| *** diogogmt has joined #openstack-lbaas | 00:03 | |
| *** minwang2 has quit IRC | 00:10 | |
| *** Varun_Lodaya has quit IRC | 00:17 | |
| *** manishg_ has joined #openstack-lbaas | 00:20 | |
| *** manishg has quit IRC | 00:24 | |
| *** manishg_ has quit IRC | 00:33 | |
| *** crc32 has joined #openstack-lbaas | 01:09 | |
| *** vivek-ebay has quit IRC | 01:13 | |
| *** vivek-ebay has joined #openstack-lbaas | 01:18 | |
| *** bharathm has joined #openstack-lbaas | 01:19 | |
| *** manishg has joined #openstack-lbaas | 01:19 | |
| *** vivek-ebay has quit IRC | 01:20 | |
| *** bharathm has quit IRC | 01:24 | |
| *** manishg has quit IRC | 01:25 | |
| *** manishg has joined #openstack-lbaas | 01:25 | |
| *** bana_k has quit IRC | 01:36 | |
| *** ajmiller has joined #openstack-lbaas | 01:40 | |
| dougwig | adam_g: default drivers do not interact with rpc agents at all | 01:45 |
|---|---|---|
| *** guest101_ has quit IRC | 01:46 | |
| dougwig | adam_g: the only driver that uses the lbaas-agent is the haproxy namespace driver. all vendor drivers and the noop do not use that mechanism. | 01:46 |
| adam_g | dougwig, just attempting to use the default devstack plugin installation /w octavia enabled as default, or later the logging noop driver, the lbaasv2 agent hits an rpc timeout on startup in sync_state(). weird? | 01:47 |
| adam_g | or is the lbaas not intended to be run with those drivers? | 01:47 |
| adam_g | *the lbaas agent | 01:47 |
| dougwig | if you don't have the haproxy driver enabled in [service_providers] in neutron_lbaas.conf, you don't need that process running. nothing will be listening for it | 01:48 |
| dougwig | neither octavia nor noop use lbaas-agent. | 01:48 |
| dougwig | so correct, it's not intended to be run in that case. | 01:48 |
| *** kiran-r has joined #openstack-lbaas | 01:49 | |
| *** kiran-r has quit IRC | 01:50 | |
| adam_g | dougwig, doh! theres my error, i was under the assumption the agent was loading those drivers.. in that case, whats actually loading them and, the case of octavia runnign this? neutron-server? https://git.openstack.org/cgit/openstack/neutron-lbaas/tree/neutron_lbaas/drivers/octavia/driver.py#n151 | 01:50 |
| dougwig | neutron-server loads them directly, via the loadbalancerv2 service plugin | 01:50 |
| dougwig | adam_g: ^^ sorry, keep forgetting to prefix. :) | 01:51 |
| *** kiran-r has joined #openstack-lbaas | 01:51 | |
| *** vivek-ebay has joined #openstack-lbaas | 01:51 | |
| *** kiranr has joined #openstack-lbaas | 01:52 | |
| adam_g | dougwig, okay awesome that clears things up for me alot. | 01:52 |
| adam_g | thanks | 01:52 |
| dougwig | adam_g: no problem, good luck. :) | 01:52 |
| *** vivek-ebay has quit IRC | 01:54 | |
| *** KunalGandhi has quit IRC | 01:56 | |
| *** kiran-r has quit IRC | 01:59 | |
| *** minwang2 has joined #openstack-lbaas | 02:05 | |
| *** minwang2 has quit IRC | 02:08 | |
| *** ajmiller has quit IRC | 02:21 | |
| *** KunalGandhi has joined #openstack-lbaas | 02:26 | |
| *** Varun_Lodaya has joined #openstack-lbaas | 02:28 | |
| *** ajmiller has joined #openstack-lbaas | 02:29 | |
| *** vivek-ebay has joined #openstack-lbaas | 02:34 | |
| *** woodster_ has quit IRC | 02:39 | |
| *** ganeshna has quit IRC | 02:45 | |
| *** diogogmt has quit IRC | 03:04 | |
| *** KunalGandhi has quit IRC | 03:08 | |
| Guest14599 | dougwig: is there any doc to enable octavia without devstack? | 03:19 |
| *** Guest14599 is now known as Brian_shang | 03:20 | |
| *** macjack has joined #openstack-lbaas | 03:22 | |
| *** kiran-r has joined #openstack-lbaas | 03:31 | |
| xgerman | not yet - sbalukoff is working on that | 03:32 |
| *** kiran-r has quit IRC | 03:39 | |
| *** minwang2 has joined #openstack-lbaas | 03:40 | |
| *** KunalGandhi has joined #openstack-lbaas | 03:42 | |
| Brian_shang | xgerman: ok, tks. | 03:43 |
| *** manishg has quit IRC | 03:44 | |
| *** KunalGandhi has quit IRC | 03:48 | |
| *** KunalGan_ has joined #openstack-lbaas | 03:48 | |
| Brian_shang | xgerman: neutron-server calls octavia api directly instead of lbaas-agent, right? | 03:48 |
| *** ganeshna has joined #openstack-lbaas | 03:53 | |
| *** kiranr has quit IRC | 04:02 | |
| *** Brian_shang has quit IRC | 04:05 | |
| *** KunalGan_ has quit IRC | 04:18 | |
| *** minwang2 has quit IRC | 04:21 | |
| *** minwang2 has joined #openstack-lbaas | 04:22 | |
| *** Brian_shang has joined #openstack-lbaas | 04:25 | |
| *** minwang2 has quit IRC | 04:33 | |
| *** numans has joined #openstack-lbaas | 04:37 | |
| *** vivek-ebay has quit IRC | 04:51 | |
| *** Brian_shang has quit IRC | 04:52 | |
| *** ganeshna has quit IRC | 04:58 | |
| *** Brian_shang has joined #openstack-lbaas | 05:07 | |
| *** kiran-r has joined #openstack-lbaas | 05:08 | |
| *** kiran-r has quit IRC | 05:10 | |
| *** ganeshna has joined #openstack-lbaas | 05:15 | |
| *** KunalGandhi has joined #openstack-lbaas | 05:16 | |
| *** KunalGandhi has quit IRC | 05:20 | |
| *** minwang2 has joined #openstack-lbaas | 05:34 | |
| *** Varun_Lodaya has quit IRC | 05:35 | |
| *** minwang2 has quit IRC | 05:39 | |
| *** macjack has quit IRC | 05:49 | |
| *** macjack has joined #openstack-lbaas | 05:58 | |
| *** Varun_Lodaya has joined #openstack-lbaas | 06:10 | |
| *** macjack has quit IRC | 06:20 | |
| *** vjay10 has joined #openstack-lbaas | 06:25 | |
| *** vjay10 has quit IRC | 06:33 | |
| *** eranra has joined #openstack-lbaas | 06:34 | |
| *** ganeshna has quit IRC | 06:52 | |
| *** evgenyf has joined #openstack-lbaas | 06:55 | |
| *** Varun_Lodaya has quit IRC | 07:00 | |
| *** sc68cal has quit IRC | 07:08 | |
| *** sc68cal has joined #openstack-lbaas | 07:10 | |
| *** akamyshnikova has joined #openstack-lbaas | 07:10 | |
| *** kiranr has joined #openstack-lbaas | 07:34 | |
| *** blallau has joined #openstack-lbaas | 07:38 | |
| *** Alex_Staf has joined #openstack-lbaas | 07:41 | |
| *** blallau has quit IRC | 07:42 | |
| *** kiran-r has joined #openstack-lbaas | 07:47 | |
| *** kobis has joined #openstack-lbaas | 07:47 | |
| *** ganeshna has joined #openstack-lbaas | 07:50 | |
| *** Brian_shang has quit IRC | 07:51 | |
| *** Brian_shang has joined #openstack-lbaas | 07:52 | |
| *** macjack has joined #openstack-lbaas | 08:06 | |
| *** rm_work has quit IRC | 08:14 | |
| *** rm_work has joined #openstack-lbaas | 08:16 | |
| *** kobis has quit IRC | 08:35 | |
| *** kobis has joined #openstack-lbaas | 08:36 | |
| kiran-r | Hi! Can anybody tell me if we can create multiple pools for the same subnet? | 08:49 |
| *** nmagnezi has joined #openstack-lbaas | 08:51 | |
| *** ganeshna has quit IRC | 09:18 | |
| *** ganeshna has joined #openstack-lbaas | 09:20 | |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: [LBaaS v2] Improve API tests performance https://review.openstack.org/226370 | 09:22 |
| *** kiran-r has quit IRC | 09:37 | |
| *** kiranr has quit IRC | 10:02 | |
| openstackgerrit | yaowei proposed openstack/neutron-lbaas: Fix lb agent reload loadbalancer https://review.openstack.org/227203 | 10:10 |
| *** ganeshna has quit IRC | 10:19 | |
| *** kiran-r has joined #openstack-lbaas | 10:22 | |
| *** ganeshna has joined #openstack-lbaas | 10:35 | |
| *** amotoki has joined #openstack-lbaas | 10:41 | |
| *** nmagnezi has quit IRC | 10:55 | |
| *** macjack has left #openstack-lbaas | 11:00 | |
| *** nmagnezi has joined #openstack-lbaas | 11:10 | |
| *** sballe has joined #openstack-lbaas | 11:31 | |
| *** ganeshna has quit IRC | 11:41 | |
| *** ganeshna has joined #openstack-lbaas | 11:42 | |
| *** ganeshna has quit IRC | 12:05 | |
| *** nishi has joined #openstack-lbaas | 12:21 | |
| *** KunalGandhi has joined #openstack-lbaas | 12:21 | |
| *** nishi has left #openstack-lbaas | 12:21 | |
| *** KunalGandhi has quit IRC | 12:26 | |
| openstackgerrit | Kobi Samoray proposed openstack/neutron-lbaas: (WIP) VMWare NSXv LBaaSv2 driver https://review.openstack.org/227266 | 12:33 |
| *** kiran-r has quit IRC | 12:43 | |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: [LBaaS v2] Improve API tests performance https://review.openstack.org/226370 | 12:47 |
| *** evgenyf has quit IRC | 12:52 | |
| *** nmagnezi_ has joined #openstack-lbaas | 13:00 | |
| *** nmagnezi has quit IRC | 13:02 | |
| *** alejandrito has joined #openstack-lbaas | 13:14 | |
| mestery | lbaas cores: Need this merged ASAP: https://review.openstack.org/227298 | 13:26 |
| mestery | Also, the issues around tagging octavia and getting it on pypi are on my mind, but are taking a backseat to getting RC1 cut, so stay tuned | 13:33 |
| mestery | johnsom xgerman dougwig blogan: ^^^ | 13:33 |
| *** nmagnezi__ has joined #openstack-lbaas | 13:34 | |
| *** nmagnezi_ has quit IRC | 13:37 | |
| *** kobis has quit IRC | 13:37 | |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: [LBaaS v2] Improve API tests performance https://review.openstack.org/226370 | 13:41 |
| *** ajmiller has quit IRC | 13:52 | |
| *** ajmiller has joined #openstack-lbaas | 13:55 | |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: [LBaaS v2] Improve API tests performance https://review.openstack.org/226370 | 14:06 |
| *** numans has quit IRC | 14:07 | |
| *** blallau has joined #openstack-lbaas | 14:08 | |
| *** diogogmt has joined #openstack-lbaas | 14:09 | |
| * mestery wakes up blogan dougwig johnsom | 14:16 | |
| mestery | Can one of you please approve this one: https://review.openstack.org/#/c/227298/ | 14:16 |
| mestery | It's needed to open mitaka for that gerrit repo | 14:16 |
| mestery | dougwig blogan johnsom: ^^^^ | 14:16 |
| xgerman | Sorry can't help... | 14:19 |
| xgerman | But those not on pacific should be awake | 14:19 |
| *** blallau has quit IRC | 14:22 | |
| mestery | We need that to land before we can cut the RC :( | 14:22 |
| mestery | I really really really hope blogan dougwig johnsom or another LBaaS core wakes up and merges that | 14:22 |
| *** diogogmt has quit IRC | 14:23 | |
| *** ganeshna has joined #openstack-lbaas | 14:29 | |
| *** ganeshna has quit IRC | 14:29 | |
| Brian_shang | Missing package name for distro/element: ubuntu/os-collect-config | 14:36 |
| Brian_shang | installing ccache_package from base | 14:36 |
| *** mlavalle has joined #openstack-lbaas | 14:48 | |
| *** minwang2 has joined #openstack-lbaas | 14:50 | |
| *** kiran-r has joined #openstack-lbaas | 14:51 | |
| *** apuimedo has quit IRC | 14:52 | |
| *** manishg has joined #openstack-lbaas | 14:56 | |
| *** crc32 has quit IRC | 14:56 | |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: [LBaaS v2] Improve API tests performance https://review.openstack.org/226370 | 14:57 |
| *** Varun_Lodaya has joined #openstack-lbaas | 15:06 | |
| *** manishg has quit IRC | 15:10 | |
| *** nmagnezi__ has quit IRC | 15:12 | |
| *** kobis has joined #openstack-lbaas | 15:15 | |
| *** diogogmt has joined #openstack-lbaas | 15:21 | |
| *** KunalGandhi has joined #openstack-lbaas | 15:22 | |
| *** manishg has joined #openstack-lbaas | 15:41 | |
| *** Touma has joined #openstack-lbaas | 15:47 | |
| *** manishg has quit IRC | 15:50 | |
| *** manishg has joined #openstack-lbaas | 15:51 | |
| *** minwang2 has quit IRC | 15:53 | |
| *** diogogmt has quit IRC | 15:55 | |
| *** diogogmt has joined #openstack-lbaas | 15:56 | |
| *** minwang2 has joined #openstack-lbaas | 15:57 | |
| *** KunalGandhi has quit IRC | 15:58 | |
| *** kiran-r has quit IRC | 15:58 | |
| *** vivek-ebay has joined #openstack-lbaas | 16:01 | |
| *** nmagnezi__ has joined #openstack-lbaas | 16:03 | |
| *** sbalukoff has quit IRC | 16:04 | |
| *** Varun_Lodaya has quit IRC | 16:08 | |
| *** madhu_ak has joined #openstack-lbaas | 16:11 | |
| johnsom | mestery Done | 16:12 |
| *** sbalukoff has joined #openstack-lbaas | 16:16 | |
| *** Touma has quit IRC | 16:16 | |
| *** minwang2 has quit IRC | 16:19 | |
| *** mlavalle_ has joined #openstack-lbaas | 16:23 | |
| *** pckizer_ has joined #openstack-lbaas | 16:24 | |
| *** pckizer_ has quit IRC | 16:24 | |
| *** pckizer_ has joined #openstack-lbaas | 16:24 | |
| *** mlavalle has quit IRC | 16:24 | |
| *** kbyrne has quit IRC | 16:24 | |
| *** pckizer has quit IRC | 16:24 | |
| *** afranc202 has quit IRC | 16:24 | |
| *** iraw- has joined #openstack-lbaas | 16:24 | |
| *** kbyrne has joined #openstack-lbaas | 16:24 | |
| *** Alex_Staf has quit IRC | 16:32 | |
| *** vivek-ebay has quit IRC | 16:32 | |
| *** sballe has quit IRC | 16:36 | |
| *** crc32 has joined #openstack-lbaas | 16:37 | |
| *** Varun_Lodaya has joined #openstack-lbaas | 16:44 | |
| *** bharathm has joined #openstack-lbaas | 16:50 | |
| *** vivek-ebay has joined #openstack-lbaas | 17:04 | |
| *** minwang2 has joined #openstack-lbaas | 17:05 | |
| *** eranra has quit IRC | 17:14 | |
| *** sbalukoff has quit IRC | 17:19 | |
| *** nmagnezi__ has quit IRC | 17:31 | |
| *** KunalGandhi has joined #openstack-lbaas | 17:32 | |
| *** sbalukoff has joined #openstack-lbaas | 17:32 | |
| *** ajmiller has quit IRC | 17:32 | |
| *** TrevorV has joined #openstack-lbaas | 17:34 | |
| *** sbalukoff has quit IRC | 17:38 | |
| *** Varun_Lodaya has quit IRC | 17:43 | |
| *** madhu_ak_ has joined #openstack-lbaas | 17:52 | |
| *** madhu_ak has quit IRC | 17:55 | |
| *** Varun_Lodaya has joined #openstack-lbaas | 17:58 | |
| *** SumitNaiksatam has joined #openstack-lbaas | 17:59 | |
| blogan | hey guys, eezhova brought up an idea we could possibly do to make the tests run even faster, especially for the pool tests so that we can unskip all the ones we did, and thats use the simulated_handler plugin in octavia. it's that old thing we used in the very beginning that basically just updated the provisioning and operating statuses of each entity, instead of going to the queue | 18:15 |
| blogan | for the api jobs only, not the scenario tests | 18:16 |
| rm_work | so, it wouldn't launch VMs? | 18:28 |
| rm_work | but we'd have tested all the launching-vms code paths in the other tests? | 18:28 |
| blogan | rm_work: no it wouldnt launch VMs, it wouldnt even need the controller worker | 18:30 |
| blogan | so it wouldnt test heart beat dependent api changes either | 18:30 |
| blogan | but those could be scenario tests though | 18:31 |
| blogan | it feels wrong to me right now, but i dont have a good reason because really the api tests just test teh api and validation layers, they shouldnt care whats being spun up on the backside | 18:31 |
| johnsom | I worry about deviating from the live code path. | 18:31 |
| johnsom | I guess we just need to have a holistic testing strategy | 18:32 |
| rm_work | yeah really the place for a lot of the tests we have in "apiv2" really belong in "scenario" | 18:33 |
| rm_work | we are testing WAY more than just the API | 18:33 |
| johnsom | Very true | 18:33 |
| rm_work | which i was having trouble finding the right words for when discussing it with infra | 18:33 |
| rm_work | which leads me to believe we may be doing it wrong ,_< | 18:33 |
| johnsom | That's what I am thinking. We need to step back and come up with a overall testing strategy. What tests, run at what layers, etc. | 18:34 |
| rm_work | yeah | 18:35 |
| rm_work | fnaval: ^^? :P | 18:35 |
| *** KunalGandhi has quit IRC | 18:35 | |
| blogan | rm_work: what tests are doign what a scenario test shoudl do? | 18:36 |
| blogan | most of what ive seen is just checking successful create, updates, and deletes, along with negative tests for testing validation | 18:37 |
| blogan | probably one or two that could be put in scenario but i dont remember them | 18:38 |
| rm_work | blogan: right but checking an API call does not require doing a whole end to end test does it? >_> | 18:38 |
| blogan | yall wanna get this merged so we can have working scenario tests now though? | 18:38 |
| blogan | https://review.openstack.org/#/c/226587/ | 18:38 |
| rm_work | I would think it would be more like what you were saying with the simulated_handler | 18:39 |
| blogan | rm_work: it doesn't and that kind of decides it for me bc if we just tested the simulated_handler, we're nto testing the controller worker code actually updtes the db status correctly | 18:40 |
| rm_work | hmm | 18:40 |
| rm_work | maybe simulated_handler needs to be further down the chain? | 18:40 |
| rm_work | like, replacing the nova and neutron amp/net drivers | 18:41 |
| rm_work | so it still goes through the queue/controller | 18:41 |
| blogan | well that would just be the noop drivers at that point | 18:41 |
| rm_work | yes | 18:41 |
| rm_work | i was thinking originally that's what it was supposed to be | 18:41 |
| rm_work | do we HAVE noop drivers? | 18:41 |
| blogan | yep | 18:41 |
| rm_work | lol | 18:41 |
| rm_work | w | 18:41 |
| rm_work | well, can we switch to those in the API tests? | 18:42 |
| blogan | that woudl be better | 18:42 |
| blogan | but still doesnt test that the nova interactions are correct :) | 18:42 |
| blogan | which is something we'll have to figure out at some point, do we need a job for every single driver and all the permutations and combinations of drivers | 18:43 |
| rm_work | right but | 18:43 |
| rm_work | we have other tests for that | 18:43 |
| blogan | well the scenario tests would test that sure | 18:43 |
| blogan | johnsom: opinion on using the noop drivers? | 18:44 |
| blogan | or if anything, the noop compute driver | 18:44 |
| blogan | although, that would probably not play well with a real amp driver | 18:44 |
| blogan | or network driver | 18:44 |
| blogan | so noop for all | 18:44 |
| pc-pothole | I still find it interesting this is being considered 'over-testing' :) | 18:45 |
| *** minwang2 has quit IRC | 18:45 | |
| blogan | pc-pothole: what do you mean? | 18:45 |
| *** bana_k has joined #openstack-lbaas | 18:45 | |
| pc-pothole | idk, the timings and what nots. even with vt-x our tests are still huge right | 18:46 |
| blogan | like they're not testing teh right things? | 18:46 |
| pc-pothole | idk, i just keep hearing were testing more then anyone else and i find it interesting that its a problem | 18:46 |
| openstackgerrit | Ihar Hrachyshka proposed openstack/neutron-lbaas: Kill HEADS file https://review.openstack.org/227466 | 18:47 |
| pc-pothole | maybe im just misunderstanding. i retract my 2 cents | 18:47 |
| blogan | pc-pothole: well its kind of a catch-22, the more we test for the longer we take | 18:47 |
| blogan | if everyone had 5 hour jobs, the gate would probably never clear up | 18:47 |
| blogan | so i understand the need for the hard limit | 18:47 |
| pc-pothole | threes got to be a better solution then ripping the tests apart to test less though | 18:48 |
| *** sbalukoff has joined #openstack-lbaas | 18:48 | |
| blogan | well there is, fast running tests :) | 18:48 |
| blogan | but that could hit the limit too with a lot of tests | 18:49 |
| pc-pothole | fast == less? | 18:49 |
| blogan | i know what you're saying but i have no idea how to solve both | 18:49 |
| blogan | no i mean, in our case, fast boots | 18:49 |
| pc-pothole | or with some of the api tests. i see that as a valid solution to clean that up | 18:49 |
| pc-pothole | ah | 18:49 |
| rm_work | yeah we could actually *test* more | 18:51 |
| pc-pothole | so much for the 'test throughly' mantra in openstack. Now it's test the minimum to be under time limit. Sounds like theres a bigger problem with the gate then with testing strats | 18:51 |
| rm_work | if we focused on just testing the API for each call | 18:51 |
| rm_work | and not testing "does an amp boot for each call" as well | 18:51 |
| rm_work | because 98% of the test is testing duplicated stuff (amp boot, network plug, flows, etc) | 18:51 |
| pc-pothole | im probably just rehashing things. can ignore me | 18:51 |
| pc-pothole | well yea, but im somewhat against the idea of one lb for all the tests too | 18:52 |
| pc-pothole | that doesnt give you true idea of the potential failures for certain calls | 18:52 |
| pc-pothole | in some cases its acceptable | 18:52 |
| pc-pothole | but not if were trying to do that for ALL the tests | 18:52 |
| pc-pothole | messy problem it is | 18:53 |
| pc-pothole | ill give example of why i see in some cases that could be potential problem. | 18:53 |
| pc-pothole | you create basic lb, it boots | 18:54 |
| pc-pothole | now what? | 18:54 |
| pc-pothole | you update stuff right? | 18:54 |
| pc-pothole | but, there happens to be a bug with configuration on create, not on update | 18:54 |
| pc-pothole | we will never hit that bug | 18:54 |
| blogan | yah i was just thinking of this | 18:54 |
| blogan | you're rgiht | 18:54 |
| rm_work | yes, you are not wrong | 18:54 |
| pc-pothole | messy fun | 18:54 |
| blogan | but since we are in the situation we are, is that edge case worth having these long running tests | 18:55 |
| pc-pothole | possible? | 18:55 |
| pc-pothole | possibly? | 18:55 |
| rm_work | I think the real solution is still moving a bunch of them to a 3rd-party CI | 18:55 |
| pc-pothole | super bad bug that we never catch | 18:55 |
| rm_work | in the long run | 18:55 |
| rm_work | we just need a bit of time to get that moving | 18:55 |
| pc-pothole | yea | 18:55 |
| blogan | well if we never catch it, that means a user never catches us, thus it doesnt exist :) | 18:56 |
| rm_work | as long as they're *working* for now... | 18:56 |
| pc-pothole | lol | 18:56 |
| rm_work | we shouldn't need to change anything yet | 18:56 |
| blogan | well the load balancer one takes 1:45 right now | 18:56 |
| blogan | that is pretty long | 18:56 |
| blogan | but at this point i dont care too much about it, just sucks we had to skip tests in the pool | 18:57 |
| pc-pothole | so all tests run in the gate right? | 18:57 |
| pc-pothole | we cant just decide on a subset, or a MVP for tests and those run in gate? | 18:58 |
| blogan | no | 18:58 |
| pc-pothole | o | 18:58 |
| blogan | we do have that for tests that neutron run for us | 18:58 |
| pc-pothole | ah | 18:58 |
| blogan | our co-gate with neutron | 18:58 |
| blogan | -minmal | 18:58 |
| blogan | is the name of the job | 18:58 |
| pc-pothole | i see | 18:58 |
| pc-pothole | well damn | 18:59 |
| pc-pothole | even our minimal is maximal | 18:59 |
| blogan | and johnsom got that to run under an hour | 18:59 |
| pc-pothole | awesome | 18:59 |
| *** barclaac|2 has quit IRC | 18:59 | |
| pc-pothole | :P | 18:59 |
| *** barclaac|2 has joined #openstack-lbaas | 18:59 | |
| johnsom | Right at an hour | 18:59 |
| johnsom | It basically calls the create tests on each of the high level parts. | 19:00 |
| pc-pothole | i suppose in a way this is a good problem. We know we have lots of coverage, and if we have to audit it, it will give us a better understanding of our tests and where we can improve. | 19:01 |
| blogan | and if we just run a few of our tests in our gate, then the other tests probably wont ever run, unless we do an honor system, which is a terrible idea | 19:01 |
| pc-pothole | messy fun it is | 19:01 |
| blogan | its a hard problem to solve | 19:01 |
| pc-pothole | indeed | 19:01 |
| blogan | and i aint got the smarts to solve it | 19:01 |
| pc-pothole | :P | 19:01 |
| blogan | im just a lowly country boy | 19:02 |
| pc-pothole | my brain has the dumbs | 19:02 |
| blogan | i think your brains has thumbs | 19:02 |
| pc-pothole | o.o | 19:02 |
| blogan | johnsom: opinions on using the noop drivers so we can re-enable the pool tests? | 19:03 |
| blogan | pretty sure the noop drivers would need some work as well | 19:03 |
| blogan | havent tested htem in a long time | 19:03 |
| xgerman | yeah, I am with pc-pothole more tests take time and infra should create a way for us to run them (e,g, nightly) | 19:03 |
| blogan | there are periodic tests | 19:03 |
| blogan | we dont haev them,b ut i see that zuul runs them | 19:04 |
| blogan | xgerman: opinion on noop driver? | 19:04 |
| xgerman | well, if we do periodic tests we can chainsaw all our tests | 19:04 |
| blogan | no! | 19:05 |
| blogan | not all of them | 19:05 |
| xgerman | no, not all of them | 19:05 |
| xgerman | but a substantial amount | 19:05 |
| blogan | its good to know when code pushed up breaks shit and stops it from being merged | 19:05 |
| blogan | yeah we just have some of the most common cases enabled for the check and gate | 19:06 |
| xgerman | yeah, so my preference is to just sue bigger boxes + wait longer for merge | 19:06 |
| xgerman | since both options have been taken from us we are grasping for straws | 19:07 |
| xgerman | now, we could set up a 4rd party CI with VT-X and run the test suite | 19:07 |
| xgerman | and have some very basic tests on the gate | 19:08 |
| xgerman | now, should those tests use the no-op driver or just be pep8 and uni tests and the rest happens on our 3rd party CI? | 19:09 |
| blogan | if we had CI that tests all the edge cases I'd feel a lot better about the noop drivers | 19:10 |
| blogan | noop drivers in a very basic set of tests | 19:10 |
| johnsom | blogan yeah, 99.999% sure the noop drivers needs some work | 19:13 |
| xgerman | ok, or we live with what we have and once container work we switch to that | 19:14 |
| blogan | im fine with that option | 19:14 |
| blogan | im pretty sure im giong to start writing a magnum compute driver soon bc it'll bother me | 19:14 |
| blogan | though im sure there will be a ton of bricks falling on my head before i get it to work | 19:15 |
| johnsom | Yeah, I hope to have a loop at the container patch today or tomorrow | 19:15 |
| xgerman | nice — I think I need to start playing with magnum more, too | 19:15 |
| blogan | well its really just a non-hotpluggable flow and network driver | 19:15 |
| blogan | but it does solve the containers issue of not being hot pluggable | 19:16 |
| johnsom | Let me noodle a bit on the test landscape. I think we need to break up the layers a bit and have each focus on testing what is needed in the layers. We still have other drivers, etc. | 19:16 |
| johnsom | I worry about neutron making a chance the breaks our hot-plug magic and we don't know about it until after the fact.... | 19:18 |
| blogan | johnsom: you mean if we switch to containers that can't hot plug the hot plug code won't be tested? | 19:18 |
| johnsom | Same thing with nova and config drive, etc | 19:18 |
| johnsom | Yes. I still assume that the hot plug issue with containers will be fixed at some point | 19:19 |
| *** minwang2 has joined #openstack-lbaas | 19:20 | |
| *** kobis has quit IRC | 19:20 | |
| blogan | johnsom: yeah i worry about all the drivers we are giong to have that we can't possibly test all the combination, but hot plug vs non hot pluggable is big enough to warrant 2 jobs in my mind | 19:20 |
| johnsom | Yep | 19:21 |
| *** barclaac has joined #openstack-lbaas | 19:23 | |
| rm_work | I feel like we need 3rd-party CI that tests EACH driver, really | 19:23 |
| rm_work | and honestly it seems there's a 1:1 mapping happening between compute and network drivers | 19:24 |
| johnsom | That is kind of the right answer | 19:24 |
| rm_work | for noop compute you really need noop network | 19:24 |
| rm_work | and same for container-compute | 19:24 |
| rm_work | and same for vm-compute | 19:24 |
| rm_work | we may have over-decomposed | 19:24 |
| blogan | network driver can be made comptue agnostic for the most part | 19:25 |
| rm_work | well, the different compute systems need different network handling <_< | 19:25 |
| blogan | it needs to handle both hot pluggable and not hot pluggable | 19:25 |
| rm_work | ah, i guess maybe it could be split out in that way | 19:25 |
| rm_work | hot-pluggable, cold-pluggable, no-plugging | 19:25 |
| blogan | ah cold-pluggable, is that a real term or did you just come up with it? | 19:26 |
| rm_work | it's the opposite of hot <_< | 19:26 |
| blogan | lol thanks | 19:26 |
| rm_work | it's a real term *now* if that's what you want to know :P | 19:26 |
| blogan | so difference between cold-pluggable and no-plugging? | 19:26 |
| *** barclaac|2 has quit IRC | 19:27 | |
| blogan | yeah ive wanted an alternative to non-hot-pluggable | 19:27 |
| rm_work | blogan: no-plugging == noop network | 19:27 |
| rm_work | which we need for example with noop-compute | 19:27 |
| blogan | but cold-pluggable implies its pluggable in some form | 19:27 |
| johnsom | Oye | 19:27 |
| blogan | lol | 19:27 |
| rm_work | or probably if we used a compute system where the neworking was just auto-handled internally | 19:27 |
| rm_work | heh | 19:27 |
| rm_work | i suppose so | 19:27 |
| blogan | that woudl be nice, but we can't just assume that | 19:27 |
| rm_work | so what you're saying is, it's probably still 1:1 | 19:28 |
| rm_work | since we can't assume anything | 19:28 |
| rm_work | lol | 19:28 |
| rm_work | or MAYBE something like ... n-1:n | 19:28 |
| rm_work | where we manage to reuse like one driver, once | 19:28 |
| rm_work | lol | 19:28 |
| blogan | there is some tight coupling right now | 19:29 |
| blogan | even without the non hot pluggable vs hot pluggable | 19:29 |
| blogan | but i cant remember | 19:29 |
| blogan | i put a note in the driver though | 19:29 |
| blogan | in the code | 19:29 |
| *** bana_k has quit IRC | 19:32 | |
| *** kobis has joined #openstack-lbaas | 19:35 | |
| *** woodster_ has joined #openstack-lbaas | 19:36 | |
| *** kobis has quit IRC | 19:53 | |
| *** kobis has joined #openstack-lbaas | 19:53 | |
| *** alejandrito has quit IRC | 19:53 | |
| *** kobis has quit IRC | 20:02 | |
| *** macjack has joined #openstack-lbaas | 20:03 | |
| *** kobis has joined #openstack-lbaas | 20:03 | |
| *** macjack has left #openstack-lbaas | 20:03 | |
| *** kobis has quit IRC | 20:14 | |
| *** bana_k has joined #openstack-lbaas | 20:40 | |
| *** SumitNaiksatam has quit IRC | 20:59 | |
| *** SumitNaiksatam has joined #openstack-lbaas | 20:59 | |
| *** mlavalle_ has quit IRC | 21:19 | |
| *** mlavalle has joined #openstack-lbaas | 21:20 | |
| *** Varun_Lodaya has quit IRC | 21:21 | |
| *** Varun_Lodaya has joined #openstack-lbaas | 21:23 | |
| *** TrevorV has quit IRC | 21:29 | |
| *** vivek-ebay has quit IRC | 21:35 | |
| *** vivek-ebay has joined #openstack-lbaas | 21:36 | |
| *** madhu_ak_ has quit IRC | 21:40 | |
| *** iraw- has quit IRC | 21:41 | |
| *** madhu_ak__ has joined #openstack-lbaas | 21:41 | |
| *** alejandrito has joined #openstack-lbaas | 21:41 | |
| *** iraw- has joined #openstack-lbaas | 21:42 | |
| *** SumitNaiksatam has quit IRC | 21:51 | |
| *** crc32 has quit IRC | 22:07 | |
| *** minwang2 has quit IRC | 22:29 | |
| *** alejandrito has quit IRC | 23:03 | |
| *** diogogmt has quit IRC | 23:16 | |
| *** crc32 has joined #openstack-lbaas | 23:34 | |
| *** vivek-ebay has quit IRC | 23:38 | |
| *** vivek-ebay has joined #openstack-lbaas | 23:38 | |
| *** sbalukoff has quit IRC | 23:47 | |
| *** xgerman has quit IRC | 23:55 | |
| *** harlowja has quit IRC | 23:56 | |
| *** rm_work has quit IRC | 23:56 | |
| *** davidlenwell has quit IRC | 23:57 | |
| *** harlowja has joined #openstack-lbaas | 23:57 | |
| *** davidlenwell has joined #openstack-lbaas | 23:58 | |
| *** rm_work has joined #openstack-lbaas | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!