Thursday, 2016-09-01

*** ducttape_ has quit IRC00:07
*** ducttape_ has joined #openstack-lbaas00:28
*** ducttape_ has quit IRC00:31
*** chenli has joined #openstack-lbaas00:41
*** chenli_ has joined #openstack-lbaas00:42
*** chenli has quit IRC00:46
*** chenli_ is now known as chenli00:46
*** ducttape_ has joined #openstack-lbaas00:49
*** bana_k has quit IRC01:01
*** kevo has quit IRC01:06
*** ducttape_ has quit IRC01:15
*** chlong has joined #openstack-lbaas01:18
*** ducttape_ has joined #openstack-lbaas01:18
*** ducttape_ has quit IRC01:42
*** hockeynut has quit IRC01:43
*** ducttape_ has joined #openstack-lbaas01:46
openstackgerritli,chen proposed openstack/octavia: Update listener stats
*** ducttape_ has quit IRC01:54
*** woodster_ has quit IRC01:59
*** chenli has quit IRC02:02
*** chenli has joined #openstack-lbaas02:06
*** cody-somerville has joined #openstack-lbaas02:13
*** yamamoto_ has joined #openstack-lbaas02:14
*** ducttape_ has joined #openstack-lbaas02:19
openstackgerritli,chen proposed openstack/octavia: Add class 'StatsMixin'
*** ducttape_ has quit IRC02:27
*** anilvenkata has joined #openstack-lbaas02:30
*** ducttape_ has joined #openstack-lbaas02:35
*** ducttape_ has quit IRC02:43
*** armax has quit IRC02:45
*** amotoki has joined #openstack-lbaas02:51
*** armax has joined #openstack-lbaas03:03
*** cody-somerville has quit IRC03:05
*** amotoki has quit IRC03:08
*** armax has quit IRC03:08
*** cody-somerville has joined #openstack-lbaas03:11
*** kevo has joined #openstack-lbaas03:12
*** cody-somerville has quit IRC03:16
*** amotoki has joined #openstack-lbaas03:21
*** fnaval has joined #openstack-lbaas03:25
*** amotoki has quit IRC03:30
*** amotoki has joined #openstack-lbaas03:35
sbalukoffjohnsom: Are you around?03:43
*** ducttape_ has joined #openstack-lbaas03:43
johnsomsbalukoff what is up?03:44
sbalukoffSo, I was looking at this patch:
sbalukoffAnd it appears to work, though the testing I'm asking for might be pretty esoteric.03:46
sbalukoffGiven this is the namespace driver, do we care?03:46
sbalukoffIn any case, it accomplishes what it sets out to do: haproxy gets respawned if you kill it, and when you alter the load balancer config, haproxy gets restarted gracefully (ie. with the "-sf <old_pid>" option added.)03:47
johnsomSo, yeah.  People are using this driver still.03:48
*** ducttape_ has quit IRC03:49
johnsomI hate haven toneat a03:49
johnsomUgh,mobile sorry03:49
johnsomI have having to babysit this in this way.  We do it with Upstart in Octavia.  Which I need to implement systemd as well.03:50
sbalukoffYep, but I have trouble finding fault with using neutron libraries intended to do this stuff in a neutron environment. :/03:50
johnsomI think if it is a functional and complete solution we shouldn't block it.  Just keep in mind the maintenance.03:51
sbalukoffYeah. Should I insist on the tests I asked for there?03:51
sbalukoff(Mostly, I think the tests I asked for wouldn't really test this in any meaningful way... though I have a hard time thinking of a meaningful test for process monitoring that isn't a scenario test.)03:52
johnsomFor drivers that the cores don't regularly use, I think we should make sure there are good tests.  You know how picky I am about tests, so personally I think you are right.03:52
sbalukoffOk, I'll update my note to continue requesting the tests.03:53
sbalukoffIn any case, as long as the process monitoring enable() method is being called with the right parameters, this should continue to work (and we should notice if something changes there in our code tree to break it).03:54
johnsomWe got bit by a revert on this once already, so good tests makes sense.03:54
sbalukoffAgreed. Ok. Thanks!03:55
sbalukoffI'm about to test Chen's octavia stats code stuff...03:56
johnsomNp.  Thanks!03:56
sbalukoffI know she's really hoping to get this in for Newton...  so any eyes we can get on that pronto would be great. (Otherwise... how would she go about trying to get a feature freeze extension).03:57
sbalukoffFWIW, I think she's been pretty responsive and prompt both getting the spec changes done and code up for review for us.03:57
sbalukoff(We just haven't had time to review it, given how busy things have been in the last few weeks.)03:57
johnsomYeah, this was a late land and I've been asking our team to review as they have been looking at metrics recently.  I think there are some open comments on those.03:59
sbalukoffOk. Thanks for that, eh!03:59
johnsomYes, we have been slammed and are still catching up.03:59
*** chenli has quit IRC04:00
johnsomIf it is good, please feel free to +2 it and push us04:00
sbalukoffOk, will do. (Will probably be around for several more hours tonight as well. I've been asked to do some relatively urgent testing on some internal code tonight as well...)04:00
*** amotoki has quit IRC04:01
johnsomOk, thanks04:01
*** cody-somerville has joined #openstack-lbaas04:05
*** amotoki has joined #openstack-lbaas04:16
*** kevo has quit IRC04:19
*** cody-somerville has quit IRC04:20
*** yamamoto_ has quit IRC04:23
*** chenli has joined #openstack-lbaas04:32
*** links has joined #openstack-lbaas04:32
chenlijohnsom, dougwig ping04:35
*** amit213 has quit IRC04:37
*** yamamoto_ has joined #openstack-lbaas04:39
*** amit213 has joined #openstack-lbaas04:40
*** cody-somerville has joined #openstack-lbaas04:40
*** ducttape_ has joined #openstack-lbaas04:45
*** bana_k has joined #openstack-lbaas04:46
*** yamamoto_ has quit IRC04:49
*** ducttape_ has quit IRC04:50
*** pcaruana has quit IRC04:57
*** dougwig has quit IRC05:03
*** vmtyler has quit IRC05:03
*** amitry has quit IRC05:04
*** vmtyler has joined #openstack-lbaas05:04
*** sbelous_ has quit IRC05:04
*** amitry has joined #openstack-lbaas05:05
*** sbelous_ has joined #openstack-lbaas05:06
*** dougwig has joined #openstack-lbaas05:07
openstackgerritli,chen proposed openstack/octavia: Add a new API to list loadbalancer statistics
*** gcheresh has joined #openstack-lbaas05:13
*** saju_m has joined #openstack-lbaas05:28
*** chenli has quit IRC05:29
*** kevo has joined #openstack-lbaas05:33
*** yamamoto_ has joined #openstack-lbaas05:34
*** kobis has joined #openstack-lbaas05:38
*** ihrachys has joined #openstack-lbaas05:42
*** cody-somerville has quit IRC05:45
*** ducttape_ has joined #openstack-lbaas05:46
*** kobis has quit IRC05:48
*** ducttape_ has quit IRC05:50
*** carrbs has quit IRC05:52
*** reedip has quit IRC06:03
*** rcernin has quit IRC06:08
*** anilvenkata has quit IRC06:15
*** nagyz has quit IRC06:22
*** nagyz has joined #openstack-lbaas06:22
*** anilvenkata has joined #openstack-lbaas06:27
*** pcaruana has joined #openstack-lbaas06:30
*** kobis has joined #openstack-lbaas06:33
*** ihrachys has quit IRC06:33
*** saju_m has quit IRC06:38
*** rcernin has joined #openstack-lbaas06:41
*** chenli has joined #openstack-lbaas06:44
*** rcernin has quit IRC06:45
*** rcernin has joined #openstack-lbaas06:45
*** bana_k has quit IRC06:46
*** ducttape_ has joined #openstack-lbaas06:47
*** ducttape_ has quit IRC06:52
*** tesseract- has joined #openstack-lbaas07:03
*** chenli has quit IRC07:06
openstackgerritRavikumar proposed openstack/octavia: Housekeeping service does not cleanup stale amphora
*** Alex_Stef has quit IRC07:24
*** kevo has quit IRC07:29
*** chenli has joined #openstack-lbaas07:37
*** ihrachys has joined #openstack-lbaas07:41
*** chenli has quit IRC07:43
*** chenli has joined #openstack-lbaas07:46
*** ducttape_ has joined #openstack-lbaas07:49
*** ducttape_ has quit IRC07:53
*** ihrachys has quit IRC07:55
*** saju_m has joined #openstack-lbaas07:59
*** nmagnezi has joined #openstack-lbaas08:09
*** duvarenkov has joined #openstack-lbaas08:17
*** duvarenkov has quit IRC08:29
*** duvarenkov has joined #openstack-lbaas08:31
*** eezhova has joined #openstack-lbaas08:38
*** anilvenkata has quit IRC08:45
*** noshankus has joined #openstack-lbaas08:46
*** ducttape_ has joined #openstack-lbaas08:49
*** Alex_Stef has joined #openstack-lbaas08:53
*** pck_ has joined #openstack-lbaas08:53
*** pck_ has quit IRC08:54
*** pck_ has joined #openstack-lbaas08:54
*** HenryG_ has joined #openstack-lbaas08:54
*** ducttape_ has quit IRC08:54
*** amotoki_ has joined #openstack-lbaas08:55
*** saju_m has quit IRC08:57
*** amotoki has quit IRC08:58
*** darrenc has quit IRC08:58
*** HenryG has quit IRC08:58
*** pck has quit IRC08:58
*** johnsom_ has quit IRC08:58
*** malos has quit IRC08:58
*** jidar has quit IRC08:58
*** _laco has quit IRC08:58
*** johnsom_ has joined #openstack-lbaas08:59
*** malos has joined #openstack-lbaas08:59
*** chenli has quit IRC09:02
*** darrenc has joined #openstack-lbaas09:03
*** jidar has joined #openstack-lbaas09:03
*** _laco has joined #openstack-lbaas09:03
*** anilvenkata has joined #openstack-lbaas09:03
*** _laco has quit IRC09:09
*** saju_m has joined #openstack-lbaas09:10
*** HenryG_ is now known as HenryG09:17
*** fnaval has quit IRC09:19
*** ihrachys has joined #openstack-lbaas09:25
*** numans has joined #openstack-lbaas09:27
*** HenryG has quit IRC09:29
*** HenryG has joined #openstack-lbaas09:29
*** ducttape_ has joined #openstack-lbaas09:50
*** ducttape_ has quit IRC09:54
*** yamamoto_ has quit IRC10:02
*** yamamoto has joined #openstack-lbaas10:43
*** anilvenkata has quit IRC10:43
*** gongysh has quit IRC10:49
*** ducttape_ has joined #openstack-lbaas10:51
*** yamamoto has quit IRC10:52
*** ducttape_ has quit IRC10:55
*** anilvenkata has joined #openstack-lbaas10:58
*** yamamoto has joined #openstack-lbaas11:00
*** reedip has joined #openstack-lbaas11:15
*** nmagnezi has quit IRC11:28
*** nmagnezi has joined #openstack-lbaas11:28
*** gongysh has joined #openstack-lbaas11:49
*** ducttape_ has joined #openstack-lbaas11:52
nmagnezisbalukoff, hi11:52
nmagnezisbalukoff, re:
nmagnezisbalukoff, first, thanks a lot for taking the time to review this. re your comment: you mean that whenever I reffer to the config option I'm just missing the default section name?11:54
nmagnezimeaning: cfg.CONF.default.foo_param ?11:54
*** ducttape_ has quit IRC11:56
*** yamamoto has quit IRC12:11
*** ducttape_ has joined #openstack-lbaas12:11
*** ducttape_ has quit IRC12:13
*** yamamoto has joined #openstack-lbaas12:13
*** yamamoto has quit IRC12:16
*** ducttape_ has joined #openstack-lbaas12:17
*** yamamoto has joined #openstack-lbaas12:26
*** cody-somerville has joined #openstack-lbaas12:34
*** ducttape_ has quit IRC12:35
*** gongysh has quit IRC12:36
*** anilvenkata has quit IRC12:39
*** anilvenkata has joined #openstack-lbaas12:39
*** cody-somerville has quit IRC12:45
*** links has quit IRC12:51
*** ducttape_ has joined #openstack-lbaas12:53
*** yamamoto has quit IRC12:57
*** numans has quit IRC13:02
*** anilvenkata_ has joined #openstack-lbaas13:06
*** anilvenkata_ is now known as anilvenkata_afk13:06
*** anilvenkata has quit IRC13:07
*** ducttape_ has quit IRC13:07
*** amotoki_ has quit IRC13:14
*** matt-borland has joined #openstack-lbaas13:15
*** gongysh has joined #openstack-lbaas13:16
openstackgerritMerged openstack/neutron-lbaas: Change assertTrue(isinstance()) by optimal assert
*** yamamoto has joined #openstack-lbaas13:26
*** yamamoto has quit IRC13:27
*** yamamoto has joined #openstack-lbaas13:30
*** amotoki has joined #openstack-lbaas13:31
*** ducttape_ has joined #openstack-lbaas13:32
*** yamamoto has quit IRC13:32
*** yamamoto has joined #openstack-lbaas13:33
*** yamamoto has quit IRC13:34
*** yamamoto has joined #openstack-lbaas13:34
*** yamamoto has quit IRC13:34
*** diogogmt has quit IRC13:39
openstackgerritMerged openstack/neutron-lbaas: Add special handling for functional tests
*** amotoki has quit IRC13:43
*** ducttape_ has quit IRC13:44
*** diogogmt has joined #openstack-lbaas13:45
*** openstackgerrit has quit IRC13:49
*** openstackgerrit has joined #openstack-lbaas13:50
*** amotoki has joined #openstack-lbaas13:50
*** diogogmt has quit IRC13:50
*** ducttape_ has joined #openstack-lbaas13:57
*** amotoki has quit IRC13:59
*** ddaskal has joined #openstack-lbaas14:00
*** yamamoto has joined #openstack-lbaas14:00
*** ddaskal has quit IRC14:04
*** ducttape_ has quit IRC14:09
*** RichB has quit IRC14:13
*** amotoki has joined #openstack-lbaas14:14
*** carrbs has joined #openstack-lbaas14:26
*** gongysh has quit IRC14:33
*** pglass has joined #openstack-lbaas14:33
*** ducttape_ has joined #openstack-lbaas14:40
*** kobis has quit IRC14:44
*** diogogmt has joined #openstack-lbaas14:51
*** saju_m has quit IRC14:56
*** Alex_Stef has quit IRC14:59
*** pcaruana has quit IRC15:03
*** rcernin has quit IRC15:06
*** ddaskal has joined #openstack-lbaas15:07
*** eezhova has quit IRC15:10
*** yamamoto has quit IRC15:10
*** armax has joined #openstack-lbaas15:26
*** rcernin has joined #openstack-lbaas15:31
TrevorVdiltram, my bro15:32
ptoohillnmagnezi: I think he was saying that it should be placed in an appropriate section, or maybe it is already in that section but the code is already looking in default, hence it working when he moved the option15:34
openstackgerritMerged openstack/octavia: Updated from global requirements
*** anilvenkata_afk is now known as anilvenkata15:38
*** BjoernT has joined #openstack-lbaas15:49
*** ddaskal_ has joined #openstack-lbaas15:52
*** ddaskal has quit IRC15:55
*** gcheresh has quit IRC16:08
*** yamamoto has joined #openstack-lbaas16:11
*** ddaskal has joined #openstack-lbaas16:15
*** yamamoto has quit IRC16:15
*** ddaskal_ has quit IRC16:19
*** bana_k has joined #openstack-lbaas16:29
*** BjoernT has left #openstack-lbaas16:39
diltramTrevorV: hey man16:41
*** woodster_ has joined #openstack-lbaas16:42
openstackgerritTrevor Vardeman proposed openstack/octavia: Add quota support to Octavia
*** yamamoto has joined #openstack-lbaas16:46
*** yamamoto has quit IRC16:50
*** amotoki has quit IRC16:51
*** anilvenkata is now known as anilvenkata_afk16:54
nmagneziptoohill, thanks Phil, I will taking a look right now.16:56
ptoohillnmagnezi: no problem. but, yeah I think he was saying once he added it to default it works, so it probably should be in the 'right' section. At least thats how I understood it16:57
nmagneziptoohill, i was under the impression that it should be under the default section :-) , i'm re-spinning a devstack to have another look at this (also, i can't recall the section names so it will help)16:58
ptoohillMaybe it is supposed to be and your question is the answer. 'setting CONF.DEFAULT...' Though, are we sure that should be a default config option?16:59
*** tesseract- has quit IRC17:01
*** ducttape_ has quit IRC17:03
diltramsbalukoff: - I'm against that code17:04
*** ducttape_ has joined #openstack-lbaas17:04
diltramthat tests are not verifying anything17:04
diltramthere is everything mocked and there is even no code tested at all17:05
diltramtake a look into that setUp17:05
diltrameverything is mocked without using original spec17:05
diltramand everything is mocked17:06
diltramI will comment there but later17:06
diltramI need to eat something because I'm starving17:06
diltramso I will comment later17:06
diltrambut if you could reconcider that +2 in this time will be great :)17:06
diltramok, food :D17:07
*** bana_k has quit IRC17:07
ptoohillstatsmixin isnt mocked and it calls it, it seems17:08
*** kevo has joined #openstack-lbaas17:08
ptoohill        ls_stats =
ptoohill            self.session, self.listener_id)17:08
ptoohillmaybe i just need more coffee17:08
ptoohilldiltram: please mark a -1 so other know this is being verified.17:09
ptoohillit does mock the other things inside of it, but its a test for the stats mixin, maybe the question/problem here is there needs to be more tests verifying the other methods functionality17:11
openstackgerritJohn Perkins proposed openstack/neutron-lbaas: Temporarily support two INTERFACE_DRIVER_OPTS
*** ducttape_ has quit IRC17:15
ptoohilland with that said, the only things that are mocked inside the statsmixin are the repos, which seems fine to me17:17
ptoohillPlease help me understand the issue youre seeing.17:17
*** bana_k has joined #openstack-lbaas17:17
ptoohillIf this were a functional/scenario, having everything mocked would be bad. This seems to mock the appropriate things for unit testing from what I see.17:18
*** ducttape_ has joined #openstack-lbaas17:23
johnsomIs anyone else having the specs tox fail on master?17:23
openstackgerritJohn Perkins proposed openstack/neutron-lbaas: Temporarily support two INTERFACE_DRIVER_OPTS
diltramjohnsom: I had with some svg files failing17:30
johnsomHmm, docs passes for me, it's specs17:30
diltramptoohill: my point is that she is even mocking data_models17:30
diltramdid you saw how she's mocking in lne 38 stats object?17:31
diltramit's not even verifying that attribute really exist in this class17:31
diltramit's all around octavia that we're using MagicMock without autospec/spec and we need to change that if we would like to have a realible tests17:33
diltramand even unit test should mock all interaction with outside17:33
diltramwith db, other services etc17:33
diltrambut I never heard that we should mock our own code17:34
diltramjohnsom: so, no it's working for me :P17:34
diltramin my opinion we should in that situation mock database access methods to returne real data_model/types objects and we're verifying that our data paths really works17:37
diltramand it's true testing will take more time but we will really test something17:37
diltramand we should have overlapping test which are verifying that paths are working and that we didn't change anything - like class methods what we can change here and bum, everything is OK in test17:38
diltramyeah, we just need to mock everywhere keystoneauth1.session.Session to return specific objects and we're good to go17:40
diltramtake a look here17:43
diltramreally I can understand what is the idea here17:43
ptoohillI suppose I still dont see an issue here. I mean I know we need to update and make testing better. But this test is for the mixin which is 'get_stats', the tests verify that if X stats are set they get returned as expected17:45
ptoohillbut, im also not the greatest tester, so have it if you will ;)17:45
*** SumitNaiksatam has joined #openstack-lbaas17:47
sbalukoffdiltram: So now that I've had a little longer to wake up, I'm not sure I follow you. But, I'll wait for your comments on that patch set.17:47
diltramsure, give me a few minutes and even I will add some additional info about what I'm talking :)17:48
ptoohilloh hrmm17:48
johnsomI have to agree with one point.  Mocking a model is strange given our models are not database bound17:48
ptoohillYeah, not sure why thats mocked there17:49
ptoohilli see17:49
ptoohillOk, i agree. Not sure whats going on with 47717:49
ptoohillwell, i suppose it could just be17:49
ptoohillno no, im confused o.o17:50
ptoohill477 shouldnt be there, thats the method that should be test.. so yeah I agree this is off a bit17:51
ptoohilli was looking at this in mixin context17:52
ptoohillthis is the other test...17:52
ptoohillwhich i think is ok, because its mocking the get_stats with the expected data for the update_db method17:53
ptoohilland yeah, no real need to mock the dmodels, but that shouldnt break anything and isnt exactly 'wrong' i dont think. In either case Im not clearing anything up obviously, so Ill leave it as I think this looks ok for unit tests.17:55
*** harlowja_ has joined #openstack-lbaas17:58
diltramo shiet I didn't thought that the result will so horrible17:59
diltramtake a look into my comment17:59
*** harlowja has quit IRC18:00
*** ihrachys has quit IRC18:00
openstackgerritMichael Johnson proposed openstack/octavia: Subnet host routes support for plug_network
sbalukoffdiltram: I'm not sure your comment there proves anything. Sure, I can add a few lines to some member API code which destroys the parent pool object in the database, and the API functional tests will never find it because those tests aren't meant to go that deep. The API functional tests (might be mis-named) are just making sure that for the proper kind of input, you get the proper kind of output.18:05
sbalukoffIn any case, I think Chen is in training for the next two days, so this might not be something she gets to for a bit.18:16
*** eezhova has joined #openstack-lbaas18:18
*** openstackgerrit has quit IRC18:18
*** openstackgerrit has joined #openstack-lbaas18:19
nmagnezisbalukoff, hey :)18:21
nmagnezisbalukoff, can you please confirm  what ptoohill said about the auto reschedule bug? I want to make sure i got you correctly so i can fix this right away18:21
sbalukoffnmagnezi: I'm not seeing any comments from ptoohill in that patch set.18:22
ptoohillah, it was here18:23
sbalukoffOh, what did you say?18:23
ptoohillI was just expressing what I understood from your comments18:23
openstackgerritJohn Perkins proposed openstack/neutron-lbaas: Temporarily support two INTERFACE_DRIVER_OPTS
ptoohillThat when you added to default it worked, So it needs to be in right section, whether that default or not, and called correctly?18:23
ptoohillwas a little more to it, but I think nmagnezi is just looking for clarification there18:24
ptoohilldont let me get in the way18:24
sbalukoffSo my comments are that the code is looking for the allow_automatic_lbaas_agent_failover option set in the [DEFAULT] section of /etc/neutron/neutron.conf, but this is actually begin set (by the template) in the [DEFAULT] section of /etc/neutron/neutron_lbaas.conf18:24
ptoohilli see, there ya go18:25
sbalukoffnmagnezi: Does that make sense?18:25
nmagnezisbalukoff, oh, really? will check that. now that i see this i think i misread your comment. so good that I asked :-) but how do you see it's reading it from neutron.conf? I remember I configured it in neutron_lbaas.conf18:27
nmagneziagain, i will recheck myself18:27
nmagneziif that actually the case than ofcourse your comment is valid18:28
sbalukoffYep, when you do your test, make sure that option is not set in /etc/neutron/neutron.conf and only in /etc/neutron/neutron_lbaas.conf18:29
*** harlowja_ has quit IRC18:31
*** harlowja has joined #openstack-lbaas18:32
carrbsheyo! Any folk available to help me get Octavia running in devstack? I’m using this guide: and have run into some issues18:37
carrbsTrying to just get a dev environment going so I can look at some bugs in launchpad18:37
diltramsbalukoff: I really can't understand why you're so determined to minimize code testing, in my previous work we never done things like this and we never loose anything because of that, because of that I'm so obsessed with testing as deep as possible without calling external services/dbs18:39
bana_kcarrbs you facing any prob?18:39
openstackgerritJohn Perkins proposed openstack/neutron-lbaas: Temporarily support two INTERFACE_DRIVER_OPTS
sbalukoffdiltram: Well, unit tests are specifically not meant to be deep. For deep testing, that's sort of what scenario tests are for-- they're supposed to give the code a complete top-to-bottom work-out. And yes, we don't really have much by way of scenario tests yet (just the neutron ones yet).18:42
carrbsbana_k: There’s a section in the guide (^) that says “Set up a simple web server…” I ran those commands and it said (in infinite loop) “address already in use” (or something of the like). I killed that window, and accidentaly killed a few of the other processes that start when you run ./devstack.sh18:44
*** eezhova has quit IRC18:44
sbalukoffdiltram: But you're right that it's possible that you can go not deep enough to the point where none of the code you're supposed to be testing is actually being tested.18:44
bana_kcarrbs : do you mean for the members ?18:44
sbalukoffIn this case, I think there's no reason to mock out the data model, so yes, we should go a little deeper than what's there.18:44
carrbsI tried and, but now I seem to be having issues with apache.18:44
diltramsbalukoff: partialy I need to disagree with you - the unit test should verify "simple" things but I never heard that unit test cannot go deep into internal projet code18:45
diltramsure should mock and block connectivity with other things18:45
diltramlike db and other services18:46
diltrambut unit tests should verify that if I'm runnig this method I'm trying to save that data to db18:46
bana_kcarrbs: I have never faced that prob till now. you can try and try and if you can try it out on the fresh VM that ll be even better18:46
diltramor like in this situation18:46
sbalukoffRight, and in this case I think I'm agreeing with you: Don't mock the model, mock where the model would interact with the DB.18:46
diltramwhen I'm requesting that stats that they will be grabbed from DB and send to me18:47
diltramsbalukoff: +118:47
carrbsbana_k: ok, I’ll try those things first. thanks!18:47
*** eezhova has joined #openstack-lbaas18:48
sbalukoffdiltram: Just as a friendly hint, please be careful with accusatory language: I'm not "so determined to minimize code testing." Though I can understand it's frustrating to encounter people with differing opinions on things you're passionate about.18:48
diltramsbalukoff: sorry for that, I didn't mean attacking you or smth18:49
diltramEnglish is my second language and sometimes I have that problems that I wrote smth not like I should18:50
sbalukoffNo problem. I appreciate that you don't want bad code to make it into our code base. That's a good thing to be passionate about, eh.18:50
sbalukoffSo yes-- please keep pointing out shit where you see it! Just be careful not to use language which implies others don't care about writing good code, eh.18:51
diltramsure :) I will work on that and again sorry18:53
sbalukoffEh, I'm the resident asshole. I can take it. ;)  (Also, I can dish it out...)18:54
sbalukoffcarrbs: You might have more luck following this guide in this patch set here:
*** harlowja_ has joined #openstack-lbaas19:13
carrbssbalukoff: thanks, I will compare the two. I was able to get ./ to run after a reboot, and I derped one of the steps I’m now realizing (thanks to bana_k for pointing out the derpage).19:16
nmagnezisbalukoff, hey19:16
nmagnezisbalukoff, so i just stacked19:16
nmagneziand is see19:16
nmagnezi$ grep allow_automatic_lbaas_agent_failover -R19:17
nmagnezineutron_lbaas.conf:#allow_automatic_lbaas_agent_failover = false19:17
nmagnezisbalukoff, it's only on neutron_lbaas19:17
*** harlowja has quit IRC19:17
nmagnezisbalukoff, how i tested: I cherry-picked the patch and ./ (RECLONE="no" , so it did not override the patch)19:19
sbalukoffnmagnezi: Yes. Again, your code does look for this option in /etc/neutron/neutron_lbaas.conf. It looks in /etc/neutron/neutron.conf19:23
sbalukoffer... sorry, your code DOESN'T look for this option in /etc/neutron/neutron_lbaas.conf19:23
nmagnezisbalukoff, the code uses cfg.CONF.param_name , isn't this point to neutron_lbaas? (sorry, i must be missing something here)19:24
nmagnezisbalukoff, we are talking about , right?19:26
sbalukoffnmagnezi: Yes, that's what I'm talking about. And ultimately, isn't that code being run by the q-svc daemon, which doesn't look in /etc/neutron/neutron_lbaas.conf ?19:27
*** gcheresh has joined #openstack-lbaas19:27
sbalukoffnmagnezi: Here's another test for you:19:27
sbalukoffnmagnezi: Uncomment that config line in /etc/neutron/neutron_lbaas.conf and set it to true. Restart your daemons, and deploy a load balancer. Make sure you've got two lbaasv2 agents running. Then kill the one to which the load balaner got scheduled.19:29
*** saju_m has joined #openstack-lbaas19:29
sbalukoffDoes the load balancer get rescheduled to the remaining (alive) lbaasv2 agent?19:29
nmagnezisbalukoff, i have done such a test while i developed this, so the answer is: 'yes'. if you want, I can test it once more.19:30
sbalukoffnmagnezi: So... I ran that test last night and it didn't work unless I put that config line in /etc/neutron/neutron.conf19:30
nmagnezisbalukoff, really? I will give it a spin :-)19:31
nmagnezisbalukoff, and re: your first question: to the best of my knowledge the q-svc indeed uses this since the lbaas is a plugin to the neutron server. but when this code runs the cfg.CONF is suppose to aim at neutron_lbaas.conf19:32
nmagnezibut again, i will test this and keep you posted19:32
nmagnezilemme see if I can add another node to my devstack19:32
sbalukoffI'll start re-stacking as well (the machine I was testing this on is set up to test your process monitor patch, which works, but your patch needs more tests.)19:35
*** gcheresh has quit IRC19:36
nmagnezisbalukoff, yes i was about to answer that comment. the process monitor has a lot of tests by itself in the neutron code base. so as an infrastructure it is well tested.19:38
nmagnezisbalukoff, any specific test you wish me to add to the neutron-lbaas codebase? I'm all up for it.19:38
nmagnezisbalukoff, and re: johnsom's comment (which I will replay as well) your answer was correct and another reason i refrain from using upstart is because it is specific to non systemd base linux flavors (such as Ubuntu). whereas this solution is OS agnostic19:40
nmagnezisbalukoff, that reminds me of the issue we have with the amphora-agent :-)19:41
sbalukoffnmagnezi: Per my comment on that patch, please make sure we're calling the pm.enable and other methods with the correct parameters.19:43
pglassso I've tried toggling admin_state_up on loadbalancers, listeners, pools, and members and it doesn't seem to have any effect.19:44
nmagnezisbalukoff, i will make sure to address that19:45
sbalukoffpglass: File a bug! I suspect we're not checking that option in places where we really should.19:46
openstackgerritMerged openstack/octavia: Corrected test_dir path in TempestPlugin
johnsompglass Hmm, We just fixed that in March:  Curious to see what it is now19:50
pglasshow is admin_state_up implemented? is it haproxy config changes, or something else?19:55
johnsomIt depends on the driver you are using19:55
johnsomIn Octavia, yes, it is mostly config changes19:56
pglassso in this case, if admin_state_up wasn't working properly, we could probably look into the haproxy config on the amp?19:57
pglassand those config changes are communicated via the amphora's rest api, right?19:57
pglassi was looking at the amphora rest api earlier. I was able to `curl -k --cert /etc/octavia/certs/client.pem` for example19:59
pglassthere was a single listener (in my case) and its status was "OFFLINE".20:00
openstackLaunchpad bug 1619449 in octavia "Toggling admin_state_up has no effect" [Undecided,New]20:00
johnsomHmm, just tested admin-state-up on a member and it stopped returning traffic, so....20:05
pglassalright. is it the controller worker that's responsible for making sure the amphora gets reconfigured when admin_state_up is changed?20:07
*** ducttape_ has quit IRC20:07
*** ducttape_ has joined #openstack-lbaas20:08
*** nmagnezi_ has joined #openstack-lbaas20:27
*** _ducttape_ has joined #openstack-lbaas20:28
*** nmagnezi has quit IRC20:28
*** carrbs has quit IRC20:31
*** ducttape_ has quit IRC20:31
*** _ducttape_ has quit IRC20:34
*** ducttape_ has joined #openstack-lbaas20:35
*** _ducttape_ has joined #openstack-lbaas20:38
*** ducttape_ has quit IRC20:42
*** _ducttape_ has quit IRC20:45
*** ducttape_ has joined #openstack-lbaas20:45
*** anilvenkata_afk has quit IRC20:50
openstackgerritCedric Shock proposed openstack/neutron-lbaas: Allow creating loadbalancer with network_id
nmagnezi_sbalukoff, devstack keeps failing for me when i try to spin, up the secondary node (for having anothe agent)20:54
nmagnezi_sbalukoff, it's midnight here. so I think i'll pick up on this tomorrow20:54
*** nmagnezi_ has quit IRC21:01
*** anilvenkata_afk has joined #openstack-lbaas21:02
*** rcernin has quit IRC21:24
*** ddaskal has quit IRC21:25
*** matt-borland has quit IRC21:27
*** anilvenkata_afk has quit IRC21:38
*** ducttape_ has quit IRC21:44
*** eezhova has quit IRC21:44
*** ducttape_ has joined #openstack-lbaas21:45
*** anilvenkata_afk has joined #openstack-lbaas21:52
johnsomOk, the routes patch is back up for review.  I fixed the type issue and had a successful run on devstack21:53
sbalukoffOk, I'll have a look.21:55
*** ducttape_ has quit IRC22:27
rm_workjust +A'd22:37
rm_workI swear I've reviewed that code 10 times now22:37
sbalukoffWell, you shouldn't have to review it any more. :)22:42
*** pglass has quit IRC22:46
*** ducttape_ has joined #openstack-lbaas22:48
rm_workHa, just found this:
rm_workDid you see that before sbalukoff? lol22:57
sbalukoffHaha! No, and I looked around pretty well before choosing that name. Besides: We got the trademark, they don't.23:04
rm_workyep :P23:05
sbalukoffHow did you even find that? It doesn't show up on the first four pages of a google search.23:07
harlowja_thats not octavia :-P23:07
rm_worksbalukoff: skills23:07
rm_workor, maybe luck23:08
*** fnaval has joined #openstack-lbaas23:16
openstackgerritMerged openstack/octavia: Subnet host routes support for plug_network
*** fnaval has quit IRC23:27
*** fnaval has joined #openstack-lbaas23:31
*** chlong has quit IRC23:33
*** gongysh has joined #openstack-lbaas23:35
*** fnaval has quit IRC23:35
*** diogogmt has quit IRC23:58

Generated by 2.14.0 by Marius Gedminas - find it at!