| *** chlong_ has quit IRC | 00:03 | |
| *** Aish has quit IRC | 00:05 | |
| *** Aish has joined #openstack-lbaas | 00:06 | |
| openstackgerrit | German Eichberger proposed openstack/neutron-lbaas: Adds Cascade Delete for LoadBalancers to Octavia Driver https://review.openstack.org/284340 | 00:07 |
|---|---|---|
| xgerman | ^^ I got end to end working... | 00:08 |
| sbalukoff | Nice! | 00:10 |
| xgerman | yeah, LBaS V2 is a mess | 00:11 |
| sbalukoff | Yes it is. | 00:11 |
| xgerman | now, we need some good reviews... | 00:12 |
| xgerman | and I need to go back to my internal fires… in which somebody poured gasoline while I was distracted | 00:12 |
| *** yamamoto_ has joined #openstack-lbaas | 00:18 | |
| *** piet has joined #openstack-lbaas | 00:20 | |
| *** yamamoto_ has quit IRC | 00:23 | |
| sbalukoff | D'oh. | 00:25 |
| sbalukoff | Good luck, xgerman. | 00:25 |
| *** SumitNaiksatam has joined #openstack-lbaas | 00:29 | |
| *** Purandar has quit IRC | 00:40 | |
| *** paco20151113 has joined #openstack-lbaas | 00:45 | |
| *** manishg has quit IRC | 00:51 | |
| *** SumitNaiksatam has quit IRC | 00:54 | |
| *** madhu_ak has quit IRC | 00:58 | |
| *** madhu_ak has joined #openstack-lbaas | 00:59 | |
| *** Aish has left #openstack-lbaas | 01:00 | |
| *** ajmiller has quit IRC | 01:01 | |
| *** minwang2 has quit IRC | 01:04 | |
| *** ducttape_ has joined #openstack-lbaas | 01:08 | |
| *** ducttape_ has quit IRC | 01:15 | |
| *** yamamoto_ has joined #openstack-lbaas | 01:20 | |
| *** yamamoto_ has quit IRC | 01:26 | |
| *** jwarendt has quit IRC | 01:29 | |
| *** madhu_ak has quit IRC | 01:34 | |
| *** piet has quit IRC | 01:37 | |
| *** chlong_ has joined #openstack-lbaas | 01:49 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/neutron-lbaas: Updated from global requirements https://review.openstack.org/285042 | 01:50 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/octavia: Updated from global requirements https://review.openstack.org/285046 | 01:51 |
| openstackgerrit | Aaron Rosen proposed openstack/neutron-lbaas: Set netscaler_ncc_password as secret to prevent it from being logged https://review.openstack.org/285079 | 02:02 |
| *** yamamoto has joined #openstack-lbaas | 02:04 | |
| *** frankie_ has quit IRC | 02:06 | |
| openstackgerrit | Stephen Balukoff proposed openstack/neutron-lbaas: Fix shared-pools lbaas_listeners migration https://review.openstack.org/283802 | 02:14 |
| *** yamamoto has quit IRC | 02:15 | |
| *** manishg has joined #openstack-lbaas | 02:27 | |
| *** bana_k has joined #openstack-lbaas | 02:29 | |
| *** kevo has quit IRC | 02:33 | |
| *** ducttape_ has joined #openstack-lbaas | 02:37 | |
| *** ducttape_ has quit IRC | 02:40 | |
| *** ducttape_ has joined #openstack-lbaas | 02:44 | |
| *** ducttape_ has quit IRC | 02:46 | |
| *** fawadkhaliq has joined #openstack-lbaas | 02:48 | |
| rm_you | sbalukoff: this is fucking ridiculous | 02:49 |
| rm_you | ^^ that patch | 02:49 |
| sbalukoff | rm_you: Yes. Yes it is. | 02:49 |
| rm_you | for the record if someone submitted a revert I would -2 it and tell them to suck it | 02:49 |
| sbalukoff | Heh! | 02:49 |
| sbalukoff | Fuck Postgresql. | 02:49 |
| sbalukoff | What a pain in the ass, eh. | 02:49 |
| rm_you | I just don't understand the people coming out of the woodwork to complain about a gate-fix | 02:50 |
| rm_you | ah this in n-lbaas though so I can't -2 :( | 02:51 |
| rm_you | where is the n-lbaas patch that actually caused this? | 02:51 |
| sbalukoff | It's the shared-pools patch. | 02:51 |
| rm_you | your sarcasm is pretty legendary though | 02:51 |
| sbalukoff | That merged on Monday. | 02:52 |
| rm_you | we'd have to back out a bit of stuff to get that out I think, wouldn't we? >_> | 02:52 |
| sbalukoff | I feel bad now though-- I don't think she meant anything by what she said and English is not her first language. | 02:52 |
| sbalukoff | rm_you: Probably. | 02:52 |
| rm_you | heh yes | 02:52 |
| sbalukoff | I think I can get this working in place. | 02:52 |
| sbalukoff | I just don't have a good way to test other than watch the Manila jobs, which are apparently dependent on this + postgresql. | 02:53 |
| sbalukoff | But... that means I get to try again at most every 4 hours. | 02:53 |
| sbalukoff | Because the queue is really, really slow today. | 02:53 |
| rm_you | IMO I would rather merge a patch that drops PG support from the repo than actually bother reverting <_< but maybe this is why i'm not a n-lbaas core | 02:53 |
| *** ducttape_ has joined #openstack-lbaas | 02:53 | |
| rm_you | I wonder if we could modify my devstack script to spin with PG? | 02:54 |
| rm_you | it's just an option isn't it? | 02:54 |
| sbalukoff | Well, technically *we* don't require it. Or we'd have gates for it. | 02:54 |
| sbalukoff | But apparently people import all of neutron (including n-lbaas) and assuming that PostgreSQL is going to work. | 02:54 |
| rm_you | yeah | 02:54 |
| rm_you | disable_service mysql | 02:54 |
| rm_you | enable_service postgresql | 02:54 |
| rm_you | <_< | 02:54 |
| rm_you | add that to localrc | 02:54 |
| sbalukoff | If you know of a good way to spin a devstack script with postgresql so I could test this locally, I'd be very appreciative! | 02:54 |
| rm_you | yeah sec let me make a branch | 02:55 |
| sbalukoff | That's it? | 02:55 |
| rm_you | lol | 02:55 |
| sbalukoff | Dammit... | 02:55 |
| sbalukoff | Ok. | 02:55 |
| sbalukoff | well... | 02:55 |
| sbalukoff | Yeah, let me try that as well. | 02:55 |
| rm_work | ah if you've got it | 02:55 |
| sbalukoff | At least then I'll be able to tell whether the damned postgresql side works at all. | 02:55 |
| *** minwang2 has joined #openstack-lbaas | 02:55 | |
| sbalukoff | Ok, we'll see if this restack succeeds. | 02:56 |
| rm_work | if your localrc is anything like mine, what you ACTUALLY want to do is change this line: | 02:57 |
| rm_work | ENABLED_SERVICES=rabbit,postgresql,key | 02:57 |
| sbalukoff | Yep. | 02:58 |
| sbalukoff | That's what I've done. | 02:58 |
| sbalukoff | And yes, my localrc is based off yours. | 02:58 |
| sbalukoff | Thanks for that, by the way! | 02:58 |
| rm_work | https://raw.githubusercontent.com/rm-you/devstack_deploy/postgres-neutron-lbaas/stackme.sh | 03:00 |
| rm_work | should do it | 03:00 |
| rm_work | but i think you're already midway through | 03:00 |
| *** yamamoto has joined #openstack-lbaas | 03:01 | |
| rm_work | WHEN do we have to have everything merged Monday? | 03:01 |
| rm_work | like | 03:02 |
| sbalukoff | Yeah, it's running. Hasn't errored out yet. | 03:02 |
| rm_work | Monday MORNING? | 03:02 |
| sbalukoff | No idea. | 03:02 |
| rm_work | *before* Monday? | 03:02 |
| rm_work | lol | 03:02 |
| sbalukoff | I would think 11:59 pm | 03:02 |
| rm_work | I may be counting on Monday daytime to actually finish merging stuff :P | 03:02 |
| sbalukoff | The deadline is an arbirary date. | 03:02 |
| sbalukoff | So... end of that date. | 03:02 |
| rm_work | yeah >_> | 03:02 |
| rm_work | ... Hawaii timezone, obviously | 03:02 |
| sbalukoff | Well, if we're slowed down by the gate, I think people won't care. | 03:02 |
| sbalukoff | I think it just has to be +A'd that day. | 03:03 |
| rm_work | yeah | 03:03 |
| sbalukoff | Haha! | 03:03 |
| rm_work | all it does is motivate me to let bugs slip through so we can fix them during bugfix weeks after the feature freeze <_< | 03:03 |
| rm_work | anywho i am going to try to crash early as i have to be in early tomorrow <_< | 03:06 |
| rm_work | good luck with the PG | 03:06 |
| *** intr1nsic has quit IRC | 03:10 | |
| *** neelashah has joined #openstack-lbaas | 03:12 | |
| sbalukoff | Thanks. | 03:17 |
| *** intr1nsic has joined #openstack-lbaas | 03:17 | |
| *** ajmiller has joined #openstack-lbaas | 03:20 | |
| *** bana_k has quit IRC | 03:23 | |
| *** links has joined #openstack-lbaas | 03:26 | |
| *** fawadkhaliq has quit IRC | 03:32 | |
| *** ajmiller has quit IRC | 03:34 | |
| *** ajmiller has joined #openstack-lbaas | 03:34 | |
| *** manishg has quit IRC | 03:39 | |
| *** amotoki has joined #openstack-lbaas | 03:41 | |
| *** Purandar has joined #openstack-lbaas | 03:42 | |
| *** piet has joined #openstack-lbaas | 03:44 | |
| *** ducttape_ has quit IRC | 03:46 | |
| *** piet has quit IRC | 03:51 | |
| *** ajmiller_ has joined #openstack-lbaas | 03:54 | |
| *** ajmiller has quit IRC | 03:57 | |
| openstackgerrit | Madhusudhan Kandadai proposed openstack/neutron-lbaas: [WIP] Neutron LBaaS: TLS Barbican Scenario Test https://review.openstack.org/164828 | 04:00 |
| *** amotoki has quit IRC | 04:05 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Neutron LBaaS leaking admin_password into logs https://review.openstack.org/284483 | 04:10 |
| *** armax has joined #openstack-lbaas | 04:11 | |
| *** woodster_ has quit IRC | 04:16 | |
| *** amotoki has joined #openstack-lbaas | 04:20 | |
| sbalukoff | Ok! I think I finally have this figured out. What a world of difference it is to be able to troubleshoot this against the actual PostgreSQL database engine. | 04:28 |
| *** amotoki has quit IRC | 04:30 | |
| *** amotoki has joined #openstack-lbaas | 04:35 | |
| *** manishg has joined #openstack-lbaas | 04:40 | |
| *** manishg has quit IRC | 04:46 | |
| openstackgerrit | Stephen Balukoff proposed openstack/neutron-lbaas: Fix shared-pools lbaas_listeners migration https://review.openstack.org/283802 | 04:48 |
| sbalukoff | Sometimes things come full circle... | 04:48 |
| *** fawadkhaliq has joined #openstack-lbaas | 04:58 | |
| *** amotoki has quit IRC | 05:00 | |
| *** ducttape_ has joined #openstack-lbaas | 05:05 | |
| *** amotoki has joined #openstack-lbaas | 05:07 | |
| *** ajmiller_ has quit IRC | 05:10 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Updated from global requirements https://review.openstack.org/285042 | 05:11 |
| *** numans has joined #openstack-lbaas | 05:19 | |
| *** manishg has joined #openstack-lbaas | 05:28 | |
| *** manishg has quit IRC | 05:32 | |
| *** ducttape_ has quit IRC | 05:37 | |
| *** neelashah has quit IRC | 05:38 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Set netscaler_ncc_password as secret to prevent it from being logged https://review.openstack.org/285079 | 05:40 |
| *** reedip is now known as outofmemory | 05:46 | |
| *** links has quit IRC | 06:26 | |
| *** manishg has joined #openstack-lbaas | 06:28 | |
| *** armax has quit IRC | 06:31 | |
| *** manishg has quit IRC | 06:33 | |
| *** numans has quit IRC | 06:43 | |
| sbalukoff | Thanks for the review and +1 on that, bharathm! | 06:45 |
| *** numans has joined #openstack-lbaas | 06:58 | |
| bharathm | Yw :-) | 07:01 |
| *** rcernin has joined #openstack-lbaas | 07:05 | |
| *** Purandar has quit IRC | 07:07 | |
| paco20151113 | can I recheck in gerrit now . is the gate recovered ? | 07:08 |
| openstackgerrit | Merged openstack/octavia: Updated from global requirements https://review.openstack.org/285046 | 07:25 |
| *** jwarendt has joined #openstack-lbaas | 07:26 | |
| *** chlong_ has quit IRC | 07:26 | |
| openstackgerrit | Stephen Balukoff proposed openstack/octavia: Fix default_pool_id type in l7policy API https://review.openstack.org/285143 | 07:27 |
| *** manishg has joined #openstack-lbaas | 07:29 | |
| *** jwarendt has quit IRC | 07:31 | |
| *** manishg has quit IRC | 07:34 | |
| *** ihrachys has joined #openstack-lbaas | 07:39 | |
| *** minwang2 has quit IRC | 07:43 | |
| *** manishg has joined #openstack-lbaas | 08:30 | |
| *** manishg has quit IRC | 08:35 | |
| openstackgerrit | Stephen Balukoff proposed openstack/octavia: Add maximum number of l7rules per l7policy https://review.openstack.org/285160 | 08:42 |
| openstackgerrit | Stephen Balukoff proposed openstack/neutron-lbaas: Fix shared-pools lbaas_listeners migration https://review.openstack.org/283802 | 08:48 |
| *** piet has joined #openstack-lbaas | 09:05 | |
| *** piet has quit IRC | 09:08 | |
| *** rcernin has quit IRC | 09:14 | |
| *** jschwarz has joined #openstack-lbaas | 09:20 | |
| *** rcernin has joined #openstack-lbaas | 09:28 | |
| *** ihrachys_ has joined #openstack-lbaas | 09:32 | |
| *** ihrachys has quit IRC | 09:35 | |
| *** fawadkhaliq has quit IRC | 09:37 | |
| openstackgerrit | Ihar Hrachyshka proposed openstack/octavia: glance: support relying on tags to extract image id https://review.openstack.org/284690 | 09:47 |
| openstackgerrit | Stephen Balukoff proposed openstack/octavia: Block deletion of pool in use by L7Policies https://review.openstack.org/285192 | 09:57 |
| *** fnaval has joined #openstack-lbaas | 10:01 | |
| ajo | do we have any docs of how the health check of amphoras is done currently ? | 10:25 |
| ajo | I thought it was health-manager, but that only seems to look at the DB and manage the amphoras.. am I right? | 10:25 |
| ajo | what agents do use the connectivity to lb-mgmt-net ? | 10:25 |
| ajo | blogan, dougwig ^ :) | 10:26 |
| *** manishg has joined #openstack-lbaas | 10:31 | |
| *** ducttape_ has joined #openstack-lbaas | 10:34 | |
| *** manishg has quit IRC | 10:36 | |
| *** jschwarz has quit IRC | 10:46 | |
| *** fnaval has quit IRC | 10:52 | |
| *** ducttape_ has quit IRC | 10:57 | |
| bharathm | ajo: health-manager does two tasks.. One receives the heartbeats from amphora and update db.. The other to check these heartbeats at configured intervals. If a heartbeat is missing beyond the conf time limit, then it initiates an amphora failover | 11:01 |
| bharathm | For your other question.. Api and Healthmanager are connected to lb-mgmt-net listening on diff ports for obvious reasons. | 11:02 |
| ajo | bharathm, and worker? | 11:03 |
| bharathm | Well Healthmanager invokes the failover lib that's actually under worker.. | 11:05 |
| bharathm | API to worker communcations happen through queue, I think.. | 11:05 |
| openstackgerrit | Ihar Hrachyshka proposed openstack/octavia: glance: support relying on tags to extract image id https://review.openstack.org/284690 | 11:11 |
| *** armax has joined #openstack-lbaas | 11:32 | |
| *** manishg has joined #openstack-lbaas | 11:32 | |
| *** armax has quit IRC | 11:36 | |
| *** manishg has quit IRC | 11:37 | |
| *** ihrachys_ is now known as ihrachys | 12:03 | |
| *** yamamoto has quit IRC | 12:08 | |
| *** amotoki_ has joined #openstack-lbaas | 12:32 | |
| *** manishg has joined #openstack-lbaas | 12:33 | |
| *** amotoki has quit IRC | 12:34 | |
| *** manishg has quit IRC | 12:37 | |
| *** fnaval has joined #openstack-lbaas | 12:45 | |
| openstackgerrit | Ihar Hrachyshka proposed openstack/octavia: glance: support relying on tags to extract image id https://review.openstack.org/284690 | 12:47 |
| *** fnaval has quit IRC | 12:50 | |
| *** yamamoto has joined #openstack-lbaas | 13:01 | |
| *** yamamoto has quit IRC | 13:05 | |
| *** ducttape_ has joined #openstack-lbaas | 13:13 | |
| *** yamamoto has joined #openstack-lbaas | 13:25 | |
| *** neelashah has joined #openstack-lbaas | 13:29 | |
| *** yamamoto has quit IRC | 13:30 | |
| *** manishg has joined #openstack-lbaas | 13:34 | |
| *** ducttape_ has quit IRC | 13:35 | |
| *** fawadkhaliq has joined #openstack-lbaas | 13:36 | |
| *** manishg has quit IRC | 13:38 | |
| *** armax has joined #openstack-lbaas | 13:57 | |
| *** Bjoern_ has joined #openstack-lbaas | 14:11 | |
| *** Bjoern_ is now known as Bjoern_zZzZzZzZ | 14:11 | |
| *** Bjoern_zZzZzZzZ is now known as Bjoern_ | 14:12 | |
| *** rtheis has joined #openstack-lbaas | 14:15 | |
| *** yamamoto has joined #openstack-lbaas | 14:16 | |
| *** paco20151113 has quit IRC | 14:18 | |
| *** piet has joined #openstack-lbaas | 14:21 | |
| *** piet has quit IRC | 14:28 | |
| *** manishg has joined #openstack-lbaas | 14:34 | |
| *** amotoki_ has quit IRC | 14:36 | |
| *** manishg has quit IRC | 14:39 | |
| *** localloop127 has joined #openstack-lbaas | 14:42 | |
| *** amotoki has joined #openstack-lbaas | 14:48 | |
| *** woodster_ has joined #openstack-lbaas | 14:54 | |
| *** fawadkhaliq has quit IRC | 14:55 | |
| *** TrevorV has joined #openstack-lbaas | 14:55 | |
| *** ducttape_ has joined #openstack-lbaas | 14:56 | |
| *** amotoki has quit IRC | 15:02 | |
| *** Purandar has joined #openstack-lbaas | 15:02 | |
| *** ajmiller has joined #openstack-lbaas | 15:20 | |
| openstackgerrit | Doug Wiegley proposed openstack/neutron-lbaas: fix mismatch between v1 keyword and v2 driver in gate_hook https://review.openstack.org/285357 | 15:25 |
| TrevorV | xgerman johnsom anyone alive yet?! | 15:29 |
| *** manishg has joined #openstack-lbaas | 15:35 | |
| *** piet has joined #openstack-lbaas | 15:38 | |
| *** manishg has quit IRC | 15:40 | |
| dougwig | no, no life here. | 15:49 |
| *** localloo1 has joined #openstack-lbaas | 15:59 | |
| *** localloop127 has quit IRC | 15:59 | |
| *** piet has quit IRC | 16:00 | |
| johnsom | TrevorV Hi | 16:02 |
| xgerman | Hi | 16:03 |
| *** Purandar has quit IRC | 16:06 | |
| TrevorV | johnsom xgerman you guys had a fix or someone did for the scenario tests upstream to be "fixed" right? | 16:06 |
| TrevorV | That intermittent issue to be cleared up? | 16:06 |
| johnsom | No, we had attempts that cleaned up a few things, but we have not completely fixed the issues. | 16:07 |
| johnsom | I'm pretty sure the gate hosts have become slower and some of the tests are timing out before the amp finishes booting. | 16:07 |
| TrevorV | oh okay, well, its the only thing that is "red" for the single-create stuff | 16:08 |
| TrevorV | I'm really excited because of my full test run | 16:08 |
| johnsom | I put up a couple of junk patches with some logging in it, but they didn't really provide the insight I had hoped | 16:08 |
| johnsom | Yeah, I'm reading through the first one now. I will ask that we pull out the commented out code, so probably an update needed today | 16:09 |
| TrevorV | johnsom where is the commented out code? | 16:12 |
| *** logan- has quit IRC | 16:14 | |
| *** logan- has joined #openstack-lbaas | 16:14 | |
| johnsom | https://review.openstack.org/#/c/256974/19/octavia/api/v1/controllers/pool.py | 16:16 |
| *** kobis has quit IRC | 16:18 | |
| ihrachys | I would appreciate if folks take a look at the glance tags patch: https://review.openstack.org/#/c/284690/ so that I can respin quickly later today in case of something | 16:21 |
| * ihrachys needs to write some tempest scenario I guess | 16:21 | |
| *** kobis has joined #openstack-lbaas | 16:22 | |
| *** jwarendt has joined #openstack-lbaas | 16:22 | |
| *** armax has quit IRC | 16:22 | |
| *** armax has joined #openstack-lbaas | 16:25 | |
| *** localloo1 has quit IRC | 16:28 | |
| *** localloo1 has joined #openstack-lbaas | 16:30 | |
| *** rcernin has quit IRC | 16:31 | |
| *** piet has joined #openstack-lbaas | 16:31 | |
| *** piet has quit IRC | 16:33 | |
| TrevorV | johnsom I'll clean those up right now. | 16:33 |
| *** piet has joined #openstack-lbaas | 16:33 | |
| TrevorV | oh, johnsom do you have any waiting comments? I can wait until I can address them before pushing a patch set | 16:34 |
| *** armax has quit IRC | 16:34 | |
| johnsom | Just those two so far | 16:34 |
| *** ihrachys has quit IRC | 16:39 | |
| *** numans has quit IRC | 16:48 | |
| *** kobis has quit IRC | 16:48 | |
| *** yamamoto has quit IRC | 16:50 | |
| *** kobis has joined #openstack-lbaas | 16:50 | |
| *** Bjoern_ is now known as BjoernT | 16:50 | |
| *** yamamoto has joined #openstack-lbaas | 16:50 | |
| TrevorV | I'll push up a change with that then | 16:50 |
| *** piet has quit IRC | 16:54 | |
| *** piet has joined #openstack-lbaas | 16:59 | |
| *** pai15- has joined #openstack-lbaas | 17:12 | |
| *** manishg has joined #openstack-lbaas | 17:14 | |
| openstackgerrit | Trevor Vardeman proposed openstack/octavia: Get me a Load Balancer API https://review.openstack.org/256974 | 17:15 |
| *** bedis_ has joined #openstack-lbaas | 17:15 | |
| openstackgerrit | Trevor Vardeman proposed openstack/octavia: Get Me A Load Balancer Controller https://review.openstack.org/257013 | 17:16 |
| *** piet has quit IRC | 17:16 | |
| *** kfox1111_ has joined #openstack-lbaas | 17:16 | |
| *** piet has joined #openstack-lbaas | 17:17 | |
| *** bana_k has joined #openstack-lbaas | 17:18 | |
| johnsom | TrevorV Ok, finished on https://review.openstack.org/#/c/256974, there are two missing branch unit tests. We can bug those like we did for L7 if you would like | 17:19 |
| *** blogan has quit IRC | 17:20 | |
| *** bedis has quit IRC | 17:20 | |
| *** crc32|znc has quit IRC | 17:20 | |
| *** pai15 has quit IRC | 17:20 | |
| *** lmiccini has quit IRC | 17:20 | |
| *** kfox1111 has quit IRC | 17:20 | |
| *** crc32|znc has joined #openstack-lbaas | 17:20 | |
| *** blogan has joined #openstack-lbaas | 17:21 | |
| *** minwang2 has joined #openstack-lbaas | 17:21 | |
| *** lmiccini has joined #openstack-lbaas | 17:23 | |
| sbalukoff | Mornin' folks! | 17:26 |
| TrevorV | johnsom I'm headed to lunch, if you think its okay to go now and I do a bug fix for the tests, then I'm good to go, but I don't mind updating accordingly as well :) | 17:26 |
| TrevorV | Up to you reviewers :D | 17:26 |
| TrevorV | Especially if we want it merged by Monday :P | 17:27 |
| johnsom | Ok. Hopefully I will be done with the other review by the time you get back. | 17:27 |
| johnsom | Should be a quick fix, then I would be good to go | 17:28 |
| *** yamamoto has quit IRC | 17:28 | |
| *** yamamoto has joined #openstack-lbaas | 17:28 | |
| *** yamamoto has quit IRC | 17:28 | |
| johnsom | sbalukoff Morning. Looks like you were busy yesterday. I was out recruiting. Five hours of talking to college kids... | 17:29 |
| sbalukoff | johnsom: Heh! Good luck on attracting that talent, eh! | 17:29 |
| sbalukoff | johnsom: Eh... those were pretty low hanging fruit. | 17:29 |
| johnsom | I'm trying to go through the get-me-an-LB now | 17:30 |
| sbalukoff | I'mma review TrevorV's patch, then I have a bit of internal stuff to take care of, then it's seeing if I can fix the couple of critical bugs around session_persistence updates. | 17:30 |
| sbalukoff | Yep, that's on my docket as well! | 17:30 |
| sbalukoff | (First thing, in fact.) | 17:30 |
| *** yamamoto has joined #openstack-lbaas | 17:30 | |
| johnsom | Cool, yeah, it would be good if we could also review the delete-me-an-lb | 17:31 |
| sbalukoff | Oh yes! Yeah-- I'll hit that up after this patch. Thanks for the reminder. | 17:32 |
| johnsom | Cool, thanks | 17:32 |
| *** BjoernT has quit IRC | 17:34 | |
| *** bana_k has quit IRC | 17:40 | |
| *** bana_k has joined #openstack-lbaas | 17:41 | |
| *** bana_k has quit IRC | 17:42 | |
| *** madhu_ak has joined #openstack-lbaas | 17:43 | |
| *** Aish has joined #openstack-lbaas | 17:47 | |
| *** crc32 has joined #openstack-lbaas | 17:50 | |
| openstackgerrit | Henry Gessau proposed openstack/neutron-lbaas: Track alembic heads https://review.openstack.org/285465 | 17:52 |
| *** piet has quit IRC | 17:56 | |
| dougwig | johnsom or others: what is this job in octavia's experimental queue? - gate-neutron-lbaasv2-octavia-dsvm-api | 17:57 |
| johnsom | No clue | 17:57 |
| *** manishg has quit IRC | 17:58 | |
| *** manishg has joined #openstack-lbaas | 17:59 | |
| madhu_ak | dougwig, I believe that job runs tests present in lbaas repo | 18:00 |
| dougwig | right, but why? seems to overlap with what already runs, but is super super long. | 18:00 |
| madhu_ak | yes, I am too confused too, I am not sure why that has been added into octavia. | 18:01 |
| dougwig | let me rephrase... i'm killing it unless someone has a reason to keep it. :) | 18:01 |
| *** piet has joined #openstack-lbaas | 18:07 | |
| dougwig | unsurprisingly the agent driver tests have rotted. this could use some eyes in my quest to get that job running again: https://review.openstack.org/#/c/285357/ | 18:07 |
| *** yamamoto has quit IRC | 18:12 | |
| *** rtheis has quit IRC | 18:12 | |
| *** yamamoto has joined #openstack-lbaas | 18:12 | |
| *** yamamoto has quit IRC | 18:12 | |
| *** yamamoto has joined #openstack-lbaas | 18:13 | |
| *** evgenyf has joined #openstack-lbaas | 18:17 | |
| *** crc32 has quit IRC | 18:17 | |
| *** kevo has joined #openstack-lbaas | 18:21 | |
| *** piet has quit IRC | 18:25 | |
| *** piet has joined #openstack-lbaas | 18:25 | |
| *** piet has quit IRC | 18:29 | |
| *** piet has joined #openstack-lbaas | 18:29 | |
| *** Aish has quit IRC | 18:33 | |
| *** ducttape_ has quit IRC | 18:34 | |
| *** armax has joined #openstack-lbaas | 18:35 | |
| sbalukoff | TrevorV: Good work on the get-me-a-load-balancer patch. I think it's really close and have one major suggestion. I need to run off for a bit but will be back on intermittently today. I'm happy to answer questions about my suggestion if you want, eh. | 18:39 |
| *** Aish has joined #openstack-lbaas | 18:45 | |
| rm_work | dougwig: i think that was OOOOLD | 18:46 |
| rm_work | if you find the blame that added it and look up that commit in gerrit, i bet it will trace back to a CR that has my name on it somewhere | 18:46 |
| rm_work | because I think i was involved in some capacity | 18:46 |
| rm_work | but it was before we had our current testing stuff in place | 18:46 |
| *** fnaval has joined #openstack-lbaas | 18:56 | |
| *** Purandar has joined #openstack-lbaas | 19:11 | |
| *** piet has quit IRC | 19:17 | |
| *** piet has joined #openstack-lbaas | 19:17 | |
| *** sbalukoff has quit IRC | 19:22 | |
| *** ducttape_ has joined #openstack-lbaas | 19:23 | |
| *** Bjoern_ has joined #openstack-lbaas | 19:31 | |
| *** fnaval has quit IRC | 19:32 | |
| *** Purandar has quit IRC | 19:42 | |
| *** piet has quit IRC | 19:43 | |
| *** piet has joined #openstack-lbaas | 19:44 | |
| *** piet has quit IRC | 19:49 | |
| *** piet has joined #openstack-lbaas | 19:52 | |
| *** bana_k has joined #openstack-lbaas | 19:54 | |
| *** Purandar has joined #openstack-lbaas | 19:58 | |
| *** piet has quit IRC | 20:01 | |
| *** yamamoto has quit IRC | 20:13 | |
| *** yamamoto has joined #openstack-lbaas | 20:13 | |
| *** yamamoto has quit IRC | 20:13 | |
| *** alejandrito has joined #openstack-lbaas | 20:20 | |
| *** Purandar has quit IRC | 20:22 | |
| *** sbalukoff has joined #openstack-lbaas | 20:28 | |
| *** madhu_ak has quit IRC | 20:37 | |
| openstackgerrit | Evgeny Fedoruk proposed openstack/neutron-lbaas: L7 capability extension implementation for lbaas v2 https://review.openstack.org/148232 | 20:44 |
| *** madhu_ak has joined #openstack-lbaas | 20:47 | |
| TrevorV | sbalukoff you around right now? | 21:03 |
| minwang2 | sbalukoff, i got one question for you | 21:07 |
| TrevorV | By the way, I just found out that "pop('list', []) | 21:11 |
| TrevorV | Doesn't work the way it looks like it works | 21:11 |
| TrevorV | Had to do " | 21:11 |
| TrevorV | "pop('list') or []" | 21:11 |
| *** yamamoto has joined #openstack-lbaas | 21:14 | |
| *** alejandrito has quit IRC | 21:16 | |
| *** yamamoto has quit IRC | 21:19 | |
| *** evgenyf has quit IRC | 21:23 | |
| sbalukoff | TrevorV: I'm around now. | 21:23 |
| sbalukoff | minwang2: What is your question? | 21:23 |
| TrevorV | sbalukoff so were you saying that you wanted me to update the patch to include parent+child traversal, or leave that for a different review? | 21:24 |
| sbalukoff | TrevorV: If you can get it done relatively quickly, let's do it now. | 21:24 |
| TrevorV | sbalukoff I don't think I can get it done and fully tested by Monday, TBH | 21:24 |
| TrevorV | I could "get it done" by Monday, but that's just putting in the code and hoping its right, you know? | 21:25 |
| minwang2 | sbalukoff, i am wondering will L7 works in this secnario: create a lb, and a listener(listener1) on port 80, create a pool (pool1) inside of listener1,create a members under pool1(10.0.0.3 on port 80 )create another pool(pool2---it is not assoicated with any listener ) ,inside this new pool (pool2)create a member(10.0.0.4 on port 81),create 3 L7poliy(reject, direct_to_pool,direct_to_url) on listener1, on each L7policy create a rule | 21:25 |
| minwang2 | (start_with,contains,REGEX) | 21:25 |
| sbalukoff | Ok, well file a bug that one-call-create doesn't handle shared pools, then. | 21:25 |
| TrevorV | sbalukoff will it ever? | 21:25 |
| sbalukoff | TrevorV: Yes, the way I described in my comment. | 21:26 |
| TrevorV | sbalukoff which comment specifically? "create_load_balancer_tree"? | 21:26 |
| sbalukoff | minwang2: That's a lot to parse, but I *thin* that'll work. | 21:26 |
| sbalukoff | TrevorV: The general comment, not on a line number, in which I first introduced the idea of the pool conditional_create method. | 21:27 |
| sbalukoff | *think* | 21:27 |
| minwang2 | i tested it out, it works with one backend node, so if the vip ip is 10.0.0.5, when i do curl 10.0.0.5 or curl 10.0.0.5 with different policies, i found somehow it just goes to one backend always | 21:27 |
| TrevorV | sbalukoff I didn't see those numbered concerns... | 21:28 |
| TrevorV | I'm reading now. | 21:28 |
| sbalukoff | minwang2: Ok, rules get ANDed together. | 21:28 |
| TrevorV | sbalukoff do we have specific l7 rule validation? | 21:28 |
| sbalukoff | So if all three of those rules get added to one policy, then all three of them have to be true in any given request for the policy to be executed. | 21:29 |
| sbalukoff | TrevorV: We do. | 21:29 |
| sbalukoff | TrevorV: I think we do rule validation at both the API and the repository layers. | 21:30 |
| TrevorV | sbalukoff well I can tell you it definitely doesn't happen in the "type" class, cuz a string is a string, but I'll follow real quick (also means that your testing was lacking, not necessarily mine, no offense) | 21:31 |
| TrevorV | Wait wait, I see the validation step | 21:31 |
| sbalukoff | TrevorV: Look for methods calls that start with "validate" in both the l7rule API controller and repository. | 21:31 |
| TrevorV | Okay, I see what you did thre. | 21:31 |
| TrevorV | there*** | 21:31 |
| TrevorV | I could add that to the "prepare" method for l7rules. | 21:32 |
| sbalukoff | TrevorV: Because it happens in the repository, I think your single-create is safe so long as it fails gracefully. | 21:32 |
| sbalukoff | TrevorV: Good idea! | 21:32 |
| TrevorV | Right, sbalukoff that's one thing I might have missed. | 21:32 |
| minwang2 | sblukoff, for my testing case, 3 policies, eevery policy i have a rule | 21:32 |
| minwang2 | sbalukoff | 21:33 |
| sbalukoff | minwang2: Ok, I think you're going to have to be more specific if you want me to be able to troubleshoot this. Can you put what you've tried (specifically) in a gist? | 21:33 |
| sbalukoff | If there are any neutron-lbaas cores around (other than dougwig), I know there are people who would like you put your stamp of approval on this: https://review.openstack.org/#/c/283802/ | 21:35 |
| sbalukoff | (Maybe ajmiller, if you've got a moment?) | 21:36 |
| johnsom | I'll take a look real quick | 21:36 |
| johnsom | I think I looked at it before | 21:36 |
| xgerman | yeah, I also like to see more love for delete_me-a-lb as Michael calls it | 21:36 |
| johnsom | Oye, if postgres | 21:37 |
| sbalukoff | xgerman: It's on my list for today. | 21:38 |
| xgerman | thanks | 21:38 |
| johnsom | sbalukoff the with thing I mentioned doesn't handle this? | 21:38 |
| sbalukoff | johnsom: That thing has a major bug in it. | 21:39 |
| sbalukoff | That I filed... | 21:39 |
| johnsom | Joy, ok | 21:39 |
| sbalukoff | Unfortunately, most of the neutron DB migration tools are half-baked. | 21:39 |
| johnsom | Well, so is sqlalchemy in my opinion | 21:39 |
| sbalukoff | They *really* ought to write a method that does a 'remove uniqueness from foreign key' that works no matter the DB engine. | 21:39 |
| sbalukoff | johnsom: +1 | 21:39 |
| *** crc32 has joined #openstack-lbaas | 21:40 | |
| sbalukoff | johnsom: In the end, what worked was to just have the migration script detect the particular DB engine dialect and "do the right thing" for that particular DB engine. | 21:40 |
| sbalukoff | Helped a lot once rm_work pointed out that deploying a devstack with postgresql was trivial. | 21:41 |
| johnsom | Yeah. I have been beat up for that in the past with a sqllite issue sqlalchemy couldn't handle. | 21:41 |
| sbalukoff | Yep. | 21:42 |
| sbalukoff | Super annoying. But... whatever, eh. | 21:42 |
| ajmiller | sbalukoff - Had some people at my desk -- I see johnsom got to it... | 21:43 |
| sbalukoff | ajmiller: Yep! No problem, eh! | 21:44 |
| ajmiller | cool! | 21:44 |
| TrevorV | Sorry sbalukoff I am ready to keep talking | 21:44 |
| sbalukoff | Uh-oh. | 21:44 |
| TrevorV | So "shared pools" already cannot be supported by single create because we have no way of identifying the pool to share between listeners | 21:44 |
| TrevorV | You can make a new pool on each (hypothetically, I didn't actually test all the way down to haproxy) | 21:45 |
| sbalukoff | If you put the 'id' parameter in your pool definition, you do. | 21:45 |
| TrevorV | You can't though, cuz you can't specify a pool on the incoming object | 21:45 |
| sbalukoff | neutron-lbaas, in its octavia driver, puts the 'id' parameter in all the time. | 21:45 |
| sbalukoff | It would be bad if it didn't because then equivalent objects would have different IDs between Neutron-LBaaS and Octavia. | 21:46 |
| TrevorV | After it makes it, sure, but in neutron-lbaas you'd already have a pool created, so you'd KNOW the id | 21:46 |
| sbalukoff | TrevorV: You're saying that I can't create a pool and know what its ID is going to be beforehand, even if I specify the 'id' in the request? | 21:47 |
| TrevorV | You *can't* specify the ID before hand... we issue it after receiving the request. | 21:48 |
| TrevorV | Unless I've COMPLETELY missed something | 21:48 |
| sbalukoff | I'm pretty sure you have. | 21:48 |
| sbalukoff | All our API controllers allow you to specify the ID as part of the request. As long as it's not taken by something else (which is unlikely, given the size of the uuid space), you're golden. | 21:48 |
| sbalukoff | Again, if it didn't work this way, we wouldn't be able to connect with Neutron-LBaaS the way we do. | 21:49 |
| sbalukoff | And we can totally take advantage of this feature to do shared pools in single-create. | 21:49 |
| madhu_ak | https://bugs.launchpad.net/octavia/+bug/1550535 | 21:52 |
| openstack | Launchpad bug 1550535 in octavia "AttributeError: 'unicode' object has no attribute 'extend' during TLS barbican scenario testing" [Critical,New] | 21:52 |
| TrevorV | sbalukoff would you be against "shared pools" being a Major bug for me to address? | 21:54 |
| sbalukoff | TrevorV: I think I suggested that, actually. :) | 21:54 |
| TrevorV | Oh okay, sorry I'm all jumbled now | 21:54 |
| sbalukoff | TrevorV: I'm willing to merge this patch without it, so long as we list that as a major bug. | 21:54 |
| sbalukoff | no worries, i know the feeling, eh. | 21:54 |
| TrevorV | Please, I'm okay with that too. | 21:54 |
| sbalukoff | Honestly, I could see this functionality rolled in with a patch that does partial-tree creates the way you've described earlier. | 21:55 |
| TrevorV | Yeah, that sounds great, I would do that all together honestly | 21:57 |
| sbalukoff | Yep. | 21:57 |
| rm_work | yes i would vote we merge as close to as-is as we can | 21:57 |
| rm_work | and fix this stuff as bugs | 21:57 |
| rm_work | otherwise there's 0% chance we get this by monday | 21:57 |
| TrevorV | Okay, well I removed the extra "l7policies" portion of the prepare class under "create_listener" | 21:57 |
| TrevorV | It works 100% without that section | 21:58 |
| rm_work | just because of the testing cycles alone for reviewers :/ | 21:58 |
| TrevorV | Will need "put back" when we do the parent+children | 21:58 |
| rm_work | madhu_ak: GAH I thought i fixed all of those | 21:58 |
| sbalukoff | I do think that being able to create a load balancer tree that includes l7rules that don't pass validation is definitely a serious bug, though: So we'll probably want to make sure that can't happen before we merge this. | 21:58 |
| TrevorV | Yeah, I'll add that tonight when I get home | 21:58 |
| sbalukoff | TrevorV: l7policies also have some validation, though that all happens in the repository layer. | 21:59 |
| rm_work | hmmmmm | 21:59 |
| madhu_ak | rm_work Oh. Can you triage it for me and comment on the same bug? | 21:59 |
| sbalukoff | TrevorV: Because the validation requires doing things like making sure that if it's a REDIRECT_TO_POOL policy, that you have actually specified a pool. | 21:59 |
| sbalukoff | TrevorV: Mostly, if you can make sure that failed attempts to pass stuff through the repository create commands fails gracefully and doesn't leave a half-built load balancer tree in the database, that's ideally what we're after, I think. | 22:00 |
| sbalukoff | Though I have no idea how difficult that would be to accomplish. (sorry!) | 22:01 |
| TrevorV | Yeah, and tbh, I'm not 100% sure that happens, though I think it does because when something fails in our SQL Alchemy sessions, they get reverted. | 22:02 |
| rm_work | madhu_ak: i DO NOT see how this is possible | 22:02 |
| TrevorV | I'm not sure about the "graceful fail" when it comes to reporting the error though | 22:02 |
| rm_work | johnsom: how did you reproduce this? | 22:02 |
| sbalukoff | TrevorV: Yeah, I'm pretty sure you can take advantage of the database transaction functionality to accomplish this. | 22:03 |
| TrevorV | I think so too. | 22:03 |
| sbalukoff | (I suspect it's already doing it, in fact, since you wrap everything in a with: session context.) | 22:03 |
| *** localloo1 has quit IRC | 22:03 | |
| rm_work | johnsom / madhu_ak: https://github.com/openstack/octavia/blob/master/octavia/common/tls_utils/cert_parser.py#L175-L178 | 22:03 |
| rm_work | literally no way | 22:03 |
| sbalukoff | Really, I just want to see you make a test that tests that so we can be sure. | 22:03 |
| TrevorV | I do have an "except Exception" block that just doesn't do anything appropriate that I meant to update. | 22:03 |
| TrevorV | ... | 22:03 |
| TrevorV | sbalukoff I'll add one soon. | 22:03 |
| johnsom | rm_work I didn't, I just pasted the log out of the pastebin in case it disappeared | 22:03 |
| rm_work | johnsom / madhu_ak: there is literally no way I see that "pem" can be anything but a list at that point | 22:03 |
| rm_work | I would be tempted to mark that bug as invalid | 22:04 |
| TrevorV | In fact sbalukoff I'll probably have to add any validation steps INTO these create methods in "prepare" that way its not doing it twice (controller and 'prepare') | 22:04 |
| rm_work | who filed it? madhu_ak were you able to reproduce this? | 22:04 |
| johnsom | I think madhu_ak has that environment up right now | 22:04 |
| sbalukoff | TrevorV: i'm less concerned about repeating validations, eh. :) | 22:05 |
| rm_work | madhu_ak: can you verify that those lines in that file are the same as https://github.com/openstack/octavia/blob/master/octavia/common/tls_utils/cert_parser.py#L175-L178 ? | 22:05 |
| rm_work | if so, this is literally not possible | 22:05 |
| TrevorV | Ha ha ha I got yoiu | 22:05 |
| TrevorV | you*** | 22:05 |
| rm_work | cosmic rays must be flipping bits on your CPU :/ | 22:05 |
| sbalukoff | TrevorV: Anything else for now? I hear xgerman's get-me-a-delete patch calling to me. | 22:06 |
| xgerman | rm_Work comic rays… and solar flares... | 22:07 |
| TrevorV | sbalukoff nah, I | 22:07 |
| madhu_ak | rm_work, checking | 22:07 |
| TrevorV | I'm about to put up a change... with the validation step, but without a test right now. | 22:07 |
| TrevorV | I can add a unit test when I get home. | 22:07 |
| sbalukoff | Cool. | 22:07 |
| sbalukoff | Poke me when you've got it, eh. | 22:07 |
| sbalukoff | johnsom: Do you have time to add bug reports for the unit test stuff that TrevorV is missing in his patch? | 22:08 |
| madhu_ak | rm_work, http://paste.openstack.org/show/488437/ | 22:09 |
| madhu_ak | rm_work, I am able to see them | 22:09 |
| johnsom | sbalukoff I don't right now, I'm dealing with an internal issue. | 22:09 |
| sbalukoff | johnsom: Ok, good luck! | 22:10 |
| sbalukoff | (I guess we don't need to add them until the patch merges anyway.) | 22:10 |
| TrevorV | sbalukoff nevermind, just wrote a unit test. :) | 22:10 |
| sbalukoff | TrevorV: Oh, cool! Does it work? (ie. did you have to change anything else in the repo methods for the test to "fail" the way we want?) | 22:11 |
| TrevorV | Threw a 400 bad request... | 22:11 |
| sbalukoff | Awesome! | 22:11 |
| TrevorV | Okay good ha ha | 22:11 |
| sbalukoff | And the DB is "clean"? | 22:11 |
| TrevorV | It never even got to adding somethign to the db | 22:11 |
| TrevorV | something*** | 22:12 |
| madhu_ak | rm_work, is there anything you think I am missing? | 22:12 |
| sbalukoff | TrevorV: Hmmm... it may not be doing what I want then... Hmmm... anyway, when you have the test uploaded I'll have a look. | 22:12 |
| TrevorV | Yeah, sounds good to me. | 22:12 |
| TrevorV | Just comment again :) | 22:12 |
| minwang2 | sbalukoff https://gist.github.com/mallow111/445e2651f13f198e4f8f | 22:12 |
| sbalukoff | Will do! | 22:12 |
| minwang2 | here is the steps that i did | 22:12 |
| sbalukoff | minwang2: Looking now. | 22:13 |
| rm_work | madhu_ak: I think maybe I actually know why this is happening | 22:13 |
| minwang2 | thank you sbalukoff | 22:13 |
| TrevorV | Oh noes sbalukoff I've just accidentally put the changes on the controller review so far... | 22:14 |
| sbalukoff | Ack! | 22:14 |
| TrevorV | Do you know how to "undo" a "git add"? | 22:14 |
| TrevorV | I haven't committed yet | 22:14 |
| sbalukoff | Uh... | 22:14 |
| sbalukoff | My git-foo is weak. I'm not sure. | 22:14 |
| madhu_ak | git rm? | 22:15 |
| TrevorV | madhu_ak I want to keep the changes though, so I can stage them on a different commit o_0 | 22:15 |
| madhu_ak | ....... | 22:15 |
| TrevorV | Yeah... YEAH | 22:15 |
| *** yamamoto has joined #openstack-lbaas | 22:16 | |
| openstackgerrit | Merged openstack/neutron-lbaas: Fix shared-pools lbaas_listeners migration https://review.openstack.org/283802 | 22:17 |
| rm_work | madhu_ak: how do you reproduce this? | 22:18 |
| *** Purandar has joined #openstack-lbaas | 22:18 | |
| rm_work | madhu_ak: the issue is that the "intermediates" should be a list of strings, but it is coming in as a SINGLE string (probably without the intermediates having been split apart) | 22:18 |
| sbalukoff | minwang2: Can you show me the haproxy.cfg that gets generated on the amphora when you follow your procedure? | 22:18 |
| minwang2 | sbaluloff, check the gist | 22:20 |
| minwang2 | i just updated it | 22:20 |
| madhu_ak | hmmm. Tried two ways.. manually tested by following https://wiki.openstack.org/wiki/Network/LBaaS/docs/how-to-create-tls-loadbalancer and run the tests as per: https://review.openstack.org/#/c/164828/ | 22:20 |
| madhu_ak | rm_work, | 22:20 |
| rm_work | madhu_ak: do you use SNI or not | 22:21 |
| rm_work | madhu_ak: if I can rule out SNI that would make tracking this down easier | 22:21 |
| *** yamamoto has quit IRC | 22:21 | |
| madhu_ak | Yep.using SNI | 22:21 |
| rm_work | ah nevermind I think I see it anyway | 22:21 |
| madhu_ak | I can try once again without using SNI and recofirm you? | 22:22 |
| minwang2 | sbalukoff, it seems only one webserver is shown inthe .cfg file | 22:22 |
| rm_work | madhu_ak: nah i see it | 22:23 |
| rm_work | madhu_ak: the problem is using intermediates at all | 22:23 |
| rm_work | the assumption was made that they would be a list | 22:23 |
| rm_work | but they aren't | 22:23 |
| rm_work | they're a huge string | 22:23 |
| rm_work | i can fix htis | 22:23 |
| sbalukoff | minwang2: Yeah... | 22:23 |
| rm_work | i'll have a patch for you to test in ~5min | 22:23 |
| madhu_ak | I see. Thaks for the conirming rm_work | 22:23 |
| madhu_ak | confirming* | 22:23 |
| sbalukoff | minwang2: If you look at the member table in the octavia database, do see both members? | 22:24 |
| minwang2 | i will post them in gitst, wait a sec | 22:24 |
| sbalukoff | Also, for case2 problem: Could you also copy-paste the exact curl command you're using? | 22:26 |
| minwang2 | sbaluoff check the gist | 22:26 |
| sbalukoff | individual_pool2 is not the same name as pool2-out-listener1 | 22:28 |
| sbalukoff | Also on this: test@test-virtual-machine:~$ curl 10.0.0.5 /cat | 22:28 |
| sbalukoff | There's a space between the 10.0.0.5 and '/cat' | 22:29 |
| minwang2 | that is contains | 22:29 |
| minwang2 | not start_with | 22:29 |
| sbalukoff | Right, but is curl doing what you expect? | 22:29 |
| sbalukoff | Maybe do this instead: curl http://10.0.0.5/some/path/with/cat | 22:29 |
| sbalukoff | Also this: test@test-virtual-machine:~$ curl 10.0.0.5 .*thisor that | 22:30 |
| sbalukoff | That's not the way to set the X-My-Header in that request. | 22:30 |
| TrevorV | sbalukoff sorry it took me, that was harder to get back to the API review than I thought... Rather, than I made it... hah | 22:31 |
| sbalukoff | Heh! | 22:31 |
| sbalukoff | minwang2 (looking up how to set custom headers with curl again...) | 22:31 |
| minwang2 | let me try | 22:31 |
| sbalukoff | Ok, maybe this: curl --header "X-My-Header: something with thisor that in it" http://10.0.0.5/ | 22:32 |
| minwang2 | also i think the problem here is why we only have one backend, is that by design | 22:32 |
| sbalukoff | The config shows two backends. But one of them doesn't have any members. | 22:33 |
| *** Purandar has quit IRC | 22:33 | |
| sbalukoff | A dump of the mysql octavia.member table would help me troubleshoot that. | 22:33 |
| sbalukoff | (ie.: select * from octavia.member on the sql command-line instead of going through the neutron cli) | 22:34 |
| minwang2 | wait a sec | 22:35 |
| TrevorV | Alright guys, now I'm heading home. Make more comments please!!! I welcome the CHALLENGE | 22:36 |
| sbalukoff | TrevorV: New patchset up? | 22:37 |
| TrevorV | waiting for tox | 22:37 |
| TrevorV | then pushin | 22:37 |
| sbalukoff | Ok. | 22:37 |
| sbalukoff | Cool. | 22:37 |
| TrevorV | yep yep | 22:37 |
| TrevorV | just about to.... | 22:37 |
| openstackgerrit | Trevor Vardeman proposed openstack/octavia: Get me a Load Balancer API https://review.openstack.org/256974 | 22:37 |
| TrevorV | there | 22:37 |
| TrevorV | done | 22:37 |
| sbalukoff | Heh! | 22:38 |
| TrevorV | see you guys online later maybe! | 22:38 |
| sbalukoff | Ok, I'll have a look later. | 22:38 |
| *** TrevorV has quit IRC | 22:38 | |
| sbalukoff | (really want to take some time for xgerman's patch today.) | 22:38 |
| openstackgerrit | Adam Harwell proposed openstack/octavia: Barbican Cert Manager fails to correctly parse intermediates https://review.openstack.org/285574 | 22:38 |
| rm_work | madhu_ak: ^^ | 22:38 |
| rm_work | madhu_ak: please apply this to your environment and test again :) | 22:38 |
| rm_work | madhu_ak: i will check back later, away for a bit | 22:38 |
| minwang2 | sbalukoff take a look at the gist | 22:39 |
| minwang2 | i have added all of the tables there | 22:39 |
| sbalukoff | Ok. | 22:39 |
| sbalukoff | minwang2: And the haproxy.cfg is still accurate? | 22:40 |
| minwang2 | the haproxy.cfg has only 1 backend node, that is my concern | 22:40 |
| sbalukoff | Right: I see the problem. | 22:40 |
| sbalukoff | The haproxy.cfg references the pool: pool2-out-listener1 | 22:40 |
| sbalukoff | Which is good, because that's what the l7policy references | 22:40 |
| sbalukoff | But you've added those extra two members to the pool: individual_pool2 | 22:41 |
| sbalukoff | Do you see what I'm seeing? | 22:41 |
| rm_work | madhu_ak: this same bug exists probably in neutron-lbaas, I will need to do the fix there too, or you can if you want -- it's a LITTLE bit different file layout and such, but I think mostly the relevant parts will be the same | 22:41 |
| rm_work | bbl | 22:42 |
| minwang2 | ok, pool2-out-listener is used to redict_to_pool | 22:43 |
| sbalukoff | Yep | 22:43 |
| minwang2 | i added two members in individual_pool2, yes | 22:43 |
| sbalukoff | But you didn't add any members to pool2-out-listener | 22:43 |
| minwang2 | so if instead i delete these 2 members and add them in pool2-out-listener1, do you think they will work | 22:43 |
| sbalukoff | Yes. | 22:43 |
| sbalukoff | In fact... | 22:43 |
| minwang2 | hmm, let me try | 22:43 |
| sbalukoff | You don't even have to delete them from individual_pool2 | 22:44 |
| sbalukoff | Since technically they'll show up as different member records in the database. | 22:44 |
| *** Purandar has joined #openstack-lbaas | 22:44 | |
| minwang2 | so i can just added them in pool2-out-listener1 i guess | 22:44 |
| sbalukoff | Yep. | 22:44 |
| minwang2 | ok, let me try | 22:44 |
| sbalukoff | Ok | 22:44 |
| *** Purandar has quit IRC | 22:45 | |
| minwang2 | now i see 2 backend nodes in cfg | 22:45 |
| sbalukoff | Good! | 22:46 |
| sbalukoff | Try your curl commands again (er... the corrected ones. ;) ) | 22:46 |
| *** Purandar has joined #openstack-lbaas | 22:47 | |
| minwang2 | wait a sec | 22:47 |
| minwang2 | curl http://10.0.0.5/some/path/with/cat | 22:48 |
| minwang2 | Welcome to 10.0.0.4 | 22:48 |
| minwang2 | finally it reaches the other node | 22:48 |
| minwang2 | thanks a lot!! | 22:48 |
| sbalukoff | Yay! | 22:48 |
| sbalukoff | Can you try the header rule check as well? | 22:48 |
| minwang2 | hang on | 22:49 |
| sbalukoff | (That's this one: curl --header "X-My-Header: something with thisor that in it" http://10.0.0.5/ ) | 22:49 |
| sbalukoff | Should redirect you to google, if we've done things correctly. | 22:50 |
| minwang2 | test@test-virtual-machine:~$ curl --header "X-My-Header: something with thisor that in it" http://10.0.0.5/ | 22:51 |
| minwang2 | test@test-virtual-machine:~$ | 22:51 |
| sbalukoff | Hmmm... | 22:51 |
| minwang2 | the header should redirect_to_url | 22:51 |
| sbalukoff | Yeah, but curl might be encoding spaces there in ways we don't expect... | 22:53 |
| minwang2 | i put the log in the gist | 22:53 |
| minwang2 | please check to see if i did the right way | 22:53 |
| sbalukoff | (Spaces are technically allowed by the specifications, but browsers and utilities like this will sometimes do URL encoding on them, which turns spaces into %20 ) | 22:53 |
| sbalukoff | Oh! | 22:54 |
| sbalukoff | It did it right. | 22:54 |
| sbalukoff | < HTTP/1.1 302 Found | 22:54 |
| sbalukoff | < Cache-Control: no-cache | 22:54 |
| sbalukoff | < Content-length: 0 | 22:54 |
| sbalukoff | < Location: http://www.google.com | 22:54 |
| sbalukoff | < Connection: close | 22:54 |
| sbalukoff | That's your redirect. | 22:54 |
| sbalukoff | So... yay! | 22:54 |
| minwang2 | check the new one, it works | 22:54 |
| sbalukoff | Nice! | 22:55 |
| sbalukoff | Ok, so curl just wasn't following the redirect. | 22:55 |
| sbalukoff | Which is to be expected, I guess. XD | 22:55 |
| sbalukoff | Ok, yay! So it works. | 22:55 |
| sbalukoff | Any other questions right now? | 22:55 |
| minwang2 | curl --header "X-My-Header: with thisorthat" 10.0.0.5 | 22:56 |
| minwang2 | Welcome to 10.0.0.3 | 22:56 |
| minwang2 | no space, not sure if it is the reason | 22:56 |
| sbalukoff | "thisorthat" != "thisor that" | 22:56 |
| sbalukoff | Yep. | 22:56 |
| minwang2 | curl --header "X-My-Header: with thisor that" 10.0.0.5 | 22:57 |
| minwang2 | this will not work | 22:57 |
| sbalukoff | Add the '-v' to see your redirect. | 22:57 |
| minwang2 | hang on | 22:58 |
| sbalukoff | Or just add '-L' to tell curl to follow the redirect. | 22:58 |
| minwang2 | let me try | 22:59 |
| sbalukoff | Also, if you're baking this into a test somewhere, then it's conventional to use 'http://www.example.com/' instead of the actual website of a big corporation. ;) | 23:00 |
| minwang2 | haha, let me change it | 23:01 |
| sbalukoff | No worries. Does the test work, though? | 23:01 |
| sbalukoff | (I'm about to get dragged away for some internal stuff. xgerman: Sorry I haven't gotten to your patch yet. If I don't get to it today, I'll get to it reeeeally late tonight or tomorrow. I'd like to see it land for M-3 as well, eh!) | 23:01 |
| xgerman | yeah, maybe I can strong-arm johnsom to have a look or rm_work | 23:02 |
| minwang2 | nope | 23:03 |
| sbalukoff | minwang2: Can you add the output of your latest test to the gist? | 23:03 |
| minwang2 | it is the same | 23:03 |
| sbalukoff | Lines 249-266? | 23:03 |
| sbalukoff | That shows that the test is working. | 23:03 |
| minwang2 | also i am curious, why for the pool —individual-pool2 did not work | 23:04 |
| *** neelashah has quit IRC | 23:04 | |
| sbalukoff | The 302 response is your REDIRECT_TO_URL response. | 23:04 |
| minwang2 | yes | 23:04 |
| sbalukoff | With the Location: header. | 23:04 |
| *** jwarendt has quit IRC | 23:04 | |
| minwang2 | so | 23:04 |
| sbalukoff | minwang2: when you created the REDIRECT_TO_POOL policy, you specified 'pool2-out-listener1' not 'individual_pool2' So you don't actually have any policy pointing at 'individual_pool2' right now. | 23:05 |
| minwang2 | interesting,so i dont have a policy which is doing REDICRECT_TO_POOL, but create an independent pool, inside of the pool,i create a new member, in that case, the load balancer will not direct to this new member, correct? | 23:07 |
| sbalukoff | I don't understand what you mean when you say "create an independent pool, inside of the pool" | 23:07 |
| sbalukoff | You *do* have a policy doing REDIRECT_TO_POOL. When you created it, you told it to redirect to the 'poo2-out-listener1' pool. | 23:08 |
| sbalukoff | pool2-out-listener1 | 23:08 |
| sbalukoff | Sorry. | 23:08 |
| *** ducttape_ has quit IRC | 23:10 | |
| minwang2 | no, i mean create an indepdent pool inside of lb while outside of listener1 | 23:10 |
| minwang2 | in this scenario, there are 2 policies with listener1: redict_to_url and reject | 23:12 |
| sbalukoff | If you create an 'independent pool' inside a load balancer and assign a member to it, the load balanacer will send requests to it if it is referenced as some listener's default_pool, or if some l7policy with REDIRECT_TO_POOL references it. | 23:12 |
| *** Bjoern_ has quit IRC | 23:12 | |
| sbalukoff | Actually, according to your database, all three of the policies you have created are on listener1 | 23:13 |
| sbalukoff | listener2 exists on the load balancer, but it doesn't have any l7policies. | 23:14 |
| minwang2 | correct | 23:15 |
| sbalukoff | And, again, according to your database, listener2 has individual_pool2 as its default pool. | 23:15 |
| minwang2 | correct | 23:16 |
| sbalukoff | Ok. | 23:16 |
| *** jwarendt has joined #openstack-lbaas | 23:17 | |
| minwang2 | so back to our previous question, the header,is there anything that i set wrong | 23:17 |
| madhu_ak | rm_work, sorry I was away. checkign now | 23:17 |
| sbalukoff | No, your header test is working. | 23:17 |
| *** yamamoto has joined #openstack-lbaas | 23:17 | |
| minwang2 | oh yeah | 23:18 |
| minwang2 | in that case, we are all set | 23:18 |
| sbalukoff | Lines 260-264 in your gist show the proper response for REDIRECT_TO_URL | 23:18 |
| sbalukoff | Ok, cool! | 23:18 |
| minwang2 | yes | 23:18 |
| sbalukoff | Ok, gonna go AFK for a bit. | 23:19 |
| minwang2 | thanks a lot sbalukoff | 23:19 |
| *** Purandar has quit IRC | 23:20 | |
| *** Purandar has joined #openstack-lbaas | 23:21 | |
| *** yamamoto has quit IRC | 23:23 | |
| *** armax has quit IRC | 23:23 | |
| *** armax has joined #openstack-lbaas | 23:25 | |
| madhu_ak | rm_work, seeing the same problem after getting your fix. I have unstacked and restacking with your fix | 23:41 |
| rm_you | hmmmmmmmmm | 23:46 |
| rm_you | k | 23:46 |
| *** yamamoto has joined #openstack-lbaas | 23:54 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!