Friday, 2015-05-29

« Thursday, 2015-05-28 Index Saturday, 2015-05-30 »
*** malini has joined #openstack-zaqar00:01
*** malini has quit IRC00:10
*** davideagnello has quit IRC00:11
*** kgriffs is now known as kgriffs|afk00:11
*** diga has joined #openstack-zaqar00:28
*** malini has joined #openstack-zaqar01:23
*** bradjones has quit IRC01:52
*** Yanyanhu has joined #openstack-zaqar01:54
*** bradjones has joined #openstack-zaqar01:56
*** csoukup has joined #openstack-zaqar02:01
*** exploreshaifali has quit IRC02:02
*** AAzza_ has joined #openstack-zaqar03:07
*** AAzza- has joined #openstack-zaqar03:08
*** davideagnello has joined #openstack-zaqar03:09
*** AAzza has quit IRC03:10
*** AAzza- is now known as AAzza03:10
*** barra204 has quit IRC03:11
*** AAzza_ has quit IRC03:11
*** barra204 has joined #openstack-zaqar03:13
*** davideagnello has quit IRC03:14
*** achanda has joined #openstack-zaqar03:53
*** malini has quit IRC04:01
*** achanda has quit IRC04:05
*** achanda has joined #openstack-zaqar04:06
*** achanda has quit IRC04:11
*** achanda has joined #openstack-zaqar05:05
*** dynarro has joined #openstack-zaqar06:02
openstackgerritOpenStack Proposal Bot proposed openstack/zaqar: Imported Translations from Transifex  https://review.openstack.org/18627106:03
*** csoukup has quit IRC06:03
*** dguerri`away is now known as dguerri06:41
*** dguerri is now known as dguerri`away06:46
*** dguerri`away is now known as dguerri07:07
*** dguerri is now known as dguerri`away07:08
*** achanda has quit IRC07:22
*** achanda has joined #openstack-zaqar07:44
*** dguerri`away is now known as dguerri08:04
*** achanda has quit IRC08:35
*** achanda has joined #openstack-zaqar08:48
*** dynarro has quit IRC08:58
*** achanda has quit IRC09:02
flaper87zigo: just to be super clear, functional tests are the ones failing for you. Right ?09:03
flaper87could you share the traceback again?09:03
flaper87:D09:03
*** Yanyanhu has quit IRC09:31
*** achanda has joined #openstack-zaqar10:28
*** achanda has quit IRC12:07
*** kgriffs|afk is now known as kgriffs12:53
zigoflaper87: No, unit tests.13:00
flaper87ok13:04
zigoflaper87: hang on, I'm doing yet another build and will give you the result "soon"13:04
zigoBut anyway, it's mostly: cannot connect to mongodb ...13:05
*** kgriffs is now known as kgriffs|afk13:11
zigoflaper87: This kind of trace: http://paste.debian.net/187255/13:12
zigoLoads of them...13:12
zigoOf course, because mongodb isn't binding on the port that the unit tests expect.13:12
*** sriram has joined #openstack-zaqar13:14
flaper87zigo: ok, patch is coming13:14
flaper87thanks13:14
*** achanda has joined #openstack-zaqar13:14
zigoGreat! :)13:14
zigoflaper87: Eventually, will zaqar-server split into zaqar-api and zaqar-somethingelse ?13:15
flaper87zigo: I don't think so. zaqar-server is zaqar-api13:16
flaper87We didn't called API because it's a "test" server13:16
flaper87people should use the wsgi app13:17
zigoflaper87: Sure, but then Zaqar only needs an API? No worker daemons?13:17
zigoflaper87: Would you recommand using uwsgi btw, like Barbican does?13:17
flaper87uwsgi/httpd/gunicorn... They all work fine. I know RAX uses uwisg though13:20
flaper87and no, it doens't need any daemons (yet)13:20
flaper87:P13:20
zigoOk.13:21
flaper87I really don't know if it'll in the future13:21
zigoI very much prefer uwsgi to a full apache, so that I can keep stuff with init sciprts and so on.13:21
zigothat's what I did with barbican, I'm thinking about doing the same for all OpenStack API servers that already support that.13:21
flaper87zigo: I also prefer uwsig over httpd13:23
*** cknight has joined #openstack-zaqar13:29
*** Yanyanhu has joined #openstack-zaqar13:39
*** Yanyanhu has quit IRC13:39
*** malini has joined #openstack-zaqar13:42
*** dynarro has joined #openstack-zaqar13:43
*** exploreshaifali has joined #openstack-zaqar14:09
*** achanda has quit IRC14:10
*** csoukup has joined #openstack-zaqar14:18
exploreshaifaliflaper87, around?14:18
flaper87exploreshaifali: heeeeeeey14:20
flaper87yes, I am here14:20
flaper87I was looking for you earlier today14:20
flaper87now I don't remember why14:20
flaper87exploreshaifali: what can I help you with?14:22
exploreshaifaliflaper87, lol that is nice that you don't remember now :P14:23
exploreshaifaliflaper87, so I want to know do you wnat me to write specs for tests stuff?14:23
flaper87i need to remember, I NEED TO REMEMBER14:23
flaper87I needed you to do some work I was supposed to do14:23
exploreshaifaliokay, what's that work?14:23
flaper87i don't remember :(14:24
exploreshaifaliso till the time your robot mind remind it again, lets discuss about refactoring about tests14:25
exploreshaifalido I need to write specs for it?14:25
flaper87yup14:25
exploreshaifaligreat!14:25
flaper87it requires a spec to make sure we cover the things we talked about at the summit14:25
flaper87https://www.youtube.com/watch?v=EYKdmo9Rg2s14:26
openstackgerritDoraly Navarro proposed openstack/zaqar-specs: Add support for missing features in zaqarclient v1.1  https://review.openstack.org/18595114:26
exploreshaifaliok, though I don't remember much what we discussed in the summit...but writing spec will make things clear for sure14:26
flaper87exploreshaifali: it should be in the etherpad14:26
flaper87no one remembers14:26
flaper87:P14:27
exploreshaifalihaha okay!14:27
exploreshaifaliI will get back to you soon than14:27
exploreshaifaliI bookmarked that video to watch it, I caught it from twitter14:28
exploreshaifali:P14:28
exploreshaifaliso first I need to watch that14:28
flaper87I'm waiting for the second one to be uploaded14:28
flaper87I gave 2 talks there and I really would like to watch the other one14:29
exploreshaifalioh it have second part too14:29
* flaper87 is a narcisist14:29
flaper87jokes apart14:29
flaper87I'm just curious to know how the second talk was14:29
exploreshaifalino no I can understand!14:30
exploreshaifaliyour narcissism14:30
exploreshaifali:P14:30
exploreshaifaliflaper87, please can you put some light on " test_insert_queue_with_metadata.tags = ['negative'] #dafuq"14:31
vkmc#dafuq lol14:31
kragnizlol14:31
vkmcexploreshaifali, hiiiiiiii!14:31
vkmckragniz, hiiiiiii!14:32
kragnizvkmc: hiii!14:32
exploreshaifalihiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiiii vkmc kragniz  :D14:32
* exploreshaifali was thinking hard to understand what #dafuq is, but now I got it after other devs reactions14:33
exploreshaifali:D14:33
flaper87LOL14:35
flaper87:D14:35
vkmcthis is hilarious14:35
flaper87sorry about that :P14:35
exploreshaifalilol14:36
exploreshaifaliI should be sorry to put it here :P14:36
flaper87In that case, #dafuq is a hash(flaper87) that is meant to say: What is this line being used for? Ask malini and see if she remembers and whether we actually need it14:36
malinidafuq?14:37
exploreshaifalilol14:37
flaper87malini: test_insert_queue_with_metadata.tags = ['negative']14:37
flaper87:D14:37
maliniu mean why we had the tags?14:38
flaper87is that a ddt thing? tsung ?14:38
flaper87yeah14:38
malinino..tht was to selectively run a few tests14:38
maliniso u cud just run the tagged test14:38
flaper87ahhh gotcha14:38
maliniI dont think we ever ended up using it14:38
maliniso shud be safe to get rid of14:38
flaper87malini: thanks for the info :)14:38
vkmc#notmoredafuq14:39
*** angvp has joined #openstack-zaqar14:39
flaper87next time, I expect you to be able to read "dafuq" and understand what I meant14:39
maliniI am the source of all useless info :D14:39
flaper87:D14:39
exploreshaifaliyes, I will understand it too.14:39
flaper87malini: no no, you are the reason why we have most of our tests so you should be proud of yourself14:39
vkmcmalini++14:39
maliniI missed the session abt test refactoring -  is tht going to be cleaned up?14:39
flaper87more like merged14:39
flaper87zaqar/tests and zaqar/zaqar/tests14:39
maliniit'll be interesting to give tempest lib a shot14:40
flaper87and well, the clean up will happen as part of that14:40
exploreshaifalimalini, https://etherpad.openstack.org/p/zaqar-liberty-testsuite14:40
vkmcexploreshaifali, you are going to work on the test refactoring, right?14:40
exploreshaifalivkmc, yes14:40
vkmcsweet14:40
exploreshaifali:D14:41
maliniexploreshaifali: ping me if you need help. I am good at telling other ppl what to do ;)14:41
exploreshaifalimalini, yes I will need guidance of few more intelligent people14:41
vkmcbtw, we have a new contributor lurking14:41
exploreshaifali:P14:41
vkmcangvp, say hi!14:41
flaper87angvp: dafuq ?14:42
vkmcdon't be shy14:42
exploreshaifaliangvp, hiiiiiiiiiii :)14:42
* angvp blushes14:42
* flaper87 saw the nicks list grow but had no iea who joined14:42
angvphi o/14:42
kragnizhi angvp!14:42
vkmcflaper87, I'm bringing hobos to the channel14:42
angvpsebasmagri: dafuq?!14:42
angvphobos? hodor!14:43
* kragniz offers angvp a sandwich14:43
flaper87angvp: sebasmagri has been lurking here for ages14:43
flaper87kragniz: dafuq ?14:43
flaper87you've NEVER offered me ANYTHING14:43
vkmcflaper87, now kragniz is feeding the hobos14:44
kragnizflaper87: don't let the hobos starve14:44
kragnizwhat sort of cruel man are you?14:44
flaper87the worst kind14:44
kragnizI thought as much14:44
flaper87Darwin would agree with me14:44
flaper87if only he was here :(14:44
* exploreshaifali offers pizza to flaper87 14:45
exploreshaifali:P14:45
flaper87exploreshaifali: thank you!14:45
* flaper87 eats it all in one bite14:45
exploreshaifali:D14:45
flaper87now, stop distracting me. I gotta get zigo's patch out there14:45
vkmcangvp, so welcome to -zaqar!14:46
angvp:) thanks for the warm welcome ;D14:47
*** mpanetta has joined #openstack-zaqar14:55
*** achanda has joined #openstack-zaqar15:00
flaper87zigo: around? it's actually not as straightforward as I'd have hoped15:03
flaper87zigo: crazy idea but don't kill me15:03
flaper87what if your patch instead of changing zaqar seds over `zaqar/tests/etc/` and replaces all uris ?15:04
flaper87:)15:04
flaper87At least until we refactor tests properly15:04
flaper87not even sure if that'll work for all the cases15:11
flaper87this thing that we require mongodb for the unittests is not good.... not good15:12
flaper87but anyway, you still need it for functional tests15:12
flaper87ok, I'll literally hack something15:16
*** exploreshaifali has quit IRC15:29
*** amitgandhinz has joined #openstack-zaqar15:34
flaper87zigo: think I found a way around it15:34
flaper87ungly but ...15:35
flaper87I'm just adding more thing that shaifali will have to clean up15:35
flaper87mmh, not sure if what I'm doing actually works15:42
therveflaper87, I guess the issue is that we're using config files in the tests, no?15:47
therveThe port seems to be all over the place15:47
flaper87therve: yeah, I'm hacking it really bad to then do a proper refactor15:50
flaper87I don't think the conf files themselves are bad15:50
flaper87it's more like there are several URLs all over the place15:50
angvpflaper87: can I ask why do you need mongo for running the tests? i mean having a factory that returns json is not enough to "simulate" mongo ?15:51
angvp(asking from my ignorance on the topic)15:52
flaper87that'd be enough. There's actually something called mockmongo but we haven't had time to adopt it. This happened because we used to have sqlite but when we removed it we simply switched to running tests assuming there was a mongodb installed15:53
flaper87ultimatedly, I think we need to get mock mongo once and for all15:53
flaper87or just keep assuming there's a mongodb installed but we gotta split these unittests better15:56
*** davideagnello has joined #openstack-zaqar15:56
*** dynarro has quit IRC15:57
angvpflaper87: this? https://github.com/vmalloc/mongomock15:58
flaper87yup15:58
flaper87angvp: ^15:58
flaper87these tests are bad :(15:59
angvpflaper87: yeah i was asking if that something called is mongomock :P15:59
*** kgriffs|afk is now known as kgriffs15:59
angvpflaper87: how hard is it to refactor the tests?15:59
flaper87there are several things to do. We have to merge zaqar/tests into zaqar/zaqar/tests - longer term task - then we need to get rid of all these mongodb:// URIs we have everywhere16:01
flaper87and we need to stop depending on a running mongodb instance16:01
flaper87or at least, we need to be able to easily switch between them16:01
flaper87I don't think it's hard, it's more like tedious but quite urgent16:02
flaper87I guess one step forward would be to have placeholders in the tests configs and replace those values16:04
angvpif you invite me ten pizzas next time i'll go to italy i offer to help16:06
angvpXD16:06
angvpbtw brb food16:07
flaper87lol, I'll be happy to16:12
vkmcI claim 4 pizzas please16:12
vkmcthanks16:12
* flaper87 should really catch up with email16:12
*** amitgandhinz has quit IRC16:48
*** amitgandhinz has joined #openstack-zaqar16:49
*** dguerri is now known as dguerri`away16:58
*** achanda has quit IRC17:10
*** achanda_ has joined #openstack-zaqar17:10
*** barra204 has quit IRC17:26
*** achanda_ has quit IRC17:32
*** achanda has joined #openstack-zaqar17:33
*** AAzza_ has joined #openstack-zaqar17:57
*** AAzza- has joined #openstack-zaqar17:58
*** AAzza has quit IRC18:00
*** AAzza- is now known as AAzza18:00
*** achanda_ has joined #openstack-zaqar18:00
*** AAzza_ has quit IRC18:02
*** achanda has quit IRC18:02
*** malini has quit IRC18:04
*** barra204 has joined #openstack-zaqar18:08
*** malini has joined #openstack-zaqar18:10
*** barra204 has quit IRC18:15
*** barra204 has joined #openstack-zaqar18:19
*** barra204_ has joined #openstack-zaqar18:27
*** barra204 has quit IRC18:31
*** malini has quit IRC18:39
*** malini has joined #openstack-zaqar18:41
*** barra204_ is now known as shakamunyi18:52
*** malini has quit IRC19:06
*** malini has joined #openstack-zaqar19:08
*** malini has quit IRC19:12
*** malini has joined #openstack-zaqar19:15
*** cknight has quit IRC19:15
*** cknight has joined #openstack-zaqar19:19
*** achanda_ has quit IRC19:23
*** exploreshaifali has joined #openstack-zaqar19:25
*** exploreshaifali has quit IRC19:37
*** malini has quit IRC20:03
*** openstack has joined #openstack-zaqar20:05
kfox1111I have some prototype code working too.20:06
kfox1111this should help with the Zaqar guest agent use case.20:06
flaper87kfox1111: hey there!20:06
flaper87awesome, will read :)20:06
flaper87thanks for putting time on this20:06
kfox1111thanks for having a look. :)20:07
kfox1111np. glad to help. :)20:07
kfox1111If zaqar could add a per queue acl like barbican is doing, then it might be easy to glue it all together.20:08
flaper87kfox1111: I've been thinking about for a bit. As in, how to implement it20:09
flaper87I think I've an idea but I gotta write it down20:09
flaper87It's more like a per-queue policy20:09
kfox1111yeah.20:11
kfox1111been talking through  the api with the barbican folks. they implemented it as a json doc per secret.20:12
kfox1111problematic though with heat.20:12
kfox1111heat really wants resources to be atomic.20:12
kfox1111with the single doc, replace api, then if two vm's are booting and wanting to gain access to the same resource, they race, and one might get dropped.20:13
kfox1111so we're talking about an Put/Delete /resource/<uuid>/acl/<userid> kind of api.20:13
*** openstack has joined #openstack-zaqar20:30
kfox1111therve: alive?20:39
thervekfox1111, Somewhat!20:40
kfox1111hehe.20:40
kfox1111so.. I asked the nova folks and didn't get a response.20:40
kfox1111I'm thinking there is maybe a second blueprint, or not, for solving the race.20:40
kfox1111an api extention to server create that creates the instance shutdown,20:40
kfox1111and optionally precreates the instance user.20:40
therve"creates the instance shutdown" ?20:41
kfox1111that would allow you to create a heat template that creates the instance, gets the userid, registers it with barbican, then starts the instance.20:41
therveOh I see what you mean20:41
kfox1111like virsh create --paused or something.20:42
therveWell yeah I wasn't even thinking about the race, just the user retrieval from heat is painful in the current proposal20:42
kfox1111yeah. could be done with wait conditions, but yeah. I was thinking about alternate ways to implement that.20:42
thervekfox1111, It seems passing the user would be easier20:42
kfox1111I was hoping they aready had a create stopped, but they dont.20:43
therveAlthough I guess it would mean passing the password which is annoying20:43
kfox1111for policy reasons, I think organizations would be much happier if users didn't get direct access to the instance users credentials.20:43
kfox1111exactly.20:43
kfox1111its more secure in that, any user can't just take the instance user and use it themselves elsewhere. They can login to the vm and get a fresh token, but it wont last forever.20:44
kfox1111I'm thinking of the user that leaves the project use case.20:44
kfox1111do you think I should put these details in the existing spec, or do a follow on spec? I'm worried if  Iput it all in one spec, it will seem overly complicated and people won't read it. :/20:45
therveI don't know... As it stands I don't think the solution is really viable.20:46
kfox1111yeah... It could be made to work, but its a lot more work. Ok. I'll amend the spec.20:46
*** dguerri`away is now known as dguerri20:56
*** openstackgerrit has quit IRC20:59
*** openstackgerrit has joined #openstack-zaqar20:59
kfox1111ok. just about done with the new spec updates...21:07
kfox1111anyone know how the spec history stuff at the bottom is intended to work?21:08
*** mpanetta has quit IRC21:13
*** sriram has quit IRC21:13
kfox1111well, I took a stab at it. :)21:16
*** amitgandhinz has quit IRC21:22
*** shakamunyi has quit IRC21:26
*** openstack has joined #openstack-zaqar21:32
*** csoukup has quit IRC21:32
*** openstackstatus has joined #openstack-zaqar21:32
*** ChanServ sets mode: +v openstackstatus21:32
*** cknight has quit IRC22:17
*** boris-42 has quit IRC22:48
openstackgerritOpenStack Proposal Bot proposed openstack/python-zaqarclient: Updated from global requirements  https://review.openstack.org/13599122:53
openstackgerritOpenStack Proposal Bot proposed openstack/zaqar: Updated from global requirements  https://review.openstack.org/17685522:54
*** kfox1111 has quit IRC23:09
*** kfox1111 has joined #openstack-zaqar23:09
*** exploreshaifali has joined #openstack-zaqar23:30
« Thursday, 2015-05-28 Index Saturday, 2015-05-30 »

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!