43
flaper87The current plan is to support: mongodb, redis and amqp 1.020:44
flaper87there's sqlalchemy but lets not talk about that20:44
tjanczuk_So what makes something internal vs external?20:45
tjanczuk_What's the principle?20:45
flaper87Right now what gets into the code base is what the team considers required and most importantly what the team is willing to/capable of maintain(ing)20:47
flaper87There are some brilliant interns working on the redis driver and the amqp one20:47
flaper87I really hope they'll maintain them afterwards20:47
flaper87but in the real awful case they won't, the team will be able to take over20:48
flaper87This rationals will likely change in the future20:48
flaper87but we definitely want to keep this set vary but small20:48
flaper87Marconi was built with the community in mind20:48
flaper87which means it has everything needed for the community to contribute without having the code in marconi's root20:49
flaper87with all that said, I need to step out for a bit20:49
flaper87I hope it makes sense for you20:49
flaper87if it doesn't, please, by all means, speak up :)20:49
flaper87malini: we haven't talked about the functional tests20:50
maliniflaper87: aah yes!20:50
flaper87but based on the few things we said, we should probably keep everything as is and focus on other things20:50
maliniflaper87: I got this idea from one of the tempest sessions20:50
maliniflaper87: The neutron folks are forking on implementing their functional tests, so tht they can just copy over their tests to tempest20:52
malinii.e instead of maintaining two sets of tests (the project level functional tests + tempest API tests), write one set of tests which can live in both the repos20:52
tjanczuk_flapper87: What would be required for the rabbit storage to become internal?20:53
maliniflaper87: this will help us avoid duplicate efforts - tht is the only advantage20:53
maliniflaper87: on the flip side we lose a lot of flexibility -eg. tempest does class level setups for tests etc.  + validations happens in the queuing client etc.20:54
maliniflaper87: So I am still split b/w if we should try doing tht or not20:55
malinialso we'll have to stop using ddt20:55
flaper87malini: I think we're better of by having our own test suite20:56
flaper87when it comes to tests, I think there's some value in duplicating the test suites20:57
flaper87there are also some drawbacks20:57
flaper87that said, I'm not really happy with the way things are done in tempest20:57
flaper87so, I'd really prefer not pulling them into marconi20:57
maliniflaper87: I dont see much value in duplication - but I like a lot of the stuff we are doing now, & we'll have to stop those if we go the tempest friendly route20:57
flaper87malini: the value in keeping both things separate is that we get more people to look deeply in our API besides us20:58
flaper87if we write tempest tests and copy them to tempest, then it'd be just us looking at it20:58
malinias is, I am the only one looking at both :D20:59
flaper87and that may hide some UX issues20:59
flaper87malini: yup :D20:59
malinihopefully tht'll change as we get more contributors20:59
flaper87that said, we should probably re-consider this during K20:59
flaper87I think we've enough things in our plate for Juno20:59
maliniLets continue the current route & re-evaluate during K20:59
flaper87awesome, sounds good21:00
malinihopefully the common tempest libs will be mature by then21:00
malinithere was also some talk abt an oslo lib for tests21:00
flaper87that one we should use21:01
*** cpallares has joined #openstack-marconi21:01
maliniflaper87: does it already exist?21:01
*** jay-atl has quit IRC21:02
flaper87it's very simple21:02
flaper87I'll migrate our code to use that library soon21:02
flaper87that's supposed to be the liasion work21:02
malinidoesnt look like they have functional tests yet21:03
malinimaybe tht'll come next21:03
flaper87oh no no, that lib is just for base clases and some common code21:04
tjanczuk_flapper87: I really care about the rabbit storage layer to be internal. What would be required to make this happen (besides me writing it)?21:04
flaper87we could probably contribute back with support for functional tests21:04
flaper87tjanczuk_: Don't get me wrong, I care about it too. You writing it is the first thing for sure. That'd be really helpful21:05
flaper87and thanked21:05
*** oz_akan_ has quit IRC21:06
flaper87We need to discuss the inclusion in the codebase but before doing that, we need to sort some other things out21:06
*** boris-42 has quit IRC21:06
flaper87for example: What is the minimum set of required features in a store driver21:06
flaper87What are the required changes in the API for an AMQP/broker-based driver?21:07
flaper87that's one of the reasons I'm recommending to start as an external driver21:07
flaper87and then we can discuss the inclusion based on the parameters I just mentioned and other things21:07
flaper87tjanczuk_: thanks for your input on this21:07
tjanczuk_My understanding is this discussion would need to happen for redis and amqp 1.0 as well, so rabbit driver would not really affect the outcome. From the perspective the difference between 1.0 and 0.9 is insubstantial.21:08
*** boris-42 has joined #openstack-marconi21:08
tjanczuk_And actually I will be happy to help flesh out some of these questions as part of the work.21:08
flaper87Absolutely, those questions are valid for the three new drivers we're targeting21:10
tjanczuk_Besides the features/APIs aspects, what are the "other things" that inclusion would be based on?21:15
*** balajiiyer has quit IRC21:23
*** igor_ has joined #openstack-marconi21:28
*** lakspace has quit IRC21:31
*** igor_ has quit IRC21:32
*** tedross has quit IRC21:35
*** AAzza has quit IRC21:35
*** vkmc has quit IRC21:43
*** vkmc has joined #openstack-marconi21:43
*** malini is now known as malini_afk21:46
*** Obulpathi has quit IRC21:50
*** mpanetta has quit IRC21:52
*** reed has quit IRC22:01
*** mkoderer has quit IRC22:02
*** jchai has quit IRC22:05
*** boris-42 has quit IRC22:17
*** boris-42 has joined #openstack-marconi22:20
*** flaper87 is now known as flaper87|afk22:24
*** igor_ has joined #openstack-marconi22:28
*** jergerber has quit IRC22:32
*** igor_ has quit IRC22:33
*** mwagner_lap has joined #openstack-marconi22:42
*** openstackgerrit has quit IRC23:19
*** openstackgerrit has joined #openstack-marconi23:21
*** cpallares has quit IRC23:25
*** oz_akan_ has joined #openstack-marconi23:26
*** igor_ has joined #openstack-marconi23:29
*** igor_ has quit IRC23:34

Generated by 2.14.0 by Marius Gedminas - find it at!