Friday, 2014-01-17

*** fifieldt has joined #openstack-marconi00:36
*** amitgandhi has joined #openstack-marconi00:38
*** amitgandhi has quit IRC00:42
*** malini has joined #openstack-marconi01:09
*** balajiiyer has joined #openstack-marconi01:15
openstackgerritA change was merged to openstack/python-marconiclient: Sync with global requirements
*** amitgandhi has joined #openstack-marconi01:39
*** amitgandhi has quit IRC01:43
*** balajiiyer has quit IRC01:43
*** nosnos has joined #openstack-marconi02:04
*** malini is now known as malini_afk02:19
*** balajiiyer1 has joined #openstack-marconi02:19
*** amitgandhi has joined #openstack-marconi02:39
*** amitgandhi has quit IRC02:43
*** flwang has quit IRC02:52
*** balajiiyer1 has quit IRC02:52
*** admiyo has joined #openstack-marconi02:58
*** flwang has joined #openstack-marconi03:29
*** amitgandhi has joined #openstack-marconi03:40
*** amitgandhi has quit IRC03:44
*** balajiiyer1 has joined #openstack-marconi04:02
*** chandankumar has joined #openstack-marconi04:32
*** amitgandhi has joined #openstack-marconi04:40
*** fifieldt has quit IRC04:41
*** fifieldt has joined #openstack-marconi04:42
*** amitgandhi has quit IRC04:45
*** balajiiyer1 has quit IRC04:48
*** amitgandhi has joined #openstack-marconi04:50
*** amitgandhi has quit IRC04:55
*** chandankumar has quit IRC05:03
*** balajiiyer1 has joined #openstack-marconi05:03
*** balajiiyer1 has quit IRC05:19
*** admiyo has quit IRC05:38
*** amitgandhi has joined #openstack-marconi05:41
*** amitgandhi has quit IRC05:46
*** flwang has quit IRC05:57
*** reed has quit IRC06:07
*** chandankumar has joined #openstack-marconi06:15
*** amitgandhi has joined #openstack-marconi06:41
*** amitgandhi has quit IRC06:45
*** amitgandhi has joined #openstack-marconi06:52
*** amitgandhi has quit IRC06:56
*** pquerna has quit IRC07:01
*** pquerna has joined #openstack-marconi07:01
*** amitgandhi has joined #openstack-marconi07:42
*** amitgandhi has quit IRC07:45
*** flaper87|afk is now known as flaper8707:55
*** jamieh has joined #openstack-marconi08:04
*** fifieldt has quit IRC08:05
*** fifieldt has joined #openstack-marconi08:07
*** fifieldt has quit IRC08:37
*** amitgandhi has joined #openstack-marconi08:42
*** amitgandhi has quit IRC08:47
*** amitgandhi has joined #openstack-marconi08:53
*** amitgandhi has quit IRC08:57
flaper87Alex_Gaynor: you are not around, are you?08:58
*** malini_afk has quit IRC09:00
*** malini_afk has joined #openstack-marconi09:06
*** malini_afk is now known as malini09:06
*** jamieh has quit IRC09:15
*** yassine has joined #openstack-marconi09:20
*** amitgandhi has joined #openstack-marconi09:43
*** amitgandhi has quit IRC09:48
*** amitgandhi has joined #openstack-marconi10:44
*** amitgandhi has quit IRC10:48
*** nosnos has quit IRC10:53
*** flwang has joined #openstack-marconi10:53
*** amitgandhi has joined #openstack-marconi10:54
*** amitgandhi has quit IRC10:58
*** amitgandhi has joined #openstack-marconi11:44
*** amitgandhi has quit IRC11:49
*** jamieh has joined #openstack-marconi11:54
*** chandankumar has quit IRC12:42
*** amitgandhi has joined #openstack-marconi12:45
*** tedross has joined #openstack-marconi12:47
*** amitgandhi has quit IRC12:49
*** alcabrera has joined #openstack-marconi12:52
alcabreraGood morning! :D12:55
*** amitgandhi has joined #openstack-marconi12:55
*** amitgandhi has quit IRC12:59
flaper87alcabrera: goood morning13:28
flaper87FYI, it's a don-deterministic failure13:28
flaper87I've no idea why that test passed13:28
alcabreraflaper87: messages/messages is weird.13:29
alcabreradefinitely non-deterministic13:29
flaper87alcabrera: TBH, I'm tempted to ninja approve that patch and see what happens but kgriffs_afk should be around soon so, I'll just wait13:30
flaper87alcabrera: do you mind +2'ng it ?13:30
*** thomasem has joined #openstack-marconi13:31
alcabreraflaper87: I +2'd it before I even saw your request. :)13:31
alcabreraI like the tesr support.13:31
alcabreraI was convinced when I saw the top 10 slowest tests listed after a run.13:31
Alex_Gaynorflaper87: I'm around now13:32
alcabreraNinja approvals are the way of the future. We should finish all the pending Marconi features while kgriffs_afk is still away, flaper87. ;)13:32
flaper87Alex_Gaynor: hey, I've 2 questions: 1) Are you going to udpate this patch? I can do it and add you as co-author if you want13:33
flaper87Alex_Gaynor: 2) We've been seeing some weird behaviors when running tests with pypy:
Alex_Gaynorflaper87: ugh, I kept meaning too, but then it was the holidays, and then I was at a conference, and now I've just finished being sick :-/ If you don't mind updating it that'd be awesome, otherwise I'll probably do it next week13:34
flaper87but that's a non-deterministic failure so, I don't think it's 100% related to pypy13:34
flaper87Alex_Gaynor: I can do it, I just need to update the tests, the rest looks good13:34
Alex_Gaynorflaper87: thanks so much -- looking at that pypy failure now13:34
flaper87I just installed an ubuntu vm with the exact same version of pypy the gate is using13:35
flaper87I'll try to replicate the issue. I haven't been able to replicate it locally13:35
flaper87Alex_Gaynor: btw, thanks for looking at the issue :)13:37
flaper87Alex_Gaynor: also, re your patch: I just need you to restore it13:38
Alex_Gaynorflaper87: done13:38
*** amitgandhi has joined #openstack-marconi13:45
*** oz_akan_ has joined #openstack-marconi13:49
*** oz_akan_ has quit IRC13:49
*** oz_akan_ has joined #openstack-marconi13:50
*** amitgandhi has quit IRC13:50
*** mpanetta has joined #openstack-marconi13:52
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Expose the stats attribute on a queue
flaper87Alex_Gaynor: ^14:06
alcabreraAlex_Gaynor, flaper87: Thanks for putting together the stats patch for the client. +2'd. :)14:08
* flaper87 is so tempted to ninja-approve the testr patch14:12
*** mpanetta_ has joined #openstack-marconi14:13
*** mpanetta has quit IRC14:13
flaper87so, I've got this ubuntu box set up14:13
flaper87pypy installed and I'm now installing marconi's dependencies14:13
flaper87lessee if I can reproduce this failure14:13
*** jcru has joined #openstack-marconi14:14
alcabreraflaper87: good luck. :D14:15
flaper87alcabrera: I'm getting that unicode error you mentioned the other day14:17
*** whenry has joined #openstack-marconi14:20
alcabreraah, to fix that, I did...14:20
alcabrerasix.binary_type(json.dumps(...)) throughout that file, and changed io.StringIO to io.BytesIO14:20
alcabrerait's like 4 places that need that change, flaper87.14:20
flaper87alcabrera: still, the tests that failed in zuul, pass in this box:
alcabreraweird. :/14:24
*** mpanetta_ is now known as mpanetta14:28
flaper87erm nevermind, I had to rebase the patch14:29
*** amitgandhi has joined #openstack-marconi14:41
*** amitgandhi has joined #openstack-marconi14:42
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Use testr instead of nosetest
flaper87alcabrera: ^ rebased14:44
flaper87and added the six.test_type things14:44
* alcabrera checks it out14:44
alcabreraflaper87: thanks!14:44
*** cpallares has joined #openstack-marconi14:45
alcabreraflaper87: +2 - looks good to me.14:46
*** balajiiyer has joined #openstack-marconi14:48
*** fifieldt has joined #openstack-marconi14:51
*** russellb is now known as rustlebee14:55
flaper87alcabrera: danke14:55
flaper87kgriffs_afk: d00d, wake up14:55
*** mpanetta has quit IRC15:02
*** mpanetta has joined #openstack-marconi15:03
*** jergerber has joined #openstack-marconi15:12
*** zyuan has joined #openstack-marconi15:14
*** ayoung has joined #openstack-marconi15:15
*** ayoung is now known as admiyo15:38
*** davidhadas has joined #openstack-marconi15:41
*** jcru has quit IRC15:44
*** jcru has joined #openstack-marconi15:45
flaper87kgriffs_afk: man16:01
alcabreraI hope he's getting plenty of rest.16:03
alcabrerahe released falcon 0.1.8 last night.16:03
*** reed has joined #openstack-marconi16:08
* flaper87 sends a hungry-killer falcon to wake kgriffs_afk up16:26
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Enforce size verification on content-length
*** kgriffs_afk is now known as kgriffs16:42
* kgriffs wakes up16:42
alcabrerathose falcons, yeah, they'll wake a person yp16:42
kgriffsthere's that patch that needs to be benchmarked16:44
*** davidhadas_ has joined #openstack-marconi16:47
*** davidhadas has quit IRC16:47
malinithanks kgriffs..I might be able to get to it only by monday16:48
kgriffsany chance alcabrera could help to make it happen today?16:49
kgriffsWe need to have everything finalized by tues for i-2 and so monday is cutting it pretty close16:49
kgriffsmalini: btw, we will want to do a comprehensive testing pass on the icehouse-2 release candidate (RC)16:50
kgriffswe will have two days to test and fix any critical bugs16:50
maliniSure..if it passes all the functional tests we should be good16:51
alcabrerakgriffs: low chance that I'll be able to help. :/16:56
kgriffsI will see how much i can get done without waiting on that patch to merge16:56
* kgriffs puts on jacket16:56
* kgriffs goes to the movies16:56
* kgriffs feels productive16:56
*** whenry has quit IRC16:58
flaper87kgriffs: you sure we want that patch for i-2 ?17:01
flaper87I'm leaning towards holding it off til i-317:02
kgriffsgood question17:02
flaper87Also, I added some comments17:02
*** davidhadas_ has quit IRC17:02
kgriffsI was hoping to handle autoreconnect since it is common for an operator to step down a node and it would be nice to have things happen transparently for the user rather than them getting 50317:03
kgriffs(i think that's what happens now?)17:03
flaper87mmh, yeah, I think it's 503 or 500, not sure.17:05
flaper87but autoreconnects are raised when something bad happens to the master node17:06
flaper87not trying to use that as an excuse, I just want us to analyse the priority of this patch17:06
flaper87which we haven't heavily tested yet. There's still time and I'm sure malini will work tomorrow and on Sunday too17:07
flaper87right ? right ?17:07
* flaper87 hides17:07
* flaper87 runs away17:07
* flaper87 uses kgriffs as a human shield17:07
malinimalini has suddenly gone blind17:07
maliniBut seriously what do we need to get done for the RC to be ready?17:08
flaper87there's another point of view. Lets get it in and fix it later if something is not right17:08
kgriffsflaper87: autoreconnect happens when you want to upgrade mongod or something too, and you step down the master, right?17:09
kgriffsbalajiiyer:  something tells me "Evaluate Pecan Framework" is going to slip into i-317:09
kgriffsor are you working on that already?17:10
flaper87kgriffs: yeah, that's another case17:12
*** fifieldt has quit IRC17:12
flaper87sorry, got distracted17:15
flaper87kgriffs: how often does one upgrades mongodb's nodes ?17:15
kgriffsmmm, probably not terribly often17:24
kgriffsI am OK letting this slip to i-3 if we end up not having enough time to test17:25
flaper87sure, lets work for getting it into i-217:28
flaper87we'll see what happens by the EOD17:28
kgriffsflaper87: I'm reviewing this:
kgriffsquestion: do we still need these settings in marconi.conf?17:33
kgriffswhy are we checking req.content_length on claim posts?17:38
flaper87  _____ ____  ___ ____    _ __   ___ _ _17:38
flaper87 |  ___|  _ \|_ _|  _ \  / \\ \ / / | | |17:38
flaper87 | |_  | |_) || || | | |/ _ \\ V /| | | |17:38
flaper87 |  _| |  _ < | || |_| / ___ \| | |_|_|_|17:38
flaper87 |_|   |_| \_\___|____/_/   \_\_| (_|_|_)17:38
*** yassine has quit IRC17:39
flaper87kgriffs: are we checking content lenght in claims ?17:39
alcabreraflaper87: awesome! :D17:39
* flaper87 stole that from rust's channel17:39
kgriffsthere is no such thing as claim "metadata"17:40
*** yassine has joined #openstack-marconi17:40
*** zyuan__ has joined #openstack-marconi17:40
kgriffslooks like some YAGNI got into our beautifurous code17:40
flaper87kgriffs: I'm confused17:40
flaper87ah ok17:41
flaper87it's already there17:41
flaper87damn gotta run17:41
kgriffsno problem17:41
*** yassine has quit IRC17:42
*** yassine has joined #openstack-marconi17:43
*** yassine has quit IRC17:43
*** zyuan has quit IRC17:44
*** yassine has joined #openstack-marconi17:44
*** yassine has joined #openstack-marconi17:45
*** yassine has quit IRC17:45
*** balajiiyer has quit IRC17:54
kgriffsalcabrera: how about "tx" for transaction ID17:56
kgriffs(instead of 'g' for group/batch)17:56
* kgriffs is open to suggestions17:57
*** amitgandhi has quit IRC18:02
alcabreratx works for me, kgriffs. I rather like it. Succinct, and it makes me think of transmissions and transactions.18:07
* kgriffs does that18:08
*** balajiiyer has joined #openstack-marconi18:15
*** amitgandhi has joined #openstack-marconi18:15
*** jamieh has quit IRC18:27
*** zyuan has joined #openstack-marconi18:30
*** zyuan__ has quit IRC18:35
kgriffspublic service announcement: zuul is crashing and burning. infra is working hard to resolve the issue.18:42
openstackgerritKurt Griffiths proposed a change to openstack/marconi: fix(mongodb): Ensure batch message posts are atomic (all or nothing)
kgriffsOpenStack is seriously considering dropping XML support entirely19:06
mpanettaIs this bad?19:06
balajiiyermpanetta: enterprise adoption will take a hit.19:10
mpanettaThat is bad then...19:10
mpanettaXML is so verbose...19:11
balajiiyerI think there are both pros and cons to having xml support, but IMO, if you want to cater to vast majority of people, I think you have to have XML.19:12
kgriffsbalajiiyer: idk19:13
kgriffsI think enterprises are mostly interesting in having a good SDK in their language of choice19:13
kgriffsin talking with some folks at OS summit, it seems that the main reason people ask for it is they have these black-box enterprisey appplicances that only speak XML19:14
balajiiyerRight. I would really like to see some numbers, idk who has them - perhaps sales folks, that try to sell products to these big enterprises. XML is the new IE6. :)19:17
kgriffsanyway, i think Rackspace is going to do some market research to find out what is needed19:17
openstackgerritA change was merged to openstack/marconi: Added a request and response file to common
kgriffssomeone on the dev list implied that rackspace's stock dip and revenue slowdown was due to private cloud not taking off19:18
kgriffshe further implied that perhaps XML was one of the blockers (since OS support for XML is haphazard)19:19
balajiiyeryeah, saw that19:19
mpanettaOH bleh19:19
kgriffslogical fallacies all over the place19:19
kgriffsanyway, it could be true, but I doubt it19:19
kgriffsI think that thread is ending on a good note, tho19:20
mpanettaHow does it make sense?19:20
mpanettaWould theynot be using the client langauage ports anyway?19:20
kgriffsyeah, that's what I don't get19:20
mpanettaThat kind of makes the transport irrelevant, no?19:20
kgriffsEnterprises say they want XML, but so far the only concrete reason I've been able to get out of anybody requesting said XML was they have this appliance thing19:21
kgriffsbut srsly19:21
kgriffsif your appliance doesn't speak JSON, you need a new vendor19:21
mpanettaIt is kind of like demanding something support SOAP these days...19:22
kgriffsOR, create a JSON <---> XML bridge19:22
balajiiyerI think part of it is because of compliance requirements.19:22
kgriffswhy the heck do you need XML for compliance?19:22
kgriffsis it because they want to rigorously validate the schema?19:23
kgriffswell, regardless, we need more info on the who and why19:23
balajiiyerFor example - one of the apps I worked on  in my previous life insisted that they should SAML to meet compliance, SAML means XML + well defined schema19:24
kgriffscan't you do that with JSON + json-schema now?19:25
balajiiyeryou could, unless your compliance inspector is forward thinking as well. ;)19:26
balajiiyer* if your compliance ....19:27
kgriffsbalajiiyer, alcabrera, flaper87: could use another round of reviews here:
* alcabrera clicks19:42
alcabrerakgriffs: +2'd. I'm pretty happy with this. I'd love to see some performance testing on this patch, and possibly some load testing, but conceptually, it LGTM.19:44
*** yassine has joined #openstack-marconi19:46
*** reed has quit IRC19:47
kgriffswe definitely need to test this rigorously!19:53
*** yassine has quit IRC19:55
notmynameI made you a graph
kgriffsthans d00d!20:06
* kgriffs bookmarks20:06
notmynamekgriffs: is everything I have, for an overall view of what's going on20:08
kgriffsah, cool. I was looking at that earlier today. Glad to have Marconi on there now!20:09
kgriffsbalajiiyer, flaper87: ^^^20:10
* flaper87 is back20:12
flaper87notmyname: wow, that's amazing, THANKS!20:13
notmynameflaper87: you misspelled "scary and horrifying"20:13
flaper87notmyname: LOOOL, I wasn't talking about the content but the tool20:14
notmyname(for the all graphs page--marconi actually is one of the best looking ones)20:14
notmynameI got an error from graphite when I tried to add gate-marconi-python33 so sorry that's not on your graph too20:15
alcabreranotmyname: thanks for the graphs! :)20:15
kgriffssilly graphite20:16
notmynamealcabrera: thanks. I figure if I don't have an answer to the problem (of slow, unstable gates), at least I can provide more info on where to look :-)20:16
alcabreradefinitely appreciated. on eager reviewing days, it's easy to wonder, "I wonder how long gate will take...?", and after enough time, "Is gate even working right now?". :)20:18
kgriffsd00t, what is up with py26 gate. Been queued forever.20:18
kgriffsI think this is a good time for me to go get a sandwich20:19
*** kgriffs is now known as kgriffs_afk20:22
flaper87kgriffs_afk: damn, :(20:23
flaper87kgriffs_afk: I hate to nitpick on these things but:
flaper87it's a very small and quick fix20:24
alcabreraflaper87: It's a valuable nitpick. Indices are such subtle things. :)20:24
*** thomasem has quit IRC21:01
*** kgriffs_afk is now known as kgriffs21:16
kgriffsok, so you wantz teh notez21:17
* kgriffs does that21:17
*** zyuan has quit IRC21:18
*** zyuan has joined #openstack-marconi21:23
*** admiyo has quit IRC21:23
openstackgerritA change was merged to openstack/marconi: Use testr instead of nosetest
openstackgerritA change was merged to openstack/marconi: Enforce size verification on content-length
alcabreraHeading home. Have a great weekend! :D21:51
*** alcabrera has quit IRC21:51
*** rustlebee is now known as russellb21:52
*** tedross has quit IRC21:53
openstackgerritKurt Griffiths proposed a change to openstack/marconi: fix(mongodb): Ensure batch message posts are atomic (all or nothing)
kgriffsflaper87: you following the XML discussion on the ML?21:54
flaper87kgriffs: no, I had to call my doctor right after I read XML in the subject21:54
kgriffsflaper87: updated comment here
flaper87then I read it was about dropping it21:54
flaper87and I relaxed a bit21:55
kgriffsit is causing quite a stir within Rackspace21:55
*** oz_akan_ has quit IRC21:55
kgriffsanyway, interesting that the elephant in the room is finally being called out21:55
flaper87kgriffs: I +2'd, If I'm not around, feel free to ninja-approve it21:56
flaper87so it gets merged before monday21:56
flaper87kgriffs: I'll read that thread just because I'm a curious cat: MEEEEEOOOWWWWWW21:57
flaper87kgriffs: and if something bad happens to me, it's your fault21:57
flaper87just sayin'21:57
kgriffsit's ok21:57
kgriffsit's not as bad as reading restructured text or being forced to read a Twilight book or anything.22:04
* kgriffs shudders at the thought of reading Twilight, formatted with rST22:04
*** reed has joined #openstack-marconi22:18
cpallaresI'm leaving. Have a good weekend! o/22:24
*** cpallares has quit IRC22:24
*** reed_ has joined #openstack-marconi22:32
*** reed__ has joined #openstack-marconi22:34
*** reed has quit IRC22:35
*** reed__ has quit IRC22:35
*** reed_ has quit IRC22:37
openstackgerritKurt Griffiths proposed a change to openstack/marconi: fix(mongodb): Ensure batch message posts are atomic (all or nothing)
flaper87kgriffs: what changed?22:42
kgriffsnothing, just rebased22:42
kgriffsyou can go back to sleep22:42
* flaper87 doesn't sleep but he appreciates the thought22:43
* kgriffs thought *he* was a light sleeper22:43
kgriffsthat explains it!22:43
*** jcru has quit IRC22:44
*** balajiiyer has left #openstack-marconi22:55
openstackgerritKurt Griffiths proposed a change to openstack/marconi: fix(wsgi): Removed duplicate, YAGNI config options for max doc sizes
flaper87kgriffs: thanks22:58
flaper87kgriffs: btw, I'd like to make queue's limit more.... queuish22:58
flaper87I mean, using metadata_size_limit is quite confusing22:59
flaper87what do you think?22:59
kgriffsi don't think I understand what you mean22:59
*** zyuan has quit IRC23:00
flaper87kgriffs: I mean, I think we should change the name of that configuration param23:04
kgriffsoh, sure23:04
flaper87queue_metadata_limit or something like that23:04
kgriffsmakes sense23:04
flaper87can we make this change part of your last patch ?23:04
flaper87sounds like a good patch to fix things23:04
kgriffswhy the heck not23:05
flaper87YAGNI violations and params inconsistency23:05
flaper87I like the way you think23:05
kgriffsit's friday, after all23:05
* flaper87 gets that ascii art again23:05
kgriffsis there a way to do this and be backwards-compatible with the old name?23:10
kgriffsdoes oslo.config support aliasing or anything?23:10
*** jergerber has quit IRC23:11
*** amitgandhi has quit IRC23:14
*** reed__ has joined #openstack-marconi23:14
kgriffsI think the world needs a new video on youtube23:20
kgriffs"EPIC RAP BATTLE: XML vs. JSON"23:20
*** oz_akan_ has joined #openstack-marconi23:22
*** reed__ has quit IRC23:25
flaper87kgriffs: LOOOL23:31
*** mpanetta has quit IRC23:46
*** flaper87 is now known as flaper87|afk23:53

Generated by 2.14.0 by Marius Gedminas - find it at!