| *** goldyfruit_ has joined #openstack-lbaas | 00:07 | |
| *** sapd1_x has joined #openstack-lbaas | 00:25 | |
| *** goldyfruit_ has quit IRC | 00:33 | |
| *** goldyfruit has joined #openstack-lbaas | 00:35 | |
| *** goldyfruit has quit IRC | 01:00 | |
| *** sapd1_x has quit IRC | 01:09 | |
| *** tkajinam has quit IRC | 01:24 | |
| *** tkajinam has joined #openstack-lbaas | 01:24 | |
| *** sapd1_x has joined #openstack-lbaas | 01:25 | |
| *** goldyfruit has joined #openstack-lbaas | 01:26 | |
| *** sapd1_x has quit IRC | 01:37 | |
| *** hongbin has joined #openstack-lbaas | 01:49 | |
| *** hongbin has quit IRC | 02:36 | |
| *** ricolin has joined #openstack-lbaas | 02:47 | |
| *** psachin has joined #openstack-lbaas | 03:32 | |
| *** ramishra has joined #openstack-lbaas | 03:44 | |
| *** goldyfruit has quit IRC | 04:24 | |
| *** AlexStaf has joined #openstack-lbaas | 05:40 | |
| *** omkartelee has joined #openstack-lbaas | 05:59 | |
| omkartelee | Hi, Does octavia support multiple provider drivers ? | 06:00 |
|---|---|---|
| omkartelee | IF it does, how can I specify different amphora images in /etc/octavia/octavia.conf | 06:00 |
| omkartelee | IF it does, how can I specify different amphora images in /etc/octavia/octavia.conf for different providers | 06:01 |
| *** AlexStaf has quit IRC | 06:03 | |
| *** luksky has joined #openstack-lbaas | 06:21 | |
| cgoncalves | omkartelee, octavia supports multiple provider drivers. you can set which one you want using the "provider" field in the API | 06:27 |
| cgoncalves | https://docs.openstack.org/api-ref/load-balancer/v2/?expanded=create-a-load-balancer-detail#create-a-load-balancer | 06:27 |
| cgoncalves | from the CLI it would be: openstack loadbalancer create --provider providerB [...] | 06:27 |
| cgoncalves | omkartelee, the amphora image is an implementation of the amphora driver. other provider drivers implement differently, although some also via a service VM | 06:28 |
| omkartelee | Thanks @cgoncalves for reply | 06:28 |
| omkartelee | in my case, I have haproxy and another provider driver, haproxy uses amphora with image1 and my driver uses amphora image2 | 06:29 |
| omkartelee | how can I specify my provider driver should use which image ? | 06:30 |
| omkartelee | like 2 different amphora image exists and 2 different provider drivers are there | 06:30 |
| omkartelee | but I dont see any field where I can provide relation of provider driver to amphora image | 06:31 |
| *** gcheresh has joined #openstack-lbaas | 06:36 | |
| cgoncalves | omkartelee, you can either create a new configuration option, specific to your provider driver, where you would set an image id to load or make use of the flavor API | 06:36 |
| cgoncalves | your provider driver would need to support flavors | 06:36 |
| omkartelee | cgoncalves, out of curiocity: can I have multiple "hot_plug_plugins" ? can there be a scenario in which there are two different controller workers are running with completely different flows, and provider driver communicates with them on different RPC channels | 06:41 |
| *** pcaruana has joined #openstack-lbaas | 06:42 | |
| *** rpittau|afk is now known as rpittau | 06:47 | |
| cgoncalves | omkartelee, I am not entirely sure. looking at the code now, it seems to be possible. however, it doesn't look to me to be a good design choice. | 06:49 |
| omkartelee | Thanks a lot cgoncalves for your inputs, can you provide a suggestion: so if I want to implement my own device as an amphora, I will have to write provider_driver and a amphora_driver? | 06:53 |
| cgoncalves | omkartelee, not sure I follow you on "own device as an amphora". do you mean run something a different load balancing backend (not haproxy) in an amphora image with the amphora driver? | 06:55 |
| cgoncalves | I think you are looking to implement your own provider driver. if that is so, it should be totally decoupled from the amphora provider driver | 06:56 |
| cgoncalves | I'd highly recommend you to read https://docs.openstack.org/octavia/latest/contributor/guides/providers.html | 06:56 |
| omkartelee | yes, we run different load balancing backend, we have virtual machine which contains different load balancing backend, we want to boot it in similar way as current haproxy amphoras are booting dynamically | 06:57 |
| *** tesseract has joined #openstack-lbaas | 07:01 | |
| cgoncalves | I see. my advice is not to hook up to the amphora driver code. if you see you want to reuse part of the taskflow flows/tasks and are compatible, maybe you could propose moving the flows/tasks code to a more provider-agnostic place | 07:06 |
| cgoncalves | however, I wouldn't support much that idea. at some point in time a provider A would want make a change that could break provider B workflow | 07:07 |
| omkartelee | You know, I am already facing the exact issue | 07:07 |
| cgoncalves | exactly. better keep provider drivers code independent from each other. no coupling | 07:08 |
| cgoncalves | check the provider driver development guide I linked before. I think the project made significant advancements in the last 2 releases to better support provider drivers | 07:09 |
| cgoncalves | I know of at least two open-source provider drivers implementing this way | 07:10 |
| omkartelee | Can you elaborate more please. I have written my own controller_worker with my custom flows, and a provider driver as well, but this implementation is not supporting multi-provider feature | 07:14 |
| omkartelee | i introduce the worker as new plugin | 07:14 |
| omkartelee | so, I want to support both haproxy and my_amphora | 07:15 |
| omkartelee | i introduce the worker as new hot plugin | 07:16 |
| *** maciejjozefczyk has joined #openstack-lbaas | 07:19 | |
| cgoncalves | omkartelee, ok, that is a good start. how does your driver pick the glance image to boot? | 07:21 |
| omkartelee | currently I am using image_id from octavia.conf, the major issue I am facing is, even if I create loadbalancer with --provider amphora, the request is being picked up by my provider driver | 07:27 |
| cgoncalves | I am trying to understand why can't you support both the amphora driver (called that way, not haproxy :)) and your own provider driver | 07:27 |
| *** luksky has quit IRC | 07:27 | |
| omkartelee | because my provider driver is putting the request in same rpc queue and topic | 07:28 |
| omkartelee | and controller worker is picking the request from the same queue | 07:29 |
| omkartelee | Can I get your email cgoncalves, I can provide much more details over email, it would be really appreciated from my side | 07:32 |
| cgoncalves | omkartelee, you should use a different rpc queue and topic | 07:37 |
| cgoncalves | omkartelee, when you try to create an amphora-based load balancer, what does "openstack loadbalancer show $your_amphora_lb_id -f value -c provider" shows? | 07:39 |
| omkartelee | The name of my provider driver | 07:39 |
| *** AlexStaf has joined #openstack-lbaas | 07:41 | |
| omkartelee | OK, what would be standard way of implementing octavia driver according to you in following scenario: 1. I have my own amphora image with different LB logic 2. I want similar lifecycle as amphora driver i.e. amphorae should get booted, get configured dynamically(using nova & neutron driver) 3. also I can use multiple providers at the same time | 07:42 |
| omkartelee | i dont want to implement something that Octavia team doesnt consider as STANDARD way of doing | 07:43 |
| omkartelee | cgoncalves, so this will help me architect the plugin properly rather than making so many changes in the octavia code and writing my own octavia | 07:45 |
| cgoncalves | omkartelee, from what I'm understanding, your provider driver needs quite a significant amount of code because there's no external load balancer controller (say, an SDN controller with LB capabilities) so you have to implement one yourself | 07:47 |
| cgoncalves | your scenario seems feasible. I'd just dig in a bit more to see why your code changes make the amphora LB create call end up in your provider driver | 07:48 |
| omkartelee | cgoncalves, it would be really great if I could get your email, i can share much more info that way | 07:49 |
| cgoncalves | omkartelee, please send it to the openstack-discuss mailing list -- http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-discuss | 07:52 |
| omkartelee | cgoncalves, Oh Yes, sure Thanks, I had forgotten about it | 07:52 |
| cgoncalves | cool | 07:53 |
| omkartelee | cgoncalves, Thanks a lot :) | 07:54 |
| *** ivve has joined #openstack-lbaas | 08:01 | |
| *** rpittau is now known as rpittau|bbl | 08:01 | |
| *** omkartelee has quit IRC | 08:14 | |
| *** luksky has joined #openstack-lbaas | 08:23 | |
| *** tkajinam has quit IRC | 08:52 | |
| *** salmankhan has joined #openstack-lbaas | 09:35 | |
| *** luksky has quit IRC | 09:54 | |
| *** rpittau|bbl is now known as rpittau | 09:57 | |
| *** pcaruana has quit IRC | 10:23 | |
| *** rcernin has quit IRC | 10:23 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Transition amphora flows to dicts https://review.opendev.org/668898 | 10:32 |
| *** luksky has joined #openstack-lbaas | 10:33 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Convert Lb flows to use provider dicts https://review.opendev.org/671725 | 10:34 |
| *** pcaruana has joined #openstack-lbaas | 10:36 | |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Fix new pep8/pylint errors https://review.opendev.org/684700 | 10:37 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: [WIP] Jobboard based controller https://review.opendev.org/647406 | 10:38 |
| openstackgerrit | sunxifa proposed openstack/octavia master: Fix confused comment in create certificate script https://review.opendev.org/684705 | 10:49 |
| *** goldyfruit has joined #openstack-lbaas | 11:32 | |
| *** goldyfruit has quit IRC | 12:16 | |
| *** psachin has quit IRC | 13:04 | |
| *** AustinR has quit IRC | 13:04 | |
| *** AustinR has joined #openstack-lbaas | 13:04 | |
| *** AustinR has quit IRC | 13:09 | |
| *** AustinR has joined #openstack-lbaas | 13:09 | |
| *** AustinR has quit IRC | 13:11 | |
| *** AustinR has joined #openstack-lbaas | 13:12 | |
| *** gcheresh has quit IRC | 13:33 | |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Fix new pep8/pylint errors https://review.opendev.org/684700 | 13:37 |
| cgoncalves | rm_work, fixes gate: https://review.opendev.org/#/c/684700/ | 13:50 |
| *** goldyfruit has joined #openstack-lbaas | 14:16 | |
| *** Vorrtex has joined #openstack-lbaas | 14:24 | |
| *** rpittau is now known as rpittau|afk | 14:34 | |
| *** luksky has quit IRC | 14:46 | |
| *** goldyfruit has quit IRC | 14:46 | |
| *** AlexStaf has quit IRC | 14:55 | |
| *** goldyfruit has joined #openstack-lbaas | 15:18 | |
| *** openstackgerrit has quit IRC | 15:21 | |
| *** dulek has joined #openstack-lbaas | 15:25 | |
| *** pcaruana has quit IRC | 15:30 | |
| *** ivve has quit IRC | 15:45 | |
| *** ajay33 has joined #openstack-lbaas | 15:58 | |
| *** AustinR has quit IRC | 15:58 | |
| *** AustinR has joined #openstack-lbaas | 15:59 | |
| cgoncalves | o/ | 16:00 |
| cgoncalves | rm_work, ping | 16:00 |
| cgoncalves | ok, it looks like rm_work is not available today, and johnsom is on PTO. I'll start the meeting. bare with me folks :) | 16:01 |
| cgoncalves | #startmeeting Octavia | 16:01 |
| openstack | Meeting started Wed Sep 25 16:01:35 2019 UTC and is due to finish in 60 minutes. The chair is cgoncalves. Information about MeetBot at http://wiki.debian.org/MeetBot. | 16:01 |
| openstack | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 16:01 |
| *** openstack changes topic to " (Meeting topic: Octavia)" | 16:01 | |
| openstack | The meeting name has been set to 'octavia' | 16:01 |
| cgoncalves | hi | 16:01 |
| gthiemonge | hi | 16:01 |
| cgoncalves | low attendance today. we'll finish the call early I think | 16:02 |
| cgoncalves | #topic Announcements | 16:02 |
| *** openstack changes topic to "Announcements (Meeting topic: Octavia)" | 16:02 | |
| haleyb | hi | 16:03 |
| cgoncalves | Train RC1 and stable/train branches are this week. Please do not merge feature patches until master branch is open for Ussuri cycle | 16:03 |
| ajay33 | Hi | 16:03 |
| *** spatel has joined #openstack-lbaas | 16:03 | |
| cgoncalves | ah, great. people joining. welcome! | 16:03 |
| cgoncalves | The TC has proposed a release schedule for Ussuri | 16:04 |
| cgoncalves | #link https://review.opendev.org/#/c/679822/ | 16:04 |
| cgoncalves | Rendered page: | 16:04 |
| cgoncalves | #link https://openstack.fortnebula.com:13808/v1/AUTH_e8fd161dc34c421a979a9e6421f823e9/zuul_opendev_logs_c4c/679822/3/check/openstack-tox-docs/c4cbc92/docs/ussuri/schedule.html | 16:04 |
| cgoncalves | Please take a look and give any feedback if you see any issues | 16:04 |
| cgoncalves | Ussuri supported runtimes are out | 16:04 |
| cgoncalves | #link https://governance.openstack.org/tc/reference/runtimes/ussuri.html | 16:05 |
| cgoncalves | CentOS 7 is out, CentOS 8 (released yesterday) in | 16:05 |
| cgoncalves | Python 2 also out | 16:05 |
| cgoncalves | This means Octavia should support all distros and Python runtimes there mentioned | 16:06 |
| cgoncalves | projects can still run CI for Python 2 jobs | 16:06 |
| cgoncalves | the results of the OpenStack User Survey are out! | 16:07 |
| cgoncalves | #link http://lists.openstack.org/pipermail/openstack-discuss/2019-September/009501.html | 16:07 |
| *** spatel has quit IRC | 16:07 | |
| cgoncalves | this time Octavia had questions added to the survey, and the results are quite interesting. I recommend everyone to have a look | 16:08 |
| cgoncalves | e.g. the top #1 requested feature is active-active load balancing | 16:08 |
| cgoncalves | also, in a Neutron question, load balancing continues to be on top #3 of features folks are actively using, interested in or looking forward to using | 16:10 |
| cgoncalves | that is all I have for announcements today. anything someone would like to add? | 16:10 |
| cgoncalves | okay, let's continue | 16:11 |
| cgoncalves | #topic Brief progress reports / bugs needing review | 16:12 |
| *** openstack changes topic to "Brief progress reports / bugs needing review (Meeting topic: Octavia)" | 16:12 | |
| *** goldyfruit_ has joined #openstack-lbaas | 16:12 | |
| cgoncalves | my main focus since last week continue to be on backporting bug fixes to stable branches as well as having them merged | 16:13 |
| gthiemonge | I've found a bug with graceful shutdown of octavia workers: #link https://review.opendev.org/#/c/684201/ I would appreciate reviews ;-) | 16:15 |
| *** goldyfruit has quit IRC | 16:15 | |
| cgoncalves | with CentOS 8 released and it being in the supported distros for Ussuri, I started looking into how to have an amphora image built and tested. there is quite a bit of work to be done in diskimage-builder. I am working with Ian from DIB on that front | 16:15 |
| gthiemonge | cgoncalves: +1 | 16:16 |
| cgoncalves | gthiemonge, awesome! great finding. thanks for sharing a fix | 16:16 |
| cgoncalves | I am adding it now to the Train priority review list | 16:17 |
| cgoncalves | btw, our gate (pep8 job fails) is currently broken due to a new release of pylint | 16:18 |
| cgoncalves | gthiemonge has a patch up for review. thank you again! | 16:18 |
| cgoncalves | #link https://review.opendev.org/#/c/684700/ | 16:18 |
| cgoncalves | so no need to recheck your jobs if you see pep8 job failing. rebase your patch on top of Greg's | 16:19 |
| cgoncalves | anything else someone would like to share on this topic? | 16:21 |
| cgoncalves | #topic Open Discussion | 16:22 |
| *** openstack changes topic to "Open Discussion (Meeting topic: Octavia)" | 16:22 | |
| cgoncalves | Other topics for today? | 16:22 |
| colin- | think we are going to proceed with haproxy 2.0 on our amps for now, havne't been able to surface any blocking issues with it. am glad for that considering all the benefits it offers | 16:22 |
| colin- | (if anyone was curious) | 16:22 |
| cgoncalves | colin-, that's really encouraging! | 16:23 |
| colin- | does anybody know if anyone had a spec or change open already for adding string searches to HTTP type healthmonitors as an alternative to expect codes? | 16:24 |
| cgoncalves | colin-, have you built haproxy yourself or used a packaged version publicly available somewhere? | 16:24 |
| colin- | the latter, based on these that i found on their site: https://haproxy.debian.net/ | 16:25 |
| cgoncalves | I don't remember seeing that. it might be somewhere in storyboard but it's not easy to look up for anything in there | 16:25 |
| colin- | so what it describes there for xenial and 2.0 stable i added to my DIB work with some help from johnsom | 16:25 |
| cgoncalves | HTTP/2 protocol, available in haproxy 2.0, was top #3 most wanted feature that came out from the user survey | 16:26 |
| cgoncalves | colin-, thanks for the link! repos for Debian and Ubuntu | 16:27 |
| cgoncalves | one could consider creating a haproxy DIB element to install it from source, giving us more flexibility to test latest and greatest haproxy versions not yet available in the distros | 16:30 |
| cgoncalves | a low-hanging fruit to folks considering to join the project :) | 16:30 |
| cgoncalves | if there are no other topics, we can close the meeting | 16:31 |
| cgoncalves | thank you everyone! let's finish preparations for Train release and open master branch for Ussuri! | 16:32 |
| cgoncalves | #endmeeting | 16:32 |
| *** openstack changes topic to "Discussions for OpenStack Octavia | Priority bug review list: https://etherpad.openstack.org/p/octavia-priority-reviews" | 16:32 | |
| openstack | Meeting ended Wed Sep 25 16:32:39 2019 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 16:32 |
| openstack | Minutes: http://eavesdrop.openstack.org/meetings/octavia/2019/octavia.2019-09-25-16.01.html | 16:32 |
| openstack | Minutes (text): http://eavesdrop.openstack.org/meetings/octavia/2019/octavia.2019-09-25-16.01.txt | 16:32 |
| openstack | Log: http://eavesdrop.openstack.org/meetings/octavia/2019/octavia.2019-09-25-16.01.log.html | 16:32 |
| gthiemonge | thank you cgoncalves! | 16:32 |
| colin- | ttyl o/ | 16:32 |
| cgoncalves | it's beer o'clock here. hope it is also in whichever timezone you are in ;) | 16:33 |
| ajay33 | :p | 16:35 |
| *** ramishra has quit IRC | 16:43 | |
| *** ivve has joined #openstack-lbaas | 16:53 | |
| *** henriqueof has joined #openstack-lbaas | 16:54 | |
| *** henriqueof1 has quit IRC | 16:55 | |
| *** maciejjozefczyk has quit IRC | 17:02 | |
| *** spatel has joined #openstack-lbaas | 17:16 | |
| *** salmankhan has quit IRC | 17:35 | |
| *** henriqueof1 has joined #openstack-lbaas | 17:53 | |
| *** henriqueof has quit IRC | 17:53 | |
| *** tesseract has quit IRC | 18:08 | |
| *** luksky has joined #openstack-lbaas | 18:20 | |
| *** ricolin has quit IRC | 18:55 | |
| rm_work | Yeah I've been thinking about putting those two things in our goals for Ussuri | 19:06 |
| rm_work | HTTP/2 (haproxy 2.0 everywhere) and Body Regex for HMs | 19:06 |
| rm_work | Both are doable I think | 19:07 |
| rm_work | And would be nice feature wins | 19:07 |
| colin- | we would wait for 2.0 to be available through traditional ubuntu package management before commiting that to master right? | 19:28 |
| * rm_work shrugs | 19:28 | |
| rm_work | I haven't before | 19:28 |
| rm_work | but that would be ideal, yes | 19:29 |
| rm_work | folks have sometimes not appreciated my ... approach | 19:29 |
| colin- | uh huh, i think folks are on vacation for a week or so :D | 19:29 |
| rm_work | excellent | 19:29 |
| rm_work | plenty of time :D | 19:29 |
| rm_work | +A'd that gatefix | 19:33 |
| *** ajay33 has quit IRC | 19:48 | |
| * johnsom glares at colin- | 20:19 | |
| johnsom | Actually I put 2.0 on the last ptg agenda. We just need to have a plan really. | 20:20 |
| johnsom | Remind me when I get back to start an etherpad so we can collect the challenges and strategy to solve them. | 20:22 |
| * johnsom vanishes into the pacific once again | 20:23 | |
| colin- | haha | 20:23 |
| colin- | outta here!! | 20:23 |
| colin- | and enjoy :) | 20:23 |
| rm_work | wait why is johnsom here, lulz | 20:34 |
| rm_work | he heard us, and now my plan is foiled | 20:35 |
| *** Vorrtex has quit IRC | 21:25 | |
| *** spatel has quit IRC | 21:35 | |
| *** openstackgerrit has joined #openstack-lbaas | 21:47 | |
| openstackgerrit | Merged openstack/octavia master: Fix new pep8/pylint errors https://review.opendev.org/684700 | 21:47 |
| *** ccamposr__ has joined #openstack-lbaas | 22:11 | |
| *** ccamposr has quit IRC | 22:13 | |
| *** rcernin has joined #openstack-lbaas | 22:15 | |
| *** ccamposr has joined #openstack-lbaas | 22:20 | |
| *** ccamposr__ has quit IRC | 22:23 | |
| *** tkajinam has joined #openstack-lbaas | 22:59 | |
| *** luksky has quit IRC | 23:00 | |
| *** ccamposr has quit IRC | 23:02 | |
| *** goldyfruit_ has quit IRC | 23:12 | |
| *** ivve has quit IRC | 23:28 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!