| *** rcernin_ has quit IRC | 00:41 | |
| *** rcernin has joined #openstack-lbaas | 00:41 | |
| *** yamamoto has joined #openstack-lbaas | 00:46 | |
| *** ricolin has joined #openstack-lbaas | 01:10 | |
| openstackgerrit | Adam Harwell proposed openstack/python-octaviaclient master: Allow creating/showing LBs with Additional VIPs https://review.opendev.org/664473 | 01:23 |
|---|---|---|
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 01:24 |
| *** yamamoto has quit IRC | 01:48 | |
| rm_work | Wtf | 02:09 |
| rm_work | Messages from my desktop client aren't going through, damnit | 02:09 |
| rm_work | I thought I was just being ignored all day lol | 02:09 |
| rm_work | test | 02:10 |
| rm_work | ok great now it works lol | 02:10 |
| rm_work | was asking if we had DDT support in tempest, i thought we did at some point, but i can't find it in the tempest docs.... so i assumed it isn't possible and just manually wrote a bunch of tests | 02:12 |
| rm_work | and some other stuff that doesn't matter i guess :D | 02:12 |
| rm_work | though i also discovered everything is continuing to fail on my stuff until https://review.opendev.org/#/c/664385/ merges | 02:12 |
| rm_work | which is ... going to be a while, the gate is failing due to something above it in the merge queue, and some other random unrelated nova failures <_< | 02:13 |
| rm_work | and apparently it takes like 3 hours for one of the tests to finish T_T | 02:13 |
| *** sapd1_x has joined #openstack-lbaas | 02:14 | |
| *** hongbin has joined #openstack-lbaas | 02:25 | |
| *** yamamoto has joined #openstack-lbaas | 02:50 | |
| johnsom | Yeah, didn't hear from you all day. | 03:19 |
| johnsom | DDT was a HUGE failure in neutron-lbaas, we long since deleted them. | 03:19 |
| rm_work | yeah k | 03:31 |
| rm_work | well i WAS talking :D | 03:31 |
| rm_work | but most of it was asking dumb questions or just opinions that i decided the path for myself when no one answered lol | 03:32 |
| rm_work | like the ridiculous thing i did for the NETNS_NAME so it wasn't different from our constants >_> | 03:32 |
| rm_work | line 6: https://review.opendev.org/#/c/660239/31/elements/haproxy-octavia/post-install.d/21-configure-netns | 03:33 |
| rm_work | i asked if you thought it was INSANE but no answer -> did it :D | 03:33 |
| rm_work | i was gonna just import it and print the var, but because i can't guarantee having octavia-lib installed, and we added an import from it in the *constants* file, I can't | 03:36 |
| rm_work | so it was 1) use ast; 2) patch the import somehow (mock?) so it didn't explode | 03:37 |
| rm_work | alternatively i could do it the other way, have a file with the name in it, and have both the script and constants.py load up the file | 03:38 |
| *** psachin has joined #openstack-lbaas | 03:40 | |
| *** ramishra has joined #openstack-lbaas | 03:59 | |
| *** hongbin has quit IRC | 04:01 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 04:24 |
| *** gcheresh has joined #openstack-lbaas | 04:25 | |
| *** gcheresh has quit IRC | 04:46 | |
| *** luksky has joined #openstack-lbaas | 05:09 | |
| *** luksky has quit IRC | 05:25 | |
| *** gcheresh has joined #openstack-lbaas | 05:29 | |
| openstackgerrit | Lingxian Kong proposed openstack/octavia-tempest-plugin master: Support more params for lb client initialization https://review.opendev.org/664497 | 05:46 |
| openstackgerrit | tonybrad proposed openstack/octavia master: Change the requirement to releases.openstack.org https://review.opendev.org/664499 | 05:50 |
| *** threestrands has joined #openstack-lbaas | 05:52 | |
| *** threestrands has quit IRC | 05:53 | |
| *** gcheresh_ has joined #openstack-lbaas | 06:17 | |
| *** gcheresh has quit IRC | 06:17 | |
| *** gcheresh has joined #openstack-lbaas | 06:21 | |
| *** gcheresh_ has quit IRC | 06:21 | |
| *** gcheresh has quit IRC | 06:33 | |
| *** gcheresh_ has joined #openstack-lbaas | 06:34 | |
| *** yamamoto has quit IRC | 06:40 | |
| *** pcaruana has joined #openstack-lbaas | 06:55 | |
| *** yamamoto has joined #openstack-lbaas | 06:58 | |
| *** rpittau|afk is now known as rpittau | 07:08 | |
| *** rcernin has quit IRC | 07:08 | |
| *** tesseract has joined #openstack-lbaas | 07:08 | |
| *** ramishra has quit IRC | 07:13 | |
| *** dayou has quit IRC | 07:23 | |
| *** ramishra has joined #openstack-lbaas | 07:30 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia-lib master: Additional VIPs is also relevant in provider_base https://review.opendev.org/664510 | 07:33 |
| rm_work | erg, missed that somehow, was it new recently? :/ | 07:35 |
| rm_work | either that or just forgot to update it | 07:35 |
| rm_work | it's not FUNCTIONALLY relevant | 07:35 |
| *** dayou has joined #openstack-lbaas | 07:37 | |
| *** yboaron has joined #openstack-lbaas | 07:42 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Use retry for AmphoraComputeConnectivityWait https://review.opendev.org/662791 | 08:00 |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Use retry for AmphoraComputeConnectivityWait https://review.opendev.org/662791 | 08:19 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 08:20 |
| *** dayou has quit IRC | 08:28 | |
| *** yboaron has quit IRC | 08:33 | |
| *** yboaron has joined #openstack-lbaas | 08:33 | |
| *** dayou has joined #openstack-lbaas | 08:41 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia-lib master: Additional VIPs is also relevant in provider_base https://review.opendev.org/664510 | 08:43 |
| *** rcernin has joined #openstack-lbaas | 09:03 | |
| *** dayou has quit IRC | 09:18 | |
| *** tobberydberg has quit IRC | 09:22 | |
| *** tobberydberg has joined #openstack-lbaas | 09:31 | |
| *** dayou has joined #openstack-lbaas | 09:31 | |
| *** yamamoto has quit IRC | 09:32 | |
| *** yamamoto has joined #openstack-lbaas | 09:37 | |
| rm_work | johnsom: that's an interesting one i haven't seen before: http://logs.openstack.org/62/664462/1/check/octavia-v2-dsvm-noop-api/589dc76/controller/logs/screen-o-api.txt.gz#_Jun_11_09_19_39_532701 | 09:45 |
| rm_work | just happened to see it while scrolling through | 09:45 |
| *** salmankhan has joined #openstack-lbaas | 09:45 | |
| *** salmankhan has quit IRC | 09:45 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 09:57 |
| *** yamamoto has quit IRC | 10:03 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 10:05 |
| *** yamamoto has joined #openstack-lbaas | 10:07 | |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 10:10 |
| *** pcaruana has quit IRC | 10:19 | |
| *** yamamoto has quit IRC | 10:24 | |
| *** yamamoto has joined #openstack-lbaas | 10:56 | |
| *** yamamoto has quit IRC | 11:00 | |
| *** pcaruana has joined #openstack-lbaas | 11:09 | |
| *** rcernin has quit IRC | 11:43 | |
| *** yamamoto has joined #openstack-lbaas | 11:56 | |
| *** yamamoto has quit IRC | 12:04 | |
| *** ccamposr has joined #openstack-lbaas | 12:15 | |
| *** goldyfruit has quit IRC | 12:17 | |
| *** ccamposr__ has quit IRC | 12:18 | |
| *** ricolin_ has joined #openstack-lbaas | 12:27 | |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 12:28 |
| *** ricolin has quit IRC | 12:29 | |
| *** boden has joined #openstack-lbaas | 12:47 | |
| *** goldyfruit has joined #openstack-lbaas | 13:26 | |
| *** yamamoto has joined #openstack-lbaas | 13:29 | |
| *** pcaruana has quit IRC | 14:04 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Use retry for AmphoraComputeConnectivityWait https://review.opendev.org/662791 | 14:08 |
| *** yamamoto has quit IRC | 14:15 | |
| *** yamamoto has joined #openstack-lbaas | 14:16 | |
| *** yamamoto has quit IRC | 14:16 | |
| *** pcaruana has joined #openstack-lbaas | 14:30 | |
| *** yamamoto has joined #openstack-lbaas | 14:30 | |
| *** pcaruana has quit IRC | 14:31 | |
| *** pcaruana has joined #openstack-lbaas | 14:31 | |
| boden | hi johnsom... if you get a few min to chat, pls let me know. thx | 14:38 |
| *** Vorrtex has joined #openstack-lbaas | 14:40 | |
| *** spatel has joined #openstack-lbaas | 14:42 | |
| spatel | can i have multiple flavor for amphora instances? | 14:43 |
| *** happyhemant has joined #openstack-lbaas | 14:44 | |
| *** yamamoto has quit IRC | 14:45 | |
| *** kklimonda has quit IRC | 14:47 | |
| *** yamamoto has joined #openstack-lbaas | 14:49 | |
| johnsom | spatel You can create as many flavors as you like, but a load balancer can only have one at a time. | 14:49 |
| spatel | can i deploy small flavor for dev and large flavor for production right? or only i can pick one and have to stick with it? | 14:50 |
| johnsom | You cannot change the flavor of a load balancer after it is created. However you can certainly have a "small" and "large" flavor as options. | 14:52 |
| *** yamamoto has quit IRC | 14:53 | |
| *** gcheresh_ has quit IRC | 14:58 | |
| johnsom | boden Hi | 15:00 |
| *** yboaron_ has joined #openstack-lbaas | 15:00 | |
| *** devfaz has quit IRC | 15:00 | |
| boden | johnsom hi... so after some munking around I think our preferred route is what I mentioned yesterday; a way for the provider to signify they want the related objects populated in the provider calls | 15:00 |
| boden | not sure what you think about that? | 15:00 |
| boden | I'm working on some stub patches to at least "show" what I'm talking about | 15:01 |
| *** devfaz has joined #openstack-lbaas | 15:01 | |
| boden | is it worth me getting those patches up so we can at least look at the code, or do you flat out think this isn't a good approach? | 15:02 |
| johnsom | boden So you want to pass a fully populated load balancer object tree to the provider driver? | 15:02 |
| johnsom | You don't see a potential need for a query interface? | 15:02 |
| *** yboaron has quit IRC | 15:02 | |
| boden | johnsom so for example a Listener data model would include a loadbalancer that's populated by the controller for providers who want it (false by default).... | 15:03 |
| boden | I think the query interface woudl work, but honestly using listener processes over sockets seems to add complexity to debug/management | 15:04 |
| johnsom | Hmm, I don't think the code would be that bad to implement. | 15:05 |
| johnsom | Here are a couple of thoughts: | 15:06 |
| johnsom | 1. I think the query option could be useful for other driver needs, such as re-sync where the driver checks the appliance configuration vs. the expected. | 15:07 |
| johnsom | 2. How do we decided what to send with each object? For example, pools could have parents of LB or listener, etc. | 15:07 |
| spatel | johnsom: thanks!! | 15:07 |
| johnsom | I think we we go down the path of optionally sending more information, we might as well send the whole thing. | 15:08 |
| johnsom | I'm just wondering if we won't want to later do the query option anyway. | 15:09 |
| johnsom | If your concern is mostly about the work to implement the query interface, I can offer to help with that. | 15:10 |
| johnsom | Mostly I just want to discuss this so we come up with the best solution. | 15:11 |
| *** yamamoto has joined #openstack-lbaas | 15:15 | |
| *** yamamoto has quit IRC | 15:15 | |
| *** yamamoto has joined #openstack-lbaas | 15:16 | |
| boden | johnsom valid points.. our main concern is just now we are adding more sockets/listeners/processes/etc. and that adds complexity to how we manage, debug, etc. octavia deployments... if there was some well defined db interface it wouldn't be a concern, it just seems we are increasing the management footprint here | 15:17 |
| *** pcaruana has quit IRC | 15:17 | |
| johnsom | Yeah, the downside is the added socket for sure. Otherwise, the code is relatively simple and maintains the separation/abstraction for the drivers. | 15:18 |
| *** gcheresh_ has joined #openstack-lbaas | 15:19 | |
| boden | johnsom I know some distros manage sockets, etc. and have "hooks" to help debug using existing openstack messaging... while I know the custom protocol is straight forward; it adds a new dimension to distro tools, etc. | 15:19 |
| johnsom | We only had the two methods defined in the driver specification, so that is how I implemented it, but I wish I had done the socket a bit different. We still have the opportunity to do that however. | 15:20 |
| *** yamamoto has quit IRC | 15:20 | |
| boden | johnsom in that respect it would probably be easier for us just to use teh SDK client since it's more "well defined" and known to such tools | 15:20 |
| *** sapd1_x has quit IRC | 15:21 | |
| johnsom | The unix domain sockets was selected to maintain a level of security. We considered doing a REST API, but we struggled to figure out how to authenticate the drivers. | 15:21 |
| *** yamamoto has joined #openstack-lbaas | 15:22 | |
| boden | johnsom sorry I'm not bashing on the current approach... I'm just trying to find some middle ground to not impact distro tooling too much | 15:22 |
| *** yamamoto has quit IRC | 15:22 | |
| johnsom | Hmm, well, I think both are "well defined" and documented. | 15:22 |
| johnsom | boden How about this, I will look at what it would take to leverage one of the existing sockets to provide the query interfaces. That way there would be no additional overhead for the distros/packagers than we already have. | 15:24 |
| *** yamamoto has joined #openstack-lbaas | 15:26 | |
| *** yamamoto has quit IRC | 15:26 | |
| boden | johnsom that may work, but I need to some investigation to confirm | 15:30 |
| johnsom | boden Ok, let me spend some time today and see if I can get an example posted for tomorrow. | 15:31 |
| *** gcheresh_ has quit IRC | 15:37 | |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 15:37 |
| openstackgerrit | Gregory Thiemonge proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 15:40 |
| *** pcaruana has joined #openstack-lbaas | 15:43 | |
| *** spatel has quit IRC | 15:46 | |
| *** luksky has joined #openstack-lbaas | 15:47 | |
| *** pcaruana has quit IRC | 15:55 | |
| *** yboaron_ has quit IRC | 16:03 | |
| *** tesseract has quit IRC | 16:18 | |
| *** spatel has joined #openstack-lbaas | 16:31 | |
| *** kklimonda has joined #openstack-lbaas | 16:32 | |
| *** yamamoto has joined #openstack-lbaas | 16:36 | |
| *** yamamoto has quit IRC | 16:42 | |
| *** pcaruana has joined #openstack-lbaas | 16:49 | |
| *** rpittau is now known as rpittau|afk | 16:55 | |
| *** spatel has quit IRC | 16:56 | |
| *** spatel has joined #openstack-lbaas | 16:56 | |
| *** luksky has quit IRC | 16:59 | |
| *** luksky has joined #openstack-lbaas | 17:05 | |
| *** ramishra has quit IRC | 17:07 | |
| *** luksky has quit IRC | 17:12 | |
| *** ccamposr__ has joined #openstack-lbaas | 17:16 | |
| *** ccamposr has quit IRC | 17:18 | |
| *** gcheresh_ has joined #openstack-lbaas | 17:52 | |
| *** KeithMnemonic has joined #openstack-lbaas | 18:10 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add project_id to all of the provider objects https://review.opendev.org/663459 | 18:14 |
| *** ricolin_ has quit IRC | 18:38 | |
| *** pcaruana has quit IRC | 18:39 | |
| rm_work | johnsom: blegh https://review.opendev.org/#/c/664473/ | 18:46 |
| rm_work | not sure how i missed that? though not in a hurry for a release or anything as it's not functionally breaking, works fine without the ABC updated | 18:46 |
| rm_work | err | 18:47 |
| rm_work | damnit wrong patch | 18:47 |
| rm_work | https://review.opendev.org/#/c/664510/ | 18:47 |
| johnsom | Ah, yeah, opps. I missed it too. | 18:49 |
| johnsom | I will be working on octavia-lib today too, so we can probably release in the next few days. | 18:49 |
| rm_work | 1.2.1? :D | 18:52 |
| *** gcheresh_ has quit IRC | 19:03 | |
| *** psachin has quit IRC | 20:02 | |
| *** ccamposr has joined #openstack-lbaas | 20:20 | |
| *** rouk has joined #openstack-lbaas | 20:21 | |
| rouk | johnsom: is it a known bug that the openstack loadbalancer cli eats --insecure arg? | 20:22 |
| *** ccamposr__ has quit IRC | 20:22 | |
| rouk | just seems to be flavorprofile subcommand doesnt listen | 20:23 |
| johnsom | rouk: no, but that is also not out code, it is provided by openstackclient. What are you seeing? | 20:31 |
| johnsom | Out->our | 20:31 |
| rouk | the following command doesnt obey --insecure, try it out. openstack --insecure loadbalancer flavorprofile create --name small --provider amphora --flavor-data '{"loadbalancer_topology":"ACTIVE_STANDBY", "compute_flavor":"1ed40309-f600-4f93-930f-d2bc39ca59cd"}' | 20:32 |
| rouk | --insecure on like... openstack --insecure server list, no problem, but that line, its not happy. | 20:33 |
| johnsom | Just to clarify, you have octavia api in https and this flag is not allowing the connection? | 20:34 |
| rouk | yep, cert verify fail | 20:34 |
| johnsom | Are the certificates on the nova api and octavia api the same? | 20:35 |
| rouk | yes. | 20:35 |
| rouk | both self sign in this particular lab setup | 20:35 |
| johnsom | Hmm, —insecure is on openstack command and handled before our code runs. This is odd | 20:36 |
| rouk | yeah, thats why im confused, i specified it in the right order, it shouldnt be getting eaten | 20:36 |
| johnsom | Would you be ok pasting a —debug output from that? | 20:36 |
| rouk | question before i do, "Provider 'amphora' reports error: SSL exception connecting to XXX" during flavor add, is this some kind of internal failure? the client is complaining, but amphra provider is serverside... | 20:38 |
| johnsom | Hmm, that is an internal error to the provider driver | 20:40 |
| johnsom | I wonder what it is calling for a flavor add | 20:40 |
| rouk | https://gist.github.com/adiana-tc/55a4bf0f088da5ac30a035540c04cf38 | 20:41 |
| rouk | lemme know when you got it | 20:42 |
| johnsom | I can see that paste, thanks | 20:42 |
| rouk | any ideas? | 20:43 |
| johnsom | Yeah, give me a minute to research | 20:43 |
| johnsom | That error is not a CLI side error, that is a API process error | 20:44 |
| rouk | yeah, i noticed once i read through the debug, all the calls got properly passed --insecure on the clientside | 20:44 |
| johnsom | I *think* Octavia is trying to contact the Nova API to validate a nova flavor and getting the SSL error. | 20:45 |
| johnsom | But let me confirm | 20:45 |
| rouk | so if the theory is right, i should be able to remove my nova flavor reference and have it work | 20:45 |
| rouk | yep, youre correct. | 20:46 |
| johnsom | rouk So, this section: https://docs.openstack.org/octavia/latest/configuration/configref.html#nova | 20:47 |
| rouk | got it. | 20:47 |
| johnsom | in the octavia.conf is not setup correctly for the nova API. Either it needs "insecure" or you need to provide a ca_certificate_file | 20:47 |
| rouk | yep, can do | 20:48 |
| johnsom | While you are there, you might want to check the other service sections too | 20:49 |
| rouk | yeah, setting insecure on neutron too, its a lab env and theres no outside access, just want to protect transit into the lab. | 20:49 |
| johnsom | Yep | 20:49 |
| rouk | i can verify CAs on the clientside and thats enough for me. | 20:50 |
| *** boden has quit IRC | 21:05 | |
| *** luksky has joined #openstack-lbaas | 21:09 | |
| *** Vorrtex has quit IRC | 21:11 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 21:21 |
| openstackgerrit | Adam Harwell proposed openstack/python-octaviaclient master: Allow creating/showing LBs with Additional VIPs https://review.opendev.org/664473 | 21:28 |
| *** luksky has quit IRC | 21:35 | |
| openstackgerrit | Adam Harwell proposed openstack/python-octaviaclient master: Allow creating/showing LBs with Additional VIPs https://review.opendev.org/664473 | 21:36 |
| openstackgerrit | Adam Harwell proposed openstack/python-octaviaclient master: Allow creating/showing LBs with Additional VIPs https://review.opendev.org/664473 | 21:38 |
| *** spatel has quit IRC | 21:39 | |
| rm_work | rouk: you could also add the self-signing (or whatever) CA for your other services to the list of accepted CAs on the octavia host, and that is a much cleaner solution (at least it verifies it's YOUR self-signed cert and not a MITM) | 21:40 |
| rm_work | then you don't need to use insecure mode for those in the octavia.conf | 21:41 |
| *** ccamposr has quit IRC | 22:06 | |
| *** rcernin has joined #openstack-lbaas | 22:10 | |
| rm_work | johnsom: lol our network noop driver is so bad | 22:30 |
| rm_work | doing some fixes now to try to get it to be not quite as painful... | 22:30 |
| rm_work | at least tracking what it has returned and being consistent about it, and having the things that return actually have the expected IDs, lolol | 22:31 |
| rm_work | hilarious example: https://github.com/openstack/octavia/blob/master/octavia/network/drivers/noop_driver/driver.py#L164-L168 | 22:31 |
| rm_work | get_subnet ... passed a subnet_id... returns a subnet object with a new random ID XD | 22:32 |
| rm_work | not sure if any of this will actually help the real issue i'm running into with these tests though T_T | 22:33 |
| johnsom | Yeah, I have fought with that noop driver myself. | 22:51 |
| *** goldyfruit has quit IRC | 23:14 | |
| *** goldyfruit has joined #openstack-lbaas | 23:40 | |
| *** ianychoi has quit IRC | 23:48 | |
| *** ianychoi has joined #openstack-lbaas | 23:48 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 23:55 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Allow multiple VIPs per LB https://review.opendev.org/660239 | 23:55 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Amphora logging https://review.opendev.org/624835 | 23:55 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Save amphora logs in gate https://review.opendev.org/626406 | 23:58 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Amphora logging https://review.opendev.org/624835 | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!