| *** kgriffs is now known as kgriffs|afk | 00:05 | |
| Eva-i | flwang1: yes, I'm still here, can't sleep. I thought you didn't liked the patch with message id in notifications. This patch was working at the moment of submission, now there's merge conflict. It's a quick solution with less refactoring, but with caveat: the cloud admin must ensure that "zaqar.notification.notifier" is the last stage in the message pipeline in zaqar.conf. | 00:13 |
|---|---|---|
| *** tqtran has quit IRC | 00:13 | |
| *** stanchan has joined #openstack-zaqar | 00:17 | |
| *** Qiming has quit IRC | 00:27 | |
| Eva-i | flwang1: hm, I think I know a solution which will be still small and it wouldn't require "zaqar.notification.notifier" to be the last stage. | 00:29 |
| Eva-i | flwang1: but it will modify the message storage controller in unusual way: it will add a job to do notifications after doing usual storage controller routine. Would you like me to upload a proof of concept of it? | 00:35 |
| flwang1 | Eva-i: TBH, i don't mind refactoring, i care about the regression and performance | 00:52 |
| flwang1 | Eva-i: that said, if we can get the function work and don't break anything, with same performance, then you got my +2 | 00:53 |
| flwang1 | Eva-i: ryansb: vkmc: flaper87: i'm reviewing our topics and we need to vote to fill our slots | 00:54 |
| Eva-i | flwang1: okay, I will upload another proof of concept. It will 99% not affect the performance. And be compatible like our previous change that added queue name to notifications. Sadly, I don't know how to measure performance loss for the case of notifications. I came up with something before, but I don't think my test was correct. | 01:05 |
| *** wxy has joined #openstack-zaqar | 01:20 | |
| *** Qiming has joined #openstack-zaqar | 01:30 | |
| *** Qiming has quit IRC | 02:03 | |
| *** Qiming has joined #openstack-zaqar | 02:03 | |
| *** Kevin_Zheng has joined #openstack-zaqar | 02:24 | |
| *** achanda has joined #openstack-zaqar | 02:34 | |
| flwang1 | Eva-i: add a test case for rally | 02:48 |
| flwang1 | Eva-i: can you help review this one https://review.openstack.org/#/c/287114/ ? | 02:50 |
| *** stanchan has quit IRC | 03:06 | |
| *** achanda has quit IRC | 03:27 | |
| *** tqtran has joined #openstack-zaqar | 03:49 | |
| *** tqtran has quit IRC | 04:02 | |
| *** achanda has joined #openstack-zaqar | 05:00 | |
| *** flwang has quit IRC | 05:03 | |
| *** GB21 has joined #openstack-zaqar | 05:14 | |
| *** exploreshaifali has joined #openstack-zaqar | 05:41 | |
| *** rcernin has joined #openstack-zaqar | 05:59 | |
| *** mdnadeem has quit IRC | 06:14 | |
| *** mdnadeem has joined #openstack-zaqar | 06:21 | |
| *** mdnadeem_ has joined #openstack-zaqar | 06:24 | |
| *** achanda has quit IRC | 06:47 | |
| *** achanda has joined #openstack-zaqar | 06:53 | |
| *** achanda has quit IRC | 07:00 | |
| *** openstackgerrit has quit IRC | 07:02 | |
| *** openstackgerrit has joined #openstack-zaqar | 07:03 | |
| *** pcaruana has joined #openstack-zaqar | 07:06 | |
| *** rcernin has quit IRC | 07:08 | |
| *** rcernin has joined #openstack-zaqar | 07:09 | |
| *** exploreshaifali has quit IRC | 07:24 | |
| *** gcb has quit IRC | 07:41 | |
| *** gcb has joined #openstack-zaqar | 07:53 | |
| *** GB21 has quit IRC | 07:58 | |
| *** achanda has joined #openstack-zaqar | 08:00 | |
| *** achanda has quit IRC | 08:05 | |
| *** exploreshaifali has joined #openstack-zaqar | 08:18 | |
| *** exploreshaifali has quit IRC | 08:40 | |
| *** exploreshaifali has joined #openstack-zaqar | 08:53 | |
| *** GB21 has joined #openstack-zaqar | 09:18 | |
| *** achanda has joined #openstack-zaqar | 09:33 | |
| *** achanda has quit IRC | 09:38 | |
| *** mdnadeem_ has quit IRC | 09:57 | |
| *** mdnadeem_ has joined #openstack-zaqar | 09:58 | |
| *** Qiming has quit IRC | 10:04 | |
| *** GB21 has quit IRC | 10:05 | |
| -openstackstatus- NOTICE: npm lint jobs are failing due to a problem with npm registry. The problem is under investigation, and we will update once the issue is solved. | 10:20 | |
| *** ChanServ changes topic to "npm lint jobs are failing due to a problem with npm registry. The problem is under investigation, and we will update once the issue is solved." | 10:20 | |
| *** mdnadeem has quit IRC | 10:32 | |
| *** mdnadeem has joined #openstack-zaqar | 10:33 | |
| *** achanda has joined #openstack-zaqar | 10:35 | |
| *** gcb has quit IRC | 10:37 | |
| *** gcb has joined #openstack-zaqar | 10:38 | |
| *** achanda has quit IRC | 10:41 | |
| *** gcb has quit IRC | 10:43 | |
| *** gcb has joined #openstack-zaqar | 10:44 | |
| *** tqtran has joined #openstack-zaqar | 10:56 | |
| *** GB21 has joined #openstack-zaqar | 10:57 | |
| *** GB21 has quit IRC | 10:59 | |
| *** exploreshaifali has quit IRC | 11:00 | |
| *** tqtran has quit IRC | 11:01 | |
| *** Qiming has joined #openstack-zaqar | 11:24 | |
| *** shu-mutou has quit IRC | 11:37 | |
| *** exploreshaifali has joined #openstack-zaqar | 11:57 | |
| *** itisha has joined #openstack-zaqar | 12:18 | |
| *** pt_15 has joined #openstack-zaqar | 12:24 | |
| *** Qiming_ has joined #openstack-zaqar | 12:27 | |
| *** Qiming has quit IRC | 12:28 | |
| *** achanda has joined #openstack-zaqar | 12:38 | |
| *** achanda has quit IRC | 12:43 | |
| *** exploreshaifali has quit IRC | 13:13 | |
| *** amitgandhinz has joined #openstack-zaqar | 13:13 | |
| *** ametts has joined #openstack-zaqar | 13:35 | |
| *** kgriffs|afk is now known as kgriffs | 13:55 | |
| *** achanda has joined #openstack-zaqar | 14:26 | |
| *** mpanetta has joined #openstack-zaqar | 14:28 | |
| *** flwang1 has quit IRC | 14:35 | |
| *** GB21 has joined #openstack-zaqar | 14:46 | |
| *** kgriffs is now known as kgriffs|afk | 14:47 | |
| *** flwang has joined #openstack-zaqar | 14:49 | |
| *** achanda has quit IRC | 14:53 | |
| *** david_cu has joined #openstack-zaqar | 14:54 | |
| *** Qiming_ has quit IRC | 15:03 | |
| *** ametts has quit IRC | 15:36 | |
| *** kgriffs|afk is now known as kgriffs | 15:41 | |
| *** kgriffs is now known as kgriffs|afk | 15:42 | |
| *** jtomasek has quit IRC | 16:33 | |
| *** amitgandhinz has quit IRC | 16:44 | |
| *** achanda has joined #openstack-zaqar | 16:53 | |
| *** david_cu_ has joined #openstack-zaqar | 16:54 | |
| *** david_cu has quit IRC | 16:55 | |
| *** davideagnello has joined #openstack-zaqar | 16:59 | |
| openstackgerrit | Ronald Bradford proposed openstack/zaqar: Define context.roles with base class https://review.openstack.org/302370 | 17:10 |
| *** jtomasek has joined #openstack-zaqar | 17:14 | |
| *** achanda has quit IRC | 17:27 | |
| *** tqtran has joined #openstack-zaqar | 17:32 | |
| *** amitgandhinz has joined #openstack-zaqar | 17:42 | |
| *** david-lyle has quit IRC | 17:44 | |
| *** rcernin has quit IRC | 17:56 | |
| *** achanda has joined #openstack-zaqar | 17:59 | |
| *** david-lyle has joined #openstack-zaqar | 18:02 | |
| *** GB21 has quit IRC | 18:32 | |
| *** flwang1 has joined #openstack-zaqar | 18:47 | |
| Eva-i | flwang: ryansb: flwang1: is it okay to merge a patch like this which doens't make a difference now, but will make a difference after other patch: https://review.openstack.org/#/c/287114. | 19:00 |
| Eva-i | ? | 19:00 |
| Eva-i | see my last comment | 19:02 |
| *** flwang1 has quit IRC | 19:04 | |
| *** flwang1 has joined #openstack-zaqar | 19:08 | |
| *** flwang1 has quit IRC | 19:21 | |
| *** ametts has joined #openstack-zaqar | 19:31 | |
| *** kgriffs|afk is now known as kgriffs | 19:36 | |
| ryansb | are you targeting these for mitaka? | 20:02 |
| Eva-i | ryansb: I think so. I think we better backport both these patches. | 20:03 |
| ryansb | hm, alright. Would like to get flwang's opinion on backporting, but I'd be OK with it | 20:06 |
| Eva-i | ryansb: aha, oki. | 20:06 |
| *** rcernin has joined #openstack-zaqar | 20:09 | |
| Eva-i | ryansb: can you please see my last comment here: https://review.openstack.org/#/c/294368/. Which options would you like to be implemented: 1 or 2 or both? Personally I would like just option 1. | 20:12 |
| Eva-i | ryansb: ah, and also paste mentioned in option 1 is this: http://paste.openstack.org/show/K3KB7gIrmwJPRer0yVUC/ | 20:13 |
| *** pcaruana has quit IRC | 20:14 | |
| ryansb | sure | 20:36 |
| ryansb | Eva-i: our update semantics right now replace the whole dict, right? | 20:43 |
| ryansb | so for a user, an update command wouldn't change much | 20:43 |
| Eva-i | ryansb: yes | 20:45 |
| ryansb | so in that case, option 1 sounds good to me as well | 20:46 |
| Eva-i | ryansb: this is option 1. | 20:46 |
| Eva-i | ryansb: okay, thanks! I think I should also get opinions of flwang and flaper87 about this. | 20:47 |
| ryansb | ok :) thanks for working on this | 20:48 |
| *** flwang1 has joined #openstack-zaqar | 21:18 | |
| *** rcernin has quit IRC | 21:27 | |
| *** david_cu_ has quit IRC | 21:29 | |
| *** david_cu has joined #openstack-zaqar | 21:30 | |
| flwang | Eva-i: ping | 21:34 |
| Eva-i | ryansb: so, can we merge this patch now to make wanghao happy? https://review.openstack.org/#/c/287114. I see you only marked it +2. | 21:34 |
| *** david_cu has quit IRC | 21:34 | |
| Eva-i | flwang: I'm here | 21:34 |
| flwang | (08:03:37) Eva-i: ryansb: I think so. I think we better backport both these patches. | 21:34 |
| flwang | what patches you were talking about for backporting? | 21:34 |
| Eva-i | flwang: this one: https://review.openstack.org/#/c/294368/. And this one: https://review.openstack.org/#/c/287114. | 21:35 |
| Eva-i | flwang: because now it's impossible for the user to update metadata in v2. | 21:36 |
| flwang | both of them are client patches | 21:36 |
| *** amitgandhinz has quit IRC | 21:37 | |
| Eva-i | flwang: so they don't require backporting? | 21:37 |
| Eva-i | flwang: they have only master version? | 21:37 |
| flwang | Eva-i: technically yes, we just need release another client version | 21:37 |
| Eva-i | flwang: oki | 21:37 |
| flwang | that's what we're going to do to fix the metadata update issue | 21:37 |
| Eva-i | flwang: today I asked Ryan for his opinion on how we should implement this patch, you can see it in log. What is your opinion, which way would you like? | 21:39 |
| flwang | Eva-i: i think i have already mentioned my idea. I prefer to keep the set metadata instead of adding new commands | 21:40 |
| Eva-i | flwang: oh, time passed and I forgot your opinion. So three core developers think the same. It's decided then? | 21:42 |
| flwang | what's the drawback of #1? | 21:42 |
| Eva-i | flwang1: drawbacks... Maybe that we don't give the user same experience in python-zaqarclient like in our new fancy update metadata method in server's API v2. Also the code is less resistant to race conditions, i.e. 2 clients decide to change queue metadata at the same time. | 21:50 |
| flwang | Eva-i: ok, personally, several new commands are not good for me since we're using openstack client, there are so many commands already | 21:54 |
| flwang | and personally, i don't want to add too many commands for zaqar, especially for one endpoint | 21:55 |
| Eva-i | About race condition. For example we have queue metadata with some keys. Client 1 wants to change key1's value. Client 2 wants to change key2's value. At the same time. So In option 1 we don't send to the server just a change to replace key, but a whole metadata dict. So in the end only one key will be changed (key1 or key2). | 21:56 |
| flwang | as for race conditions, even though we go for the #2 option, we still can't guarantee much, right? | 21:56 |
| Eva-i | flwang: yeah, I agree with you with everything. | 21:56 |
| flwang | Eva-i: ok, so let's go for #1 | 21:56 |
| * flwang is happy to have Eva-i as our guard | 21:58 | |
| *** mpanetta has quit IRC | 21:59 | |
| Eva-i | =) | 21:59 |
| flwang | Eva-i: ryansb: we need to vote our session topics | 22:00 |
| flwang | vkmc: flaper87: wxy: wanghao: ^ | 22:00 |
| Eva-i | flwang: vote here https://etherpad.openstack.org/p/newton-zaqar-summit-planning, right? | 22:01 |
| flwang | yes, thanks for the link | 22:01 |
| *** ametts has quit IRC | 22:21 | |
| *** achanda has quit IRC | 22:32 | |
| *** achanda has joined #openstack-zaqar | 22:38 | |
| *** gcb has quit IRC | 22:45 | |
| *** gcb has joined #openstack-zaqar | 22:50 | |
| Eva-i | flwang: I see now the new nova way for api-ref(there's also link to rendered nova api-ref): https://etherpad.openstack.org/p/api-site-in-rst | 22:56 |
| flwang | Eva-i: is it the way we can follow? | 22:57 |
| Eva-i | flwang: From the api-ref I see that two approaches are welcomed: 1. RST+swagger approach 2. Full swagger approach. | 22:58 |
| Eva-i | flwang: nova and other main openstack projects decided to go with approach №1, because some things can't be implemented in full swagger. But the output api-ref is a bit poor because of it. | 22:59 |
| Eva-i | *some their things | 22:59 |
| Eva-i | flwang: I think for Zaqar full swagger approach is better solution. This is a quote from the etherpad: "The rest of OpenStack APIs that do not have multiple POST /actions resources and are more RESTful and have no plans to support API microversions can certainly handle a swagger conversion and does not have the difficulty nor complexity that the Compute, Network, Volume, Database, Baremetal, Shared Filesystem APIs offer." | 23:00 |
| flwang | Eva-i: for the full swagger way, what we should do? | 23:02 |
| *** AAzza has quit IRC | 23:06 | |
| *** AAzza has joined #openstack-zaqar | 23:06 | |
| Eva-i | flwang: I don't know now. Magnum has swagger files here: https://github.com/jamiehannaford/swagger-magnum. But I don't see that these files are used for magnum api-ref. (And will they be used?) | 23:07 |
| Eva-i | flwang: hm, maybe 1st approach is just RST approach, without swagger. I misunderstood, because I found json file in nova's api-ref: https://github.com/openstack/nova/tree/master/api-ref/source | 23:09 |
| Eva-i | Maybe I should ask magnum team what are they going to do with their swagger api-ref. | 23:10 |
| Eva-i | flwang: ryansb: okay, thanks for today. I'm going to sleep. | 23:11 |
| *** gcb has quit IRC | 23:11 | |
| *** AAzza has quit IRC | 23:11 | |
| *** gcb has joined #openstack-zaqar | 23:11 | |
| Eva-i | Forgot to say: that's how api-ref with RST approach renders: http://docs-draft.openstack.org/71/298671/1/check/gate-nova-docs/c9e7f66//doc/build/html/rest_api/. That's how api-ref with Swagger approach renders (on the right):http://editor.swagger.io/#/ | 23:13 |
| *** AAzza has joined #openstack-zaqar | 23:14 | |
| Eva-i | I think swagger approach is nicer. | 23:15 |
| *** achanda has quit IRC | 23:15 | |
| *** Qiming has joined #openstack-zaqar | 23:17 | |
| flwang | Eva-i: yep, now the questions are : 1. which way to go 2. how to publish it on docs.openstack.org | 23:17 |
| *** itisha has quit IRC | 23:19 | |
| *** achanda has joined #openstack-zaqar | 23:22 | |
| *** gcb has quit IRC | 23:33 | |
| *** gcb has joined #openstack-zaqar | 23:34 | |
| *** pt_15 has quit IRC | 23:35 | |
| openstackgerrit | Fei Long Wang proposed openstack/zaqar: Support Tempest test as gate job https://review.openstack.org/302479 | 23:55 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!