| bar_ | johnsom, looking. thanks. regarding the capital letters, I figured it would ease typing. Do you mind if I'll treat 'A' and 'a', just the same? | 00:03 |
|---|---|---|
| johnsom | That is fine | 00:03 |
| johnsom | We just have constants and API ref that uses capitals, so they are documented as uppercase | 00:04 |
| openstackgerrit | Merged openstack/octavia master: Fix health monitor type "PING" to use icmp ping https://review.openstack.org/518632 | 00:10 |
| *** fnaval has joined #openstack-lbaas | 00:35 | |
| *** yamamoto has quit IRC | 00:36 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Add unit tests for neutron utils, add model/util for floating_ip https://review.openstack.org/525353 | 00:38 |
| rm_work | johnsom: check out them sweet sweet new tests ^^ | 00:38 |
| johnsom | grin | 00:38 |
| *** yamamoto has joined #openstack-lbaas | 00:38 | |
| johnsom | Trying to test german's proxy plugin at the moment | 00:38 |
| rm_work | all it'll cost you is adding a model/conversion-util for floating-ip | 00:38 |
| xgerman_ | thanks — appreciated | 00:40 |
| johnsom | devstack is not being your friend at the moment. Last run neutron-db-manage puked on some agent issue. Very odd | 00:41 |
| *** dayou has joined #openstack-lbaas | 00:42 | |
| rm_work | err, missed a couple, fixing | 00:43 |
| *** yamamoto has quit IRC | 00:44 | |
| openstackgerrit | Merged openstack/neutron-lbaas master: Update pool member operating status for haproxy https://review.openstack.org/521250 | 00:44 |
| *** armax has quit IRC | 00:47 | |
| openstackgerrit | Michael Johnson proposed openstack/neutron-lbaas master: Octavia Proxy Plugin https://review.openstack.org/418530 | 00:49 |
| johnsom | rebase, I think it was a neutron-lib thing. | 00:50 |
| *** dayou has quit IRC | 00:52 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Add unit tests for neutron utils, add model/util for floating_ip https://review.openstack.org/525353 | 00:52 |
| johnsom | Yep, the rebase fixed it | 00:59 |
| bar_ | johnsom, the list --loadbalancer breaks in the octavia patch, not the client. unfortunately, I cannot use _type_to_model_map out of the box for the conversion. | 01:02 |
| bar_ | I'll fix it tomorrow. | 01:02 |
| johnsom | Yeah, I expected it was in the octavia patch, but easier to mark the place there. | 01:03 |
| johnsom | Ok | 01:03 |
| *** bar_ has quit IRC | 01:04 | |
| *** dayou has joined #openstack-lbaas | 01:11 | |
| *** yamamoto has joined #openstack-lbaas | 01:41 | |
| *** yamamoto has quit IRC | 01:46 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Update compute noop driver/tests (many were disabled/useless) https://review.openstack.org/525295 | 01:51 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: WIP: Floating IP Network Driver (spans L3s) https://review.openstack.org/435612 | 01:55 |
| rm_work | k | 01:56 |
| rm_work | k handled it | 01:57 |
| *** armax has joined #openstack-lbaas | 02:08 | |
| rm_work | err johnsom you've been participating in this vrrp_ rename thing? | 02:13 |
| rm_work | just actually *looked* at it ... | 02:14 |
| rm_work | we can't be doing schema changes like that O_o | 02:14 |
| *** annp has joined #openstack-lbaas | 02:33 | |
| *** links has joined #openstack-lbaas | 03:07 | |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/neutron-lbaas master: Updated from global requirements https://review.openstack.org/523762 | 03:16 |
| openstackgerrit | OpenStack Proposal Bot proposed openstack/octavia master: Updated from global requirements https://review.openstack.org/525387 | 03:20 |
| *** sanfern has quit IRC | 03:20 | |
| johnsom | rm_work Participating, no, aware of the discussion from the PTG, yes. | 03:22 |
| rm_work | i saw some comments of yours on it | 03:22 |
| johnsom | I think our current naming is a mess | 03:22 |
| rm_work | yeah | 03:22 |
| rm_work | i would love to see that changed | 03:22 |
| rm_work | i don't think we are allowed to do schema changes like this -- and/or just thinking about the pain hurts my brain a little | 03:23 |
| rm_work | do you think we can do it? | 03:23 |
| rm_work | i don't think anyone just ... renames DB fields | 03:23 |
| rm_work | maybe i'm overreacting | 03:23 |
| johnsom | I don’t think I have looked at it close enough yet. | 03:23 |
| johnsom | Well, we abstract the DB. As long as there is clean migration path it shouldn’t be a problem. | 03:24 |
| *** threestrands has joined #openstack-lbaas | 03:27 | |
| *** yamamoto has joined #openstack-lbaas | 03:49 | |
| openstackgerrit | ZhaoBo proposed openstack/octavia master: WIP:UDP jinja template https://review.openstack.org/525420 | 04:08 |
| *** sanfern has joined #openstack-lbaas | 04:26 | |
| sanfern | hi johnsom | 04:47 |
| *** ianychoi has joined #openstack-lbaas | 05:07 | |
| *** krypto has joined #openstack-lbaas | 05:25 | |
| *** threestrands has quit IRC | 05:25 | |
| *** sanfern has quit IRC | 05:52 | |
| *** Alex_Staf has joined #openstack-lbaas | 06:10 | |
| Alex_Staf | xgerman_, johnsom Hi, tnx for the answers , added some other question regarding the namespace and allowed pairs | 06:10 |
| *** oanson has quit IRC | 06:20 | |
| *** oanson has joined #openstack-lbaas | 06:20 | |
| *** krypto has quit IRC | 06:43 | |
| *** armax has quit IRC | 06:57 | |
| *** jcross has quit IRC | 07:05 | |
| *** pcaruana has joined #openstack-lbaas | 07:32 | |
| openstackgerrit | Guoqiang Ding proposed openstack/neutron-lbaas master: Choose correct log option by listener protocol https://review.openstack.org/453612 | 07:41 |
| *** knsahm has joined #openstack-lbaas | 07:43 | |
| *** rcernin has quit IRC | 07:48 | |
| *** krypto has joined #openstack-lbaas | 08:00 | |
| *** krypto has quit IRC | 08:00 | |
| *** krypto has joined #openstack-lbaas | 08:00 | |
| Alex_Staf | xgerman_, Hi, could you please send me provider and flavor docs? tnx | 08:06 |
| *** krypto has quit IRC | 08:10 | |
| *** krypto has joined #openstack-lbaas | 08:10 | |
| *** b_bezak has joined #openstack-lbaas | 08:11 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Add unit tests for neutron utils, add model/util for floating_ip https://review.openstack.org/525353 | 08:13 |
| Alex_Staf | rm_work, Hi, could you please send me provider and flavor docs? tnx | 08:13 |
| *** tesseract has joined #openstack-lbaas | 08:24 | |
| *** rcernin has joined #openstack-lbaas | 08:33 | |
| *** eN_Guruprasad_Rn has joined #openstack-lbaas | 08:37 | |
| *** pcaruana has quit IRC | 08:40 | |
| oanson | Hi all: Running dragonflow+kuryr+octavia (or DVR+kuryr+octavia) and the VIP port is in admin_state_down. | 08:44 |
| oanson | I see that that's how it's created here: http://git.openstack.org/cgit/openstack/neutron-lbaas/tree/neutron_lbaas/db/loadbalancer/loadbalancer_dbv2.py#n145 | 08:44 |
| *** irenab has joined #openstack-lbaas | 08:44 | |
| oanson | And set to up by calling plug_vip_port here: https://github.com/openstack/neutron-lbaas/blob/e890821ba07deb4561f4712eccd5b62f09f5cec5/neutron_lbaas/drivers/haproxy/namespace_driver.py#L319 | 08:44 |
| oanson | The latter is the agent. | 08:44 |
| oanson | (I think it's the HAProxy agent, but I'm not sure) | 08:45 |
| oanson | What's the reason behind this behaviour? | 08:45 |
| rm_work | i don't believe that state actually matters | 08:50 |
| rm_work | mine always say down, even when they're operational | 08:50 |
| rm_work | but, i'm not using DVR, so not positive | 08:50 |
| rm_work | Alex_Staf: what docs? | 08:50 |
| Alex_Staf | rm_work, something regarding the flavors and providers for octavia | 08:51 |
| nmagnezi | oanson, to the best of my knowledge, that VIP port is not bound to any host (which is a must from the neutron L2 agent POV). the reason is that we configure that IP address in the haproxy-amphora namespace inside the amphora. this is done by the amphora agent. | 08:51 |
| rm_work | uhh | 08:51 |
| rm_work | not sure anything has merged yet | 08:51 |
| nmagnezi | so the port kina reserves the ip address.. | 08:51 |
| nmagnezi | Alex_Staf, both not merged. but you can read and comment on them | 08:52 |
| rm_work | Alex_Staf: maybe this? https://review.openstack.org/#/c/509957/ | 08:52 |
| oanson | I guessed as much. But an SDN controller that respects that field will not wire it | 08:52 |
| oanson | rm_work, nmagnezi thanks. and ^^^ | 08:52 |
| Alex_Staf | nmagnezi, tnx | 08:53 |
| oanson | In Dragonflow, we detect the behaviour of the VIP port, but it goes through the general port-handling code. That code doesn't process ports that are 'disabled' | 08:53 |
| rm_work | and maybe this? https://review.openstack.org/#/c/486499/ | 08:53 |
| oanson | If the field is ignored in most cases, would there be a problem to set that field to up? | 08:53 |
| rm_work | oanson: maybe we could patch it to be enabled? I honestly don't think it matters for us, so yeah, maybe that's fine too? | 08:53 |
| nmagnezi | oanson, not sure about the usecase you mention. I don't know anything about Dragonflow. when Octavia will have providers built-in it might look differently for those other drivers | 08:53 |
| oanson | rm_work, I opened a bug. Patch should come in soon | 08:54 |
| rm_work | it might just have to be a downstream patch depending on how it interacts with various network systems | 08:54 |
| nmagnezi | rm_work, we'll have to provide that port with a host to bind to. that would mean the neutron L2 agent will try to wire it, no? | 08:54 |
| rm_work | but the patch seems VERY simple, so you could always put that up and test it in your environment | 08:54 |
| oanson | nmagnezi, I think any implementation respecting port admin state would ignore the port if it is down. | 08:54 |
| rm_work | and let us know if it solves your issue -- and we could try it out and see if it negatively impacts things | 08:55 |
| oanson | nmagnezi, leaving the binding information blank is fine by me | 08:55 |
| oanson | rm_work, I did test it in my environment :) Seems to fix my issue | 08:55 |
| nmagnezi | oanson, if and when you propose the patch. let me know and I'd be happy to have look :) | 08:55 |
| rm_work | nmagnezi: i think it's literally just a bool we pass in | 08:55 |
| rm_work | i looked at it before, it didn't seem to matter for me | 08:55 |
| rm_work | oanson: if you put up a patch, our CI will test it and see if it interferes with devstack | 08:55 |
| rm_work | if so, we will have to look at what's going on | 08:56 |
| oanson | mnaser, rm_work, will do. Thanks | 08:56 |
| openstackgerrit | Omer Anson proposed openstack/neutron-lbaas master: Set VIP port to be enabled https://review.openstack.org/525504 | 08:57 |
| oanson | rm_work, nmagnezi, ^^^ | 08:58 |
| nmagnezi | oanson, not sure about this ^. you changed it in a patch that is relevant for *all* drivers. not only Octavia. let's see how it runs in CI | 09:01 |
| rm_work | ah this is n-lbaas | 09:01 |
| rm_work | yeah we'll see what happens when CI runs | 09:01 |
| nmagnezi | yeah I was under the impression oanson meant for Octavia as it's own endpoint | 09:02 |
| oanson | nmagnezi, rm_work, this is the only location I found in the code. I didn't find the octavia-specific location. | 09:02 |
| rm_work | oanson: but, this means you're running neutron-lbaas | 09:02 |
| oanson | If you prefer a different location (especially if CI fails) then I can move the change there | 09:02 |
| oanson | Should be running octavia. I understood this flow is in the path. | 09:03 |
| oanson | When I changed this on my environment, I saw a difference. And I am pretty sure I'm using Octavia (e.g. I see amphora VMs) | 09:03 |
| rm_work | what version of neutron-lbaas and octavia are you running? | 09:04 |
| rm_work | as of Pike, we recommend running Octavia standalone, without running neutron-lbaas in front | 09:04 |
| oanson | I'm running master | 09:04 |
| rm_work | in which case, your change would be in the octavia project, not the neutron-lbaas project | 09:04 |
| rm_work | yeah in that case, HIGHLY RECOMMEND you don't run neutron-lbaas | 09:04 |
| rm_work | run Octavia only | 09:05 |
| rm_work | neutron-lbaas is a huge headache | 09:05 |
| oanson | Yes, I see both in my local.conf | 09:05 |
| oanson | Let me try and fix my local.conf and re-run my environment, and see if that solves the issue. | 09:05 |
| oanson | Thanks! | 09:05 |
| rm_work | ok, so you are just doing devstack right now? | 09:05 |
| rm_work | in that case, yeah, definitely remove any reference to neutron-lbaas | 09:05 |
| rm_work | if you show me your local.conf i can tell you what you can remove | 09:05 |
| rm_work | (censored for passwords/etc of course) | 09:06 |
| oanson | Yes | 09:06 |
| oanson | local.conf is here: https://paste.fedoraproject.org/paste/~s8kKduywnfibmj5tIbWrw | 09:07 |
| oanson | (I work fully upstream (: ) | 09:07 |
| oanson | This also includes kuryr | 09:07 |
| rm_work | aaah | 09:07 |
| rm_work | you are using kuryr k8s | 09:07 |
| rm_work | i wonder if they fixed their stuff to work directly with octavia yet (not that they should really have to change anything, but it may not be tested) | 09:08 |
| rm_work | well, anyway, you can TRY this: | 09:08 |
| rm_work | comment out 88-90 | 09:08 |
| oanson | I'll try that. | 09:09 |
| rm_work | actually yeah it even says, "in case Octavia is older than pike" | 09:09 |
| rm_work | which it is not :P | 09:09 |
| rm_work | amp_active_retries=9999 | 09:09 |
| rm_work | THAT is interesting, lol | 09:09 |
| rm_work | maybe a *little bit* high :P | 09:09 |
| rm_work | hmmm also that cirros image looks a little old, but it shouldn't really matter | 09:10 |
| rm_work | I assume KURYR_K8S_LBAAS_USE_OCTAVIA is True | 09:10 |
| irenab | rm_work, you cannot imagine how much time it takes to get the environment up | 09:10 |
| rm_work | i mean... | 09:10 |
| rm_work | maybe :P | 09:10 |
| irenab | so the retries are essential :-) | 09:10 |
| rm_work | err so... | 09:11 |
| rm_work | it may take a while | 09:11 |
| rm_work | like 10m to boot them without nested virt? | 09:11 |
| rm_work | I just mean, when you get to production, remember 9999 is probably not a good choice :P but you're in devstack still so whatever | 09:12 |
| rm_work | if you DO get a legitimate failure though, you're going to be waiting a while | 09:12 |
| *** cody-somerville has joined #openstack-lbaas | 09:22 | |
| *** knsahm has quit IRC | 09:41 | |
| *** dokua has joined #openstack-lbaas | 09:42 | |
| *** dokua has quit IRC | 09:42 | |
| *** dokua has joined #openstack-lbaas | 09:44 | |
| *** salmankhan has joined #openstack-lbaas | 09:53 | |
| openstackgerrit | Numan Siddique proposed openstack/neutron-lbaas master: Support OVN driver for lbaas v2 https://review.openstack.org/510921 | 09:54 |
| *** krypto has quit IRC | 10:06 | |
| openstackgerrit | Numan Siddique proposed openstack/neutron-lbaas master: Support OVN driver for lbaas v2 https://review.openstack.org/510921 | 10:08 |
| *** krypto has joined #openstack-lbaas | 10:13 | |
| *** krypto has joined #openstack-lbaas | 10:13 | |
| *** krypto has quit IRC | 10:17 | |
| *** krypto has joined #openstack-lbaas | 10:18 | |
| *** krypto has quit IRC | 10:18 | |
| *** krypto has joined #openstack-lbaas | 10:18 | |
| *** rcernin has quit IRC | 10:30 | |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: Change how path munging works for sphinx docs https://review.openstack.org/521211 | 10:31 |
| *** annp has quit IRC | 10:33 | |
| *** salmankhan has quit IRC | 10:52 | |
| *** ianychoi has quit IRC | 10:55 | |
| *** ianychoi has joined #openstack-lbaas | 10:55 | |
| *** salmankhan has joined #openstack-lbaas | 10:57 | |
| *** yamamoto has quit IRC | 11:28 | |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: ignore api-ref/build directory https://review.openstack.org/522385 | 11:50 |
| *** sanfern has joined #openstack-lbaas | 12:04 | |
| *** yamamoto has joined #openstack-lbaas | 12:28 | |
| openstackgerrit | Hengqing Hu proposed openstack/octavia-dashboard master: Remove build_sphinx section from setup.cfg https://review.openstack.org/525580 | 12:29 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: Change how path munging works for sphinx docs https://review.openstack.org/521211 | 12:30 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: ignore api-ref/build directory https://review.openstack.org/522385 | 12:30 |
| *** yamamoto has quit IRC | 12:31 | |
| *** yamamoto has joined #openstack-lbaas | 12:31 | |
| openstackgerrit | Hengqing Hu proposed openstack/python-octaviaclient master: Remove build_sphinx section from setup.cfg https://review.openstack.org/525586 | 12:36 |
| *** yamamoto has quit IRC | 12:38 | |
| *** yamamoto has joined #openstack-lbaas | 12:40 | |
| openstackgerrit | Hengqing Hu proposed openstack/neutron-lbaas master: Remove build_sphinx section from setup.cfg https://review.openstack.org/525591 | 12:46 |
| dokua | Hi colleagues, I have a problem creating pool members using Heat. When creating two members, I'm getting the following problem: | 13:07 |
| dokua | - Neutron calls for both members creation: | 13:07 |
| dokua | INFO neutron_lbaas.services.loadbalancer.plugin [...] Calling driver operation MemberManager.create | 13:07 |
| dokua | INFO neutron_lbaas.services.loadbalancer.plugin [...] Calling driver operation MemberManager.create | 13:07 |
| dokua | - while Octavia creates one and refuses creation of another: | 13:07 |
| dokua | INFO octavia.api.v1.controllers.member [...] Member cannot be created or modified because the Load Balancer is in an immutable state | 13:07 |
| dokua | INFO octavia.controller.queue.endpoint [-] Creating member 'aed824c6-753e-42a7-ab24-9f38f225b2e5'... | 13:07 |
| dokua | - and haproxy.conf contains only one member in pool: | 13:07 |
| dokua | backend d50fc8b3-93da-488f-b235-4d2316d8700b | 13:07 |
| dokua | [ ... ] | 13:07 |
| dokua | server aed824c6-753e-42a7-ab24-9f38f225b2e5 10.1.1.11:8080 weight 1 check inter 5s fall 3 rise 3 | 13:07 |
| dokua | When I do the same using CLI, no problems. It seems that Heat issues requests much faster than I can do this from CLI and there is kind of "deadlock" somewhere. Any ideas on this? Thanks! | 13:07 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: Change how path munging works for sphinx docs https://review.openstack.org/521211 | 13:18 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: ignore api-ref/build directory https://review.openstack.org/522385 | 13:18 |
| *** links has quit IRC | 13:22 | |
| openstackgerrit | Guoqiang Ding proposed openstack/neutron-lbaas master: Fix the deprecated usage of "get_transport" https://review.openstack.org/525604 | 13:22 |
| openstackgerrit | Numan Siddique proposed openstack/neutron-lbaas master: Support OVN driver for lbaas v2 https://review.openstack.org/510921 | 13:23 |
| *** krypto has quit IRC | 13:23 | |
| *** yamamoto has quit IRC | 13:47 | |
| *** eN_Guruprasad_Rn has quit IRC | 13:50 | |
| *** fnaval has quit IRC | 14:02 | |
| *** yamamoto has joined #openstack-lbaas | 14:11 | |
| openstackgerrit | Bar RH proposed openstack/octavia master: Fix filtering in list API calls https://review.openstack.org/522689 | 14:12 |
| openstackgerrit | Bar RH proposed openstack/octavia master: Fix filtering in list API calls https://review.openstack.org/522689 | 14:15 |
| *** sanfern has quit IRC | 14:42 | |
| *** fnaval has joined #openstack-lbaas | 14:52 | |
| *** armax has joined #openstack-lbaas | 14:56 | |
| *** Alex_Staf has quit IRC | 15:09 | |
| *** jniesz has joined #openstack-lbaas | 15:22 | |
| openstackgerrit | Bar RH proposed openstack/python-octaviaclient master: Add Quota client API and OSC support https://review.openstack.org/518767 | 15:24 |
| openstackgerrit | Bar RH proposed openstack/python-octaviaclient master: Extend loadbalancer_create valid VIP parameters combinations https://review.openstack.org/519439 | 15:24 |
| openstackgerrit | Bar RH proposed openstack/python-octaviaclient master: Complement Octavia client with a set of features https://review.openstack.org/522666 | 15:24 |
| *** bar_ has joined #openstack-lbaas | 15:24 | |
| *** rm_work has quit IRC | 15:32 | |
| *** bar_ has quit IRC | 15:34 | |
| *** rm_work has joined #openstack-lbaas | 15:36 | |
| *** longstaff has joined #openstack-lbaas | 15:44 | |
| *** krypto has joined #openstack-lbaas | 15:48 | |
| *** openstackgerrit has quit IRC | 15:48 | |
| *** krypto has quit IRC | 15:50 | |
| *** dokua has quit IRC | 15:55 | |
| *** dokua1 has joined #openstack-lbaas | 15:55 | |
| *** longstaff has quit IRC | 15:57 | |
| *** dokua1 has quit IRC | 15:59 | |
| *** longstaff has joined #openstack-lbaas | 15:59 | |
| *** salmankhan has quit IRC | 16:03 | |
| *** longstaff has quit IRC | 16:05 | |
| *** salmankhan has joined #openstack-lbaas | 16:06 | |
| *** longstaff has joined #openstack-lbaas | 16:11 | |
| *** longstaff has quit IRC | 16:14 | |
| *** longstaff has joined #openstack-lbaas | 16:14 | |
| *** pck is now known as pckizer | 16:29 | |
| *** bar_ has joined #openstack-lbaas | 16:32 | |
| *** bar_ has quit IRC | 17:11 | |
| *** longstaff has quit IRC | 17:11 | |
| *** longstaff has joined #openstack-lbaas | 17:14 | |
| *** openstackgerrit has joined #openstack-lbaas | 17:18 | |
| openstackgerrit | German Eichberger proposed openstack/neutron-lbaas master: [WIP] Adds the missing stats command https://review.openstack.org/525704 | 17:18 |
| *** pcaruana has joined #openstack-lbaas | 17:22 | |
| *** sanfern has joined #openstack-lbaas | 17:25 | |
| *** pcaruana has quit IRC | 17:26 | |
| *** pcaruana has joined #openstack-lbaas | 17:27 | |
| *** b_bezak has quit IRC | 17:30 | |
| *** gcheresh has joined #openstack-lbaas | 17:43 | |
| *** dokua has joined #openstack-lbaas | 17:46 | |
| sanfern | Hi johnsom , | 17:52 |
| *** gcheresh has quit IRC | 17:52 | |
| johnsom | sanfern Hi | 17:52 |
| sanfern | Adam wants some clarification on backwards compatibility - https://review.openstack.org/#/c/521138/ | 17:53 |
| sanfern | We have discussed this once, can we put this discussion on tomorrow's agenda | 17:54 |
| johnsom | Yes, we can add it to the agenda | 17:54 |
| sanfern | yeah do you have the link to update | 17:55 |
| johnsom | It is here but I can do it for you: https://wiki.openstack.org/wiki/Octavia/Weekly_Meeting_Agenda | 17:55 |
| sanfern | ok great thanks. | 17:56 |
| sanfern | I want this to be merged | 17:56 |
| sanfern | Also https://review.openstack.org/#/c/490164/20 | 17:56 |
| johnsom | Yeah, I spent all day doing reviews yesterday. Still more to do. We will get to it. | 17:57 |
| johnsom | Thanks for the reminder | 17:57 |
| *** pcaruana has quit IRC | 17:58 | |
| sanfern | Thank you johnsom | 17:58 |
| johnsom | It would be nice to see any reviews on it... | 17:58 |
| sanfern | I have asked xgerman_ to do it. | 18:16 |
| *** tesseract has quit IRC | 18:21 | |
| *** salmankhan has quit IRC | 18:24 | |
| *** longstaff has quit IRC | 18:33 | |
| *** tongl has joined #openstack-lbaas | 18:37 | |
| *** Alex_Staf has joined #openstack-lbaas | 18:38 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: DNM: Revert "Disable kvm on OVH infra instances" https://review.openstack.org/525729 | 18:45 |
| rm_work | ^^ testing OVH stuff | 18:47 |
| xgerman_ | rm_work: https://review.openstack.org/#/c/418530/28 | 18:48 |
| rm_work | xgerman_: it works? | 18:48 |
| rm_work | I'll review it now | 18:48 |
| rm_work | ugh i'm not in a great place for spinning a devstack though so I will trust johnsom assuming he tested it? | 18:49 |
| johnsom | I did a test pass. The basic stuff seems to work. Certainly not an exhaustive test. I found that status and stats don't work. I had him open a bug and deferred to the other cores to decide if we merge as-is and fix or wait for those to work. | 18:50 |
| rm_work | yeah that's fine | 18:50 |
| rm_work | a bug we can even backport, too | 18:50 |
| rm_work | hate to be thinking like that, but | 18:50 |
| rm_work | we need to get some features in | 18:51 |
| rm_work | xgerman_: i have some tiny ones for you, just some easy fixes/additions that make things easier for me, if you had a min to look while i'm reviewing that | 18:54 |
| rm_work | https://review.openstack.org/525295 | 18:54 |
| rm_work | https://review.openstack.org/525302 | 18:54 |
| rm_work | https://review.openstack.org/525353 <-- this one is admittedly a little more controversial, just want to add a thing because it belongs there, isn't used yet in the main code, but comes with bonus tests! | 18:55 |
| johnsom | rm_work I hope you can make tomorrows IRC meeting. We want to talk about the -2 | 18:59 |
| *** bar_ has joined #openstack-lbaas | 18:59 | |
| openstackgerrit | Merged openstack/octavia master: Updated from global requirements https://review.openstack.org/525387 | 19:01 |
| johnsom | mugsie How would it break API compatibility other than the amphora API change (which has not yet gone out in a release)? | 19:01 |
| rm_work | johnsom: yes, there's another one too | 19:01 |
| rm_work | johnsom: yeah i commented to that effect | 19:01 |
| rm_work | xgerman_: quick -1, maybe you can answer my questions? ignore the deep/shallow copy thing, that was just an offhand nit | 19:02 |
| johnsom | Yeah, upgrades, I see two issues. 1. it would be control plane downtime. 2. the amphora API probably needs some compat work so you don't have to change out the amps. | 19:02 |
| rm_work | johnsom: i wanted to discuss this as well tomorrow: https://review.openstack.org/#/c/505158/ | 19:03 |
| rm_work | ah which links into #2 there | 19:03 |
| johnsom | Ok, throw it on the agenda | 19:03 |
| rm_work | k | 19:04 |
| rm_work | added | 19:05 |
| johnsom | Thanks. It should be a busy one | 19:05 |
| rm_work | last week i was having jetlag issues and also alarm issues, combined i missed ... a lot of stuff | 19:06 |
| *** longstaff has joined #openstack-lbaas | 19:07 | |
| johnsom | bar_ are we good for a re-test on the client / filter patches? | 19:09 |
| xgerman_ | rm_work commented - let me know if I need to re-roll | 19:10 |
| xgerman_ | and this was sitting there for almost a year… | 19:10 |
| johnsom | broken | 19:10 |
| *** gcheresh has joined #openstack-lbaas | 19:24 | |
| *** yamamoto has quit IRC | 19:27 | |
| rm_work | xgerman_: this was the first time I heard it was "ready/working" :P | 19:28 |
| *** SumitNaiksatam has joined #openstack-lbaas | 19:31 | |
| xgerman_ | yeah, it’s just hard for me to conext-switch | 19:32 |
| bar_ | johnsom, yes, the filter patches should do the job. | 19:36 |
| *** bar_ has quit IRC | 19:43 | |
| *** Alex_Staf has quit IRC | 19:52 | |
| *** salmankhan has joined #openstack-lbaas | 19:54 | |
| *** salmankhan has quit IRC | 19:58 | |
| *** csomerville has joined #openstack-lbaas | 20:00 | |
| mugsie | johnsom: technically as the API has been put to master, it has been "released" - but that is a philisophical debate :) | 20:15 |
| mugsie | API compatibility was supposed to be patch to patch (part of the reason tempest is in a separate repo) | 20:15 |
| *** kobis has joined #openstack-lbaas | 20:18 | |
| xgerman_ | I won the battle on the neutron side apparently, too | 20:19 |
| *** gcheresh has quit IRC | 20:24 | |
| *** yamamoto has joined #openstack-lbaas | 20:28 | |
| xgerman_ | rm_work you want a re-roll or you are fine with my answers | 20:31 |
| xgerman_ | ? | 20:31 |
| *** b_bezak has joined #openstack-lbaas | 20:31 | |
| *** yamamoto has quit IRC | 20:35 | |
| *** b_bezak has quit IRC | 20:35 | |
| *** kobis has quit IRC | 20:47 | |
| *** longstaff has quit IRC | 21:09 | |
| rm_work | xgerman_: sorry went to lunch, will look now | 21:11 |
| xgerman_ | let me know happy to-reroll | 21:11 |
| *** longstaff has joined #openstack-lbaas | 21:15 | |
| rm_work | xgerman_: followup question | 21:15 |
| xgerman_ | sure | 21:15 |
| rm_work | posted | 21:15 |
| xgerman_ | ok, answered | 21:20 |
| xgerman_ | let me know what you want me to do next with this | 21:20 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Update devstack to properly configure nova-lxd img https://review.openstack.org/453237 | 21:21 |
| rm_work | xgerman_: so the tests that were in tests/unit/db/loadbalancer/test_db_loadbalancerv2.py | 21:22 |
| rm_work | went to tests/unit/db/loadbalancer/util.py | 21:22 |
| rm_work | ? | 21:22 |
| rm_work | it's hard to tell | 21:23 |
| xgerman_ | thos weren’t tests - they were test helpers | 21:23 |
| rm_work | it looked like a bunch of code moved to test_proxy_plugin | 21:23 |
| rm_work | but | 21:23 |
| rm_work | i guess yeah, i see it | 21:23 |
| xgerman_ | yeah, I felt refactoring was best… | 21:23 |
| *** armax has quit IRC | 21:25 | |
| rm_work | alright, good enough for me | 21:26 |
| rm_work | posted final comments, but +A | 21:26 |
| xgerman_ | THANK YOU! This was haunting me for almost a year… I will run a follow up patch with stats and status — as soon as my devstack becomes agreeable | 21:27 |
| *** bar_ has joined #openstack-lbaas | 21:35 | |
| openstackgerrit | Bar RH proposed openstack/octavia master: Fail-proof VIP deallocation task https://review.openstack.org/523931 | 21:39 |
| johnsom | FYI, these tempest updates are just adding the story tag so we track our Queens community goal... | 21:41 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Add basic tests for listeners https://review.openstack.org/492311 | 21:42 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Add tests for pools https://review.openstack.org/492673 | 21:43 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Create scenario tests for members https://review.openstack.org/497139 | 21:43 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: Create scenario tests for health monitors https://review.openstack.org/497643 | 21:44 |
| bar_ | johnsom, ^ I have added retry for Exception. | 21:44 |
| openstackgerrit | Michael Johnson proposed openstack/octavia-tempest-plugin master: WIP: Failover test https://review.openstack.org/501559 | 21:44 |
| johnsom | bar_ Thanks! | 21:45 |
| johnsom | Again, sorry for the channel spam. Just getting things tracked so I can post progress on the Queens community goal for the tempest plugin. | 21:45 |
| johnsom | I now return you to your regularly scheduled programming. | 21:46 |
| *** sshank has joined #openstack-lbaas | 21:54 | |
| *** threestrands has joined #openstack-lbaas | 22:05 | |
| *** threestrands has quit IRC | 22:05 | |
| *** threestrands has joined #openstack-lbaas | 22:05 | |
| *** rcernin has joined #openstack-lbaas | 22:05 | |
| johnsom | Any cores have a minute for a dashboard patch updating the auth code? Want to get this in Q2 if we can. https://review.openstack.org/#/c/524011/ | 22:05 |
| rm_work | sure why not | 22:10 |
| rm_work | johnsom: i'm about to submit the code for the amphora-api /failover | 22:12 |
| johnsom | Ok | 22:12 |
| rm_work | johnsom: should i do it as a patch on top of https://review.openstack.org/#/c/525302/ or just merge it in | 22:12 |
| rm_work | I was trying to keep things simple | 22:13 |
| rm_work | but | 22:13 |
| rm_work | i guess it's impossible to test that code without this code | 22:13 |
| johnsom | That is what parent/child is for... | 22:14 |
| johnsom | I like just checking out the end of the chain for things like this when testing. | 22:15 |
| *** armax has joined #openstack-lbaas | 22:20 | |
| openstackgerrit | Merged openstack/octavia-dashboard master: Update openstacksdk construction to be forward compatible https://review.openstack.org/524011 | 22:21 |
| rm_work | k | 22:29 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Amphora API Failover call https://review.openstack.org/525778 | 22:36 |
| rm_work | there it is | 22:36 |
| bar_ | Does the order of the attributes in json response have significance? lb_stats_response.json example differs from the response description in the documentation (loadbalancer.inc) | 22:37 |
| openstackgerrit | Michael Johnson proposed openstack/python-octaviaclient master: Complement Octavia client with a set of features https://review.openstack.org/522666 | 22:40 |
| *** sshank has quit IRC | 22:44 | |
| bar_ | johnsom, If you have reservations from the octavia server-side dependency (https://review.openstack.org/#/c/522689/) please comment on it today so I'll have time to correct it. | 22:44 |
| johnsom | bar_ on my last test with it, then +2 | 22:44 |
| rm_work | bar_: order doesn't matter | 22:46 |
| rm_work | the assertion stuff in tempest shouldn't test | 22:46 |
| rm_work | err, shouldn't test order | 22:46 |
| rm_work | if you do assertEquals on dicts, it ignores order | 22:47 |
| rm_work | since they are inherently unordered | 22:47 |
| rm_work | and we're fine with that | 22:47 |
| bar_ | oh, I was reviewing the client patch set for loadbalancer stats, so it's all user facing. And in this patch, the order of the user views the response is not the same as manifested in loadbalancer.inc | 22:48 |
| bar_ | rm_work, ^ | 22:48 |
| rm_work | doesn't matter | 22:49 |
| rm_work | feel free to fix docs if you want i guess | 22:49 |
| bar_ | nah | 22:49 |
| rm_work | but order REALLY shouldn't matter | 22:49 |
| johnsom | I am going to comment that the CLI output should be in alpha order though | 22:50 |
| bar_ | johnsom, asked me yesterday to add columns for amphora client in a certain order | 22:50 |
| rm_work | heh | 22:50 |
| johnsom | Yeah, the CLI table output should have an order that makes sense | 22:51 |
| johnsom | That is different than JSON | 22:51 |
| rm_work | yep it's true | 22:51 |
| rm_work | tho alpha? | 22:51 |
| johnsom | Well, in this case: | 22:51 |
| johnsom | https://review.openstack.org/#/c/523306/1/octaviaclient/osc/v2/constants.py | 22:51 |
| rm_work | don't things normally do some intelligent sorting, like id-first? | 22:51 |
| johnsom | bar_ Yep, the filtering patch works for me | 22:54 |
| johnsom | I think we might need to fix the quota patch ourselves to get this chain in. | 22:54 |
| bar_ | johnsom, I'm glad to here | 22:54 |
| bar_ | johnsom, I'll approach Carlos tomorrow, it doesn't need much work, and I know he's on to it (among other assignments) | 22:55 |
| bar_ | *to hear | 22:56 |
| johnsom | Yeah, tomorrow makes me nervous given I need to do the release tomorrow. If no one else has time I make poke at it. | 22:56 |
| johnsom | I want to review some dashboard stuff too, but client takes priority | 22:57 |
| openstackgerrit | German Eichberger proposed openstack/neutron-lbaas master: Adds the missing stats command and fixes status https://review.openstack.org/525704 | 22:57 |
| xgerman_ | ^^ now if pep8, etc. are happy we have stats and status… | 22:58 |
| xgerman_ | johnsom anyway I can help with the client? | 22:58 |
| *** sshank has joined #openstack-lbaas | 22:59 | |
| rm_work | johnsom: got a weird one ... member update died in the worker because of a connection timeout to neutron looking for the extensions | 23:03 |
| rm_work | and it ... didn't roll back | 23:03 |
| rm_work | O_o | 23:03 |
| rm_work | because it was before the flow maybe? | 23:03 |
| rm_work | trying to init the neutron driver? | 23:03 |
| rm_work | sec i'll post logs | 23:03 |
| rm_work | http://paste.openstack.org/show/VlRFt3vjI17ZaAJJXJDs/ | 23:05 |
| johnsom | xgerman I think bar_ has it covered | 23:08 |
| bar_ | +1 | 23:08 |
| johnsom | rm_work Yep, that check for neutron extensions is outside the flow in the driver __init__()... | 23:13 |
| rm_work | T_T | 23:13 |
| johnsom | https://github.com/openstack/octavia/blob/master/octavia/network/drivers/neutron/base.py#L47 | 23:13 |
| rm_work | so that happens if connectivity goes bad to neutron temporarily | 23:13 |
| rm_work | (our neutron was down for 5min) | 23:13 |
| johnsom | Yeah, we should move that into the flow. It's not like it's a shared driver and only loaded once.... | 23:15 |
| rm_work | do you know where you'd do that? | 23:19 |
| xgerman_ | mmh, why are we optimizing for. abroken neutron? | 23:19 |
| rm_work | we're trying to remove cases where it's possible for things to go to error | 23:20 |
| rm_work | connectivity issues happen | 23:20 |
| johnsom | because it leaves an LB in PENDING... | 23:20 |
| johnsom | It's a fix our state machine issue IMO | 23:21 |
| rm_work | yes | 23:21 |
| xgerman_ | I don’t like covering every edge case and neutron being not available is pretty edgy | 23:21 |
| johnsom | Frankly I would probably just move those check calls into the network driver methods that care. Those are neutron driver specific issues, so I think it should be a detail hidden. | 23:22 |
| rm_work | yeah | 23:22 |
| johnsom | xgerman But we should take the LB to ERROR to give a clear indication to the user and operator. | 23:23 |
| johnsom | If it's outside the flow we don't have the ERROR fail safe | 23:23 |
| xgerman_ | ok, at best I see that as low priority… | 23:24 |
| xgerman_ | but I like the plan moving things to the methods which care - though we should then cache the response so we are not going over the wire all the time | 23:25 |
| rm_work | yeah basically the flows should be airtight -- doesn't matter if it's neutron or anything else, connectivity blips for any reason should not cause LBs to be forever stuck in PENDING. *NOTHING* should cause LBs to get stuck in PENDING. | 23:30 |
| xgerman_ | given that we are having Q-2 and are in no position to deprecate lbaasv2 we have bigger fish to fry than all the edge cases | 23:31 |
| johnsom | I think supporting our teammates that have this in customer hands has some priority in my book.... | 23:33 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: Change how path munging works for sphinx docs https://review.openstack.org/521211 | 23:34 |
| xgerman_ | I don’t think Neutron outages are very common… | 23:36 |
| xgerman_ | but we can also keep supporting lbaasv2 another cycle | 23:36 |
| rm_work | i think i could prolly fix it in approximately the time it took to argue about this :P | 23:39 |
| xgerman_ | most things are that way… | 23:39 |
| xgerman_ | reviews are our bottle neck | 23:40 |
| rm_work | kinda? | 23:40 |
| rm_work | a lot of stuff just doesn't look ready to review, to me | 23:40 |
| rm_work | i guess there are a few | 23:40 |
| xgerman_ | a couple of people are new to OpenStack and would benefit from early reviews | 23:41 |
| rm_work | at this point it might be good to have an etherpad or something where people put links to reviews that they actually think are *ready to go* | 23:41 |
| rm_work | hmmm | 23:41 |
| *** tongl has quit IRC | 23:41 | |
| rm_work | i guess that's a fair point | 23:41 |
| rm_work | sometimes reviews can be good even when stuff isn't done | 23:41 |
| rm_work | though that might be the luxury | 23:41 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia master: Change how path munging works for sphinx docs https://review.openstack.org/521211 | 23:42 |
| xgerman_ | yeah, it’s the teach people to fish thing and then they get put on another project | 23:42 |
| *** tongl has joined #openstack-lbaas | 23:49 | |
| *** sshank has quit IRC | 23:51 | |
| openstackgerrit | Michael Johnson proposed openstack/python-octaviaclient master: Add loadbalancer stats client api and osc https://review.openstack.org/523306 | 23:52 |
| openstackgerrit | Merged openstack/python-octaviaclient master: Updated from global requirements https://review.openstack.org/525400 | 23:53 |
| johnsom | Ok, that client patch works for me. I just fixed the two minor things I found. | 23:53 |
| *** sshank has joined #openstack-lbaas | 23:53 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Move loading the network driver into the flows https://review.openstack.org/525790 | 23:55 |
| rm_work | ^^ fixed the network driver loading badly | 23:55 |
| rm_work | xgerman_: what about this: you're doing maintenance on neutron to upgrade it -- should octavia let LBs get stuck immutable because of that? :/ | 23:58 |
| rm_work | not sure if your neutron is zero-downtime-upgrades, but ours isn't | 23:58 |
| xgerman_ | if neutron is not zero downtime I would schedule a maintenance and take down all the APIs… | 23:59 |
| xgerman_ | it’s not like nova will be working | 23:59 |
| *** longstaff has quit IRC | 23:59 | |
| rm_work | yeah but I would also say that we should do the same for the compute driver | 23:59 |
| rm_work | if there's somewhere we needed to do that | 23:59 |
| rm_work | we have to be resilient | 23:59 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!