*** diogogmt has joined #openstack-lbaas | 00:00 | |
*** ducttape_ has joined #openstack-lbaas | 00:02 | |
dougwig | sbalukoff: where are the patches that need a look? | 00:03 |
---|---|---|
dougwig | sbalukoff: found it | 00:05 |
sbalukoff | Don't have them handy (on the road). johnsom? Got the etherpad link handy? | 00:05 |
sbalukoff | Ok | 00:05 |
sbalukoff | Thanks for going over them, dougwig! | 00:05 |
johnsom | https://etherpad.openstack.org/p/LBaaSOctaviaNewtonPatches | 00:06 |
johnsom | Just to keep poking | 00:06 |
openstackgerrit | Doug Wiegley proposed openstack/neutron-lbaas: DO NOT MERGE - functional job debugging https://review.openstack.org/370494 | 00:29 |
johnsom | Ok, so the inconsistent LB provisioning status patch is not fixing the real root cause which is an un handled exception in a revert task (can't delete the security group). Working on that now. | 00:41 |
*** gongysh has joined #openstack-lbaas | 00:55 | |
*** itsuugo has quit IRC | 01:08 | |
*** itsuugo has joined #openstack-lbaas | 01:10 | |
*** bana_k has quit IRC | 01:14 | |
*** itsuugo has quit IRC | 01:15 | |
*** itsuugo has joined #openstack-lbaas | 01:16 | |
*** itsuugo has quit IRC | 01:22 | |
*** itsuugo has joined #openstack-lbaas | 01:22 | |
*** kevo has quit IRC | 01:27 | |
*** itsuugo has quit IRC | 01:27 | |
*** itsuugo has joined #openstack-lbaas | 01:28 | |
*** bana_k has joined #openstack-lbaas | 01:35 | |
*** ducttape_ has quit IRC | 01:37 | |
openstackgerrit | Banashankar k proposed openstack/neutron-lbaas: validate that delay and timeout Make sure delay must be greater than timeout https://review.openstack.org/355115 | 01:46 |
*** gongysh has quit IRC | 01:47 | |
*** itsuugo has quit IRC | 01:48 | |
*** woodster_ has quit IRC | 01:50 | |
*** itsuugo has joined #openstack-lbaas | 01:50 | |
*** ducttape_ has joined #openstack-lbaas | 01:55 | |
*** itsuugo has quit IRC | 01:55 | |
*** ducttape_ has quit IRC | 01:56 | |
*** itsuugo has joined #openstack-lbaas | 01:57 | |
*** itsuugo has quit IRC | 02:08 | |
*** diogogmt has quit IRC | 02:09 | |
*** itsuugo has joined #openstack-lbaas | 02:09 | |
*** itsuugo has quit IRC | 02:16 | |
*** itsuugo has joined #openstack-lbaas | 02:17 | |
*** itsuugo has quit IRC | 02:22 | |
*** itsuugo has joined #openstack-lbaas | 02:23 | |
*** itsuugo has quit IRC | 02:29 | |
*** ducttape_ has joined #openstack-lbaas | 02:30 | |
*** itsuugo has joined #openstack-lbaas | 02:30 | |
*** itsuugo has quit IRC | 02:35 | |
*** itsuugo has joined #openstack-lbaas | 02:36 | |
*** itsuugo has quit IRC | 02:43 | |
*** itsuugo has joined #openstack-lbaas | 02:44 | |
*** yuanying has quit IRC | 02:47 | |
*** ducttape_ has quit IRC | 02:47 | |
*** itsuugo has quit IRC | 02:48 | |
*** itsuugo has joined #openstack-lbaas | 02:50 | |
*** gongysh has joined #openstack-lbaas | 02:50 | |
*** itsuugo has quit IRC | 03:01 | |
*** itsuugo has joined #openstack-lbaas | 03:03 | |
*** itsuugo has quit IRC | 03:08 | |
*** fnaval has quit IRC | 03:09 | |
*** itsuugo has joined #openstack-lbaas | 03:10 | |
*** itsuugo has quit IRC | 03:17 | |
*** itsuugo has joined #openstack-lbaas | 03:18 | |
*** gongysh has quit IRC | 03:21 | |
*** itsuugo has quit IRC | 03:23 | |
*** itsuugo has joined #openstack-lbaas | 03:24 | |
*** fnaval has joined #openstack-lbaas | 03:25 | |
*** gongysh has joined #openstack-lbaas | 03:25 | |
*** itsuugo has quit IRC | 03:33 | |
*** itsuugo has joined #openstack-lbaas | 03:36 | |
*** itsuugo has quit IRC | 03:41 | |
*** itsuugo has joined #openstack-lbaas | 03:42 | |
*** ducttape_ has joined #openstack-lbaas | 03:48 | |
*** yuanying has joined #openstack-lbaas | 03:51 | |
*** itsuugo has quit IRC | 03:51 | |
*** itsuugo has joined #openstack-lbaas | 03:52 | |
*** ducttape_ has quit IRC | 03:53 | |
*** kevo has joined #openstack-lbaas | 03:58 | |
*** itsuugo has quit IRC | 03:59 | |
*** itsuugo has joined #openstack-lbaas | 04:00 | |
*** itsuugo has quit IRC | 04:10 | |
*** itsuugo has joined #openstack-lbaas | 04:12 | |
*** itsuugo has quit IRC | 04:16 | |
*** itsuugo has joined #openstack-lbaas | 04:18 | |
*** itsuugo has quit IRC | 04:23 | |
*** yamamoto_ has joined #openstack-lbaas | 04:24 | |
*** itsuugo has joined #openstack-lbaas | 04:24 | |
*** Alex_Stef has joined #openstack-lbaas | 04:26 | |
*** itsuugo has quit IRC | 04:29 | |
*** itsuugo has joined #openstack-lbaas | 04:29 | |
*** itsuugo has quit IRC | 04:34 | |
*** itsuugo has joined #openstack-lbaas | 04:35 | |
*** itsuugo has quit IRC | 04:40 | |
*** itsuugo has joined #openstack-lbaas | 04:41 | |
*** ducttape_ has joined #openstack-lbaas | 04:49 | |
*** itsuugo has quit IRC | 04:53 | |
*** itsuugo has joined #openstack-lbaas | 04:54 | |
*** ducttape_ has quit IRC | 04:55 | |
*** itsuugo has quit IRC | 05:03 | |
*** itsuugo has joined #openstack-lbaas | 05:04 | |
*** kevo has quit IRC | 05:07 | |
*** yamamoto_ has quit IRC | 05:07 | |
*** yamamoto_ has joined #openstack-lbaas | 05:09 | |
*** itsuugo has quit IRC | 05:10 | |
*** itsuugo has joined #openstack-lbaas | 05:10 | |
*** rcernin has joined #openstack-lbaas | 05:17 | |
*** yamamoto_ has quit IRC | 05:19 | |
*** itsuugo has quit IRC | 05:20 | |
*** itsuugo has joined #openstack-lbaas | 05:21 | |
*** itsuugo has quit IRC | 05:26 | |
*** itsuugo has joined #openstack-lbaas | 05:28 | |
*** itsuugo has quit IRC | 05:37 | |
*** itsuugo has joined #openstack-lbaas | 05:38 | |
*** itsuugo has quit IRC | 05:43 | |
*** itsuugo has joined #openstack-lbaas | 05:43 | |
*** itsuugo has quit IRC | 05:51 | |
*** itsuugo has joined #openstack-lbaas | 05:51 | |
*** itsuugo has quit IRC | 05:56 | |
*** itsuugo has joined #openstack-lbaas | 05:58 | |
*** itsuugo has quit IRC | 06:03 | |
*** yamamoto_ has joined #openstack-lbaas | 06:03 | |
*** itsuugo has joined #openstack-lbaas | 06:03 | |
*** pcaruana has joined #openstack-lbaas | 06:04 | |
*** kobis has joined #openstack-lbaas | 06:13 | |
*** pcaruana is now known as pcaruana|afk| | 06:13 | |
*** itsuugo has quit IRC | 06:18 | |
*** itsuugo has joined #openstack-lbaas | 06:19 | |
*** kobis has quit IRC | 06:25 | |
*** kobis has joined #openstack-lbaas | 06:28 | |
*** itsuugo has quit IRC | 06:28 | |
*** itsuugo has joined #openstack-lbaas | 06:30 | |
*** itsuugo has quit IRC | 06:37 | |
*** itsuugo has joined #openstack-lbaas | 06:38 | |
*** pcaruana|afk| is now known as pcaruana | 06:40 | |
*** itsuugo has quit IRC | 06:44 | |
*** itsuugo has joined #openstack-lbaas | 06:45 | |
*** bana_k has quit IRC | 06:49 | |
*** ducttape_ has joined #openstack-lbaas | 06:52 | |
*** ducttape_ has quit IRC | 06:56 | |
*** itsuugo has quit IRC | 07:00 | |
*** itsuugo has joined #openstack-lbaas | 07:02 | |
*** itsuugo has quit IRC | 07:17 | |
*** itsuugo has joined #openstack-lbaas | 07:18 | |
*** nmagnezi_ has joined #openstack-lbaas | 07:22 | |
*** nmagnezi_ has quit IRC | 07:22 | |
*** itsuugo has quit IRC | 07:23 | |
*** nmagnezi has joined #openstack-lbaas | 07:24 | |
*** itsuugo has joined #openstack-lbaas | 07:24 | |
*** atul has joined #openstack-lbaas | 07:26 | |
*** itsuugo has quit IRC | 07:29 | |
*** itsuugo has joined #openstack-lbaas | 07:30 | |
*** itsuugo has quit IRC | 07:35 | |
*** itsuugo has joined #openstack-lbaas | 07:35 | |
*** cody-somerville has quit IRC | 07:37 | |
*** cody-somerville has joined #openstack-lbaas | 07:38 | |
*** gongysh has quit IRC | 07:40 | |
*** itsuugo has quit IRC | 07:47 | |
*** eezhova has joined #openstack-lbaas | 07:49 | |
*** itsuugo has joined #openstack-lbaas | 07:49 | |
*** Alex_Stef has quit IRC | 07:53 | |
*** basilAB has quit IRC | 08:03 | |
*** basilAB has joined #openstack-lbaas | 08:05 | |
*** amotoki has joined #openstack-lbaas | 08:14 | |
*** Alex_Stef has joined #openstack-lbaas | 08:15 | |
*** itsuugo has quit IRC | 08:32 | |
*** itsuugo has joined #openstack-lbaas | 08:34 | |
*** eezhova has quit IRC | 08:37 | |
*** itsuugo has quit IRC | 08:46 | |
*** Alex_Stef has quit IRC | 08:46 | |
*** itsuugo has joined #openstack-lbaas | 08:46 | |
*** itsuugo has quit IRC | 08:51 | |
*** itsuugo has joined #openstack-lbaas | 08:52 | |
*** ihrachys has joined #openstack-lbaas | 08:55 | |
openstackgerrit | Stephen Balukoff proposed openstack/octavia: Add L7 user guides https://review.openstack.org/360934 | 08:55 |
*** itsuugo has quit IRC | 09:06 | |
*** itsuugo has joined #openstack-lbaas | 09:08 | |
*** Alex_Stef has joined #openstack-lbaas | 09:11 | |
* ajo fights project-config and devstack-gate to setup multinode basics for octavia | 09:16 | |
ajo | round 1... | 09:16 |
ajo | nmagnezi, ^ :P :) | 09:16 |
*** itsuugo has quit IRC | 09:18 | |
*** itsuugo has joined #openstack-lbaas | 09:19 | |
*** Alex_Stef has quit IRC | 09:19 | |
nmagnezi | ajo, may the force be with you :-) | 09:26 |
nmagnezi | sbalukoff, hey, re https://review.openstack.org/#/c/363713 | 09:27 |
nmagnezi | sbalukoff, i wanted to add a tiny tempest scenario, but i see there are no scenarios in the octavia code base | 09:27 |
nmagnezi | sbalukoff, is there a plan to add such scenarios? | 09:27 |
nmagnezi | sbalukoff, i noticed that patch: https://review.openstack.org/#/c/172199/ which still WIP by ptoohill | 09:28 |
*** eezhova has joined #openstack-lbaas | 09:39 | |
openstackgerrit | Nir Magnezi proposed openstack/octavia: _extract_amp_image_id_by_tag should return last created image https://review.openstack.org/363713 | 09:43 |
*** Satya_ has joined #openstack-lbaas | 09:57 | |
*** itsuugo has quit IRC | 10:00 | |
*** itsuugo has joined #openstack-lbaas | 10:01 | |
*** Satya_ has quit IRC | 10:09 | |
*** gongysh has joined #openstack-lbaas | 10:10 | |
*** itsuugo has quit IRC | 10:11 | |
*** itsuugo has joined #openstack-lbaas | 10:12 | |
*** itsuugo has quit IRC | 10:25 | |
*** itsuugo has joined #openstack-lbaas | 10:27 | |
*** itsuugo has quit IRC | 10:32 | |
*** itsuugo has joined #openstack-lbaas | 10:32 | |
*** itsuugo has quit IRC | 10:39 | |
*** itsuugo has joined #openstack-lbaas | 10:41 | |
*** itsuugo has quit IRC | 10:46 | |
*** itsuugo has joined #openstack-lbaas | 10:46 | |
*** anilvenkata has joined #openstack-lbaas | 10:56 | |
*** yamamoto_ has quit IRC | 11:01 | |
*** Alex_Stef has joined #openstack-lbaas | 11:06 | |
*** itsuugo has quit IRC | 11:23 | |
*** itsuugo has joined #openstack-lbaas | 11:25 | |
*** atul has quit IRC | 11:35 | |
*** yamamoto has joined #openstack-lbaas | 11:38 | |
*** yamamoto has quit IRC | 11:41 | |
*** itsuugo has quit IRC | 11:43 | |
*** itsuugo has joined #openstack-lbaas | 11:44 | |
*** yamamoto has joined #openstack-lbaas | 11:47 | |
*** yamamoto has quit IRC | 11:51 | |
*** yamamoto has joined #openstack-lbaas | 11:52 | |
*** yamamoto has quit IRC | 11:57 | |
*** anilvenkata has quit IRC | 11:58 | |
*** gongysh has quit IRC | 12:12 | |
*** ducttape_ has joined #openstack-lbaas | 12:14 | |
*** itsuugo has quit IRC | 12:23 | |
*** itsuugo has joined #openstack-lbaas | 12:23 | |
*** itsuugo has quit IRC | 12:28 | |
*** itsuugo has joined #openstack-lbaas | 12:29 | |
*** yamamoto has joined #openstack-lbaas | 12:30 | |
*** yamamoto has quit IRC | 12:31 | |
*** yamamoto has joined #openstack-lbaas | 12:31 | |
*** yamamoto has quit IRC | 12:37 | |
*** yamamoto has joined #openstack-lbaas | 12:38 | |
*** ducttape_ has quit IRC | 12:42 | |
*** itsuugo has quit IRC | 12:45 | |
*** itsuugo has joined #openstack-lbaas | 12:46 | |
*** yamamoto has quit IRC | 12:56 | |
*** reedip has quit IRC | 12:56 | |
*** yamamoto has joined #openstack-lbaas | 12:58 | |
*** itsuugo has quit IRC | 13:01 | |
*** itsuugo has joined #openstack-lbaas | 13:02 | |
*** yamamoto has quit IRC | 13:07 | |
*** yamamoto has joined #openstack-lbaas | 13:07 | |
*** reedip has joined #openstack-lbaas | 13:08 | |
*** matt-borland has joined #openstack-lbaas | 13:23 | |
*** BjoernT has joined #openstack-lbaas | 13:36 | |
*** yamamoto has quit IRC | 13:38 | |
*** ducttape_ has joined #openstack-lbaas | 13:42 | |
*** itsuugo has quit IRC | 13:46 | |
*** kobis has quit IRC | 13:47 | |
*** ducttape_ has quit IRC | 13:47 | |
*** itsuugo has joined #openstack-lbaas | 13:48 | |
*** Frito has joined #openstack-lbaas | 13:51 | |
*** gongysh has joined #openstack-lbaas | 13:51 | |
*** Frito has quit IRC | 13:54 | |
*** gongysh has quit IRC | 13:55 | |
*** yamamoto has joined #openstack-lbaas | 13:59 | |
*** pglass has joined #openstack-lbaas | 13:59 | |
*** ducttape_ has joined #openstack-lbaas | 14:05 | |
*** ducttape_ has quit IRC | 14:09 | |
*** nmagnezi has quit IRC | 14:16 | |
*** ducttape_ has joined #openstack-lbaas | 14:18 | |
*** ducttape_ has quit IRC | 14:23 | |
*** yamamoto has quit IRC | 14:28 | |
*** yamamoto has joined #openstack-lbaas | 14:29 | |
*** yamamoto has quit IRC | 14:30 | |
*** yamamoto has joined #openstack-lbaas | 14:34 | |
*** yamamoto has quit IRC | 14:35 | |
*** eezhova has quit IRC | 14:45 | |
*** eezhova has joined #openstack-lbaas | 14:46 | |
*** ducttape_ has joined #openstack-lbaas | 14:47 | |
*** pglass has quit IRC | 14:50 | |
*** diogogmt has joined #openstack-lbaas | 15:03 | |
*** pglass has joined #openstack-lbaas | 15:08 | |
*** brad_behle has joined #openstack-lbaas | 15:15 | |
*** csomerville has joined #openstack-lbaas | 15:16 | |
*** cody-somerville has quit IRC | 15:19 | |
*** pcaruana has quit IRC | 15:19 | |
*** diogogmt has quit IRC | 15:20 | |
*** yamamoto has joined #openstack-lbaas | 15:35 | |
*** woodster_ has joined #openstack-lbaas | 15:36 | |
*** ducttape_ has quit IRC | 15:37 | |
*** yamamoto has quit IRC | 15:42 | |
*** itsuugo has quit IRC | 15:48 | |
*** ducttape_ has joined #openstack-lbaas | 15:49 | |
*** itsuugo has joined #openstack-lbaas | 15:49 | |
openstackgerrit | Michael Johnson proposed openstack/octavia: Fixes a revert failure in AllocateVIP task https://review.openstack.org/370968 | 15:53 |
johnsom | That one should be an easy review | 15:54 |
*** nmagnezi has joined #openstack-lbaas | 16:03 | |
*** rcernin has quit IRC | 16:26 | |
*** itsuugo has quit IRC | 16:32 | |
*** itsuugo has joined #openstack-lbaas | 16:34 | |
*** yamamoto has joined #openstack-lbaas | 16:44 | |
*** yamamoto has quit IRC | 16:44 | |
diltram | johnsom: +1 | 16:51 |
*** Alex_Stef has quit IRC | 16:53 | |
johnsom | Ok, my self concious is getting me. I'm adding a test and scanning for other unhandled exceptions in the revert calls. Probably another patchset coming | 16:56 |
*** _ducttape_ has joined #openstack-lbaas | 17:00 | |
*** nmagnezi has quit IRC | 17:03 | |
*** ducttape_ has quit IRC | 17:03 | |
*** eezhova has quit IRC | 17:20 | |
*** _ducttape_ has quit IRC | 17:22 | |
*** ducttape_ has joined #openstack-lbaas | 17:23 | |
openstackgerrit | Merged openstack/neutron-lbaas: Allow creating loadbalancer with network_id https://review.openstack.org/363302 | 17:23 |
openstackgerrit | Michael Johnson proposed openstack/octavia: Fixes a revert failure in AllocateVIP task https://review.openstack.org/370968 | 17:24 |
johnsom | Ok, got the other two. | 17:24 |
openstackgerrit | Merged openstack/octavia: Fix admin_state_up for loadbalancer and listener https://review.openstack.org/364707 | 17:27 |
*** ducttape_ has quit IRC | 17:34 | |
*** SumitNaiksatam has joined #openstack-lbaas | 17:38 | |
openstackgerrit | Merged openstack/octavia: Add tox 'specs' test to pep8 environment https://review.openstack.org/367710 | 17:38 |
openstackgerrit | Merged openstack/octavia: Set haproxy global connection limit https://review.openstack.org/369089 | 17:38 |
openstackgerrit | Merged openstack/octavia: Set fullconn parameter in haproxy backends https://review.openstack.org/369102 | 17:38 |
openstackgerrit | Merged openstack/octavia: Adding a unit test for graph create failure case https://review.openstack.org/369627 | 17:38 |
*** ihrachys has quit IRC | 17:43 | |
*** yamamoto has joined #openstack-lbaas | 17:45 | |
*** bana_k has joined #openstack-lbaas | 17:50 | |
*** yamamoto has quit IRC | 17:52 | |
*** ducttape_ has joined #openstack-lbaas | 17:55 | |
*** brad_behle has quit IRC | 18:01 | |
*** brad_behle has joined #openstack-lbaas | 18:01 | |
*** bana_k has quit IRC | 18:08 | |
sbalukoff | johnsom: I know you had a look at this yesterday, but I liked Dustin's suggestion to add an example of how to do A/B testing with a cookie so much that I went ahead and added it. If you could have another gander at that section, I'd appreciate it: https://review.openstack.org/#/c/360934/ | 18:12 |
johnsom | Sure, NP | 18:13 |
johnsom | sbalukoff Do you have thoughts on the comments here: https://review.openstack.org/#/c/367039/ | 18:14 |
sbalukoff | Let me look... | 18:14 |
johnsom | I am thinking it is a backward compat issue/guessing at the operator setup | 18:14 |
johnsom | I'm heating up some food, back in a few... | 18:15 |
sbalukoff | johnsom: Yeah, was going to say: the whole reason not to enforce the use of this setting by default was to avoid breaking backward compatibility. | 18:16 |
sbalukoff | Especially this late in a cycle. | 18:17 |
sbalukoff | I'll add those comments to that patch set. | 18:17 |
johnsom | Thanks | 18:19 |
openstackgerrit | Merged openstack/neutron-lbaas: haproxy ns driver: set global haproxy maxconn https://review.openstack.org/369088 | 18:21 |
*** bana_k has joined #openstack-lbaas | 18:21 | |
sbalukoff | It's great to see lots of stuff merging today. | 18:22 |
*** pcaruana has joined #openstack-lbaas | 18:22 | |
sbalukoff | johnsom: Does the latest patch set of this have the test you wanted to add? https://review.openstack.org/#/c/370968/ | 18:24 |
johnsom | Yes. There will be follow on patches, but that one is good for review | 18:26 |
sbalukoff | Ok, I'll check it out. | 18:26 |
dlundquist | The haproxy global maxconn fixes have merged for both Octavia and the Namespace driver, so bug 1544861 is fixed. | 18:41 |
openstack | bug 1544861 in neutron "LBaaS: connection limit does not work with HA Proxy" [Medium,In progress] https://launchpad.net/bugs/1544861 - Assigned to Dustin Lundquist (dlundquist) | 18:41 |
*** itsuugo has quit IRC | 18:42 | |
dlundquist | Oh, I can change status, just not the assignee | 18:43 |
*** itsuugo has joined #openstack-lbaas | 18:44 | |
johnsom | dlundquist We use fix released now as the release process is automated and doesn't flip this bit. I updated them for you. | 18:57 |
dlundquist | yes, but these both marked partial bug, so I didn't think automation worked correctly. And I just added also effects Octavia the other day. | 18:58 |
openstackgerrit | Michael Johnson proposed openstack/octavia: Option to restrict amp glance image owner https://review.openstack.org/367039 | 19:03 |
johnsom | ^^^ fixed the merge conflict | 19:03 |
*** eezhova has joined #openstack-lbaas | 19:05 | |
sbalukoff | johnsom: Got another merge conflict headed your way. I just put a +A on this: https://review.openstack.org/#/c/363713/ | 19:11 |
johnsom | Ugh | 19:12 |
sbalukoff | Sorry! | 19:13 |
johnsom | Probably should have had a relase note on that | 19:13 |
johnsom | Since it does change behavior | 19:14 |
sbalukoff | Well... for most operators I don't think it does. But I could be wrong. On that note though, any code change changes behavior (which is the whole point of the code change). :P | 19:15 |
johnsom | Well, in an upgrade scenario, as soon as you upload a new image it will start being used. Previously it wasn't until the tag was removed from the previous image they were using. | 19:16 |
johnsom | So, yeah, surprise the image is live! grin | 19:17 |
sbalukoff | Really? I thought the original code already sorted by created_at, and took the last image returned. Isn't that patch really just an optimization? (ie. sort on create_at:desc, grab a maximum of two images so we can warn if the Operator made a mistake, but use the first image.) | 19:18 |
johnsom | Nope, this is inverting the way it worked before | 19:18 |
sbalukoff | What's the difference if you sort on create_at and take the last image versus created_at:desc and take the first image? | 19:19 |
johnsom | That is one reason the issue with the image owner didn't immediately trigger with a user uploaded image, it was still pulling the oldest | 19:20 |
johnsom | Maybe he changed it as the patch went along | 19:21 |
sbalukoff | Oh, I see... from the bug report, when you tell the glance client to sort based on 'created_at', it doesn't actually do that unless you give it a direction. So... this patch is to work around glance client boneheadedness. | 19:22 |
johnsom | It's desc by default | 19:23 |
*** matt-borland has quit IRC | 19:23 | |
sbalukoff | Ok, I get it now. | 19:24 |
johnsom | Maybe I am confused with what came out of this stuff. I'm deep in updating a bunch of unhandled exceptions. | 19:24 |
sbalukoff | No worries. | 19:24 |
johnsom | It's going to be a large patch, but it is just wrapping a bunch of DB calls to make sure the revert doesn't stop | 19:24 |
sbalukoff | In any case, assuming this one finishes merging, it's a merge conflict that should be really easy to resolve for you, eh. | 19:25 |
johnsom | yeah | 19:25 |
johnsom | Don't forget me.... It will need another +2... grin | 19:25 |
sbalukoff | Yep, I'm hovering over the button... ;) | 19:26 |
sbalukoff | Feel free to poke at me here once you have the new patch set up. | 19:26 |
*** pcaruana has quit IRC | 19:32 | |
*** davidlenwell has quit IRC | 19:34 | |
openstackgerrit | Merged openstack/octavia: _extract_amp_image_id_by_tag should return last created image https://review.openstack.org/363713 | 19:41 |
openstackgerrit | Michael Johnson proposed openstack/octavia: Option to restrict amp glance image owner https://review.openstack.org/367039 | 19:46 |
*** davidlenwell has joined #openstack-lbaas | 19:47 | |
johnsom | sbalukoff ^^^^ | 19:47 |
sbalukoff | Thanks! | 19:48 |
*** rcernin has joined #openstack-lbaas | 19:53 | |
*** itsuugo has quit IRC | 19:53 | |
*** itsuugo has joined #openstack-lbaas | 19:55 | |
TrevorV | Hey sbalukoff | 19:56 |
TrevorV | Was it you or evgeny that worked on the octavia side of L7 Policy stuffs? | 19:56 |
sbalukoff | Uh-oh. | 19:56 |
sbalukoff | It was me. | 19:56 |
sbalukoff | Evgeny hasn't touched Octavia. | 19:56 |
TrevorV | Is there a specific or particular reason you left off "project_id" on the l7policy object? | 19:57 |
sbalukoff | L7 Policies are necessarily linked to Listeners, which have a project_id. | 19:57 |
sbalukoff | If you want to know which project an L7 policy is connected to, look at l7_policy.listener.project_id | 19:58 |
sbalukoff | (I thought it was more DRY) | 19:58 |
TrevorV | Alright, so that just differs from all the other objects. LB, Listener, Pool, and HM all have project_id | 19:59 |
TrevorV | I say "all" | 19:59 |
TrevorV | I just meant those others | 19:59 |
TrevorV | Idk about others beyond that | 19:59 |
TrevorV | sbalukoff, just to give a little background, its the only "odd" object for quotas. I can just do a filter by project_id on all the other quota'd objects because they all have those project_ids associated in their tables | 20:01 |
TrevorV | sbalukoff, are you "against" adding it as a column? | 20:01 |
sbalukoff | We're going to support quotas on L7Policies now? | 20:01 |
sbalukoff | I'm not sure why we'd do that... but if we are, I guess I'm not against it. | 20:02 |
TrevorV | They exist in Neutron LBaaS currently | 20:02 |
sbalukoff | Really, though, we have a lot of redundancy in our models around project_id... I think we really ought to limit it down to the objects that really need it. (like, I dunno... loadbalancer only). | 20:02 |
*** ducttape_ has quit IRC | 20:03 | |
TrevorV | That would complicate all the quota checks ha ha ha | 20:03 |
sbalukoff | It ought to be possible to augment the model with something that returns a project_id attribute (that just looks up the chain at the load balancer), and thus not store the same information in 8 different tables in the DB. :P | 20:03 |
TrevorV | I'm not against filtering all the way back to the LB to get the project ID, just wanted to be consistent, and was curious what your thoughts were originally | 20:04 |
TrevorV | If we want to make the change to only one table with project_id, we should actually file a bug in launchpad to have that done | 20:04 |
sbalukoff | I didn't think we needed to worry about quotas around L7 Policies. | 20:04 |
TrevorV | Because right now I have quite a bit of logic around them existing | 20:04 |
TrevorV | I mean I'll have to write new logic around them not existing if that's a thing | 20:05 |
sbalukoff | I agree, we ought to file a bug and have that done. | 20:05 |
TrevorV | johnsom, ? Scrollback and thoughts? | 20:05 |
sbalukoff | TrevorV: I don't think we need to change the logic, just the model. Look at how I'm generating the pool.listeners property, for example of what I mean. | 20:06 |
johnsom | sbalukoff Didn't you guys add a quota for L7? | 20:06 |
johnsom | in neutron? | 20:06 |
sbalukoff | johnsom: I don't think so. | 20:06 |
sbalukoff | johnsom: I'm less familiar with the neutron-lbaas code. | 20:06 |
sbalukoff | Since that was all Evgeny. | 20:06 |
sbalukoff | The N-LbaaS API doesn't presently support quotas for L7 stuff. | 20:08 |
johnsom | TrevorV do you remember where we saw that? | 20:10 |
TrevorV | I'm trying to find it right now johnsom | 20:11 |
TrevorV | I can't even find the right quotas documentation | 20:11 |
TrevorV | johnsom, its just in the neutron quota-show | 20:12 |
TrevorV | CLI in devstack | 20:12 |
TrevorV | just ran it | 20:12 |
sbalukoff | Looking in neutron-lbaas/neutron_lbaas/extensions, I'm seeing only the following have quota resources associated with them: | 20:12 |
johnsom | Yeah, it is | 20:12 |
sbalukoff | loadbalancerv2.py: cfg.IntOpt('quota_loadbalancer', | 20:12 |
sbalukoff | loadbalancerv2.py: cfg.IntOpt('quota_listener', | 20:12 |
sbalukoff | loadbalancerv2.py: cfg.IntOpt('quota_pool', | 20:12 |
sbalukoff | loadbalancerv2.py: cfg.IntOpt('quota_member', | 20:12 |
sbalukoff | loadbalancerv2.py: cfg.IntOpt('quota_healthmonitor', | 20:12 |
johnsom | So, did we somehow get a ghost in the client that isn't implemented in the API???? | 20:13 |
TrevorV | sbalukoff, that's amusing, because the CLI doesn't even show a member quota | 20:13 |
sbalukoff | johnsom: I suspect so. | 20:13 |
sbalukoff | Again, quota testing is pretty shit right now. | 20:13 |
sbalukoff | Having just had to hobble in a "fix" for the gate on this stuff last Friday. :P | 20:13 |
TrevorV | Alright, so I'm going to go ahead and say we should file a bug that changes the CLI to have member quota and not l7policy quota | 20:15 |
TrevorV | Sound right to you two? | 20:15 |
sbalukoff | Well... the question to me is: Should we be supporting quotas around L7 objects? | 20:16 |
sbalukoff | Technically evaluating L7 rules does lead to a small increase in CPU usage by haproxy. | 20:16 |
johnsom | neutron_lbaas/extensions/l7.py | 20:16 |
sbalukoff | It also seems silly to me that we would put quotas on members or pools as well. | 20:17 |
sbalukoff | Or health monitors. | 20:17 |
TrevorV | sbalukoff, well to be fair I'm only talking about what I found implemented by neutron lbaas, for feature parity | 20:17 |
johnsom | line 204 registers it as a quota | 20:17 |
sbalukoff | johnsom: When you follow that code back, you find that it actually doesn't/. | 20:17 |
sbalukoff | You need the config options named quota_ thing to do that. | 20:18 |
sbalukoff | And yes, following that code back is a spaghetti mess. | 20:18 |
johnsom | Well, I am suspecting this is how it is getting in the client output | 20:19 |
sbalukoff | Does setting those quota values actually work? Or save a value? | 20:20 |
TrevorV | sbalukoff, I admit, I didn't actually test quotas for neutronlbaas specifically, I just knew I had to get it working in octavia... | 20:20 |
johnsom | CLI doesn't have it in the quota update list | 20:21 |
TrevorV | that's amusing | 20:21 |
johnsom | So, TrevorV Can you do us a favor, file a neutron bug tagged with lbaas that L7policy is showing up as quota in the CLI quota-show. Then forget about L7 for the Octavia quotas.... | 20:22 |
TrevorV | Why does it show l7policy and NOT member in the "show", but shows member in the update...... | 20:22 |
johnsom | Because this stuff was slapped together poorly is my non-PC answer | 20:22 |
TrevorV | Okay, yeah. I'll file the bug. | 20:23 |
johnsom | Thank you sir | 20:23 |
sbalukoff | It is possible that you could set the quotas for l7policy stuff, I support-- the testing code for quotas specifically behaves differently than production code, thanks to a recent quotas patch (the one that broke us last Friday, actually). | 20:23 |
TrevorV | Then I'll need to update my quotas stuff to include member quotas though ha ha | 20:23 |
sbalukoff | And yes, what I just said is patently absurd. | 20:23 |
sbalukoff | Er... s/I support/I suppose/ | 20:24 |
sbalukoff | Have I mentioned how much I dislike the Neutron quotas code yet? | 20:24 |
johnsom | Make sure to review the work TrevorV is doing for Octavia. That is our way out.... | 20:25 |
sbalukoff | Good! | 20:25 |
sbalukoff | TrevorV: Are you using neutron quotas at all in your code? (pulling in *any* part of their code base?) | 20:26 |
TrevorV | sbalukoff, its probably actually bad practice from me, but I really really dislike copy-pasta code./ | 20:26 |
TrevorV | code.**** | 20:26 |
sbalukoff | So... you're not? Good man! | 20:26 |
TrevorV | I even typically re-type URLs that are already completed in my browser.... | 20:26 |
TrevorV | ha ha ha | 20:26 |
sbalukoff | Haha! | 20:26 |
TrevorV | No, I didn't copy anything, just the "look" its supposed to have | 20:27 |
*** ducttape_ has joined #openstack-lbaas | 20:27 | |
sbalukoff | Great! That's what I was hoping you'd say. | 20:27 |
TrevorV | I'll put it this way sbalukoff... I was one of the biggest proponents AGAINST neutron lbaas, mostly because I hate the code base... so its far more likely I'll write something myself :) | 20:28 |
sbalukoff | And this is why we love you, TrevorV/. | 20:28 |
TrevorV | by the way johnsom I'm also including the bit that the quota show doesn't include member in the details. | 20:28 |
TrevorV | So its a "neutron quota-show is broken" sort of bug | 20:29 |
johnsom | Yeah, probably a second bug? | 20:29 |
TrevorV | https://bugs.launchpad.net/neutron/+bug/1624097 | 20:32 |
openstack | Launchpad bug 1624097 in neutron "Neutron LBaaS CLI quota show includes l7policy and doesn't include member" [Undecided,New] | 20:32 |
TrevorV | Did I do it right? | 20:32 |
TrevorV | I mean, the change should probably be done in the same place, I figured it wouldn't hurt to be the same bug... | 20:33 |
* TrevorV has never reported a neutron lbaas bug before... shocker | 20:33 | |
johnsom | TrevorV Looks good enough to me | 20:34 |
TrevorV | sweet | 20:34 |
sbalukoff | Yep, looks good to me, too. | 20:34 |
*** eezhova has quit IRC | 20:39 | |
*** itsuugo has quit IRC | 20:47 | |
TrevorV | sbalukoff, one piece of good news... member in octavia has "project_id" on it, so the quota logic is still clean ;) | 20:48 |
*** itsuugo has joined #openstack-lbaas | 20:50 | |
sbalukoff | Heh! | 20:52 |
openstackgerrit | Michael Johnson proposed openstack/octavia: Fixes the revert methods with unhandled exceptions https://review.openstack.org/371154 | 20:56 |
johnsom | ^^^^ WIP | 20:56 |
*** catintheroof has joined #openstack-lbaas | 21:15 | |
sbalukoff | Nice! | 21:17 |
sbalukoff | Feel free to tell me I'm smoking crack on this one: https://bugs.launchpad.net/octavia/+bug/1624113 | 21:18 |
openstack | Launchpad bug 1624113 in octavia "RFE: project_id database field is redundant on many Octavia objects" [Undecided,New] | 21:18 |
johnsom | Hmm, yeah, there are arguments on both sides. If we truly have our foreign key strategy setup such that every object relates, I agree with this (health/stats is a valid exception to allow those table to be external). That said, having it on each table makes it easier for operator debug and potentially filtering results based on the project_id. | 21:29 |
johnsom | They can also speed up queries... | 21:30 |
diltram | using that project_id from load balancer object will speed up queries? | 21:33 |
diltram | how? | 21:33 |
diltram | johnsom: ^^ | 21:33 |
johnsom | No, the inverse | 21:33 |
diltram | I don't understand | 21:34 |
johnsom | Having them on each table can speed up queries | 21:34 |
diltram | aaaaa, ok yeah | 21:34 |
diltram | thats my point | 21:35 |
johnsom | Sorry, train of thought typing. Knee deep in fixing the database tasks reverts/unit tests | 21:35 |
diltram | removing that property will slow down all queries and will require joining tables | 21:36 |
diltram | so maybe we should add something what will remove way of overriding that property | 21:36 |
diltram | and when you're creating new object they will somehow use lb.project_id to populate own data | 21:37 |
diltram | and it will be unchangable | 21:37 |
sbalukoff | It comes down to whether storing that data redundantly (and the risk of associated problems with redundant fields getting out of sync) is better than keeping things DRY and making some database queries take longer, and making project_id related problems somewhat more difficult to troubleshoot for operators. | 21:40 |
diltram | sbalukoff: some queries? it looks like all queries except load balancer will take longer | 21:41 |
diltram | and also they cannot be out of sync | 21:42 |
johnsom | Redundant data also leads to IO/caching overhead that typically exceeds the join overhead on modern databases, but that is mostly an issue on very large DBs | 21:42 |
diltram | project_id is not changebale thing | 21:42 |
diltram | or I'm not seeing smth? | 21:42 |
sbalukoff | our API allows specifying the project_id on descendant objects on creation (which is probably meant to keep things simpler for our neutron-lbaas driver, but opens the opportunity for a bug to create a descendant object with a different project_id than its load balancer parent.) | 21:43 |
diltram | true | 21:44 |
sbalukoff | If we close that, that would probably eliminte a lot of potential for error. | 21:44 |
johnsom | Well, soon project_id should only come from keystone in our API | 21:45 |
diltram | that because of that I'm asking about maybe adding something to remove that feature :) | 21:45 |
sbalukoff | Yep, I would support removing that "feature". :) | 21:45 |
diltram | also does anyone worked with keystone domains? | 21:45 |
diltram | how they work? | 21:45 |
sbalukoff | No idea. | 21:46 |
johnsom | About the extent of my knowledge is they were added for ldap/active directory support (if I am remembering right) | 21:46 |
*** rm_work has quit IRC | 21:46 | |
*** mhayden has quit IRC | 21:46 | |
diltram | but it's also from our perspective | 21:46 |
diltram | there is no projects supporting domains right now, just keystone | 21:47 |
*** amit213 has quit IRC | 21:47 | |
diltram | you got my messages? | 21:50 |
diltram | I lost connection for a few sec | 21:50 |
diltram | so yeah, I would like to know what will be performance impact | 21:51 |
*** amit213 has joined #openstack-lbaas | 21:51 | |
*** rm_work has joined #openstack-lbaas | 21:51 | |
diltram | and even how will work just simplest query get all listeners belonging to project_id = 10 | 21:52 |
sbalukoff | Well, in any case, for now, I would be happy with the happy medium of just removing the ability to specify the project_id on descendant objects. (or rather, ignore it for now if it's specified, and instead, get the project_id from the ancestor load_balancer). I'll file another bug report with suggestions here. | 21:52 |
diltram | +1 | 21:53 |
diltram | going sleep, it's midnight for me :P | 21:55 |
diltram | cu guys | 21:55 |
*** csomerville has quit IRC | 21:58 | |
rm_work | ^^ that sounds good to me | 21:58 |
*** csomerville has joined #openstack-lbaas | 21:59 | |
*** mhayden has joined #openstack-lbaas | 22:00 | |
*** yamamoto has joined #openstack-lbaas | 22:02 | |
*** yamamoto has quit IRC | 22:02 | |
sbalukoff | Trying to submit bug report... launchpad timing out. :P | 22:09 |
*** yamamoto has joined #openstack-lbaas | 22:13 | |
*** itsuugo has quit IRC | 22:15 | |
*** itsuugo has joined #openstack-lbaas | 22:17 | |
*** pglass has quit IRC | 22:18 | |
*** yamamoto has quit IRC | 22:18 | |
*** yamamoto has joined #openstack-lbaas | 22:24 | |
*** yamamoto has quit IRC | 22:24 | |
*** yamamoto has joined #openstack-lbaas | 22:24 | |
*** yamamoto has quit IRC | 22:28 | |
sbalukoff | Ok, there's my bug report: https://bugs.launchpad.net/octavia/+bug/1624145 | 22:30 |
openstack | Launchpad bug 1624145 in neutron "Octavia should ignore project_id on API create commands (except load_balancer)" [Undecided,New] | 22:30 |
johnsom | That is against neutron, shouldn't it be octavia | 22:31 |
johnsom | ? | 22:31 |
openstackgerrit | Trevor Vardeman proposed openstack/octavia: Add quota support to Octavia https://review.openstack.org/360794 | 22:31 |
*** itsuugo has quit IRC | 22:34 | |
*** itsuugo has joined #openstack-lbaas | 22:37 | |
*** itsuugo has quit IRC | 22:41 | |
*** brad_behle has quit IRC | 22:42 | |
*** brad_behle has joined #openstack-lbaas | 22:42 | |
*** itsuugo has joined #openstack-lbaas | 22:42 | |
*** ducttape_ has quit IRC | 22:42 | |
sbalukoff | Sorry, tried editing it to list both Octavia and Neutron (dealing with timeouts). It should list both projects now. | 22:44 |
*** brad_behle has quit IRC | 22:46 | |
*** itsuugo has quit IRC | 23:08 | |
*** itsuugo has joined #openstack-lbaas | 23:09 | |
*** ducttape_ has joined #openstack-lbaas | 23:12 | |
*** catintheroof has quit IRC | 23:21 | |
*** itsuugo has quit IRC | 23:23 | |
*** itsuugo has joined #openstack-lbaas | 23:24 | |
*** ducttape_ has quit IRC | 23:29 | |
bana_k | getting this error DeprecatedConfig: Fatal call to deprecated config: Deprecated: Implicit loading of service providers from neutron_*.conf files is deprecated and will be removed in Ocata release | 23:29 |
bana_k | when neutron.con has fatal_deprecations = True | 23:30 |
bana_k | but out devstack still has this file | 23:30 |
sbalukoff | I recommend not setting fatal_depreciations in neutron.conf right now, and filing a bug for us to fix that if one hasn't been filed already. | 23:34 |
bana_k | sure sbalukoff | 23:34 |
*** itsuugo has quit IRC | 23:35 | |
bana_k | thanks | 23:36 |
*** itsuugo has joined #openstack-lbaas | 23:37 | |
*** armax has quit IRC | 23:37 | |
*** BjoernT has quit IRC | 23:44 | |
*** itsuugo has quit IRC | 23:49 | |
*** itsuugo has joined #openstack-lbaas | 23:51 | |
*** armax has joined #openstack-lbaas | 23:54 | |
*** yamamoto_ has joined #openstack-lbaas | 23:58 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!