| *** armax has quit IRC | 00:40 | |
| sorrison | johnsom: Would you be able to reconsider your -1 on https://review.opendev.org/#/c/711127/ (have replied to your comment) | 00:40 |
|---|---|---|
| johnsom | sorrison I can take a look | 00:40 |
| sorrison | thanks | 00:41 |
| sorrison | just trying to make it easier and more standard to deploy it and bring inline with other openstack projects | 00:41 |
| johnsom | sorrison What other projects? A quick search shows only Ironic and keystone have it | 00:41 |
| sorrison | All projects use it | 00:42 |
| sorrison | most projects you can configure it using paste | 00:42 |
| johnsom | I can guarantee it's not all projects. | 00:42 |
| johnsom | http://codesearch.openstack.org/?q=from%20oslo_middleware%20import%20healthcheck&i=nope&files=&repos= | 00:42 |
| johnsom | I only see two | 00:43 |
| sorrison | Well we use 19 projects in production and they all have it | 00:43 |
| *** vishalmanchanda has joined #openstack-lbaas | 00:43 | |
| sorrison | some have it coded in but the majority it's configured with paste.ini | 00:43 |
| johnsom | The reason I flagged this is I have been in the meetings talking about starting that library over from scratch. | 00:44 |
| johnsom | Because right now, it doesn't provide much more than the API discovery documents on / provide for all of the APIs. | 00:44 |
| sorrison | yeah I see that as a separate issue | 00:44 |
| sorrison | I'm after consistency | 00:45 |
| johnsom | The sad state of things is everyone was supposed to drop paste as well. sigh. | 00:45 |
| johnsom | Ok, let me look into this a bit more. | 00:45 |
| johnsom | Yeah, the / health check is going to be consistent too. It's required. | 00:45 |
| sorrison | ti could be enabled with a config option which defaults to false if you like | 00:45 |
| sorrison | hitting / is also inconsistent between projects | 00:47 |
| sorrison | That's the reason a lot of operators use /healthcheck | 00:47 |
| johnsom | It can't be. It is a requirement for OpenStack API projects. It's heavily used by OSC and SDK | 00:47 |
| sorrison | ok maybe that's changed, last I looked some required you be authenticated to hit / | 00:48 |
| sorrison | but it's been a while | 00:48 |
| johnsom | Yeah, we got bugged about it in Pike because we didn't have a OpenStack "standard" format discovery document. We had to fix it for OSC and SDK. | 00:49 |
| johnsom | Like I said on the patch, I'm on the fence a bit. I will do some digging and re-evaluate. You might also bug the others to chime in too. | 00:49 |
| sorrison | ok thanks, I wasn't aware the / was mandated so I don't really care about it that much. Although oslo healthcheck seems to be the obvious thing to use when determining if a service is ok for a frontend LB to use | 00:53 |
| sorrison | And if in the future this is going to be more indicative of the health then shouldn't octavia support it> | 00:54 |
| johnsom | Yeah, the problem with that library (other than it's lightly maintained) is that it doesn't actually check the API app, code, or DB. So it gives false positives | 00:54 |
| sorrison | ok, fair enough | 00:54 |
| johnsom | Yeah, it was a total replace with a new library, not an extension of this current code | 00:54 |
| johnsom | Frankly, I might be more comfortable adding it to our code without the library. lol That way at least it is a decent test. | 00:58 |
| johnsom | sorrison Do you use the "disable by file" part of that? | 01:00 |
| johnsom | Or is it really just for API LBs to check if the API instance is healthy? | 01:00 |
| sorrison | We don't use the disable by file think, just used by our F5 which all our openstack services are behind | 01:02 |
| johnsom | Yeah, ok. That is what I have heard from others as well. | 01:02 |
| sorrison | We just have a common template for each service and it would mean octavia would need to be different | 01:02 |
| johnsom | So, there might be more value in having it be a really test. | 01:02 |
| sorrison | (horizon is also different so it's not a massive deal) | 01:02 |
| *** spatel has joined #openstack-lbaas | 01:28 | |
| *** spatel has quit IRC | 01:32 | |
| rm_work | yeah i feel like it wouldn't be too hard to make that endpoint exist within octavia but actually *DO* something | 01:37 |
| rm_work | to johnsom's point. | 01:37 |
| rm_work | or even rewrite /healthcheck to / | 01:38 |
| rm_work | could do that in your wsgi deployment? | 01:38 |
| rm_work | in uwsgi that'd be like: | 01:39 |
| rm_work | route-uri = ^/healthcheck$ rewrite:/ | 01:39 |
| *** armax has joined #openstack-lbaas | 01:40 | |
| *** yamamoto has quit IRC | 01:41 | |
| *** yamamoto has joined #openstack-lbaas | 01:42 | |
| *** yamamoto has quit IRC | 01:49 | |
| *** yamamoto has joined #openstack-lbaas | 02:04 | |
| *** yamamoto has quit IRC | 02:05 | |
| *** psachin has joined #openstack-lbaas | 02:43 | |
| *** spatel has joined #openstack-lbaas | 03:02 | |
| *** spatel has quit IRC | 03:13 | |
| johnsom | I added some additional comments to my review. There is a "plugin" model that we may be able to use to make the Oslo code more useful. It really just comes down to whether we are ok with that library or not. | 03:22 |
| *** psachin has quit IRC | 03:30 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Support hacking 2.0.0 https://review.opendev.org/699302 | 03:53 |
| johnsom | rm_work Can you re-+2 https://review.opendev.org/#/c/699302/ ? I had to add a new line to another file to clear a pep8. | 03:53 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Use sqlalchemy isnot(None) https://review.opendev.org/710144 | 03:57 |
| rm_work | done :) | 05:56 |
| johnsom | I think I have a working plugin for that healthcheck library to check the Octavia DB connection. I just need to figure out how to get the functional test to use the healthcheck library so I can verify. | 06:52 |
| johnsom | However, calling it a night. | 06:53 |
| *** rpittau|afk is now known as rpittau | 07:24 | |
| *** yamamoto has joined #openstack-lbaas | 07:31 | |
| openstackgerrit | Merged openstack/octavia master: Support hacking 2.0.0 https://review.opendev.org/699302 | 07:47 |
| *** maciejjozefczyk has joined #openstack-lbaas | 08:08 | |
| *** gcheresh has joined #openstack-lbaas | 08:11 | |
| *** tesseract has joined #openstack-lbaas | 08:12 | |
| openstackgerrit | Merged openstack/octavia master: Use sqlalchemy isnot(None) https://review.opendev.org/710144 | 08:25 |
| *** rcernin has quit IRC | 08:27 | |
| *** gcheresh has quit IRC | 08:28 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Jobboard based controller https://review.opendev.org/647406 | 08:29 |
| *** tkajinam has quit IRC | 08:31 | |
| *** gcheresh has joined #openstack-lbaas | 09:11 | |
| *** yamamoto has quit IRC | 09:49 | |
| *** gcheresh has quit IRC | 09:55 | |
| *** yamamoto has joined #openstack-lbaas | 10:05 | |
| openstackgerrit | Ann Taraday proposed openstack/octavia master: Jobboard based controller https://review.opendev.org/647406 | 10:33 |
| *** yamamoto has quit IRC | 10:40 | |
| *** nicolasbock has joined #openstack-lbaas | 11:25 | |
| *** rpittau is now known as rpittau|bbl | 11:28 | |
| *** nicolasbock has quit IRC | 12:07 | |
| *** nicolasbock has joined #openstack-lbaas | 12:07 | |
| *** yamamoto has joined #openstack-lbaas | 12:35 | |
| *** vishalmanchanda has quit IRC | 13:03 | |
| *** yamamoto has quit IRC | 13:04 | |
| *** rpittau|bbl is now known as rpittau | 13:04 | |
| *** yamamoto has joined #openstack-lbaas | 13:14 | |
| *** yamamoto has quit IRC | 13:23 | |
| *** TrevorV has joined #openstack-lbaas | 13:35 | |
| *** rcernin has joined #openstack-lbaas | 13:41 | |
| *** yamamoto has joined #openstack-lbaas | 13:41 | |
| *** yamamoto has quit IRC | 14:33 | |
| *** yamamoto has joined #openstack-lbaas | 15:15 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Fix listener update with SNI certificates https://review.opendev.org/712790 | 15:27 |
| *** yamamoto has quit IRC | 15:28 | |
| *** yamamoto has joined #openstack-lbaas | 16:04 | |
| johnsom | FYI, there is a new cirros out: cirros 0.5.1 proposed to be the default in devstack. If someone has the cycles we should probably test the webserver | 16:07 |
| *** yamamoto has quit IRC | 16:07 | |
| johnsom | devstack proposal: https://review.opendev.org/#/c/711492/ | 16:08 |
| johnsom | Downside is it eats more RAM.... sigh | 16:09 |
| cgoncalves | 128MB -> 129MB RAM? that's odd | 16:11 |
| gthiemonge | 64->128 and 128->129 | 16:12 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Refactor the failover flows https://review.opendev.org/705317 | 16:13 |
| johnsom | Yeah, not sure what the alt image thing at 129 is about, but the main one is 64->128 | 16:14 |
| johnsom | FYI, the pep8 issue on master should be fixed with last nights merges, so rebase if you need. | 16:16 |
| johnsom | Cheers to haleyb for having that ready to go | 16:17 |
| openstackgerrit | Michael Johnson proposed openstack/python-octaviaclient master: Remove the dependency on the "mock" package https://review.opendev.org/713016 | 16:28 |
| johnsom | I might as well finish the job so all our repos are the same for ussuri | 16:29 |
| *** openstackgerrit has quit IRC | 16:31 | |
| *** yamamoto has joined #openstack-lbaas | 16:42 | |
| *** rpittau is now known as rpittau|afk | 16:58 | |
| *** maciejjozefczyk has quit IRC | 17:03 | |
| *** tesseract has quit IRC | 17:07 | |
| cgoncalves | the SNI db model has a position field but I don't see where it is being used. the REST API does not mention it nor does the provider driver API. am I missing something? | 17:15 |
| *** openstackgerrit has joined #openstack-lbaas | 17:16 | |
| openstackgerrit | Michael Johnson proposed openstack/python-octaviaclient master: Align python-octaviaclient to octavia coding style https://review.opendev.org/713032 | 17:16 |
| johnsom | Hmm, yeah, so position was in neutron-lbaas at the request of a vendor, but it goes against the standards, so it may have been dropped along the way. | 17:17 |
| cgoncalves | thanks. that helped | 17:18 |
| *** maciejjozefczyk has joined #openstack-lbaas | 17:21 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia-lib master: Remove the dependency on the "mock" package https://review.opendev.org/713039 | 17:33 |
| *** TrevorV has quit IRC | 17:34 | |
| openstackgerrit | Michael Johnson proposed openstack/octavia-lib master: Remove the dependency on the "mock" package https://review.opendev.org/713039 | 17:34 |
| johnsom | forgot to remove it from lower. | 17:35 |
| johnsom | Ok, that should update all of our repos for using the built in mock. | 17:35 |
| johnsom | Some easy reviews there, hint, hint | 17:36 |
| cgoncalves | all references to -> all references from | 17:40 |
| cgoncalves | :P | 17:40 |
| johnsom | ha, well, we didn't catch that in the master patch, so my copy/paste of the commit message propagated the issue. | 17:46 |
| *** maciejjozefczyk_ has joined #openstack-lbaas | 18:16 | |
| *** maciejjozefczyk has quit IRC | 18:16 | |
| *** yamamoto has quit IRC | 18:33 | |
| haleyb|away | johnsom: that's one way to get my patches merged :) | 19:01 |
| *** gcheresh has joined #openstack-lbaas | 19:16 | |
| *** gcheresh has quit IRC | 20:00 | |
| *** gcheresh has joined #openstack-lbaas | 20:55 | |
| *** maciejjozefczyk_ has quit IRC | 20:59 | |
| *** gcheresh has quit IRC | 21:04 | |
| johnsom | Sigh: [ERROR] /opt/stack/devstack/lib/glance:364 g-api did not start | 21:24 |
| *** rcernin has quit IRC | 22:58 | |
| *** rcernin has joined #openstack-lbaas | 22:59 | |
| *** numans has quit IRC | 23:04 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Fix listener update with SNI certificates https://review.opendev.org/712790 | 23:10 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!