| *** openstackgerrit_ has quit IRC | 00:23 | |
| *** openstackgerrit_ has joined #openstack-zaqar | 00:24 | |
| flwang | Eva-i: for the py34 issue, did you figure out a good way to fix it? | 00:28 |
|---|---|---|
| Eva-i | flwang: no, still not | 00:29 |
| flwang | Eva-i: ok | 00:29 |
| Eva-i | flwang: no one answered me in #openstack-oslo. I guess we have to instantiate conf from a scratch. But I'm going to inspect the problem further. | 00:30 |
| Eva-i | flwang: I was asking there a simple way to copy 'ConfigOpts' object | 00:31 |
| flwang | Eva-i: i did some investigation yesterday based on the question from ryansb | 00:31 |
| Eva-i | flwang: we may partially revert this patch https://github.com/openstack/zaqar/commit/188632806c2fdb6e9306708484c71153a49fd248 | 00:31 |
| flwang | Eva-i: i would suggest review the recent changes about that part | 00:31 |
| flwang | Eva-i: not revert | 00:32 |
| flwang | but rewrite a bit | 00:32 |
| Eva-i | flwang: that's why I said "partially" | 00:32 |
| * Eva-i still writing functional tests | 00:35 | |
| flwang | :) | 00:35 |
| flwang | Eva-i: when you're free, we can discuss about the plan of Newton :) | 00:36 |
| Eva-i | flwang: sure | 00:36 |
| *** openstackgerrit_ has quit IRC | 00:48 | |
| *** openstackgerrit_ has joined #openstack-zaqar | 00:49 | |
| flwang | ryansb: ping | 00:53 |
| openstackgerrit | Merged openstack/zaqar: Forbid update subscription to duplicate https://review.openstack.org/279946 | 02:01 |
| *** kgriffs is now known as kgriffs|afk | 02:31 | |
| *** tqtran-afk has quit IRC | 02:45 | |
| *** mdnadeem_ has joined #openstack-zaqar | 03:21 | |
| *** barra204 has joined #openstack-zaqar | 03:21 | |
| *** shakamunyi has quit IRC | 03:30 | |
| *** mdnadeem has quit IRC | 03:30 | |
| openstackgerrit | wangxiyuan proposed openstack/python-zaqarclient: Fix the get/set queue metadata https://review.openstack.org/289179 | 03:40 |
| *** tqtran-afk has joined #openstack-zaqar | 03:42 | |
| Eva-i | flwang: I created two functional tests. I tried also to speed up unit tests, but realized that it's not so easy. I can't imagine how to mock database things, maybe I will come up with something later. I guess for now functional tests will be enough. | 03:42 |
| *** tqtran-afk has quit IRC | 03:46 | |
| flwang | ok, i will happy to review them | 03:50 |
| *** shu-mutou-AFK is now known as shu-mutou | 04:02 | |
| *** wxy has joined #openstack-zaqar | 04:08 | |
| openstackgerrit | Eva Balycheva proposed openstack/zaqar: Renew subscription on update with TTL https://review.openstack.org/289665 | 04:09 |
| Eva-i | flwang: oki, uploaded new patch set ^^ I feel a bit awkward about adding such slow tests. =/ | 04:15 |
| shu-mutou | Eva-i, flwang: I could not find command for creating subscription. Let me know that please. | 04:15 |
| Eva-i | shu-mutou: you use python-zaqarclient as library, right? | 04:17 |
| shu-mutou | Eva-i: I use as library for UI development. but I want to use the command on devstack. | 04:18 |
| shu-mutou | Eva-i: for Zaqar-UI development, I want to confirm how python-zaqarclient's works. | 04:22 |
| Eva-i | shu-mutou: you probably mean requests by using curl program, right? Or CLI python-zaqarclient's commands? | 04:25 |
| Eva-i | shu-mutou: I can share my curl requests | 04:25 |
| shu-mutou | Eva-i: I'd like to know in CLI. python-zaqarclient's comands. | 04:26 |
| shu-mutou | Eva-i: if command doesn't exists, please share requests by curl. | 04:28 |
| Eva-i | shu-mutou: alright, run "openstack help --os-queues-api-version=2" and you'll see all available "subscription ..." commands ;) | 04:28 |
| shu-mutou | Ah!! I found it! Thanks! (/･ω･)/ | 04:29 |
| Eva-i | shu-mutou: no problem ^‿^v | 04:36 |
| *** tqtran-afk has joined #openstack-zaqar | 04:58 | |
| *** tqtran-afk has quit IRC | 05:02 | |
| *** GB21 has joined #openstack-zaqar | 05:25 | |
| *** GB21 has quit IRC | 06:00 | |
| *** GB21 has joined #openstack-zaqar | 06:02 | |
| *** barra204 has quit IRC | 06:04 | |
| *** openstackgerrit_ has quit IRC | 06:17 | |
| *** openstackgerrit_ has joined #openstack-zaqar | 06:17 | |
| *** shu-mutou is now known as shu-mutou-AFK | 06:46 | |
| *** rcernin has joined #openstack-zaqar | 06:49 | |
| *** GB21 has quit IRC | 07:18 | |
| *** GB21 has joined #openstack-zaqar | 07:18 | |
| *** jtomasek has joined #openstack-zaqar | 07:27 | |
| *** pcaruana has joined #openstack-zaqar | 08:41 | |
| *** GB21 has quit IRC | 08:48 | |
| *** GB21 has joined #openstack-zaqar | 08:48 | |
| *** GB21 has quit IRC | 09:06 | |
| *** GB21 has joined #openstack-zaqar | 09:33 | |
| *** mpanetta has joined #openstack-zaqar | 09:37 | |
| *** mpanetta has quit IRC | 10:07 | |
| openstackgerrit | Neerja Narayan proposed openstack/zaqar: Adding Negative tests cases to Zaqar v1.0 https://review.openstack.org/291643 | 10:15 |
| -openstackstatus- NOTICE: Gerrit is going to be restarted due to bad performance | 10:55 | |
| *** ChanServ changes topic to "Gerrit is going to be restarted due to bad performance" | 10:55 | |
| *** ChanServ changes topic to "OpenStack Queuing and Notification Service || Smile :D || Meetings every Monday @ 21:00 UTC/15:00 UTC || Wiki: https://wiki.openstack.org/wiki/Zaqar || Paste: http://paste.openstack.org/ || Send messages and make some noise :D" | 11:01 | |
| -openstackstatus- NOTICE: Gerrit has been restarted successfully | 11:01 | |
| Eva-i | flwang: hello | 11:26 |
| ryansb | good morning | 11:28 |
| ryansb | Eva-i: I'm looking at that dynamic_conf issue - I'm planning to write code that loops over the conf object and adds those to a new one | 11:28 |
| ryansb | in the Oslo channel they said the ConfigOpts object isn't designed to be copied, and we should instantiate a new one instead | 11:29 |
| Eva-i | ryansb: aha, I see. But how are you going to extract all options before adding them to a new conf? You will iterate over attributes? :o | 11:30 |
| ryansb | Still working on that bit | 11:30 |
| Eva-i | ryansb: you may take the bug assigned to me, I don't mind ;) | 11:31 |
| ryansb | the easiest would be if, instead of passing a config object, the caller could pass parameters | 11:31 |
| Eva-i | ryansb: you can assign yourself, I mean | 11:31 |
| *** tesseract has joined #openstack-zaqar | 11:33 | |
| Eva-i | ryansb: I don't know fully how oslo_config works. Maybe a patch like this was made only for convenience: https://github.com/openstack/zaqar/commit/188632806c2fdb6e9306708484c71153a49fd248. And there was nothing wrong with instantiating config from a scratch. | 11:33 |
| *** jtomasek has quit IRC | 11:34 | |
| *** tesseract has quit IRC | 11:34 | |
| *** Guest16009 has joined #openstack-zaqar | 11:34 | |
| ryansb | yeah, the commit doesn't explicitly mention a problem | 11:34 |
| flaper87 | Eva-i: TBH, I hate the dynamic_conf think | 11:35 |
| flaper87 | https://bugs.launchpad.net/zaqar/+bug/1372335 | 11:35 |
| openstack | Launchpad bug 1372335 in zaqar "Zaqar does not enforce writes to majority" [Critical,Fix released] - Assigned to Flavio Percoco (flaper87) | 11:35 |
| flaper87 | that was the problem | 11:35 |
| flaper87 | When you create a new pool, it won't use the global config but a dynamic instance of ConfigOpts | 11:36 |
| flaper87 | we need to inherit some of the configs in the Zaqar's main ConfigOpts object | 11:36 |
| flaper87 | so that we don't skip some of the system-wide values | 11:36 |
| flaper87 | dynamic_conf is one of the things I've always wanted to get rid off | 11:36 |
| flaper87 | but it was just not possible at the time | 11:37 |
| flaper87 | hope that helps | 11:37 |
| *** Guest16009 has quit IRC | 11:37 | |
| ryansb | hm. You don't happen to know what values, do you? | 11:38 |
| flaper87 | ryansb: I can dig into it. It'll likely come back to my head | 11:38 |
| flaper87 | ryansb: Eva-i what problems do you have? | 11:38 |
| ryansb | sometimes the copy raises an exception because the ConfigOpt object isn't designed to be copied | 11:39 |
| ryansb | (in Python 3) | 11:39 |
| *** jtomasek has joined #openstack-zaqar | 11:39 | |
| *** tesseract- has joined #openstack-zaqar | 11:40 | |
| flaper87 | oh, mmh. Interesting | 11:40 |
| flaper87 | one of the reasons copying was easier is that it already has all the options registered | 11:40 |
| Eva-i | flaper87: hello, thank you for the information. I still don't understand few things. Let me subscribe you to this bug, it's hidden now. | 11:40 |
| flaper87 | but that was even before we had list_opts, I think | 11:40 |
| ryansb | I see, so if we could iterate over the options and save them to the newly-instantiated object, would that cover the needs copy currently handles? | 11:41 |
| flaper87 | ryansb: yes, that would, I think. | 11:42 |
| flaper87 | ConfigOpts can be converted to a dict | 11:42 |
| flaper87 | but that has nested ConfigGroups | 11:42 |
| flaper87 | and who knows what else | 11:43 |
| ryansb | hrm, I'll test that out | 11:43 |
| ryansb | I see we have the dict_to_conf util | 11:43 |
| flaper87 | Eva-i: FWIW, some contents of that bug have been posted in a public resource. That makes it disclosed :( | 11:45 |
| Eva-i | flwang: do you mean paste.openstack.org? | 11:45 |
| flaper87 | Eva-i: which I avoided to mention in the channel on purpose :P | 11:46 |
| flaper87 | Eva-i: but yes, that | 11:46 |
| Eva-i | flwang: when I posted information there, I clicked on "private" checkbox. So it's not really private? =/ | 11:46 |
| Eva-i | "Paste private" | 11:47 |
| flaper87 | Eva-i: (you're mentioning flwang not flaper87 :P) | 11:48 |
| flaper87 | Eva-i: oh mmh, maybe you're right | 11:48 |
| flaper87 | oh, and there's a dropbox link | 11:48 |
| flaper87 | mmh | 11:48 |
| * ryansb gets breakfast, be back shortly | 11:48 | |
| flaper87 | Eva-i: we'll worry about that later. Let's try to fix this bug | 11:49 |
| Eva-i | flaper87: the dropbox link is also "private". It may be visible to dropbox admins, but personally I'm not sure if the information posted there is usable without context. | 11:50 |
| flaper87 | Eva-i: it's always better to keep everything in one place and undisclosed. | 11:51 |
| Eva-i | flaper87: yes, my bad. | 11:51 |
| flaper87 | Eva-i: no worries. Not big deal in this case. Let's focus on the bug | 11:52 |
| flaper87 | :) | 11:52 |
| Eva-i | flaper87: next time I'll make things right | 11:52 |
| Eva-i | flaper87: yeah | 11:52 |
| Eva-i | flaper87: hm. Few months ago I noticed this thing: we instantiate Zaqar by global ConfigOpts object. https://github.com/openstack/zaqar/blob/master/zaqar/transport/wsgi/app.py#L39. Then we register all options to this ConfigOpts object. Does it make passing 'conf' object to constructors pointless? Does it also make copying 'conf' in dynamic_conf pointless? | 12:10 |
| flaper87 | no, it's not pointless. The only reason we're *forced* to use the global config object is because other oslo libraries use a global config object | 12:11 |
| flaper87 | Otherwise, we'd be using a local one everywhere | 12:11 |
| flaper87 | that's why it was all designed as if there was no global config object and we don't use it anywhere | 12:12 |
| openstackgerrit | Neerja Narayan proposed openstack/zaqar: Adding Negative tests cases to Zaqar v1.0 https://review.openstack.org/291643 | 12:18 |
| Eva-i | flwang: seems like it was designed for passing some local object as 'conf' and I like this design. But the reference to the global config is always accessible from "oslo_config.cfg.CONF". Maybe we should use two configs: one local for zaqar things only (and copies of it for zaqar's pools), and the global one for oslo libraries? | 12:21 |
| Eva-i | flaper87: ^ | 12:22 |
| Eva-i | flaper87: oh.. I need to sleep, I make mistakes and maybe not thinking right. | 12:22 |
| *** jtomasek has quit IRC | 12:25 | |
| flaper87 | Eva-i: no no, the design is right. The only place where the global config object is accessed is in that file. (ok, maybe there are two placeS). We just use it to create the bootstrap instance | 12:26 |
| flaper87 | then it all goes through local access | 12:26 |
| *** jtomasek has joined #openstack-zaqar | 12:27 | |
| *** GB21 has quit IRC | 12:28 | |
| Eva-i | flaper87: hm. I'll try to check "tomorrow" if oslo_config.cfg.CONF from anywhere does not reference the same 'conf' that we pass in constructors. | 12:31 |
| Eva-i | flaper87: thanks for your answers | 12:31 |
| *** mdnadeem has joined #openstack-zaqar | 12:46 | |
| *** mdnadeem_ has quit IRC | 12:48 | |
| *** itisha has joined #openstack-zaqar | 13:01 | |
| *** akanksha_ has joined #openstack-zaqar | 13:52 | |
| *** amitgandhinz has joined #openstack-zaqar | 14:44 | |
| *** JAHoagie has joined #openstack-zaqar | 16:49 | |
| *** pcaruana has quit IRC | 16:50 | |
| *** david_cu has joined #openstack-zaqar | 16:51 | |
| *** tesseract- has quit IRC | 16:55 | |
| *** rcernin has quit IRC | 17:01 | |
| *** pcaruana has joined #openstack-zaqar | 17:36 | |
| *** david_cu has quit IRC | 18:01 | |
| *** wanghao has quit IRC | 18:42 | |
| *** wanghao has joined #openstack-zaqar | 18:43 | |
| *** david_cu has joined #openstack-zaqar | 19:02 | |
| *** tqtran-afk has joined #openstack-zaqar | 19:06 | |
| *** david_cu has quit IRC | 19:06 | |
| *** tqtran-afk is now known as tqtran | 19:11 | |
| *** david_cu has joined #openstack-zaqar | 19:26 | |
| *** ksheedlo has quit IRC | 20:03 | |
| *** wanghao has quit IRC | 20:24 | |
| *** itisha has quit IRC | 20:59 | |
| *** AAzza has quit IRC | 21:19 | |
| *** AAzza has joined #openstack-zaqar | 21:20 | |
| *** david_cu has quit IRC | 21:57 | |
| *** akanksha_ has quit IRC | 22:07 | |
| *** amitgandhinz has quit IRC | 22:15 | |
| *** david-lyle has quit IRC | 22:37 | |
| *** jtomasek has quit IRC | 22:38 | |
| *** rcernin has joined #openstack-zaqar | 22:47 | |
| *** rcernin has quit IRC | 23:06 | |
| *** rcernin has joined #openstack-zaqar | 23:07 | |
| *** rcernin has quit IRC | 23:31 | |
| *** david_cu_ has quit IRC | 23:41 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!