*** achanda has joined #openstack-zaqar | 00:16 | |
*** achanda_ has quit IRC | 00:19 | |
*** achanda has quit IRC | 00:21 | |
*** openstackgerrit has quit IRC | 00:22 | |
*** openstackgerrit has joined #openstack-zaqar | 00:22 | |
*** sriram has quit IRC | 00:33 | |
*** openstack has joined #openstack-zaqar | 00:36 | |
*** amitgandhinz has joined #openstack-zaqar | 00:38 | |
*** davideagnello has quit IRC | 01:18 | |
flwang | flaper87: ping | 01:34 |
---|---|---|
*** liusheng has quit IRC | 02:16 | |
*** achanda has joined #openstack-zaqar | 02:35 | |
*** exploreshaifali has quit IRC | 02:46 | |
*** exploreshaifali has joined #openstack-zaqar | 02:47 | |
*** cpallares has quit IRC | 02:55 | |
*** exploreshaifali has quit IRC | 03:30 | |
*** achanda has quit IRC | 03:41 | |
*** achanda has joined #openstack-zaqar | 04:17 | |
*** davideagnello has joined #openstack-zaqar | 04:19 | |
*** davideagnello has quit IRC | 04:24 | |
*** davideagnello has joined #openstack-zaqar | 04:33 | |
*** flaper87 has quit IRC | 05:01 | |
*** vkmc has quit IRC | 05:02 | |
*** flaper87 has joined #openstack-zaqar | 05:04 | |
*** vkmc has joined #openstack-zaqar | 05:08 | |
*** rsFF has quit IRC | 05:12 | |
*** rsFF has joined #openstack-zaqar | 05:12 | |
*** davideagnello has quit IRC | 05:19 | |
*** achanda has quit IRC | 05:52 | |
*** achanda has joined #openstack-zaqar | 05:57 | |
*** achanda has quit IRC | 06:03 | |
*** pcaruana has quit IRC | 07:11 | |
*** flaper87 has quit IRC | 07:23 | |
*** flaper87 has joined #openstack-zaqar | 07:23 | |
*** dynarro has joined #openstack-zaqar | 08:13 | |
*** flwang1 has joined #openstack-zaqar | 09:55 | |
flwang1 | flaper87: ping | 09:55 |
flwang1 | zigo: ping | 09:59 |
*** boris-42 has quit IRC | 10:05 | |
*** dynarro has quit IRC | 10:06 | |
*** boris-42 has joined #openstack-zaqar | 10:08 | |
*** flwang1 has quit IRC | 10:53 | |
*** dynarro has joined #openstack-zaqar | 11:50 | |
*** therve has joined #openstack-zaqar | 11:56 | |
therve | Heya | 11:56 |
flaper87 | flwang: pong | 12:02 |
flaper87 | therve: hey | 12:02 |
therve | I started to play a bit with zaqar. I have some questions about general usage, but first I noticed that queue.pop may not work. | 12:03 |
therve | I get "zaqarclient.transport.errors.MalformedRequest: The "ids" query parameter is required." when I try to use pop in the client library | 12:04 |
therve | Any idea? | 12:04 |
flaper87 | therve: mind sharing the code ? I'm working on the client now | 12:07 |
therve | flaper87, http://paste.openstack.org/show/204967/ | 12:09 |
therve | Using latest master AFAICT | 12:10 |
vkmc | hey! | 12:11 |
zigo | flaper87: pong | 12:31 |
vkmc | therve, I'm testing it :) | 12:33 |
therve | vkmc, Does the usage look correct to you? | 12:58 |
vkmc | therve, seems right mmh | 13:10 |
*** amitgandhinz has quit IRC | 13:11 | |
*** sriram has joined #openstack-zaqar | 13:14 | |
zigo | Zaqar is using python-autoban which has lots of non-free files. | 13:17 |
zigo | I'm talking about the example folder. | 13:17 |
zigo | For example: examples/twisted/websocket/echo_wsfallbacks/web/WebSocketMain.swf | 13:17 |
zigo | I hope this can be fixed. | 13:17 |
zigo | Also, adding Python3 support to autobhan needs quite some work, as dependencies don't have support for it (yet). | 13:18 |
vkmc | therve, it seems there is a bug server side, you shouldn't need ids | 13:18 |
zigo | Note: I already uploaded python-txaio to Sid. | 13:19 |
therve | vkmc, I think it should be using version 1.1 ? | 13:19 |
therve | Default is 1 | 13:20 |
flaper87 | zigo: autobahn is not a strong requirement, it's part of a feature that is considered experimental. Is it possible to avoid adding it as a dependency in the package? | 13:20 |
flaper87 | zigo: what would be the right process there? (sorry for my ignorance in the process) | 13:20 |
zigo | flaper87: Well, then remove it from requirements.txt ! | 13:20 |
zigo | flaper87: Is there some tests with it? | 13:20 |
flaper87 | we can move simply move it to test-requirements | 13:21 |
flaper87 | s/move simply/simply/ | 13:21 |
flaper87 | zigo: would that be enough for you? | 13:21 |
*** mpanetta has joined #openstack-zaqar | 13:21 | |
zigo | flaper87: No. | 13:21 |
zigo | I do run unit tests ... | 13:21 |
flaper87 | :( | 13:21 |
zigo | (at package build time) | 13:22 |
flaper87 | mmh, ok. | 13:22 |
flaper87 | right | 13:22 |
vkmc | zigo, would you suggest we change the websocket server library? | 13:22 |
vkmc | not sure if we have other options | 13:22 |
zigo | Why not using websockify? | 13:22 |
vkmc | because atm autobahn seemed a good fit for several reasons | 13:22 |
zigo | It's already in use by other OpenStack projects (like Nova) and it's already maintained. | 13:22 |
flaper87 | zigo: we'll take a better look at websockify | 13:23 |
flaper87 | but I think we did already | 13:23 |
zigo | Well, yeah, we have many of such things in OpenStack, and then we end up maintaining 3 times the amount of libs. | 13:23 |
therve | zigo, I don't think it serves the same purpose | 13:23 |
zigo | If autobhan is better, why not contributing to Nova to switch to it then? | 13:23 |
zigo | Ah, really? | 13:23 |
zigo | therve: Hi there! Do you contribute to Zaqar now? :) | 13:24 |
therve | zigo, Not there yet :) | 13:24 |
zigo | :) | 13:24 |
zigo | therve: Are you going to Vancouver? | 13:25 |
vkmc | zigo, I just took a look, and indeed, it doesn't serve the same purpose | 13:25 |
* flaper87 did the same | 13:25 | |
zigo | vkmc: Ok, fair enough then. | 13:25 |
therve | zigo, Yep | 13:25 |
flaper87 | websockitify translates from ws -> socket | 13:25 |
zigo | therve: Cool, see you there then! :) | 13:25 |
zigo | In such case, could you talk with upstream for autobhan, and ask that they remove the non-free files? | 13:26 |
flaper87 | zigo: that I can do | 13:26 |
vkmc | we could do that yes | 13:26 |
flaper87 | thanks for pointing it out, | 13:26 |
zigo | Let me list them, hang on... | 13:26 |
therve | Looked removed from here already: http://packages.ubuntu.com/trusty/all/python-autobahn/filelist | 13:26 |
zigo | therve: What you are pointing at is a binary, not the sources. | 13:27 |
zigo | Oh, is this already in Ubuntu? | 13:27 |
zigo | Interesting... | 13:27 |
zigo | Oh, it's already in Debian... | 13:27 |
zigo | I think the issue here is between the keyboard and chair: I can't spell german correctly! :) | 13:28 |
therve | Heh :) | 13:28 |
zigo | Ah no, we have version 0.5.14, and you guys declare to want 0.10.1... | 13:29 |
zigo | Outch, the package isn't team maintained, and the maintainer has this as a single package. | 13:29 |
zigo | (I mean, he doesn't maintain anything else...) | 13:29 |
vkmc | maybe we can ask together to autobahn folks to remove the non-free files | 13:30 |
zigo | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=780061 | 13:30 |
openstack | Debian bug 780061 in python-autobahn "Uploading python-autobahn >= 0.10.1 to Experimental" [Wishlist,Open] | 13:30 |
zigo | Please read on ... | 13:30 |
zigo | The current maintainer doesn't know he needs to package more than one package for maintaining autobahn. | 13:31 |
zigo | :/ | 13:31 |
openstackgerrit | Flavio Percoco proposed openstack/zaqar: Fix pymongo 3 https://review.openstack.org/175860 | 13:31 |
zigo | I'll upgrade the package to Sid and take it within the OpenStack packaging team then. | 13:31 |
vkmc | zigo, keep us on the loop | 13:32 |
vkmc | and let me know if I can help with something | 13:32 |
zigo | Sure. | 13:32 |
zigo | Let me list the non-free files. | 13:33 |
vkmc | k | 13:33 |
zigo | http://paste.debian.net/168036/ | 13:33 |
zigo | vkmc: swfobject was also in websockify, and finally got removed... | 13:35 |
zigo | I mean, the minified version was replaced. | 13:36 |
zigo | The flash object is a no-go. | 13:36 |
zigo | I can work-out a +dfsg version though, without the example folder. | 13:36 |
zigo | But that's really annoying... :( | 13:37 |
vkmc | therve, you were right, the client version was missing from the client init | 13:41 |
vkmc | therve, it should look like cli = client.Client(url="http://127.0.0.1:8888", version=1.1, conf=conf) | 13:41 |
vkmc | zigo, not really sure why they need flash | 13:41 |
therve | vkmc, I get "The header X-PROJECT-ID was missing" | 13:42 |
vkmc | therve, yes, you have to add another param to the conf | 13:42 |
vkmc | like... | 13:42 |
vkmc | https://github.com/openstack/python-zaqarclient/blob/master/examples/keystone_auth.py#L30-L39 | 13:42 |
vkmc | os_project_id is the one missing | 13:43 |
therve | vkmc, Aha, thanks | 13:43 |
therve | OK I get further, but still an error | 13:44 |
vkmc | therve, which error this time? | 13:44 |
therve | vkmc, 'NoneType' object has no attribute 'pop' | 13:45 |
therve | When I try to iterate on the result | 13:45 |
vkmc | ok then... that looks like a bug | 13:45 |
therve | message_pop doesn't return anything, probably the issue | 13:46 |
*** amitgandhinz has joined #openstack-zaqar | 13:46 | |
vkmc | I'll try to reproduce here | 13:46 |
*** cpallares has joined #openstack-zaqar | 13:49 | |
zigo | Oh, also, autobahn embedds a version of pbkdf2 | 13:50 |
zigo | This should also be fixed. | 13:50 |
therve | vkmc, I don't think 1.1 is well supported by the client | 13:50 |
therve | I get other errors even without pop | 13:51 |
zigo | Hum... probably already fixed in the latest version. | 13:51 |
vkmc | therve, certainly | 13:53 |
vkmc | therve, I'll take a look more in deep and debug what is going on | 13:54 |
therve | vkmc, Cool thanks | 13:54 |
vkmc | therve, if you have a moment, please submit the bug reports | 13:54 |
vkmc | therve, thanks for taking it for a spin | 13:54 |
therve | No problem, will do | 13:54 |
zigo | I uploaded autobahn 0.10.3 to Debian Experimental, it should be fine IMO. | 14:21 |
zigo | Could someone tell me what daemon needs to run with Zaqar? | 14:23 |
zigo | I got the impression that there's a zaqar-server that is in fact an API server, am I right? | 14:24 |
zigo | Then what? | 14:24 |
zigo | Could you describe how to setup zaqar? | 14:24 |
zigo | vkmc: flaper87: ^ | 14:24 |
flaper87 | zigo: that's correct. That's pretty much all you need | 14:24 |
flaper87 | Let me elaborate | 14:24 |
zigo | Thanks! | 14:26 |
flaper87 | zaqar-server is the api service (the only one, for now). A production deployment shouldn't run zaqar-server directly as it uses the python built-in wsgi container. Therefore, a production environment should pick gunicorn, uwsgi or whatever they prefer | 14:26 |
flaper87 | That last part is probably not relevant for you but I guess it's always good to know these things | 14:26 |
zigo | Sure! | 14:26 |
flaper87 | Other than that, zaqar needs a storage which can either be mongodb or redis | 14:26 |
zigo | Where's the WSGI definition file then? | 14:27 |
zigo | (so that I can put it in /usr/share/zaqar-common ...) | 14:27 |
zigo | Also, do I need mongodb service to run when running the unit tests? | 14:27 |
zigo | I'm getting lots of "Connection refused" errors when running the tests... | 14:28 |
zigo | (267 failures...) | 14:28 |
flaper87 | zigo: zaqar.transport.wsgi.app <- this module contains the wsgi app. | 14:28 |
flaper87 | zigo: unfortunately, yes. We're working on mocking the mongodb tests. We used to have sqlite for that but as of kilo, we removed support for sqlalchemy on the data plane. | 14:28 |
zigo | flaper87: Yeah, but could you give me a .wsgi file that I could put for the Apache config? | 14:29 |
flaper87 | mocking mongodb and redis* | 14:29 |
zigo | flaper87: There's no need to mock mongodb, I can run it before running the unit tests, I've done that in the past (for example, in the Ceilometer package). | 14:29 |
flaper87 | zigo: ah, I cna write one, yes. there's not such file in the code base now | 14:29 |
zigo | flaper87: Please do, that's helpful ! :) | 14:29 |
zigo | Like this, I could provide a zaqar-apache or something, which would configure apache directly. | 14:30 |
flaper87 | zigo: that's indeed helpful for deployers using apache | 14:30 |
* flaper87 gets to that | 14:30 | |
*** kgriffs|afk is now known as kgriffs | 14:31 | |
zigo | Have a look into Ceilometer's setup-test-env-mongodb.sh | 14:31 |
zigo | That's quite helpful. | 14:31 |
zigo | Then I do, in the Ceilometer package: bash -x ./setup-test-env-mongodb.sh testr run --subunit 'tests\.(?!.*test_bin.*)' | tee $$TEMP_REZ | subunit2pyunit | 14:31 |
flaper87 | I was precisely looking at that, they also have an apache dir under etc | 14:32 |
zigo | Which is wrong. | 14:34 |
zigo | You shouldn't define a new virtualhost for * | 14:34 |
zigo | You should *only* define an URL for /zaqar on the existing default Apache setup. | 14:35 |
*** achanda has joined #openstack-zaqar | 14:35 | |
zigo | In other words: do not hijack all of Apache please, it's probably used for other things... | 14:35 |
zigo | So, do: | 14:36 |
zigo | WSGIScriptAlias / [...] | 14:36 |
zigo | I mean, don't do that. Instead, do: | 14:36 |
zigo | WSGIScriptAlias /zaqar [...] | 14:36 |
openstackgerrit | Flavio Percoco proposed openstack/zaqar: Move devstack's Zaqar code into the codebase https://review.openstack.org/169305 | 14:36 |
flaper87 | vkmc: ^ pls :) | 14:36 |
vkmc | sure | 14:36 |
flaper87 | zigo: yes, I was just looking at it :D | 14:36 |
flaper87 | I'll add you to the review if you don't mind | 14:36 |
flaper87 | :) | 14:36 |
zigo | Sure, my pleasure ! :) | 14:39 |
zigo | I've imported the setup script of Ceilometer into my debian/ folder, in order to run MONGO. | 14:39 |
zigo | Is it ok to run only with Mongo, or is Redis also mandatory for unit tests? | 14:40 |
zigo | I have zero experience with redis... | 14:40 |
flaper87 | zigo: redis is not mandatory, either is good | 14:40 |
zigo | Cool. | 14:41 |
zigo | Hum.. I still get a shitload of connection refused... :( | 14:41 |
zigo | Where does Zaqar reads its connection info for the unit tests? | 14:42 |
zigo | In Ceilometer, we have: export CEILOMETER_TEST_MONGODB_URL="mongodb://localhost:${MONGO_PORT}/ceilometer" | 14:42 |
zigo | Could we have something like this in Zaqar? | 14:42 |
zigo | So that I would just export the URL, and Zaqar would use that ... | 14:43 |
*** achanda has quit IRC | 14:43 | |
flaper87 | zigo: it currently reads config files that are shipped in the codebase, those point to localhost:27017. We can definitely have an env var like that. We already have one to enable/disable some tests and to change the dir where the test config files are | 14:45 |
* flaper87 starts writing all these things down | 14:46 | |
flaper87 | zigo: thanks for taking the time and for your feedback | 14:46 |
therve | So as pop is not available on 1.0, what's the best way to achieve the same thing without it? | 14:51 |
therve | Does it involve claiming messages? | 14:51 |
flaper87 | therve: I'd really recommend using 1.1 | 14:52 |
flaper87 | 1.1 is 1.0 with few improvements | 14:52 |
therve | flaper87, Can I get a working client? :) | 14:52 |
flaper87 | therve: oh mmh, is it broken? | 14:53 |
flaper87 | ah the error you mentioned | 14:53 |
flaper87 | so, yeah, sorry about that. | 14:53 |
therve | flaper87, Another one bug #1446677 | 14:53 |
openstack | bug 1446677 in Python client library for Zaqar "Can't get a message from 1.1" [Undecided,New] https://launchpad.net/bugs/1446677 | 14:53 |
therve | Maybe it's something on my side though | 14:53 |
* flaper87 fixes all the things | 14:53 | |
therve | \o/ | 14:53 |
flaper87 | therve: TBH, the client needs some love, I'll focus on it in the next couple of days | 14:54 |
flaper87 | therve: do you have feedback w.r.t the API and structure? | 14:54 |
therve | flaper87, Could, would love to test it | 14:54 |
flaper87 | we've tried to keep the structure flexible enough to give access to the internals and/or just use it at a high level | 14:54 |
therve | flaper87, I don't have much just yet. Trying to implement a basic use-case in heat for now | 14:55 |
therve | pop seems to be what I want to do. I need to see how it behaves when messages aren't present. | 14:56 |
flaper87 | therve: you should be able to get all the info you need about the API from here: https://wiki.openstack.org/wiki/Zaqar/specs/api/v1.1 | 14:57 |
flaper87 | The client should respect that | 14:57 |
zigo | flaper87: What happens is that, under a build env, I'd run mongodb on a non-standard port manually on the shell, so it can't be on the standard port that the mongodb package uses, as it wouldn't bind if someone just builds without a pbuilder. | 15:03 |
zigo | Julien Danjou well knows this kind of stuff, as he's a Debian Developer himself... | 15:04 |
zigo | I'm trying with the standard port for now, but this will need to be fixed, somehow. | 15:04 |
zigo | What's the default db name? | 15:04 |
zigo | And do I need to create it before running the tests? | 15:05 |
zigo | Now I get a "connection closed" instead of "connection refused", so i'm guessing Zaqar doesn't do the db setup by itself in the unit tests. | 15:06 |
zigo | message_zaqar_test ? | 15:08 |
zigo | Is this the one? | 15:08 |
*** kgriffs is now known as kgriffs|afk | 15:22 | |
*** kgriffs|afk is now known as kgriffs | 15:30 | |
zigo | flaper87: vkmc: ^ | 15:46 |
*** cpallares has quit IRC | 15:57 | |
openstackgerrit | Doug Hellmann proposed openstack/python-zaqarclient: Update README to work with release tools https://review.openstack.org/175967 | 16:07 |
*** kgriffs is now known as kgriffs|afk | 16:14 | |
*** davideagnello has joined #openstack-zaqar | 16:20 | |
*** exploreshaifali has joined #openstack-zaqar | 16:30 | |
*** exploreshaifali has quit IRC | 16:40 | |
*** dynarro has quit IRC | 17:20 | |
*** achanda has joined #openstack-zaqar | 17:26 | |
*** kgriffs|afk is now known as kgriffs | 17:50 | |
*** kgriffs is now known as kgriffs|afk | 18:03 | |
*** kgriffs|afk is now known as kgriffs | 18:09 | |
*** achanda has quit IRC | 18:15 | |
*** achanda has joined #openstack-zaqar | 18:20 | |
*** openstackgerrit has quit IRC | 18:37 | |
*** openstackgerrit has joined #openstack-zaqar | 18:37 | |
*** davideagnello has quit IRC | 18:41 | |
*** davideagnello has joined #openstack-zaqar | 18:44 | |
*** kgriffs is now known as kgriffs|afk | 19:00 | |
*** kgriffs|afk is now known as kgriffs | 19:25 | |
*** achanda has quit IRC | 19:41 | |
*** ametts has quit IRC | 19:56 | |
*** kgriffs is now known as kgriffs|afk | 20:08 | |
*** kgriffs|afk is now known as kgriffs | 20:11 | |
*** achanda has joined #openstack-zaqar | 20:14 | |
*** openstackgerrit has quit IRC | 20:52 | |
*** openstackgerrit has joined #openstack-zaqar | 20:52 | |
*** exploreshaifali has joined #openstack-zaqar | 21:07 | |
*** mpanetta has quit IRC | 21:23 | |
flwang | flaper87: ping | 21:30 |
flwang | re the puppet zaqar work | 21:30 |
vkmc | flwang, o/ | 21:30 |
flwang | vkmc: i'm wondering if the work is still ongoing... https://github.com/jasontclark/puppet-zaqar | 21:31 |
flwang | we want to deploy zaqar, but seems we need some work to puppet zaqar | 21:32 |
vkmc | flwang, doesn't seem so | 21:32 |
vkmc | I haven't seen the one who was working on that for weeks now | 21:35 |
vkmc | IIRC there is a module for puppet in RDO | 21:35 |
flwang | vkmc: really? | 21:38 |
flwang | puppet for zaqar in RDO? | 21:38 |
flwang | is it opensource? | 21:38 |
flwang | where can I find it? | 21:38 |
vkmc | let me check, I might be wrong | 21:39 |
vkmc | https://github.com/redhat-openstack/openstack-puppet-modules indeed, I was wrong | 21:40 |
flwang | wow, it even includes gnocchi but not zaqar | 22:00 |
flwang | i can't believe my eyes | 22:00 |
flwang | btw, do you have any charts about the architecture of zaqar? | 22:01 |
vkmc | like... for a prez? | 22:02 |
*** kgriffs is now known as kgriffs|afk | 22:10 | |
flwang | yep | 22:14 |
flwang | i need to give a prez for my team to get some ideas how to adopt zaqar in our prod | 22:14 |
*** kgriffs|afk is now known as kgriffs | 22:17 | |
vkmc | I made one... and there are also some prior prez kgriffs, flaper87 and alej gave in the past | 22:18 |
*** amitgandhinz has quit IRC | 22:18 | |
vkmc | let me find it :) | 22:20 |
*** sriram has quit IRC | 22:31 | |
flwang | vkmc: it would be nice if you can share it :) | 23:23 |
flwang | flaper87: the jenkins of zaqar client is always failed | 23:38 |
*** exploreshaifali has quit IRC | 23:44 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!