| rm_work | xgerman_: you SHOULD probably review https://review.openstack.org/#/c/504175/ though | 00:00 |
|---|---|---|
| rm_work | that's going to change how we use TLS Certs | 00:01 |
| johnsom | Yeah, that is on my review short list too | 00:01 |
| rm_work | my guess is, there will be some -1s the first round ;P | 00:03 |
| rm_work | brb | 00:04 |
| xgerman_ | Ok. Will look at it when kids are in bed ;-) | 00:07 |
| *** sanfern has quit IRC | 00:31 | |
| rm_work | lolk | 00:33 |
| *** sanfern has joined #openstack-lbaas | 00:34 | |
| *** cody-somerville has joined #openstack-lbaas | 00:36 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 00:42 | |
| *** AlexeyAbashkin has quit IRC | 00:47 | |
| *** cody-somerville has quit IRC | 00:51 | |
| *** cody-somerville has joined #openstack-lbaas | 01:02 | |
| *** cody-somerville has quit IRC | 01:02 | |
| *** cody-somerville has joined #openstack-lbaas | 01:02 | |
| *** cody-somerville has quit IRC | 01:07 | |
| *** cody-somerville has joined #openstack-lbaas | 01:11 | |
| *** tongl has joined #openstack-lbaas | 01:17 | |
| *** cody-somerville has quit IRC | 01:39 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 01:45 | |
| *** AlexeyAbashkin has quit IRC | 01:49 | |
| *** sanfern has quit IRC | 01:57 | |
| *** sanfern has joined #openstack-lbaas | 01:58 | |
| *** sanfern has joined #openstack-lbaas | 01:59 | |
| *** sanfern has quit IRC | 01:59 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 02:43 | |
| *** AlexeyAbashkin has quit IRC | 02:48 | |
| *** links has joined #openstack-lbaas | 02:51 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 03:03 |
| rm_work | johnsom: we still having issues with certain hosts spawning instances with kvm? | 03:08 |
| rm_work | http://logs.openstack.org/45/511045/4/check/gate-octavia-v1-dsvm-scenario-ubuntu-xenial/3e67c3f/ | 03:08 |
| johnsom | rm_work yep, ovh still has kvm problems | 03:10 |
| rm_work | :/ | 03:10 |
| rm_work | it's been like, a year | 03:10 |
| rm_work | though i guess the danger is that if we complain too much they might just *disable it* instead of fixing it, lol | 03:11 |
| johnsom | It was fixed for a while but started again | 03:11 |
| rm_work | bleh | 03:11 |
| johnsom | It could be the infra image too, not sure | 03:12 |
| johnsom | But I only see it fail on the ovh hosts | 03:14 |
| openstackgerrit | Chenghui Yu proposed openstack/neutron-lbaas master: Fix usage of method ensure_dir https://review.openstack.org/502932 | 03:54 |
| openstackgerrit | Chenghui Yu proposed openstack/neutron-lbaas master: Update link in the README when using devstack to install neutron-lbaas https://review.openstack.org/495202 | 04:01 |
| *** pcaruana has joined #openstack-lbaas | 04:27 | |
| *** aojea has joined #openstack-lbaas | 04:33 | |
| *** aojea has quit IRC | 04:37 | |
| *** armax has quit IRC | 04:50 | |
| *** armax has joined #openstack-lbaas | 04:51 | |
| *** armax has quit IRC | 04:51 | |
| *** armax has joined #openstack-lbaas | 04:51 | |
| *** armax has quit IRC | 04:52 | |
| *** armax has joined #openstack-lbaas | 04:52 | |
| *** armax has quit IRC | 04:53 | |
| *** pcaruana has quit IRC | 04:56 | |
| *** cody-somerville has joined #openstack-lbaas | 05:00 | |
| *** cody-somerville has quit IRC | 05:00 | |
| *** cody-somerville has joined #openstack-lbaas | 05:00 | |
| *** cody-somerville has quit IRC | 05:04 | |
| *** avtar_ has joined #openstack-lbaas | 05:05 | |
| *** cristicalin has joined #openstack-lbaas | 05:10 | |
| *** chlong has joined #openstack-lbaas | 05:20 | |
| *** cristicalin has quit IRC | 05:23 | |
| *** pcaruana has joined #openstack-lbaas | 05:30 | |
| *** slaweq_ has joined #openstack-lbaas | 05:32 | |
| *** slaweq_ has quit IRC | 05:33 | |
| *** pcaruana has quit IRC | 05:39 | |
| *** rcernin has joined #openstack-lbaas | 05:41 | |
| *** cristicalin has joined #openstack-lbaas | 05:43 | |
| *** gcheresh_ has joined #openstack-lbaas | 05:53 | |
| *** spectr has quit IRC | 05:58 | |
| *** spectr has joined #openstack-lbaas | 05:58 | |
| *** slaweq_ has joined #openstack-lbaas | 06:09 | |
| *** cristicalin has quit IRC | 06:17 | |
| *** gcheresh_ has quit IRC | 06:23 | |
| *** gcheresh_ has joined #openstack-lbaas | 06:37 | |
| *** slaweq_ has quit IRC | 06:40 | |
| *** slaweq_ has joined #openstack-lbaas | 06:44 | |
| *** gcheresh_ has quit IRC | 06:47 | |
| *** Alex_Staf_ has joined #openstack-lbaas | 06:54 | |
| *** pcaruana has joined #openstack-lbaas | 07:07 | |
| *** slaweq_ has quit IRC | 07:17 | |
| *** slaweq_ has joined #openstack-lbaas | 07:20 | |
| *** gcheresh_ has joined #openstack-lbaas | 07:22 | |
| *** tesseract has joined #openstack-lbaas | 07:22 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 07:24 | |
| *** spectr has quit IRC | 07:26 | |
| *** tesseract has quit IRC | 07:26 | |
| *** rcernin has quit IRC | 07:26 | |
| *** pcaruana has quit IRC | 07:26 | |
| *** spectr has joined #openstack-lbaas | 07:39 | |
| *** pcaruana has joined #openstack-lbaas | 07:40 | |
| *** rcernin has joined #openstack-lbaas | 07:40 | |
| *** gcheresh_ has quit IRC | 07:42 | |
| *** spectr has quit IRC | 07:43 | |
| *** pcaruana has quit IRC | 07:44 | |
| *** spectr has joined #openstack-lbaas | 07:44 | |
| *** pcaruana has joined #openstack-lbaas | 07:44 | |
| *** slaweq_ has quit IRC | 07:53 | |
| *** tongl has quit IRC | 07:58 | |
| *** slaweq_ has joined #openstack-lbaas | 08:02 | |
| *** AlexeyAbashkin has quit IRC | 08:10 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 08:11 | |
| *** gcheresh_ has joined #openstack-lbaas | 08:11 | |
| *** Alex_Staf_ has quit IRC | 08:21 | |
| *** slaweq_ has quit IRC | 08:35 | |
| *** slaweq_ has joined #openstack-lbaas | 08:44 | |
| *** aojea has joined #openstack-lbaas | 08:47 | |
| *** salmankhan has joined #openstack-lbaas | 08:53 | |
| *** Alex_Staf_ has joined #openstack-lbaas | 08:54 | |
| *** salmankhan1 has joined #openstack-lbaas | 08:57 | |
| *** salmankhan has quit IRC | 08:57 | |
| *** salmankhan1 is now known as salmankhan | 08:57 | |
| *** rcernin has quit IRC | 08:59 | |
| *** spectr has quit IRC | 08:59 | |
| *** pcaruana has quit IRC | 08:59 | |
| *** AlexeyAbashkin has quit IRC | 09:03 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 09:03 | |
| *** rcernin has joined #openstack-lbaas | 09:12 | |
| *** spectr has joined #openstack-lbaas | 09:12 | |
| *** pcaruana has joined #openstack-lbaas | 09:13 | |
| *** AlexeyAbashkin has quit IRC | 09:15 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 09:15 | |
| *** slaweq_ has quit IRC | 09:17 | |
| *** slaweq_ has joined #openstack-lbaas | 09:21 | |
| *** salmankhan has quit IRC | 09:26 | |
| *** salmankhan has joined #openstack-lbaas | 09:31 | |
| *** chlong has quit IRC | 09:40 | |
| *** slaweq_ has quit IRC | 09:55 | |
| *** slaweq_ has joined #openstack-lbaas | 10:00 | |
| *** aojea has quit IRC | 10:22 | |
| *** aojea has joined #openstack-lbaas | 10:23 | |
| *** aojea has quit IRC | 10:27 | |
| *** salmankhan has quit IRC | 10:45 | |
| *** aojea has joined #openstack-lbaas | 10:46 | |
| *** aojea has quit IRC | 10:47 | |
| *** pcaruana has quit IRC | 10:48 | |
| *** yamamoto has quit IRC | 10:48 | |
| *** salmankhan has joined #openstack-lbaas | 11:02 | |
| *** slaweq_ has quit IRC | 11:03 | |
| *** slaweq_ has joined #openstack-lbaas | 11:10 | |
| *** aojea has joined #openstack-lbaas | 11:18 | |
| *** aojea has quit IRC | 11:23 | |
| *** salmankhan has quit IRC | 11:38 | |
| *** salmankhan has joined #openstack-lbaas | 11:39 | |
| *** yamamoto has joined #openstack-lbaas | 11:40 | |
| *** gcheresh_ has quit IRC | 11:41 | |
| *** slaweq_ has quit IRC | 11:43 | |
| *** yamamoto has quit IRC | 11:45 | |
| *** yamamoto has joined #openstack-lbaas | 11:48 | |
| *** yamamoto has quit IRC | 11:54 | |
| *** isantosp has quit IRC | 12:15 | |
| *** isantosp has joined #openstack-lbaas | 12:18 | |
| *** aojea has joined #openstack-lbaas | 12:19 | |
| *** slaweq_ has joined #openstack-lbaas | 12:23 | |
| *** aojea has quit IRC | 12:23 | |
| isantosp | Is it possible in Pike release or future plans to create a loadBalancer without VIP? I think u had a work around but some inconsistencies of this, is that right? | 12:23 |
| *** yamamoto has joined #openstack-lbaas | 12:28 | |
| *** yamamoto has quit IRC | 12:29 | |
| *** links has quit IRC | 12:45 | |
| *** pcaruana has joined #openstack-lbaas | 12:48 | |
| *** slaweq_ has quit IRC | 12:56 | |
| *** spectr has quit IRC | 12:57 | |
| *** slaweq_ has joined #openstack-lbaas | 13:01 | |
| *** spectr has joined #openstack-lbaas | 13:01 | |
| *** aojea has joined #openstack-lbaas | 13:20 | |
| *** aojea has quit IRC | 13:24 | |
| *** yamamoto has joined #openstack-lbaas | 13:30 | |
| *** sanfern has joined #openstack-lbaas | 13:32 | |
| *** chlong has joined #openstack-lbaas | 13:33 | |
| *** slaweq has quit IRC | 13:33 | |
| *** slaweq_ has quit IRC | 13:34 | |
| openstackgerrit | Dong Jun proposed openstack/octavia master: import _ from octavia.i18n https://review.openstack.org/511846 | 13:36 |
| openstackgerrit | Dong Jun proposed openstack/octavia master: Valuable change reported by pylint https://review.openstack.org/511847 | 13:36 |
| openstackgerrit | Dong Jun proposed openstack/octavia master: Trivial change to adapt pylint rules https://review.openstack.org/511848 | 13:36 |
| openstackgerrit | Dong Jun proposed openstack/octavia master: Support pylint https://review.openstack.org/511849 | 13:36 |
| *** yamamoto has quit IRC | 13:40 | |
| *** slaweq has joined #openstack-lbaas | 13:41 | |
| *** rcernin has quit IRC | 13:44 | |
| *** yamamoto has joined #openstack-lbaas | 13:50 | |
| *** reedip has quit IRC | 13:50 | |
| *** reedip has joined #openstack-lbaas | 14:02 | |
| *** armax has joined #openstack-lbaas | 14:11 | |
| *** slaweq has quit IRC | 14:13 | |
| *** leitan has joined #openstack-lbaas | 14:18 | |
| *** aojea has joined #openstack-lbaas | 14:20 | |
| *** yamamoto has quit IRC | 14:21 | |
| *** slaweq has joined #openstack-lbaas | 14:23 | |
| *** catintheroof has joined #openstack-lbaas | 14:23 | |
| *** aojea has quit IRC | 14:25 | |
| *** sanfern has quit IRC | 14:25 | |
| *** sanfern has joined #openstack-lbaas | 14:26 | |
| *** yamamoto has joined #openstack-lbaas | 14:41 | |
| *** KeithMnemonic1 has joined #openstack-lbaas | 14:43 | |
| *** rcernin has joined #openstack-lbaas | 14:45 | |
| *** spectr has quit IRC | 14:46 | |
| *** KeithMnemonic has quit IRC | 14:47 | |
| *** reedip has quit IRC | 14:53 | |
| *** slaweq has quit IRC | 14:56 | |
| *** rcernin has quit IRC | 15:03 | |
| *** slaweq has joined #openstack-lbaas | 15:03 | |
| *** reedip has joined #openstack-lbaas | 15:05 | |
| *** tongl has joined #openstack-lbaas | 15:07 | |
| *** fnaval has joined #openstack-lbaas | 15:19 | |
| *** fnaval_ has joined #openstack-lbaas | 15:20 | |
| *** aojea has joined #openstack-lbaas | 15:21 | |
| *** fnaval has quit IRC | 15:23 | |
| *** aojea has quit IRC | 15:25 | |
| *** krypto has joined #openstack-lbaas | 15:35 | |
| *** slaweq has quit IRC | 15:36 | |
| *** slaweq has joined #openstack-lbaas | 15:41 | |
| *** chlong has quit IRC | 15:53 | |
| *** AlexeyAbashkin has quit IRC | 15:55 | |
| *** fnaval_ has quit IRC | 16:02 | |
| *** slaweq has quit IRC | 16:14 | |
| *** slaweq has joined #openstack-lbaas | 16:16 | |
| *** slaweq has quit IRC | 16:18 | |
| *** slaweq has joined #openstack-lbaas | 16:18 | |
| *** aojea has joined #openstack-lbaas | 16:22 | |
| *** krypto has quit IRC | 16:23 | |
| *** pcaruana has quit IRC | 16:23 | |
| *** aojea has quit IRC | 16:27 | |
| *** krypto has joined #openstack-lbaas | 16:28 | |
| *** yamamoto has quit IRC | 16:29 | |
| *** krypto has quit IRC | 16:39 | |
| *** salmankhan has quit IRC | 16:42 | |
| *** slaweq has quit IRC | 17:06 | |
| *** yamamoto has joined #openstack-lbaas | 17:09 | |
| *** yamamoto has quit IRC | 17:09 | |
| *** aojea has joined #openstack-lbaas | 17:23 | |
| *** catintheroof has quit IRC | 17:25 | |
| *** catintheroof has joined #openstack-lbaas | 17:25 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 17:27 | |
| *** aojea has quit IRC | 17:27 | |
| *** AlexeyAbashkin has quit IRC | 17:28 | |
| *** catintheroof has quit IRC | 17:29 | |
| *** sshank has joined #openstack-lbaas | 17:33 | |
| *** slaweq has joined #openstack-lbaas | 17:37 | |
| *** armax has quit IRC | 17:42 | |
| *** dosaboy has quit IRC | 17:49 | |
| *** fnaval has joined #openstack-lbaas | 17:50 | |
| *** armax has joined #openstack-lbaas | 17:52 | |
| *** salmankhan has joined #openstack-lbaas | 18:04 | |
| *** salmankhan has quit IRC | 18:08 | |
| *** yamamoto has joined #openstack-lbaas | 18:10 | |
| *** sanfern has quit IRC | 18:10 | |
| *** yamamoto has quit IRC | 18:19 | |
| *** aojea has joined #openstack-lbaas | 18:23 | |
| *** aojea has quit IRC | 18:28 | |
| *** catintheroof has joined #openstack-lbaas | 18:33 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 18:35 |
| *** armax_ has joined #openstack-lbaas | 18:40 | |
| rm_work | johnsom: are L7Rules put together with AND? | 18:42 |
| *** armax has quit IRC | 18:42 | |
| *** armax_ is now known as armax | 18:42 | |
| *** slaweq has quit IRC | 18:43 | |
| rm_work | ok yeah, docs say yes | 18:43 |
| *** slaweq has joined #openstack-lbaas | 18:48 | |
| *** dosaboy has joined #openstack-lbaas | 18:48 | |
| *** dosaboy has quit IRC | 18:48 | |
| rm_work | johnsom: comment on ERD | 18:49 |
| *** dosaboy has joined #openstack-lbaas | 18:50 | |
| johnsom | I was just going to paste a link to the L7 doc | 18:55 |
| johnsom | https://docs.openstack.org/octavia/latest/user/guides/l7.html | 18:55 |
| johnsom | Thanks, will take a look after lunch | 18:55 |
| *** fnaval has quit IRC | 19:02 | |
| rm_work | xgerman_: https://review.openstack.org/#/c/511848/1/octavia/controller/worker/tasks/network_tasks.py | 19:03 |
| rm_work | johnsom: so I feel like ... did we support pylint before, and remove it? | 19:03 |
| rm_work | because I SWEAR we did | 19:03 |
| *** gcheresh_ has joined #openstack-lbaas | 19:07 | |
| *** gcheresh_ has quit IRC | 19:14 | |
| *** slaweq has quit IRC | 19:21 | |
| *** slaweq has joined #openstack-lbaas | 19:21 | |
| *** aojea has joined #openstack-lbaas | 19:24 | |
| *** sshank has quit IRC | 19:26 | |
| *** aojea has quit IRC | 19:29 | |
| *** leitan has quit IRC | 19:34 | |
| *** sshank has joined #openstack-lbaas | 19:35 | |
| *** leitan has joined #openstack-lbaas | 19:35 | |
| *** sshank has quit IRC | 19:39 | |
| *** leitan has quit IRC | 19:40 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 19:40 | |
| *** fnaval has joined #openstack-lbaas | 19:40 | |
| *** Swami has joined #openstack-lbaas | 19:41 | |
| johnsom | rm_work: you are correct | 19:41 |
| rm_work | well, whatever, pylint++ | 19:42 |
| johnsom | I thought all of OpenStack pulled it due to stability issues | 19:42 |
| rm_work | maybe it's fixed ;P | 19:42 |
| rm_work | i don't know how it could be *unstable* | 19:42 |
| *** fnaval has quit IRC | 19:42 | |
| rm_work | it's a freaking linter | 19:42 |
| johnsom | I mean, I like it in principle, but yeah, we pulled it in the past | 19:42 |
| rm_work | did it seriously crash a lot? T_T | 19:42 |
| rm_work | well, i'm +2 on most of that chain | 19:42 |
| *** fnaval has joined #openstack-lbaas | 19:43 | |
| johnsom | I don’t remember what the issue was | 19:43 |
| johnsom | Ok, I will review and +w when I get back from lunch | 19:44 |
| *** AlexeyAbashkin has quit IRC | 19:44 | |
| rm_work | there's one fix i want on the third one, i think | 19:45 |
| rm_work | oh, just wanted xgerman_ to look | 19:45 |
| xgerman_ | yo | 19:46 |
| *** blallau has joined #openstack-lbaas | 19:49 | |
| *** blallau has quit IRC | 19:49 | |
| rm_work | xgerman_: do you know what this was about: https://review.openstack.org/#/c/511848/1/octavia/controller/worker/tasks/network_tasks.py | 19:53 |
| *** Swami has quit IRC | 19:53 | |
| xgerman_ | ok, we ignore the error and I am worried that’s not right since we will still be connected to that network —- | 19:54 |
| xgerman_ | chewing up ports, etc. — but haven’t come up with what else to do (e.g. a retry) | 19:55 |
| xgerman_ | and us dropping the ball will make it necessary for people to go to greater lengths during clean up | 19:58 |
| xgerman_ | or you just run neutron purge | 19:58 |
| rm_work | so, should we preserve the comment? | 19:58 |
| *** Swami has joined #openstack-lbaas | 19:58 | |
| xgerman_ | yes | 19:58 |
| rm_work | k | 19:58 |
| xgerman_ | I will file an issue so we have some more visibility | 19:58 |
| *** sshank has joined #openstack-lbaas | 20:16 | |
| *** KeithMnemonic1 has quit IRC | 20:19 | |
| *** aojea has joined #openstack-lbaas | 20:25 | |
| rm_work | johnsom: ugh, my change is not having a good time | 20:28 |
| rm_work | https://review.openstack.org/#/c/511045/ | 20:28 |
| rm_work | alternating gates are failing | 20:28 |
| rm_work | we may be at or near critical mass again for hosts that fail * number of gates | 20:29 |
| *** aojea has quit IRC | 20:30 | |
| *** fnaval has quit IRC | 20:34 | |
| *** fnaval has joined #openstack-lbaas | 20:34 | |
| johnsom | rm_work It's probably related to the gates issue this week and zuulv3. The current pool of working gate hosts may have a higher number of OVH hosts... | 20:37 |
| johnsom | BTW, there was one voting gate that failed on that patch... | 20:37 |
| johnsom | rm_work On the ERD thing, is it a good enough start or do you really want me fiddling with graphviz settings for the rest of the day instead of reviews, etc.? | 20:39 |
| johnsom | Personally I hope there is someone with better graphviz fu that will come along and feel pity on use. | 20:39 |
| johnsom | us | 20:39 |
| rm_work | k its fine | 20:42 |
| rm_work | i think i did +1 not -1 so | 20:42 |
| rm_work | i was checking in case it was easy | 20:42 |
| johnsom | I personally hate graphviz, but having always update diagrams out weighs my distaste | 20:43 |
| *** sshank has quit IRC | 20:53 | |
| *** sshank has joined #openstack-lbaas | 20:56 | |
| *** slaweq has quit IRC | 21:04 | |
| *** catintheroof has quit IRC | 21:10 | |
| *** armax has quit IRC | 21:25 | |
| *** aojea has joined #openstack-lbaas | 21:26 | |
| *** aojea has quit IRC | 21:30 | |
| *** armax has joined #openstack-lbaas | 21:36 | |
| *** slaweq has joined #openstack-lbaas | 21:37 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 21:39 | |
| *** AlexeyAbashkin has quit IRC | 21:44 | |
| *** slaweq has quit IRC | 22:10 | |
| *** colby_ has joined #openstack-lbaas | 22:13 | |
| colby_ | johsom: Thanks for letting me know about this channel. So Im looking at the octavia docs. It says the controll services need to be able to reach the lb network (which is a network in openstack). Our current openstack setup is 2 controller nodes, 2 network nodes (l3 ha agents), and the compute nodes. Since it sounds like the octavia controller processes need access to the lb network would I need to run those on the network nodes? | 22:15 |
| colby_ | Our controllers dont have access to the tenant networks | 22:16 |
| colby_ | johnsom: oops mispelled | 22:17 |
| johnsom | There is not requirement that they be on the network nodes. The LB mgmt network is just a private network used to control the "amphora" and collect health and statistics. It can be a provider network or any other neutron visible network | 22:17 |
| colby_ | So this would be seperate form the management, and tenant networks? | 22:18 |
| johnsom | We tell nova to attach it when the service VM boots, so it needs to be connected into neutron on the service VM side. Some deployers just make the controllers have access to a VLAN provider network | 22:18 |
| johnsom | The lb mgmt net is only between the controllers and the service VMs. It is isolated from the tenant networks with a network namespace in the service VMs | 22:18 |
| *** slaweq has joined #openstack-lbaas | 22:19 | |
| colby_ | would the compute nodes need to have this network plumbed in (on the physical side) or is that all managed by openstack accross the tenant network? | 22:20 |
| johnsom | It can be a routed path too. | 22:20 |
| johnsom | neutron on the compute nodes needs to be able to plug this network into the service VM instances when nova boots them. It doesn't have to be physical though. We just create a neutron network for it in the Octavia service account, then either connect the controllers into that or bridge/route it to a provider network the controllers have access to. | 22:22 |
| *** cody-somerville has joined #openstack-lbaas | 22:24 | |
| colby_ | ok so our setup. We have a 10.2.x.x vlan setup for management, a 10.3.x.x vlan setup for ceph, and a 10.4.x.x network setup as the tenant network. Neutron uses GRE tunnels between the hypervisors, and the network nodes over the 10.4.x.x network. | 22:25 |
| johnsom | I guess I should say, could. Basically there is IP connectivity from the controller process to the amphora and from the amphora back to the health manager processes | 22:25 |
| *** aojea has joined #openstack-lbaas | 22:26 | |
| johnsom | No other traffic crosses that network. | 22:27 |
| colby_ | I can create another vlan to link the hypervisors, network nodes, and controllers if necessary. then just set that up as a providor network in openstack? | 22:27 |
| johnsom | Sure, that would work | 22:27 |
| colby_ | Is there a way to do it without having to create new vlan in our switches and plumming that into the trunk? | 22:28 |
| johnsom | The part you need to figure out is how to connect your controllers. | 22:28 |
| johnsom | For the service VMs (amphora) you can just use a neutron network like you do for any other VM or tenant network. It's just owned by the octavia service account instead of a tenant. | 22:29 |
| johnsom | Then from that neutron network, you need to figure out how you want to plumb it to he controller processes. You could create a provider network for the controllers, then add a router between that provider network and the lb-mgmt-net, or you could bridge it, etc. | 22:30 |
| *** aojea has quit IRC | 22:30 | |
| johnsom | If you run the octavia controller processes on your network node, you can just locally bridge it like we do in devstack. | 22:31 |
| colby_ | ok that may be easier then | 22:31 |
| johnsom | https://github.com/openstack/octavia/blob/master/devstack/plugin.sh#L315 | 22:32 |
| johnsom | There is a lot of flexibility in how you deploy it. | 22:32 |
| johnsom | If you use openstack ansible it sets this network up for you | 22:33 |
| colby_ | we use puppet | 22:33 |
| johnsom | There is a puppet module for octavia (probably the old neutron-lbaas way though) | 22:34 |
| johnsom | I'm just not sure how they set that up. | 22:34 |
| johnsom | https://github.com/openstack/puppet-octavia | 22:34 |
| colby_ | ok Ill dig more on this and let you know if I have any more questions. | 22:34 |
| colby_ | thanks! | 22:35 |
| johnsom | Since you are starting a fresh install, you probably want to use the native Octavia v2 API and not use neutron-lbaas | 22:35 |
| johnsom | You can run Octavia Pike release even if the rest of your cloud is older. | 22:35 |
| colby_ | ok cool. Ill take a look at that. | 22:36 |
| johnsom | colby_ Sure, NP. Folks are usually here in the channel, so don't be shy | 22:36 |
| colby_ | So is that code you linked creating the connection to lb network. So if we ran the controllers on the network nodes, it could just use openvswitch to create the bridge to the network? | 22:39 |
| johnsom | Correct | 22:39 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 22:39 | |
| johnsom | That is exactly what we are doing there | 22:39 |
| colby_ | So it would do that automatically then? | 22:39 |
| colby_ | If so then that is much simpler and will make sure its always setup correctl | 22:40 |
| johnsom | No, that is our devstack plugin script. It's a developer deployment script | 22:40 |
| colby_ | oh I see. Ok we I can just run the openvswitch command I see there | 22:40 |
| johnsom | Yeah, you could basically copy what it is doing into your puppet that deploys the octavia controller processes. | 22:41 |
| *** AlexeyAbashkin has quit IRC | 22:43 | |
| *** Swami has quit IRC | 22:45 | |
| *** slaweq has quit IRC | 22:52 | |
| rm_work | colby_: for the record, we use one public network for both the VIP (we require users provide a specific network for the VIP...) and for management | 22:54 |
| rm_work | it works fine | 22:54 |
| rm_work | honestly the management traffic is secure enough that it's not a huge issue to run it over a public network -- just recommended to avoid it if possible (for us it would be extremely difficult) | 22:54 |
| colby_ | Everything in our setup is on private network except our api load balancers, and network nodes have external net | 22:55 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Trivial change to adapt pylint rules https://review.openstack.org/511848 | 22:56 |
| rm_work | colby_: ok, so just saying you can share any other network too, if you have some private provider network as well | 22:57 |
| rm_work | really it doesn't matter | 22:57 |
| colby_ | rm_work: great thanks! | 22:57 |
| rm_work | it just needs *a network* that can be plugged into every service VM | 22:57 |
| rm_work | johnsom: did you weigh in on https://review.openstack.org/#/c/494407/ | 22:57 |
| rm_work | resurrected, trivial change | 22:57 |
| rm_work | these gate failures T_T | 22:58 |
| johnsom | On the fence on that one. I get Nir's perspective. I'm also half tempted to just point the governance stance on this | 22:58 |
| rm_work | what is the governance stance? | 22:58 |
| johnsom | And drop the 16.x all together | 22:58 |
| rm_work | lol | 22:58 |
| rm_work | it's just a freaking "recommended install guide" | 22:59 |
| rm_work | we can't be opinionated? | 22:59 |
| johnsom | https://governance.openstack.org/tc/reference/project-testing-interface.html | 22:59 |
| johnsom | I guess, really, my stance is it shouldn't be there unless there are gates.... | 23:00 |
| *** slaweq has joined #openstack-lbaas | 23:00 | |
| johnsom | So, 16.04 is fine. If we want to add others, there should be gates | 23:00 |
| rm_work | it's for telling people how to contribute | 23:00 |
| rm_work | i'll wait and try to convince Nir | 23:00 |
| rm_work | not like it's a rush :P | 23:00 |
| johnsom | I commented | 23:01 |
| johnsom | But will wait for Nir as he -1 | 23:01 |
| rm_work | yep | 23:01 |
| johnsom | rm_work the OVH thing... I don't have a good path forward on that. I mean I don't want to put in code that checks if it's OVH and drops qemu to TCG.... | 23:03 |
| johnsom | I did open a but upstream with KVM, but they wanted lower level details than I have, so it's just sitting open | 23:04 |
| johnsom | Part of it is probably that they are running artful qemu on a xenial host... | 23:06 |
| johnsom | ii qemu-system-x86 1:2.10+dfsg-0ubuntu1~cloud0 amd64 QEMU full system emulation binaries (x86) | 23:07 |
| johnsom | https://packages.ubuntu.com/search?keywords=qemu-system-x86&searchon=names&suite=all§ion=all | 23:07 |
| *** m-greene has quit IRC | 23:09 | |
| rm_work | bleh | 23:10 |
| johnsom | Yeah, that is most likely the issue. Similar to the problem I have with devstack pulling in the ocata apt repo and a newer qemu. | 23:10 |
| rm_work | yeah i don't suppose we can ... downgrade | 23:10 |
| rm_work | as part of our gate | 23:10 |
| rm_work | lol | 23:10 |
| johnsom | Hmmm, | 23:10 |
| *** m-greene_ has quit IRC | 23:10 | |
| johnsom | We *might* be able to trick nova into running the right code.... | 23:10 |
| johnsom | Like I do | 23:10 |
| rm_work | yeah | 23:10 |
| rm_work | i mean | 23:10 |
| rm_work | the same thing i put in my devstack config | 23:10 |
| johnsom | Right. | 23:11 |
| johnsom | Problem is we need to wait for zuul v3. They won't merge a project-config change now. | 23:11 |
| rm_work | https://github.com/rm-you/devstack_deploy/blob/master/local.conf#L29-L37 | 23:11 |
| rm_work | eugh | 23:11 |
| johnsom | Yep, that | 23:11 |
| rm_work | why does it only break SOME of the time tho | 23:12 |
| rm_work | it's ... weird | 23:12 |
| johnsom | It's pretty much always on the OVH hosts | 23:12 |
| johnsom | I don't know what the combination is, but it's those hosts | 23:12 |
| rm_work | augh cause ... we're failing over half the time | 23:13 |
| rm_work | i mean, at least one voting check it | 23:13 |
| rm_work | *is | 23:13 |
| johnsom | Not quite that much | 23:13 |
| rm_work | ... look at the gates | 23:13 |
| rm_work | i had to recheck my change 4 times | 23:13 |
| rm_work | and of the changes we just +Ad | 23:13 |
| rm_work | already 3 of them failed | 23:13 |
| johnsom | It all depends on the percentage of OVH hosts in the pool | 23:13 |
| rm_work | yeah and i'm saying that statistically it ends up being over 50% | 23:14 |
| rm_work | that at least one of our tests will get on a host that fails | 23:14 |
| johnsom | Maybe today, but not overall | 23:14 |
| rm_work | T_T | 23:14 |
| johnsom | Hmm, we actually might be able to do this in a gate hook..... | 23:15 |
| rm_work | yeah, i was wondering | 23:15 |
| rm_work | as long as we have something that runs early enough on the host | 23:15 |
| rm_work | before devstack (which is where those hooks run, IIRC) | 23:15 |
| johnsom | Yeah, let me cook something up | 23:16 |
| rm_work | k | 23:16 |
| *** sshank has quit IRC | 23:22 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Fix mis-matched qemu and nova hw_machine_type https://review.openstack.org/511980 | 23:26 |
| johnsom | rm_work Let's watch that ^^^^ | 23:26 |
| *** aojea has joined #openstack-lbaas | 23:27 | |
| *** aojea has quit IRC | 23:31 | |
| *** slaweq has quit IRC | 23:33 | |
| rm_work | kk | 23:34 |
| johnsom | Hmm, new google searches indicate it could be an issue with AMD hosts. | 23:34 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 23:39 | |
| rm_work | :3 | 23:42 |
| *** slaweq has joined #openstack-lbaas | 23:43 | |
| *** AlexeyAbashkin has quit IRC | 23:43 | |
| openstackgerrit | Merged openstack/octavia master: Adds automatic ERD generation for the Octavia DB https://review.openstack.org/509319 | 23:48 |
| openstackgerrit | Merged openstack/octavia master: Add dependent packages to config reference https://review.openstack.org/509879 | 23:49 |
| johnsom | See, important things are merging.... Grin | 23:49 |
| rm_work | T_T | 23:50 |
| rm_work | 2 of 5 | 23:50 |
| rm_work | err, 2 of 6 | 23:50 |
| johnsom | Docs jobs | 23:50 |
| rm_work | lots of rechecks running | 23:50 |
| *** tongl has quit IRC | 23:51 | |
| rm_work | johnsom: the issue i'm stuck on for tempest, is that for some reason the ssh client doesn't return after running the nc webserver | 23:56 |
| rm_work | linux_client.exec_command('sudo %s 80 &' % script) | 23:56 |
| rm_work | never returns | 23:56 |
| rm_work | because it gets stuck in a loop waiting for some polling thing | 23:56 |
| rm_work | probably because it's running a command that doesn't exit? | 23:56 |
| johnsom | Try what I do: | 23:57 |
| johnsom | "screen -d -m sh webserver.sh" | 23:57 |
| rm_work | no screen on cirros | 23:57 |
| johnsom | Yes | 23:57 |
| johnsom | That is how I run my cirros | 23:57 |
| rm_work | ? | 23:57 |
| rm_work | err | 23:57 |
| rm_work | i swear i just checked this | 23:57 |
| rm_work | or i would have done that already | 23:57 |
| johnsom | ssh -o UserKnownHostsFile=/dev/null -i /etc/octavia/.ssh/octavia_ssh_key -o StrictHostKeyChecking=no -q cirros@${IP1} "screen -d -m sh webserver.sh" | 23:57 |
| rm_work | wut. k | 23:57 |
| johnsom | cirros-0.3.4-x86_64-disk.img | 23:58 |
| johnsom | is what I pull down, but newer should work | 23:58 |
| rm_work | hmmm | 23:58 |
| rm_work | well, ok then | 23:58 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!