| *** crc32 has quit IRC | 00:02 | |
| *** Brian has joined #openstack-lbaas | 00:13 | |
| *** minwang2 has quit IRC | 00:13 | |
| *** Brian is now known as Guest69615 | 00:13 | |
| *** Guest69615 has quit IRC | 00:18 | |
| *** Varun_Lodaya has quit IRC | 00:39 | |
| *** denghui has quit IRC | 00:44 | |
| *** denghui has joined #openstack-lbaas | 00:45 | |
| openstackgerrit | Michael Johnson proposed openstack/neutron-lbaas: Set Octavia as the reference LBaaSv2 driver https://review.openstack.org/223265 | 00:48 |
|---|---|---|
| johnsom | Ok, ^^^ sets up a minimal "happy path" set of tests for the neutron gate. Should run about an hour including the painful devstack setup time. | 00:49 |
| *** Brian has joined #openstack-lbaas | 01:04 | |
| *** Brian is now known as Guest44166 | 01:05 | |
| *** bana_k has quit IRC | 01:07 | |
| *** ganeshna has joined #openstack-lbaas | 01:07 | |
| *** ganeshna has quit IRC | 01:07 | |
| *** ganeshna has joined #openstack-lbaas | 01:08 | |
| *** ganeshna has quit IRC | 01:11 | |
| *** ganeshna has joined #openstack-lbaas | 01:12 | |
| *** ganeshna has quit IRC | 01:13 | |
| *** ganeshna has joined #openstack-lbaas | 01:14 | |
| *** ganeshna has quit IRC | 01:19 | |
| *** vivek-ebay has quit IRC | 01:27 | |
| xgerman | dougwig: wonder if what johnsom did is enough to get you to remove the -2 | 01:31 |
| rm_you | hmm | 01:31 |
| rm_you | johnsom: is that 5 classes? | 01:32 |
| xgerman | He said happy path and he measured the time on his box and calculated up | 01:36 |
| *** manishg has joined #openstack-lbaas | 01:41 | |
| *** denghui has quit IRC | 01:41 | |
| *** denghui has joined #openstack-lbaas | 01:42 | |
| *** manishg has quit IRC | 01:48 | |
| *** vivek-ebay has joined #openstack-lbaas | 01:58 | |
| *** woodster_ has quit IRC | 01:59 | |
| *** vivek-ebay has quit IRC | 02:06 | |
| *** diogogmt has joined #openstack-lbaas | 02:28 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Use admin context in octavia driver, improve logs https://review.openstack.org/225542 | 02:32 |
| xgerman | yeah… so if we want to be merged by tomorrow we better start now ;-) | 02:33 |
| blogan | the load balancer job still took 1 hr 40 mins | 02:35 |
| blogan | going to do a recheck for sanity though | 02:35 |
| blogan | since that last merge just happened | 02:36 |
| *** minwang2 has joined #openstack-lbaas | 02:37 | |
| blogan | i'm willing to bet 1.666 hrs is still not going to be good enough | 02:38 |
| xgerman | johnsom did a new minimal set | 02:39 |
| xgerman | the long lb test would be for things in the labs project | 02:39 |
| blogan | that just completed | 02:39 |
| xgerman | yep | 02:39 |
| blogan | what do you mean things in the labs project? | 02:39 |
| xgerman | aurocorrect lbaas | 02:40 |
| blogan | ahh | 02:40 |
| blogan | well i mean that still took 1.66 hrs | 02:40 |
| blogan | seems inconsistent though | 02:40 |
| xgerman | I though our queue can stay slow but we want the neutron queue to be peppy | 02:40 |
| xgerman | yeah, but I also think zulu runs on different hardware — fanatical RAX machines and well... | 02:41 |
| xgerman | zuul | 02:41 |
| blogan | im not sure which lbaas test will run for neutron gate | 02:42 |
| blogan | have to ask tomorrow though | 02:42 |
| rm_you | I feel like this a lot when I'm trying to debug something but the gate is getting in the way: https://www.youtube.com/watch?v=lg7MAacSPNM | 02:42 |
| blogan | gozer | 02:43 |
| xgerman | our HP internal CI is called Zuul | 02:44 |
| xgerman | I guess we are done for today | 02:45 |
| xgerman | :-) | 02:45 |
| blogan | well i did a recheck for sanity too | 02:53 |
| johnsom | There is no Dana | 02:54 |
| johnsom | Cool, so the subset I put together is in the minimal job, which I think is the upstream neutron gate. I targeted about an hour total run time (counting devstack) based on the previous runs. | 02:56 |
| johnsom | That should not change the neutron-lbaas gate(s) run time. | 02:56 |
| blogan | johnsom: was the previous minimal only running the load_balancers tests? | 02:57 |
| blogan | johnsom: or was it running lb, listeners, pools, hms, and members? | 02:57 |
| johnsom | Yes, it was running all of the tests that matched "load_balancer" | 02:57 |
| johnsom | Ah, before Adam changed it, good question. | 02:58 |
| blogan | matched in file name? dotted path? | 02:58 |
| johnsom | test name | 02:58 |
| johnsom | I think | 02:58 |
| blogan | yeah bc if you've changed it to run all of thsoe then i'd expect it to have a greater run time | 02:58 |
| johnsom | NO! | 02:58 |
| blogan | if previously it was only running test_load_balancers_admin and test_load_balancers_non_admin | 02:58 |
| johnsom | I changed it to just run certain tests. Basically one happy path test from each part of the LBaaS api | 02:59 |
| johnsom | Just specific tests | 02:59 |
| blogan | oh duh you're specifying method too | 02:59 |
| blogan | sorry | 02:59 |
| johnsom | Yep. It should be ~25-30 minutes | 02:59 |
| blogan | okay, i thought it was just the module being specified, but glad its not :) | 03:00 |
| blogan | if this recheck passes ill feel a bit better | 03:01 |
| johnsom | Yeah, Adam's minimal ran just tests with load_balancer in the name. It ran 24m under namespace | 03:02 |
| johnsom | Prior to that, with namespace and full test suite we ran ~57m so in theory, the new minimal will match the old full test suite | 03:03 |
| *** vivek-ebay has joined #openstack-lbaas | 03:04 | |
| *** crc32 has joined #openstack-lbaas | 03:04 | |
| johnsom | Ok, I'm going to check back in a bit | 03:05 |
| *** ganeshna has joined #openstack-lbaas | 03:06 | |
| rm_work | correct | 03:07 |
| *** ganeshna has quit IRC | 03:21 | |
| *** ganeshna has joined #openstack-lbaas | 03:23 | |
| *** minwang2 has quit IRC | 03:28 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/neutron-lbaas: Updated from global requirements https://review.openstack.org/225227 | 03:34 |
| *** ganeshna_ has joined #openstack-lbaas | 03:43 | |
| *** ganeshna has quit IRC | 03:43 | |
| *** vivek-ebay has quit IRC | 03:46 | |
| *** ganeshna_ has quit IRC | 03:49 | |
| *** minwang2 has joined #openstack-lbaas | 03:49 | |
| *** minwang2 has quit IRC | 04:02 | |
| *** Guest44166 has quit IRC | 04:04 | |
| *** KunalGandhi has quit IRC | 04:06 | |
| *** numans has joined #openstack-lbaas | 04:10 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Removed +x flag from alembic script https://review.openstack.org/222660 | 04:15 |
| *** Brian has joined #openstack-lbaas | 04:50 | |
| *** Brian is now known as Guest19801 | 04:50 | |
| *** Guest19801 has quit IRC | 04:55 | |
| johnsom | Well, that passed | 04:58 |
| *** Alex_Staf has joined #openstack-lbaas | 04:58 | |
| johnsom | It looks like increasing the retry count on the neutron call to delete a security group helped. I can see in the log where it hit six attempts before succeeding. The old retry limit was five. | 05:06 |
| johnsom | I don't see the database error in the log for the lb either. | 05:06 |
| *** crc32 has quit IRC | 05:17 | |
| *** ganeshna has joined #openstack-lbaas | 05:21 | |
| *** Brian has joined #openstack-lbaas | 05:25 | |
| *** Brian is now known as Guest38248 | 05:25 | |
| *** ganeshna has quit IRC | 05:26 | |
| openstackgerrit | venkata anil proposed openstack/neutron-lbaas: Update status of members from Haproxy statistics https://review.openstack.org/215027 | 06:28 |
| *** vjay10 has joined #openstack-lbaas | 07:05 | |
| *** kiran-r has joined #openstack-lbaas | 07:07 | |
| *** vjay10 has quit IRC | 07:09 | |
| *** kiran-r has quit IRC | 07:18 | |
| *** kiran-r has joined #openstack-lbaas | 07:23 | |
| *** codekobe has quit IRC | 07:41 | |
| *** dougwig has quit IRC | 07:41 | |
| *** xgerman has quit IRC | 07:41 | |
| *** ctracey has quit IRC | 07:41 | |
| openstackgerrit | Rishabh Das proposed openstack/neutron-lbaas: Fix typo in test names for test_health_monitors_non_admin.py Closes-Bug: 1498301 https://review.openstack.org/226192 | 07:45 |
| openstack | bug 1498301 in neutron "[neutron-lbaas] typo in health monitor test names" [Undecided,In progress] https://launchpad.net/bugs/1498301 - Assigned to Rishabh Das (rishabh5290) | 07:45 |
| openstackgerrit | venkata anil proposed openstack/neutron-lbaas: lbaasv2: validate tenant-id is non-empty in api https://review.openstack.org/219250 | 07:48 |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Adding new network driver for containers https://review.openstack.org/197858 | 08:10 |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Adding new network driver for containers https://review.openstack.org/197858 | 08:11 |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Adding new network driver for containers https://review.openstack.org/197858 | 08:11 |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Updates for containers functionality https://review.openstack.org/199954 | 08:11 |
| rm_work | pc-pothole: you on? | 08:12 |
| *** blallau has joined #openstack-lbaas | 08:13 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/neutron-lbaas: Updated from global requirements https://review.openstack.org/225227 | 09:19 |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: [LBaaS v2] Validate http_method and url_path for health monitor https://review.openstack.org/223651 | 09:21 |
| openstackgerrit | Rishabh Das proposed openstack/neutron-lbaas: Fix typo in test names for test_health_monitors_non_admin.py in API v2 https://review.openstack.org/226192 | 09:50 |
| *** xgerman has joined #openstack-lbaas | 09:52 | |
| *** dougwig has joined #openstack-lbaas | 09:53 | |
| *** bharathm has quit IRC | 09:54 | |
| *** ctracey has joined #openstack-lbaas | 09:54 | |
| *** codekobe has joined #openstack-lbaas | 09:55 | |
| *** Alex_Staf has quit IRC | 10:04 | |
| *** ganeshna has joined #openstack-lbaas | 10:27 | |
| *** ganeshna has quit IRC | 11:07 | |
| *** vjay10 has joined #openstack-lbaas | 11:37 | |
| *** vjay10 has quit IRC | 11:42 | |
| *** kiran-r has quit IRC | 12:45 | |
| *** _kiran_ has joined #openstack-lbaas | 12:45 | |
| *** _kiran_ is now known as kiran-r | 12:45 | |
| *** alejandrito has joined #openstack-lbaas | 12:47 | |
| *** manishg has joined #openstack-lbaas | 12:55 | |
| *** manishg has quit IRC | 13:01 | |
| *** Guest38248 has quit IRC | 13:16 | |
| *** _kiran_ has joined #openstack-lbaas | 13:16 | |
| *** kiran-r has quit IRC | 13:17 | |
| *** _kiran_ is now known as kiran-r | 13:17 | |
| *** kiran-r has quit IRC | 13:40 | |
| *** amotoki has joined #openstack-lbaas | 13:53 | |
| *** mlavalle has joined #openstack-lbaas | 14:02 | |
| *** manishg has joined #openstack-lbaas | 14:04 | |
| *** Brian has joined #openstack-lbaas | 14:17 | |
| *** Brian is now known as Guest48827 | 14:17 | |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: Improve resource cleanup in API v2 tests https://review.openstack.org/226344 | 14:22 |
| *** Guest48827 has quit IRC | 14:22 | |
| *** vivek-ebay has joined #openstack-lbaas | 14:34 | |
| *** denghui has quit IRC | 14:38 | |
| *** bharathm has joined #openstack-lbaas | 14:38 | |
| *** vivek-ebay has quit IRC | 14:38 | |
| *** vivek-ebay has joined #openstack-lbaas | 14:43 | |
| johnsom | blogan What is your thought on the ref-driver patch? Should we give it another go with the merge gate? The recheck after our patches merged looked good last night. | 14:45 |
| *** minwang2 has joined #openstack-lbaas | 14:55 | |
| xgerman | we better get that show on the road with the gate queue what it is | 14:55 |
| *** vivek-ebay has quit IRC | 14:58 | |
| *** blallau has quit IRC | 15:01 | |
| *** kiran-r has joined #openstack-lbaas | 15:12 | |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: [LBaaS v2] Improve API tests performance https://review.openstack.org/226370 | 15:12 |
| openstackgerrit | Elena Ezhova proposed openstack/neutron-lbaas: Improve resource cleanup in API v2 tests https://review.openstack.org/226344 | 15:14 |
| *** ganeshna has joined #openstack-lbaas | 15:16 | |
| *** blogan_ has joined #openstack-lbaas | 15:18 | |
| *** TrevorV has joined #openstack-lbaas | 15:19 | |
| blogan_ | johnsom: to answer your question, has dougwig or mestery said anything? | 15:19 |
| johnsom | Yeah, we need to make a decision today | 15:20 |
| xgerman | blogan is there something which will increase our confidence in the gate | 15:23 |
| xgerman | or should e just wait until our name comes up? | 15:23 |
| xgerman | (I think there is only one way to find out if the gate is better and the earlier we start the better) | 15:24 |
| *** akamyshnikova has quit IRC | 15:24 | |
| *** akamyshnikova has joined #openstack-lbaas | 15:25 | |
| johnsom | Agreed. We need to get in the gate queue early if we want to try this out | 15:25 |
| blogan_ | well dougwig has it -2'ed right now anyway, he needs to chime in on it | 15:25 |
| xgerman | so paging dougwig... | 15:25 |
| *** diogogmt has quit IRC | 15:26 | |
| xgerman | also we can always back the patch out if it doesn’t work as expected | 15:26 |
| blogan_ | is there anyway to test the minimal job out without merging this? so we know exactly how long it takes for neutron merges? | 15:26 |
| xgerman | if we split the minimal job out and merge we still won’t know if it works with octavia | 15:27 |
| johnsom | You can run it local. We also run those same jobs during the main gate run, so if you want to see run times just look at our last gate run. | 15:27 |
| *** manishg has quit IRC | 15:36 | |
| *** diogogmt has joined #openstack-lbaas | 15:36 | |
| johnsom | My guess, the "minimal" that runs for neutron will run somewhere around 3919.85 seconds or 65.33 minutes including devstack setup time. | 15:36 |
| johnsom | Based on our last gate run times | 15:37 |
| *** manishg has joined #openstack-lbaas | 15:39 | |
| *** manishg has quit IRC | 15:42 | |
| *** KunalGandhi has joined #openstack-lbaas | 15:42 | |
| johnsom | That is pretty much inline with what we were running previously | 15:44 |
| blogan_ | well from what i can tell it currently runs at 30 mins | 15:46 |
| blogan_ | but 65 minutes wouldn't make us the long pole in the tent anyway | 15:46 |
| *** vivek-ebay has joined #openstack-lbaas | 15:46 | |
| xgerman | yep | 15:46 |
| blogan_ | bc other jobs take just as long if not longer | 15:46 |
| xgerman | yeah, now are we happy with just running the happy path? | 15:47 |
| xgerman | I am but I also expect us to get containers going before M ends so we can run the full test suite | 15:48 |
| *** KunalGan_ has joined #openstack-lbaas | 15:48 | |
| *** KunalGandhi has quit IRC | 15:49 | |
| blogan_ | yeah i expect that too | 15:52 |
| johnsom | yeah, I took a cross section of happy path tests. I think that is good for the neutron gate | 15:52 |
| johnsom | We can optimize test run times in M | 15:53 |
| blogan_ | yeah i'm fine with that too, just worried about the run time, bc there are tests for each entity, so thats 5 lbs that get created, each taking 5-8 mins boot time, so worst case scenario it takes 40 mins for the boot, plus 30 mins for devstack, +5 minutes for child entity creation, around 75 minutes worst case scenario? | 15:54 |
| *** f13o has joined #openstack-lbaas | 15:55 | |
| blogan_ | does hot plugging a VM without the vt-x enabled take longer? | 15:55 |
| johnsom | Not that I have noticed | 15:55 |
| *** woodster_ has joined #openstack-lbaas | 15:56 | |
| johnsom | I think 75 minutes would be the absolute maximum. I put some cushion in my numbers too | 15:56 |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Adding new network driver for containers https://review.openstack.org/197858 | 15:58 |
| openstackgerrit | Phillip Toohill proposed openstack/octavia: Updates for containers functionality https://review.openstack.org/199954 | 15:58 |
| xgerman | the Neutron gate is really to protect us from crazy changes over there — so it’s how much we want to play with fire ;-) | 15:58 |
| *** Aish has joined #openstack-lbaas | 16:00 | |
| *** blogan_ has quit IRC | 16:01 | |
| *** blogan_ has joined #openstack-lbaas | 16:02 | |
| blogan_ | grrr | 16:02 |
| blogan_ | yall get any of my messages 10:59 and later? | 16:02 |
| blogan_ | well *:59 for yall | 16:02 |
| *** numans has quit IRC | 16:03 | |
| *** whydidyoustealmy is now known as shakamunyi | 16:03 | |
| xgerman | no | 16:04 |
| blogan_ | welp i feel less apprehensive about it, but i won't +A something with a -2 on it until they've got a chance to comment | 16:05 |
| blogan_ | so dougwig needs to come out of hibernation | 16:05 |
| * dougwig stretches. | 16:05 | |
| dougwig | the -2 was just because i wanted to hear from all the NO votes today. this needs to be a team decision to move forward. | 16:06 |
| blogan_ | well im still a bit nervous, but im sure i'd be nervous if it all went smoothly anyway | 16:09 |
| *** vivek-eb_ has joined #openstack-lbaas | 16:14 | |
| dougwig | so i'll take that as a yes from johnsom and blogan_ . rm_work ? pc-pothole ? who else voted no? | 16:15 |
| johnsom | mestery was the other no | 16:16 |
| pc-pothole | I was no only because of the testing | 16:16 |
| pc-pothole | The code itself is solid enough to be comparable to the current ref driver IMO. | 16:16 |
| *** vivek-ebay has quit IRC | 16:17 | |
| *** Brian has joined #openstack-lbaas | 16:19 | |
| *** Brian is now known as Guest56742 | 16:20 | |
| dougwig | mestery: ping, i think we might have a reversal here. | 16:20 |
| *** Guest56742 has quit IRC | 16:24 | |
| *** enikanorov_ has quit IRC | 16:26 | |
| xgerman | well, he is a lame duck... | 16:26 |
| *** enikanorov_ has joined #openstack-lbaas | 16:26 | |
| *** Aish has quit IRC | 16:27 | |
| johnsom | He is probably out partying now that his last neutron meeting is over.... | 16:27 |
| xgerman | wasn’t he fond of gin? | 16:30 |
| *** kiran-r is now known as _kiran_ | 16:35 | |
| *** manishg has joined #openstack-lbaas | 16:35 | |
| *** _kiran_ has quit IRC | 16:40 | |
| *** kiran-r has joined #openstack-lbaas | 16:40 | |
| *** kiran-r has quit IRC | 16:40 | |
| *** kiranr_ has joined #openstack-lbaas | 16:40 | |
| *** kiranr_ has quit IRC | 16:40 | |
| blogan_ | well rm_work is preoccupied, mestery appears to be as well | 16:40 |
| *** vivek-eb_ has quit IRC | 16:40 | |
| blogan_ | dougwig: i gotta leave here in a second, im thinking everyone will be okay with this | 16:41 |
| blogan_ | and in the interest of getting this thing in the gate | 16:41 |
| blogan_ | i can +A it right now | 16:41 |
| blogan_ | if someone objects we can cancel it | 16:41 |
| mestery | dougwig blogan_: I'm back :) | 16:42 |
| mestery | Octavia is in now? | 16:42 |
| blogan_ | mestery: i think we all feel more confident in it now | 16:43 |
| xgerman | yep, we cut the Neutron gate down + increased retries/timeouts helped a lot | 16:44 |
| *** Aish has joined #openstack-lbaas | 16:51 | |
| *** johnsom_ has joined #openstack-lbaas | 16:52 | |
| blogan_ | welp i gotta go, dougwig has to remove his -2 anyway so he can +A it or mestery | 16:54 |
| mestery | blogan_: Ack | 16:54 |
| *** bana_k has joined #openstack-lbaas | 17:01 | |
| *** vivek-ebay has joined #openstack-lbaas | 17:03 | |
| *** ganeshna has quit IRC | 17:10 | |
| *** ganeshna has joined #openstack-lbaas | 17:10 | |
| *** Brian has joined #openstack-lbaas | 17:20 | |
| *** Brian is now known as Guest64351 | 17:20 | |
| *** KunalGan_ has quit IRC | 17:21 | |
| *** Guest64351 has quit IRC | 17:25 | |
| *** kiran-r has joined #openstack-lbaas | 17:32 | |
| *** kiranr has joined #openstack-lbaas | 17:33 | |
| *** crc32 has joined #openstack-lbaas | 17:34 | |
| *** johnsom_ has quit IRC | 17:41 | |
| davidlenwell | So I'm looking for something and thought I'd just ask here and see if anyone can point me in the right direction.. where is the code that removes an lb from neutron? does octavia just set the status to GONE or whatever and neutron cleans it up on its own or is some other action required to make neutron clean things up? | 17:51 |
| johnsom | We call nova delete once we are done with an instance | 17:52 |
| johnsom | Oh, you are talking about the neutron DB side of the shop | 17:54 |
| davidlenwell | yes | 17:54 |
| davidlenwell | johnsom: that is what I mean.. sorry I wasn't clear.. | 17:55 |
| johnsom | No, I'm sorry, distracted so thought you were asking about the instance deletes. | 17:55 |
| johnsom | Once they call delete in lbaas it will pass that back to Octavia. Octavia will delete and set status to DELETED which the status sync will pickup | 17:56 |
| johnsom | Not that I'm really familiar with that part of the code, but I'm pretty sure that is how it works. | 17:57 |
| davidlenwell | well that really does answer my question.. if i can trust that neutron status sync will clean it up once the status is set to deleted | 17:58 |
| davidlenwell | thanks johnsom | 17:58 |
| johnsom | Yeah, we are working to improve the status sync in the M cycle. There is code already in progress | 17:59 |
| *** ganeshna has quit IRC | 18:06 | |
| openstackgerrit | min wang proposed openstack/octavia: Add cert tracking and rotating in Housekeeping https://review.openstack.org/215359 | 18:14 |
| bana_k | I was checking out the create listener and we have this option | 18:19 |
| bana_k | --connection-limit | 18:19 |
| bana_k | but if I give any values to this I don't see any difference on the create rest call body | 18:19 |
| bana_k | -d '{"listener": {"protocol": "HTTP", "name": "test_listener", "admin_state_up": true, "tenant_id": "asdadsgfad", "protocol_port": "80", "loadbalancer_id": "e3e0b1d2-12b2-4855-a5cb-6faf121b2d10"}}' | 18:20 |
| mestery | Looks like the LBaaS Octavia patch is a go once dougwig removes his -2. | 18:27 |
| xgerman | yeah, waiting for dougwig to remove his -2 is almost a abased as waiting for the gate queue | 18:27 |
| johnsom | Yeah, it is all resting in dougwig's hands.... | 18:27 |
| xgerman | abased = as bad | 18:28 |
| *** kiran-r has quit IRC | 18:29 | |
| dougwig | As bad as the gate? Man. | 18:29 |
| dougwig | 30 minutes. Almost back to office. | 18:30 |
| *** manishg has quit IRC | 18:32 | |
| *** manishg has joined #openstack-lbaas | 18:34 | |
| *** minwang2 has quit IRC | 18:34 | |
| mestery | rofl | 18:37 |
| *** ganeshna has joined #openstack-lbaas | 18:41 | |
| pc-pothole | For those interested: https://review.openstack.org/#/c/199954/ This can be tested with devstack now. Just need to update config for the plugin and network driver since its not default. But any eyes on it is a good thing. | 18:43 |
| *** ganeshna has quit IRC | 18:43 | |
| *** ganeshna has joined #openstack-lbaas | 18:44 | |
| *** ganeshna has quit IRC | 18:47 | |
| *** ganeshna has joined #openstack-lbaas | 18:47 | |
| *** ganeshna has quit IRC | 18:51 | |
| *** vivek-ebay has quit IRC | 18:55 | |
| *** manishg has quit IRC | 19:01 | |
| johnsom | pc-pothole Definitely interested. I probably won't have time to take a pass over it until Thursday, but will put it on my list | 19:01 |
| *** KunalGandhi has joined #openstack-lbaas | 19:02 | |
| pc-pothole | Thats fine, thank you | 19:02 |
| *** manishg has joined #openstack-lbaas | 19:10 | |
| blogan_ | davidlenwell: the octavia driver spins up a thread whenever a delete comes in and after sending the delete action to octavia's api, the thread polls that load balancer until its deleted from octavia. once it sees its deleted, it then deletes it from neutron's db | 19:14 |
| *** vivek-ebay has joined #openstack-lbaas | 19:19 | |
| *** bharathm has quit IRC | 19:20 | |
| *** Brian has joined #openstack-lbaas | 19:21 | |
| *** Brian is now known as Guest22207 | 19:22 | |
| *** Aish has quit IRC | 19:26 | |
| *** mlavalle has quit IRC | 19:26 | |
| *** Guest22207 has quit IRC | 19:26 | |
| *** mlavalle has joined #openstack-lbaas | 19:28 | |
| blogan_ | gate is running pretty fast today | 19:30 |
| johnsom | I hope so. We have a true trial by fire here. There is a minimal gate going too | 19:31 |
| blogan_ | well if this gate run fails it will make my vote a no on pushing forward with this | 19:31 |
| johnsom | Yeah, me too. That is a given | 19:32 |
| *** vivek-ebay has quit IRC | 19:32 | |
| * blogan_ knocks on wood while crossing fingers around a lucky rabbits food with a four leaf clover glued to it | 19:33 | |
| * johnsom hands blogan a rabbits foot | 19:33 | |
| TrevorV | all while he jumps around black cats | 19:33 |
| xgerman | don’t jinx us | 19:37 |
| *** manishg has quit IRC | 19:38 | |
| *** manishg has joined #openstack-lbaas | 19:38 | |
| *** vivek-ebay has joined #openstack-lbaas | 19:39 | |
| *** blogan_ has quit IRC | 19:39 | |
| johnsom | How many songs does it take to build a devstack? | 19:43 |
| xgerman | 99 bottles of beer on the wall;? | 19:43 |
| *** blogan_ has joined #openstack-lbaas | 19:43 | |
| johnsom | I like yours better | 19:43 |
| * xgerman glued to the zulu status | 19:44 | |
| xgerman | and autocorrect stroke again | 19:44 |
| *** manishg has quit IRC | 19:44 | |
| * johnsom watching lb, listener, minimal, and zuul across the screens | 19:45 | |
| *** manishg has joined #openstack-lbaas | 19:45 | |
| xgerman | I am more of a big picture kind-a guy | 19:45 |
| *** bradjones has joined #openstack-lbaas | 19:46 | |
| *** bradjones has quit IRC | 19:46 | |
| *** bradjones has joined #openstack-lbaas | 19:46 | |
| johnsom | lb devstack is built, going into the tests | 19:47 |
| johnsom | minimal too | 19:47 |
| johnsom | ~20 minutes just to build the devstack | 19:49 |
| *** vivek-ebay has quit IRC | 19:57 | |
| *** minwang2 has joined #openstack-lbaas | 19:57 | |
| *** bharathm has joined #openstack-lbaas | 19:58 | |
| *** manishg_ has joined #openstack-lbaas | 19:59 | |
| *** manishg has quit IRC | 19:59 | |
| *** Aish has joined #openstack-lbaas | 20:01 | |
| *** blogan_ has quit IRC | 20:04 | |
| *** manishg_ has quit IRC | 20:04 | |
| *** manishg has joined #openstack-lbaas | 20:04 | |
| *** blogan_ has joined #openstack-lbaas | 20:06 | |
| johnsom | member tests passed | 20:07 |
| johnsom | healthmonitor tests passed | 20:07 |
| blogan_ | now just the last 2, the ones that were failing consistently | 20:09 |
| johnsom | listener passed | 20:10 |
| blogan_ | load balancer is creating the final lb | 20:11 |
| blogan_ | grr nvm, it creates a few in the actual tests | 20:11 |
| blogan_ | so itll still be a while | 20:11 |
| blogan_ | yeah it'll be another 45 mins before it completes | 20:14 |
| blogan_ | at least listener passed though, that was failing consistently | 20:15 |
| *** vivek-ebay has joined #openstack-lbaas | 20:19 | |
| johnsom | What? aborted? | 20:20 |
| xgerman | hbuh? | 20:20 |
| johnsom | All of the jobs aborted | 20:20 |
| xgerman | back to square 1? we broke the VM? | 20:20 |
| johnsom | No, not us, someone aborted the jobs | 20:21 |
| xgerman | this thriller has twists | 20:21 |
| johnsom | The whole thing just started over | 20:21 |
| johnsom | Jenkins says Build was aborted, Aborted by anonymous | 20:22 |
| blogan_ | wtf | 20:23 |
| johnsom | Our run and the minimal run both got killed | 20:23 |
| blogan_ | maybe something merged in neutron that required these to run again | 20:24 |
| johnsom | Ah, yeah, there were some above us in zuul that are now gone | 20:24 |
| blogan_ | well thats unfortunate | 20:25 |
| blogan_ | but being the optimistic person that i am, its an even great sanity check :) | 20:25 |
| xgerman | just another roadblock | 20:25 |
| xgerman | this will be a nail biter | 20:25 |
| blogan_ | more like a speed bump | 20:25 |
| blogan_ | that is on a fault line | 20:25 |
| xgerman | I know all about fault lines ;-) | 20:26 |
| blogan_ | if you didnt, you'd be sinking into the ocean when you found out about them | 20:27 |
| xgerman | right now I am like 5 blocks from the ocean but I have it on good authority I am living on future ocean front property | 20:28 |
| *** kiranr has quit IRC | 20:32 | |
| blogan_ | sounds like a wise investment | 20:34 |
| blogan_ | other than hoping that the people and homes closer to the ocean sink into it | 20:35 |
| dougwig | that's the integrated gate. when someone ahead of you fails, it resets. | 20:48 |
| *** diogogmt has quit IRC | 20:52 | |
| blogan_ | so if ours fails it fails everyone behind us? or just those that have an gate job on us? | 20:55 |
| blogan_ | dougwig: ^ | 20:55 |
| dougwig | it stalls everyone behind us, yes. | 20:55 |
| dougwig | and every reset has to wait for our jobs, so if we're >60 mins, we're making the overall gate throughput much slower. | 20:56 |
| blogan_ | welp we are >60 mins | 20:58 |
| blogan_ | why 60 mins? | 20:58 |
| *** diogogmt has joined #openstack-lbaas | 20:58 | |
| johnsom | listener passed again | 21:03 |
| dougwig | blogan_: that's about the longest job now. | 21:04 |
| xgerman | we are competitive — we can be longer | 21:05 |
| blogan_ | dougwig: okay, well we shall take that record now, though we may look at skipping some of the tests until we can get containers implemented | 21:05 |
| blogan_ | some of the load balancer tests | 21:05 |
| blogan_ | but obviously that can wait | 21:06 |
| xgerman | yeah, babysteps | 21:06 |
| dougwig | blogan_: https://review.openstack.org/#/c/226105/ | 21:07 |
| xgerman | those people… we will never catch a break | 21:07 |
| xgerman | also see how one core can +A thjere | 21:07 |
| blogan_ | dougwig: if that merges, that'll fail our job then no? | 21:08 |
| johnsom | Yeah, I can optimize the minimal more. | 21:08 |
| johnsom | We don't have anything in front of us now, so I don't think anything else should block us | 21:08 |
| blogan_ | minimal should be easy to get under 1 hr, we're going to have to skip a good amoutn of tests for our load_balancer tests | 21:09 |
| xgerman | also that idea to restrict testing is sort of trying to solve the wrong problem | 21:10 |
| johnsom | Oh, wow, yeah, that would impact our neutron-lbaas tempest jobs too I think. | 21:10 |
| xgerman | where does infra think tests should run? | 21:10 |
| johnsom | We can meet that if we either re-work the lb tests or split them more | 21:11 |
| xgerman | well, we can always get rid of tests but even with the namespace driver we were pushing an hour | 21:12 |
| johnsom | Yeah, namespace was over an hour | 21:12 |
| johnsom | It definitely would have failed with that change | 21:13 |
| blogan_ | i know restricting tests is solving the wrong problem, but its a short term solution | 21:13 |
| xgerman | well, the patch doesn’t sat short term | 21:13 |
| *** TrevorV has quit IRC | 21:13 | |
| blogan_ | i kinda assume any skipped tests is considered temporary | 21:14 |
| blogan_ | unless the test is just deleted, which these would not be | 21:14 |
| xgerman | yep, but if they restrict us to 60 minutes we will never be able to run all our tests | 21:14 |
| blogan_ | this is true | 21:14 |
| xgerman | (unless we make dozens of jobs) and I wonder how that helps them | 21:14 |
| blogan_ | i suppose they would like to keep the gate flowing though, which i understand | 21:14 |
| blogan_ | well i bet they're hoping this makes developers rethink how they write their tests, to make more efficient tests | 21:15 |
| blogan_ | which we coudl certainly do, but then again we'd be modifying our in tree tempest as well | 21:15 |
| johnsom | Well, I'm going to throw in a -1 and ask for nested virtualization. | 21:15 |
| blogan_ | and its a lot of work | 21:15 |
| blogan_ | johnsom: thats a good idea | 21:15 |
| blogan_ | at least they'll see your request while +A'ing it | 21:16 |
| johnsom | Heck over half of the 60 minutes is just getting the host ready to test on it. | 21:16 |
| blogan_ | yeah no kidding | 21:16 |
| blogan_ | so really our tests need to run in 30 mins | 21:16 |
| xgerman | yep, and I think having more slower tests outweighs having less tests and taking a long time | 21:17 |
| johnsom | blogan_ actually a lot less than 30 mins to account for the host variability | 21:17 |
| johnsom | They have to be crazy enabling that before Liberty is done. Just looking at a typical neutron test run they would all fail. | 21:19 |
| xgerman | well, I added a nastygram in comments as well | 21:21 |
| johnsom | Just look at this run: https://review.openstack.org/#/c/213563/ | 21:22 |
| johnsom | All of these would fail: gate-tempest-dsvm-neutron-full, gate-grenade-dsvm-neutron, gate-tempest-dsvm-neutron-dvr, gate-tempest-dsvm-neutron-linuxbridge, gate-neutron-lbaasv2-dsvm-api(namespace), gate-tempest-dsvm-neutron-multinode-full, and gate-tempest-dsvm-neutron-dvr-multinode-full | 21:23 |
| xgerman | yeah, quality has a cost | 21:23 |
| blogan_ | isnt nodepool a neutron thing? so will this only affect nodepool? | 21:24 |
| blogan_ | i dont know the jobs well | 21:24 |
| xgerman | imagine we wouldn’t have tests (like say other *aaS) we would be done | 21:24 |
| blogan_ | so im probably dumb on this | 21:24 |
| xgerman | yeah, maybe dougwig just wanted to distract us ;-) | 21:25 |
| dougwig | simmer down, let's let jeblair answer the questions. :) | 21:26 |
| xgerman | heh, I was getting my pitchfork ready | 21:26 |
| blogan_ | i feel like this is only for nodepool tests, not everyones | 21:27 |
| blogan_ | but i could be wrong | 21:27 |
| blogan_ | and yeah nodepool is an infra thing, that i barely read about that i hardly know anything about | 21:27 |
| dougwig | nodepool dsvm == devstack-virtual-machine | 21:27 |
| blogan_ | yeah | 21:27 |
| blogan_ | but they have tests for that right? and it looks like that is just modifying nodepools tests, or does that yaml control all jobs? | 21:27 |
| *** alejandrito has quit IRC | 21:32 | |
| *** alejandrito has joined #openstack-lbaas | 21:33 | |
| xgerman | it seems we don’t need the pitchfork today | 21:33 |
| dougwig | blogan_ is correct. | 21:36 |
| xgerman | now I feel like a fool | 21:36 |
| dougwig | xgerman: welcome to my everyday. | 21:37 |
| blogan_ | yes i dont look like the fool this time! | 21:43 |
| blogan_ | well other than my normal foolish look | 21:43 |
| *** manishg has quit IRC | 21:48 | |
| *** manishg has joined #openstack-lbaas | 21:48 | |
| xgerman | only load balancer left | 21:52 |
| johnsom | Passed! | 21:56 |
| johnsom | Twice even | 21:56 |
| openstackgerrit | Merged openstack/neutron-lbaas: Set Octavia as the reference LBaaSv2 driver https://review.openstack.org/223265 | 21:56 |
| * xgerman dances in joy | 21:57 | |
| johnsom | Ok, bad news(though not that bad) is the minimal regexed in a bad way and is running way too many tests | 21:58 |
| johnsom | I can fix that quickly | 21:58 |
| minwang2 | we finally get this merged, it is huge!!! | 21:59 |
| openstackgerrit | Michael Johnson proposed openstack/neutron-lbaas: The minimal test suite is running too many tests https://review.openstack.org/226574 | 22:00 |
| johnsom | blogan rm_work dougwig ^^^ | 22:00 |
| *** Aish has quit IRC | 22:03 | |
| xgerman | it’s really just pretty_ox vs. to syntax | 22:05 |
| *** Aish has joined #openstack-lbaas | 22:06 | |
| *** jerrygb has joined #openstack-lbaas | 22:08 | |
| *** davidlenwell has quit IRC | 22:08 | |
| johnsom | Yeah, that is annoying. tox, the way I originally tested it isn't greedy, but when run through the full apiv2 pretty_tox stuff it is greedy | 22:09 |
| *** davidlenwell has joined #openstack-lbaas | 22:09 | |
| blogan_ | im tempted to +A it | 22:15 |
| xgerman | yep, infra style | 22:15 |
| blogan_ | of course this means it'll be another 2 hours before it merges | 22:16 |
| blogan_ | at beast | 22:16 |
| blogan_ | at best | 22:16 |
| xgerman | but I won’t be watching this time — too much excitement already ;-) | 22:16 |
| johnsom | We aren't failing the minimal jobs, it's just still running 111 out of the 217 instead of the 7 intended | 22:17 |
| blogan_ | and there's a neutron patch in the gate | 22:17 |
| blogan_ | yeah but wont it take longer than 2 hours? | 22:17 |
| johnsom | Yeah, I caught it by watching another neutron job | 22:17 |
| blogan_ | oh it completed? | 22:17 |
| johnsom | yes | 22:17 |
| blogan_ | oh okay well thats good then | 22:17 |
| johnsom | Just too darn slow | 22:18 |
| blogan_ | was afraid it'd cause failures all around | 22:18 |
| johnsom | No, it didn't fail | 22:18 |
| johnsom | https://review.openstack.org/#/c/225927/ | 22:18 |
| blogan_ | ha pretty close! | 22:19 |
| xgerman | well, nobody should me merging stuff anyway | 22:20 |
| johnsom | Yeah. I was not happy when I saw that it was running more than the seven | 22:20 |
| *** mlavalle_ has joined #openstack-lbaas | 22:34 | |
| *** mlavalle has quit IRC | 22:36 | |
| *** mlavalle has joined #openstack-lbaas | 22:36 | |
| *** mlavalle_ has quit IRC | 22:39 | |
| rm_work | lol | 22:40 |
| rm_work | yes it is very open | 22:40 |
| rm_work | remember I just put "load_balancer" originally | 22:41 |
| rm_work | it basically shoves .* on both sides | 22:41 |
| rm_work | for great lulz | 22:41 |
| xgerman | well, we fixed one end | 22:41 |
| johnsom | tox working with just the full method path threw me off | 22:41 |
| *** rm_you has quit IRC | 22:55 | |
| *** alejandrito has quit IRC | 22:55 | |
| *** blogan_ has quit IRC | 22:56 | |
| openstackgerrit | Brandon Logan proposed openstack/neutron-lbaas: Fix scenario tests for octavia driver https://review.openstack.org/226587 | 23:02 |
| *** manishg has quit IRC | 23:03 | |
| *** manishg has joined #openstack-lbaas | 23:03 | |
| *** blogan_ has joined #openstack-lbaas | 23:11 | |
| blogan | that should fix the tempest tests ^ | 23:27 |
| blogan | but we dont have to merge | 23:27 |
| *** crc32 has quit IRC | 23:34 | |
| blogan | grrr vt-x strikes the scenario tests now! | 23:38 |
| openstackgerrit | Brandon Logan proposed openstack/neutron-lbaas: Fix scenario tests for octavia driver https://review.openstack.org/226587 | 23:41 |
| rm_work | T_T | 23:41 |
| dougwig | stupid question, can we use containers for the amps in these tests? | 23:41 |
| rm_work | dougwig: we'd need a containers driver | 23:42 |
| rm_work | <_< | 23:42 |
| rm_work | and then we would be testing a different code path | 23:42 |
| rm_work | so ... yes? | 23:42 |
| rm_work | but i don't know that it solves our problem | 23:42 |
| dougwig | if containers was our default, it would. :) | 23:42 |
| rm_work | heh | 23:42 |
| rm_work | but then how would we test the other drivers? | 23:42 |
| rm_work | we just ... don't | 23:43 |
| rm_work | ? | 23:43 |
| johnsom | dougwig I think we all agree that containers are an answer. We just need more time to make containers work. | 23:43 |
| dougwig | johnsom: gotcha. | 23:43 |
| johnsom | The issue is working around the network hot-plug limitations containers bring | 23:44 |
| dougwig | if you go to any con lately, containers are the question, answer, and religious nirvana, too. | 23:44 |
| rm_work | i mean, do we spin up Magnum too? | 23:44 |
| rm_work | or is there like ... nova-containers or something? | 23:44 |
| xgerman | yep, hence we are really excited about them | 23:44 |
| johnsom | Yep. | 23:44 |
| rm_work | because obviously nova needs their own version of EVERYTHING | 23:44 |
| rm_work | dug | 23:44 |
| rm_work | *duh | 23:44 |
| xgerman | magnum? | 23:44 |
| rm_work | Magnum is the Openstack CaaS project | 23:44 |
| johnsom | More to the point, we are told we are really excited about them... | 23:44 |
| rm_work | not sure what nova-containers is | 23:45 |
| rm_work | I think nova-networking : neutron :: nova-containers : magnum | 23:45 |
| xgerman | docker!! | 23:45 |
| blogan | yes we spin up magnum | 23:46 |
| dougwig | and then there's kuryr | 23:46 |
| blogan | rm_work: even the CaaS we are goin got use internally will move to be magnum | 23:46 |
| johnsom | Yeah, I started to read about kuryr but didn't learn much | 23:46 |
| xgerman | + magnum will have VT_X enabled boxes in infra | 23:47 |
| xgerman | so they are clearly the answer ;-) | 23:47 |
| rm_work | heh | 23:50 |
| johnsom | dougwig blogan rm_work Can you guys take a look at the release note section I added: https://wiki.openstack.org/wiki/ReleaseNotes/Liberty | 23:50 |
| johnsom | Under Neutron "Performance Considerations" about the vt-x issue | 23:51 |
| johnsom | Anyone can comment, I just wanted to highlight it for those folks | 23:53 |
| rm_work | looks good i think | 23:54 |
| *** ganeshna has joined #openstack-lbaas | 23:56 | |
| *** manishg has quit IRC | 23:57 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!