| *** sapd has quit IRC | 00:19 | |
| *** mburrows has quit IRC | 00:45 | |
| *** mburrows has joined #openstack-lbaas | 00:48 | |
| *** mburrows has quit IRC | 00:50 | |
| *** mburrows has joined #openstack-lbaas | 00:51 | |
| *** dayou has quit IRC | 01:06 | |
| *** dayou has joined #openstack-lbaas | 01:07 | |
| *** jaff_cheng has joined #openstack-lbaas | 01:56 | |
| *** sapd has joined #openstack-lbaas | 01:58 | |
| *** jaff_cheng has quit IRC | 02:16 | |
| *** jaff_cheng has joined #openstack-lbaas | 02:16 | |
| *** sapd has quit IRC | 02:55 | |
| *** sapd has joined #openstack-lbaas | 02:55 | |
| openstackgerrit | wangqi proposed openstack/neutron-lbaas-dashboard master: Update links in README https://review.openstack.org/551829 | 02:59 |
|---|---|---|
| *** sapd1 has joined #openstack-lbaas | 03:15 | |
| *** jaff_cheng has quit IRC | 04:38 | |
| *** jaff_cheng has joined #openstack-lbaas | 04:38 | |
| *** jaff_cheng has quit IRC | 04:49 | |
| *** jaff_cheng has joined #openstack-lbaas | 04:49 | |
| *** jaff_cheng has quit IRC | 05:51 | |
| *** jaff_cheng has joined #openstack-lbaas | 05:52 | |
| *** sapd1 has quit IRC | 05:52 | |
| *** kobis has joined #openstack-lbaas | 05:54 | |
| *** kobis has quit IRC | 05:56 | |
| *** annp has joined #openstack-lbaas | 06:01 | |
| openstackgerrit | weizj proposed openstack/octavia master: The MarkL7PolicyActiveInDB is not necessary when l7policy deleted https://review.openstack.org/551846 | 06:06 |
| *** threestrands_ has joined #openstack-lbaas | 06:26 | |
| *** amotoki_ has joined #openstack-lbaas | 06:27 | |
| *** kobis has joined #openstack-lbaas | 06:28 | |
| *** threestrands has quit IRC | 06:34 | |
| *** amotoki has quit IRC | 06:34 | |
| *** kobis has quit IRC | 06:51 | |
| *** annp has quit IRC | 06:57 | |
| *** annp has joined #openstack-lbaas | 06:57 | |
| *** kobis has joined #openstack-lbaas | 06:58 | |
| *** rcernin has quit IRC | 07:09 | |
| *** threestrands_ has quit IRC | 07:13 | |
| *** pcaruana has joined #openstack-lbaas | 07:56 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 07:59 | |
| *** b_bezak has joined #openstack-lbaas | 08:00 | |
| *** b_bezak has quit IRC | 08:01 | |
| *** tesseract has joined #openstack-lbaas | 08:10 | |
| *** ispp has joined #openstack-lbaas | 08:27 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Rename q- to neutron- services https://review.openstack.org/544281 | 09:08 |
| *** dmellado has joined #openstack-lbaas | 09:29 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Rename q- to neutron- services https://review.openstack.org/544281 | 09:48 |
| *** oanson has quit IRC | 09:51 | |
| *** irenab has quit IRC | 09:51 | |
| *** irenab has joined #openstack-lbaas | 09:56 | |
| *** oanson has joined #openstack-lbaas | 09:57 | |
| *** yamamoto has quit IRC | 10:02 | |
| *** annp has quit IRC | 10:05 | |
| rm_work | cgoncalves: AHA! So THAT'S what made our gates all wonky! It was youuuuuuuu~ | 10:11 |
| cgoncalves | rm_work: soRRRRRRy | 10:11 |
| rm_work | ah and i was actually about to write the patch to move that lxd test to experimental also ;P | 10:11 |
| rm_work | glad i read through the channel first | 10:12 |
| dmellado | lol | 10:12 |
| rm_work | ok need nir or someone to merge | 10:14 |
| cgoncalves | https://review.openstack.org/#/c/543034/20/zuul.d/jobs.yaml@142 | 10:14 |
| rm_work | maybe xgerman_ is awake about now | 10:14 |
| cgoncalves | rm_work: ^ seems familiar to you? :D | 10:14 |
| cgoncalves | rm_work: Nir is out this week | 10:14 |
| rm_work | ah | 10:14 |
| rm_work | i didn't actually finish reviewing that patch yet | 10:15 |
| openstackgerrit | Hengqing Hu proposed openstack/octavia-dashboard master: Add l7 support https://review.openstack.org/551947 | 10:15 |
| rm_work | ah he is removing barbican :( | 10:15 |
| rm_work | I was going to finally add some tests for that :( | 10:15 |
| rm_work | oh nm he just moved them out of the base | 10:15 |
| rm_work | cgoncalves: are you just getting on or just finishing up | 10:16 |
| rm_work | i forget | 10:16 |
| cgoncalves | rm_work: reviewing that tempest patch? I haven't reviewed it yet, only somehow managed to spot that concurrenty flag there | 10:18 |
| rm_work | yeah i hadn't gotten that far yet on the tempest patch | 10:18 |
| cgoncalves | but I've just left a comment about it before I forgot | 10:18 |
| rm_work | only got through that big file | 10:18 |
| rm_work | it MIGHT be ok in octavia | 10:18 |
| rm_work | the concurrency model for us is less shitty than the neutron one | 10:18 |
| rm_work | like, significantly | 10:18 |
| cgoncalves | yeah | 10:21 |
| rm_work | hmm when does johnsom get back again | 10:22 |
| cgoncalves | the grenade patch is in a good shape now for reviews https://review.openstack.org/#/c/549654/ | 10:22 |
| rm_work | i have about 25 reviews for him to do | 10:22 |
| cgoncalves | dunno. he might have gotten stuck in dublin due to "adverse weather" ;) | 10:23 |
| rm_work | cgoncalves: ah cool, is it passing? I took a look yesterday but realized i have no idea how grenade actually works, and most of that patch is just being like "turn on grenade, set some grenade options, throw grenade" :P | 10:23 |
| *** salmankhan has joined #openstack-lbaas | 10:23 | |
| cgoncalves | rm_work: that's grenade in a nutshell :D | 10:23 |
| rm_work | what does it DO though? | 10:23 |
| rm_work | install on n-1 cloudver and then upgrade? | 10:23 |
| cgoncalves | it's passing, yes. I had to work on another patch (check depends-on) | 10:23 |
| cgoncalves | https://docs.openstack.org/grenade/latest/readme.html#basic-flow | 10:24 |
| rm_work | yep this looks like what i figured it would do | 10:25 |
| rm_work | but it actually has to "shut down all services"? :/ I thought we were specifically trying to prove zero-downtime | 10:25 |
| cgoncalves | so I still need to investigate how we can prove zero-downtime. all grenade plugins I've seen do not have such thing | 10:26 |
| cgoncalves | for now we prove zero dataplane downtime | 10:26 |
| rm_work | k | 10:30 |
| rm_work | yeah step 1, sure | 10:30 |
| rm_work | cgoncalves: -1 on grenade with comments (I do see it is still marked WIP so maybe you were still getting to some of those) | 10:34 |
| cgoncalves | rm_work: thanks! I left some old bits behind on purpose while patch receives first round of reviews | 10:35 |
| rm_work | cgoncalves: err do you remember the version on pyopenssl/cryptography we need to set in global requirements? | 10:38 |
| rm_work | you are also correct that we really SHOULD push global-requirements to update that stuff | 10:38 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Reduce runtime for l7rule too_many_rules test https://review.openstack.org/548991 | 10:39 |
| cgoncalves | rm_work: cryptography >= 1.9 and pyopenssl >= 17.1.0 | 10:42 |
| *** jaff_cheng has quit IRC | 10:53 | |
| *** yamamoto has joined #openstack-lbaas | 11:03 | |
| *** gcheresh has joined #openstack-lbaas | 11:09 | |
| *** yamamoto has quit IRC | 11:11 | |
| *** yamamoto has joined #openstack-lbaas | 11:14 | |
| *** gcheresh has quit IRC | 11:14 | |
| cgoncalves | uqq | 11:30 |
| cgoncalves | :q | 11:30 |
| cgoncalves | oops! | 11:30 |
| *** fnaval has quit IRC | 11:33 | |
| cgoncalves | rm_work: re: running golang httpd in member (cirros img). what about python SimpleHTTPServer instead? otherwise we'd need to install golang (either in gate host (+ copy to member) or member), build, run it in screen | 11:36 |
| *** salmankhan has quit IRC | 12:12 | |
| *** salmankhan has joined #openstack-lbaas | 12:16 | |
| *** yamamoto has quit IRC | 12:26 | |
| *** atoth has joined #openstack-lbaas | 12:26 | |
| *** fnaval has joined #openstack-lbaas | 13:17 | |
| *** fnaval has quit IRC | 13:20 | |
| *** yamamoto has joined #openstack-lbaas | 13:23 | |
| *** yamamoto has quit IRC | 13:28 | |
| *** andreykurilin has joined #openstack-lbaas | 13:34 | |
| andreykurilin | hi folks! can anyone help me with launching octavia in gates? | 13:34 |
| *** fnaval has joined #openstack-lbaas | 13:45 | |
| *** fnaval has quit IRC | 13:46 | |
| *** fnaval has joined #openstack-lbaas | 13:50 | |
| *** yamamoto has joined #openstack-lbaas | 13:53 | |
| *** beagles is now known as beagles|brb | 13:54 | |
| rm_work | cgoncalves: it's a tiny static-compiled binary | 13:55 |
| rm_work | we already use it in our testing, it's checked in | 13:55 |
| *** salmankhan has quit IRC | 13:55 | |
| rm_work | https://github.com/openstack/octavia-tempest-plugin/tree/master/octavia_tempest_plugin/contrib/httpd | 13:56 |
| rm_work | and we were discussing adding the binary to the octavia tree as well, since the code is there also: https://github.com/openstack/octavia/tree/cebfd319aef9e82b23f4af4a2f676d4cff4941e3/octavia/tests/contrib | 13:57 |
| rm_work | cgoncalves: ^^ so feel free to copy httpd.bin from our tempest repo there into our main repo's contrib (next to the source for it) | 13:58 |
| cgoncalves | rm_work: I noticed it being compiled in tests. I don't understand the reason for that. why not using python simplehttpserver? | 13:58 |
| *** yamamoto has quit IRC | 13:58 | |
| rm_work | cgoncalves: eh, the go one is significantly faster | 13:58 |
| rm_work | we were using it for stress testing | 13:58 |
| rm_work | c10k | 13:59 |
| rm_work | and we usually don't compile it in tests, just copy the binary | 13:59 |
| rm_work | is something actually compiling it? | 13:59 |
| *** ianychoi has joined #openstack-lbaas | 13:59 | |
| cgoncalves | rm_work: hmm ok in that case +1, also because we started working on rally support ;) | 13:59 |
| rm_work | ah i guess the c10k test script | 13:59 |
| rm_work | but that doesn't run in a gate or anything | 14:00 |
| *** pcaruana has quit IRC | 14:01 | |
| *** pcaruana has joined #openstack-lbaas | 14:15 | |
| rm_work | cgoncalves: anywho, here's how we put it on a member cirros: https://github.com/openstack/octavia-tempest-plugin/blob/master/octavia_tempest_plugin/tests/server_util.py#L278-L326 | 14:17 |
| *** salmankhan has joined #openstack-lbaas | 14:17 | |
| cgoncalves | rm_work: I've seen it, thanks | 14:18 |
| rm_work | kk | 14:18 |
| cgoncalves | rm_work: I'm not in favor of adding the binary in-tree though | 14:18 |
| *** pcaruana has quit IRC | 14:30 | |
| rm_work | I think we were already planning to in another patch | 14:31 |
| rm_work | but if you are also grabbing the tempest tests, you could get it from there | 14:31 |
| *** beagles|brb is now known as beagles | 14:32 | |
| *** yamamoto has joined #openstack-lbaas | 14:38 | |
| *** yamamoto has quit IRC | 14:43 | |
| *** pcaruana has joined #openstack-lbaas | 14:43 | |
| andreykurilin | hey! any chance to get help for configuring octavia to be launched in gates? | 14:43 |
| *** yamamoto has joined #openstack-lbaas | 14:53 | |
| *** yamamoto has quit IRC | 14:53 | |
| *** yamamoto has joined #openstack-lbaas | 14:53 | |
| rm_work | andreykurilin: which project's gates? | 14:57 |
| rm_work | andreykurilin: also ... are you on my team at GD? :P | 14:57 |
| andreykurilin | hah. possible possible ) | 14:58 |
| *** yamamoto has quit IRC | 14:58 | |
| andreykurilin | for rally ;) we have an old legacy job which works fine. I'm trying to setup a new one for Zuul V3 | 14:58 |
| rm_work | hmm | 14:59 |
| rm_work | I mean, I think you still just need to enable our devstack plugin? | 15:00 |
| rm_work | and everything should just work | 15:00 |
| andreykurilin | rm_work: not so easy :) | 15:00 |
| rm_work | hmm | 15:00 |
| rm_work | I mean, we use for example, Barbican, in our gate | 15:00 |
| rm_work | and the usage didn't change much between v2 and v3 of zuul | 15:00 |
| rm_work | but I don't know exactly how rally works | 15:01 |
| andreykurilin | rm_work: by default it will fail due to unavailable `q-svc` service. But it is easy to fix | 15:01 |
| rm_work | ahh well | 15:01 |
| rm_work | you may for that want to look at: https://review.openstack.org/#/c/544281/ | 15:02 |
| rm_work | which should fix this issue | 15:02 |
| andreykurilin | rm_work: when I enabled it, other neutron interfaces became unavailable (listing floating ips, for examle) | 15:02 |
| dmellado | andreykurilin: actually I meant to speak with you as we were thinking about creating a rally job for kuryr too | 15:02 |
| dmellado | o/ | 15:02 |
| rm_work | enabling octavia broke floating-ips?! | 15:02 |
| rm_work | that's ... not something that should be possible to happen, lol | 15:02 |
| dmellado | but if it's not zuulv3 compliant then we'll wait | 15:02 |
| andreykurilin | rm_work: turning on just `q-svc` | 15:03 |
| rm_work | is that because of q-svc being neutron-legacy? | 15:03 |
| rm_work | or something else | 15:03 |
| andreykurilin | I guess so | 15:03 |
| rm_work | because maybe that will be fixed with the patch i linked | 15:03 |
| andreykurilin | I suppose it had happened since new neutron devstack plugin found enabling old service and expects that I will turn on other q-* manually as well | 15:03 |
| andreykurilin | rm_work: ok, will wait :) | 15:04 |
| rm_work | we should rely on "neutron" not "q-svc" :P | 15:04 |
| andreykurilin | sure :) | 15:04 |
| rm_work | yeah try with a "depends-on" for that patch | 15:04 |
| rm_work | and see if it helps | 15:04 |
| andreykurilin | done. last attempt took ~40 minutes, so the result will not be so quick:) | 15:06 |
| rm_work | yeah | 15:06 |
| andreykurilin | dmellado: hi! we already have a basic rally-task-on-devstack job and it is fully Zuul V3 native (a bunch of ansible roles and so) | 15:07 |
| dmellado | andreykurilin: oh, cool, any link where I can fetch some info? | 15:09 |
| andreykurilin | dmellado: I did not make any annonce yet. hope will find a time this week. | 15:11 |
| andreykurilin | dmellado: https://github.com/openstack/rally-openstack/tree/master/.zuul.d#what-do-we-have-in-this-this-dir This dir includes all job definitiones . | 15:11 |
| dmellado | andreykurilin: thanks, looking forward to it ;) | 15:11 |
| andreykurilin | dmellado: and here is an example of re-using one of our jobs(which we launch in our gates) in another project - https://review.openstack.org/#/c/550003 | 15:11 |
| johnsom | o/ Hi folks, I am back from travelling and vacation... | 15:23 |
| dayou | o/, jason is waiting for you for many days :-D | 15:24 |
| *** yamamoto has joined #openstack-lbaas | 15:26 | |
| xgerman_ | o/ | 15:26 |
| xgerman_ | guess everyone gets a number to talk to johnsom | 15:27 |
| johnsom | Ha, yeah. My welcome back was an 8am video meeting... | 15:28 |
| johnsom | A lot of stuff to catch up on | 15:28 |
| *** yamamoto has quit IRC | 15:34 | |
| *** irenab has quit IRC | 15:35 | |
| *** oanson has quit IRC | 15:38 | |
| rm_work | ton of changes sitting around the top of the octavia queue awaiting +2/+a | 15:49 |
| *** amotoki_ is now known as amotoki | 15:50 | |
| *** kobis has quit IRC | 15:56 | |
| xgerman_ | yeah, also we have the periodic for the image build we can merge — | 16:06 |
| *** irenab has joined #openstack-lbaas | 16:11 | |
| *** oanson has joined #openstack-lbaas | 16:11 | |
| *** pcaruana has quit IRC | 16:12 | |
| rm_work | xgerman_: one question on that | 16:13 |
| xgerman_ | yes | 16:13 |
| rm_work | i posted it | 16:14 |
| xgerman_ | saw it | 16:15 |
| xgerman_ | commented | 16:18 |
| rm_work | xgerman_: i'm referring to our requirements file *in* our diskimage-create folder, which does include diskimage-builder | 16:20 |
| rm_work | but also some other stuff | 16:20 |
| rm_work | and usually --constraint is used for upper-constraints.txt | 16:21 |
| rm_work | which probably we should be doing? | 16:21 |
| rm_work | dunno | 16:21 |
| rm_work | https://github.com/openstack/octavia/blob/master/diskimage-create/requirements.txt | 16:21 |
| rm_work | of course, that may be hilariously out of date | 16:21 |
| xgerman_ | yeah, this doesn’t contrain DIB at all | 16:22 |
| rm_work | maybe your approach is fine | 16:22 |
| rm_work | i mean, when we are setting up to build an amp, we don't need to install the main octavia requirements file, usually JUST this one | 16:22 |
| xgerman_ | mmh, I just want to install DIB… | 16:23 |
| rm_work | well, what you want to do is build an amp image | 16:23 |
| xgerman_ | yes, but once DIB runs it will chroot and honor it’s own requirements | 16:23 |
| rm_work | but yeah, i think this file is a little wonky anyway, as I don't know why we are pulling in Babel explicitly, or PyYAML | 16:24 |
| xgerman_ | yep, the only two changes I am happy with 1) remove constraints 2) pull in requirements project | 16:24 |
| rm_work | i might just drop the constraint altogether | 16:25 |
| xgerman_ | ok | 16:26 |
| rm_work | when we build for devstack we pull in *master* | 16:26 |
| rm_work | of dib | 16:26 |
| rm_work | https://github.com/openstack/octavia/blob/master/devstack/settings#L5-L6 | 16:27 |
| openstackgerrit | German Eichberger proposed openstack/octavia master: Periodic job to build + publish diskimage https://review.openstack.org/549259 | 16:28 |
| xgerman_ | yes, but I am not testing diskimagebuilder but more making some image | 16:28 |
| rm_work | heh yes | 16:29 |
| rm_work | it should be fine | 16:29 |
| xgerman_ | yeah, let’s see how johnsom likes that patch. I see changes… | 16:30 |
| rm_work | eh I +2'd | 16:30 |
| rm_work | it won't harm anything | 16:30 |
| rm_work | and i want to see what happens when it runs | 16:30 |
| *** yamamoto has joined #openstack-lbaas | 16:32 | |
| xgerman_ | if it runs periodics have a bad rep | 16:38 |
| *** yamamoto has quit IRC | 16:38 | |
| rm_work | you are literally the only person i have heard say that, though you've said it 10+ times I think :P | 16:39 |
| openstackgerrit | Merged openstack/octavia master: Move scenario-lxd over to experimental https://review.openstack.org/551805 | 16:40 |
| *** irenab has quit IRC | 16:56 | |
| *** sshank has joined #openstack-lbaas | 16:57 | |
| *** oanson has quit IRC | 16:57 | |
| rm_work | wtb this https://review.openstack.org/551801 and https://review.openstack.org/551789 | 17:05 |
| rm_work | and https://review.openstack.org/#/c/551796/ | 17:05 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Reduce stestr concurrency to 1 for tempest jobs https://review.openstack.org/551801 | 17:08 |
| cgoncalves | rm_work: ^ minor update. please revisit and +2/-1 | 17:11 |
| *** Swami has joined #openstack-lbaas | 17:12 | |
| *** dmellado has quit IRC | 17:12 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Add experimental scenario-ovsfw job https://review.openstack.org/550431 | 17:17 |
| *** pcaruana has joined #openstack-lbaas | 17:18 | |
| *** AlexeyAbashkin has quit IRC | 17:23 | |
| *** pcaruana has quit IRC | 17:27 | |
| johnsom | I see you guys review my patches... I just deleted the lxd job in one of my gate fix patches. | 17:31 |
| johnsom | There really isn't anything to that job, so no real point at the moment | 17:31 |
| *** salmankhan has quit IRC | 17:32 | |
| *** irenab has joined #openstack-lbaas | 17:33 | |
| *** oanson has joined #openstack-lbaas | 17:35 | |
| *** salmankhan has joined #openstack-lbaas | 17:35 | |
| *** irenab has quit IRC | 17:42 | |
| *** imacdonn has quit IRC | 17:43 | |
| *** imacdonn has joined #openstack-lbaas | 17:44 | |
| *** oanson has quit IRC | 17:44 | |
| *** oanson_ has joined #openstack-lbaas | 17:44 | |
| *** irenab has joined #openstack-lbaas | 17:44 | |
| *** oanson_ is now known as oanson | 17:46 | |
| rm_work | johnsom: ummm yeah totally <_< | 17:52 |
| rm_work | ah i think it was in the one where i gave up after like 30 comments and slept, and then never got back to | 17:52 |
| johnsom | Yeah, part of that chain. | 17:53 |
| johnsom | I plan to work on that today, after digging out a bit here | 17:53 |
| *** pcaruana has joined #openstack-lbaas | 17:53 | |
| johnsom | FYI, I am -1ing the OVSFW gate. I don't see why WE should be testing OVSFW | 17:54 |
| rm_work | i asked him | 17:55 |
| rm_work | it's experimental | 17:55 |
| * rm_work shrugs | 17:55 | |
| johnsom | Seems like that should be over on the OVSFW patches, not here. | 17:56 |
| rm_work | hmm yeah i guess that's true | 17:56 |
| rm_work | though i bet 99% what is going to happen is that this whole test will just get dropped | 17:57 |
| johnsom | I have faith that Carlos will follow through. It's just not a good development process to push patches to fix OVSFW, merge them, and then run this experimental gate to see if it fixed it. | 17:58 |
| rm_work | yeah so actually i'm wondering if it even matters if this merges | 17:59 |
| rm_work | really it can be used as-is without merging | 17:59 |
| rm_work | to test whether the patch in depends-on fixes things | 17:59 |
| johnsom | True | 17:59 |
| rm_work | really, we need to get the last few gate stabilization things through | 18:06 |
| rm_work | https://review.openstack.org/#/c/551789/ | 18:06 |
| rm_work | https://review.openstack.org/#/c/551796/ | 18:06 |
| rm_work | those | 18:06 |
| *** tesseract has quit IRC | 18:06 | |
| rm_work | the intermittent failures from the tempest concurrency thing in n-lbaas have been making me cry | 18:06 |
| *** Swami has quit IRC | 18:11 | |
| *** Swami_ has joined #openstack-lbaas | 18:11 | |
| *** Swami_ has quit IRC | 18:11 | |
| *** Swami has joined #openstack-lbaas | 18:12 | |
| -openstackstatus- NOTICE: Most jobs in zuul are currently failing due to a recent change to zuul; we are evaluating the issue and will follow up with a recommendation shortly. For the moment, please do not recheck. | 18:17 | |
| *** ChanServ changes topic to "Most jobs in zuul are currently failing due to a recent change to zuul; we are evaluating the issue and will follow up with a recommendation shortly. For the moment, please do not recheck." | 18:17 | |
| johnsom | Yeah, I am looking at the functional patch. I had a few more changes in mine, so evaluating what I should do there. | 18:22 |
| johnsom | Trying to dig up what the story is with this stestr stuff. We just migrated to ostester like a year ago... | 18:27 |
| rm_work | yeah >_< | 18:32 |
| rm_work | we had to quickly swap over because everything broke when someone did something | 18:32 |
| *** kobis has joined #openstack-lbaas | 18:34 | |
| *** gcheresh has joined #openstack-lbaas | 18:35 | |
| johnsom | Darn, there are no pre-patch logs. This nlbaas patch output looks wrong | 18:37 |
| rm_work | oh? | 18:37 |
| rm_work | i was trawling through the outputs to figure out why everything was breaking, and i did not notice any significant issues with it | 18:37 |
| *** ChanServ changes topic to "Discussion of OpenStack Load Balancing (Octavia) | Rocky is open for development!" | 18:40 | |
| -openstackstatus- NOTICE: Zuul has been restarted without the breaking change; please recheck any changes which failed tests with the error "Accessing files from outside the working dir ... is prohibited." | 18:40 | |
| *** sshank has quit IRC | 18:40 | |
| johnsom | I just wish we had a run output from before that patch. | 18:41 |
| rm_work | hmmmm | 18:41 |
| openstackgerrit | iain MacDonnell proposed openstack/octavia master: Amphora agent should recognise Oracle Linux https://review.openstack.org/548729 | 18:43 |
| rm_work | johnsom: i mean, what makes you think it's wrong? | 18:43 |
| johnsom | I just wanted to confirm the file paths, output formatting, and the # of jobs run. | 18:44 |
| johnsom | I compared to octavia pre-change and I feel better about it. | 18:44 |
| rm_work | really all logs from before that are wiped? | 18:45 |
| johnsom | Yeah, we aren't turning patches on nlbaas really | 18:45 |
| johnsom | They all expired out | 18:45 |
| rm_work | https://review.openstack.org/#/c/545767/ | 18:46 |
| rm_work | there? | 18:46 |
| rm_work | it's on stable but | 18:46 |
| rm_work | the tests should be the same for the same reason you just mentioned, lol | 18:46 |
| johnsom | Yeah, hmm, 554->611 tests | 18:47 |
| johnsom | Seems odd | 18:47 |
| rm_work | or here https://review.openstack.org/#/c/549697/ | 18:47 |
| johnsom | Oh, that is ocata | 18:47 |
| rm_work | that one is master | 18:47 |
| rm_work | maybe better | 18:47 |
| *** harlowja has joined #openstack-lbaas | 18:50 | |
| johnsom | Did you test tox -e cover with his change? | 18:51 |
| *** kobis has quit IRC | 18:53 | |
| *** kobis has joined #openstack-lbaas | 18:55 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 18:56 | |
| *** kobis has quit IRC | 19:00 | |
| *** AlexeyAbashkin has quit IRC | 19:00 | |
| johnsom | Yeah, that patch broke the cover environment badly | 19:02 |
| rm_work | i don't run cover locally :( | 19:02 |
| rm_work | just functional and py35 and pep8 | 19:02 |
| *** openstackgerrit has quit IRC | 19:04 | |
| * rm_work looks | 19:05 | |
| *** eandersson has joined #openstack-lbaas | 19:06 | |
| rm_work | hmm yes it appears to die | 19:06 |
| rm_work | how is it "passing" in the gate then | 19:07 |
| rm_work | oh nm i think these are local fails | 19:07 |
| rm_work | mostly passing | 19:08 |
| *** salmankhan has quit IRC | 19:08 | |
| rm_work | forgot i was in the middle of a patch | 19:08 |
| *** gcheresh has quit IRC | 19:09 | |
| johnsom | We don't seem to have a coverage gate for nlbaas | 19:20 |
| cgoncalves | ah, coverage test right | 19:25 |
| cgoncalves | I intended to fix it today but I got a load of internal stuff | 19:25 |
| johnsom | cgoncalves Did you see my comments on the OVSFW gate? | 19:27 |
| cgoncalves | johnsom: reading now | 19:28 |
| cgoncalves | johnsom: we needed a way to test the neutron patch in a fullstack and since octavia isn't working when firewall=openvswitch I thought of adding that experimental job | 19:29 |
| cgoncalves | as for the zuul v2/v2, I thought about it and my take on it was that once we do it for one we do it for all | 19:30 |
| johnsom | Yeah, but that gate should be over on the OVSFW project (neutron?). Here, they have to change the patch to depend on their fix patches each time they push an attempted fix. | 19:30 |
| cgoncalves | which should be a dedicated review imo | 19:31 |
| johnsom | Well, "legacy" is purely for jobs converted with the migration scripts. | 19:31 |
| johnsom | This patch: https://review.openstack.org/#/c/541039/ does them all except for the multinodes, but now needs some additional work. | 19:32 |
| cgoncalves | ok, that one was not under my radar | 19:33 |
| johnsom | My biggest issue is that it shouldn't be in our project, but theirs to test their patches. In ours it's not very useful, they would have to merge there stuff or add depends-on for each of their patches | 19:33 |
| cgoncalves | ok, I get your point and is valid. bottom line is that I'm happy now with ovsfw being fixed and octavia working fine with ovs as firewall | 19:35 |
| johnsom | Ah, so they have fixed it? then maybe this can just be abandoned? | 19:35 |
| cgoncalves | it has not been merged yet. the patch I submitted verifies that 1) the neutron patch works as expected and 2) octavia works now when firewall=ovs | 19:36 |
| cgoncalves | so I propose to keep it open until the neutron patch gets merged, and then we can abandon this one | 19:37 |
| johnsom | Ok, so maybe just WIP it until they merge | 19:37 |
| cgoncalves | I'd rather keep it as experimental since it doesn't cost us anything, though | 19:38 |
| cgoncalves | k, will set WIP | 19:38 |
| *** openstackgerrit has joined #openstack-lbaas | 19:39 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: [DNM] Add experimental scenario-ovsfw job https://review.openstack.org/550431 | 19:39 |
| johnsom | It's more job bloat IMO. If it's not used it will just rot and take up space. I wouldn't mind seeing it as an experimental on their side though. It's a good test for OVSFW | 19:40 |
| cgoncalves | -W'ed it and left a comment | 19:44 |
| johnsom | Thanks | 19:44 |
| cgoncalves | so now onward to fixing the coverage test I broke... | 19:45 |
| openstackgerrit | Merged openstack/octavia master: Updated from global requirements https://review.openstack.org/551539 | 19:48 |
| cgoncalves | johnsom: not sure you're already up-to-date on the requirements.txt issues? | 19:50 |
| johnsom | No | 19:51 |
| cgoncalves | https://github.com/openstack/octavia/blob/stable/queens/requirements.txt#L35 | 19:52 |
| cgoncalves | minimum required version should have been 17.1.0 | 19:52 |
| cgoncalves | https://github.com/openstack/octavia/blob/stable/queens/requirements.txt#L37 | 19:53 |
| johnsom | Looks like upper has 17.5.0 | 19:53 |
| cgoncalves | and minimum for jinja should have been 2.10 | 19:53 |
| cgoncalves | right | 19:53 |
| johnsom | Same, upper has 2.10 | 19:53 |
| cgoncalves | my understanding is that if the our minimum does not match with our requirements.txt we should propose bumping it in global-requirements | 19:54 |
| johnsom | jinja is for the namespace support, what needs that version of pyopenssl? | 19:54 |
| cgoncalves | pyopenssl: https://bugzilla.redhat.com/show_bug.cgi?id=1553520 | 19:55 |
| openstack | bugzilla.redhat.com bug 1553520 in openstack-octavia "Cannot create listener with TLS termination" [Urgent,Assigned] - Assigned to cgoncalves | 19:55 |
| johnsom | Yeah, jinja2 is my bad. I should have pushed a bump for that, however I think the patch landed after the freeze | 19:55 |
| johnsom | I.e the bug and fix came in after requirements wouldn't accept a change any longer. | 19:56 |
| cgoncalves | whoever's installing octavia from source may ran into these issues in case has jinja>=2.8,<2.9 installed | 19:58 |
| johnsom | Have you talked with the requirements folks about these? | 19:59 |
| cgoncalves | no, I've been waiting for you plus busy with other urgent stuff | 20:00 |
| johnsom | Ha, thanks? grin | 20:00 |
| *** atoth has quit IRC | 20:03 | |
| * cgoncalves bows to johnsom | 20:03 | |
| johnsom | So G-R for jinja2 is already up to date in master requirements. | 20:03 |
| johnsom | I don't see us importing pyOpenSSL at all, so... | 20:03 |
| johnsom | It hasn't been bumped yet. I need more info on my we need 17.1.0 or greater. | 20:04 |
| cgoncalves | https://review.openstack.org/#/c/549913/ | 20:04 |
| cgoncalves | imacdonn, who found the issue and reported here on IRC, confirmed that python-cryptography>=1.9 *and* python-pyOpenSSL>=17.1.0 fixed the listener with TLS termination | 20:06 |
| imacdonn | johnsom: The specific reason is at https://pyopenssl.org/en/stable/api/crypto.html#OpenSSL.crypto.X509.to_cryptography | 20:07 |
| cgoncalves | right, in octavia/certificates/common/pkcs12.py | 20:08 |
| johnsom | Ah, ok. That I can work with. | 20:09 |
| johnsom | Not sure what to do about the queens requirements.txt though, I don't think we can update that now... | 20:09 |
| cgoncalves | I don't think so too | 20:10 |
| cgoncalves | what we need, though, is a way to make sure our requirements.txt is good enough | 20:11 |
| johnsom | Well, I think this is really an issue for the requirements team. We currently test using upper-constraints so it won't catch this. Either that needs to come off, or we need packagers to ONLY use upper-constraints | 20:12 |
| cgoncalves | for the latter, forcing packages to have latest and greatest versions would prove to be problematic | 20:14 |
| johnsom | Well, if that's what the community is testing with... | 20:14 |
| cgoncalves | also if the latter, then it wouldn't make no sense still having requirements.txt | 20:16 |
| cgoncalves | it would be u-c for all, everyone | 20:17 |
| cgoncalves | *everywhere | 20:17 |
| imacdonn | it becomes a bit grey when you consider that some of the requirements may be satisfied by packages that are part of the OS distro, not part of the OpenStack distro | 20:17 |
| cgoncalves | right, specially when it touches security-related packages (e.g. pyopenssl and cryptography) | 20:18 |
| johnsom | Yeah, it's pip's nature to always pull in the latest version that matches requirements.txt | 20:18 |
| johnsom | Ok, I think I see. We need a gate using the lower constraints file to catch these | 20:24 |
| johnsom | Though interesting that nobody else seems to have one | 20:25 |
| cgoncalves | johnsom: l-c as is now would install pyOpenSSL==16.2.0 which is not good enough for us | 20:46 |
| *** fnaval has quit IRC | 21:03 | |
| *** fnaval has joined #openstack-lbaas | 21:18 | |
| *** fnaval has quit IRC | 21:18 | |
| *** fnaval has joined #openstack-lbaas | 21:18 | |
| *** sshank has joined #openstack-lbaas | 21:30 | |
| *** pcaruana has quit IRC | 21:36 | |
| johnsom | pyOpenSSL G-R: https://review.openstack.org/552147 | 21:40 |
| *** Swami has quit IRC | 21:50 | |
| *** Swami has joined #openstack-lbaas | 21:51 | |
| openstackgerrit | Carlos Goncalves proposed openstack/neutron-lbaas master: Reduce stestr concurrency to 1 for tempest jobs https://review.openstack.org/551796 | 21:54 |
| cgoncalves | johnsom: ^ fixes coverage test | 21:54 |
| nmagnezi_ | cgoncalves, I start to think you relocated to rm_work's timezone.. :D | 21:54 |
| johnsom | Yeah, it's not like he had as much jet-lag as the rest of us.... | 21:55 |
| cgoncalves | nmagnezi_: haha! I have to hurry up, my gf is coming back in 1h :P | 21:55 |
| nmagnezi_ | lol | 21:55 |
| cgoncalves | nmagnezi_: shouldn't you be be playing call of duty in real life? :) | 21:56 |
| nmagnezi_ | cgoncalves, kinda :) | 21:57 |
| johnsom | cgoncalves Hmm, this runs local for you? | 21:58 |
| cgoncalves | johnsom: it did. double-check nevertheless :) | 22:03 |
| johnsom | Running now. I thought maybe we needed a more explicit path, but I see we ran the whole package before, so... | 22:04 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Reduce stestr concurrency to 1 for tempest jobs https://review.openstack.org/551801 | 22:20 |
| cgoncalves | johnsom: I'd like to get your view on rm_work's comment: https://review.openstack.org/#/c/549654/22/devstack/upgrade/vm_user_data.sh | 22:22 |
| cgoncalves | if nc causes trouble, going the goland server way is a bit overkilling IMO | 22:23 |
| cgoncalves | I'm thinking of python's SimpleHttpServer | 22:23 |
| *** rcernin has joined #openstack-lbaas | 22:24 | |
| johnsom | cgoncalves Yeah, the issue with the netcat approach is it doesn't handle concurrent requests well and causes random errors. The golang server was built to have a lightweight webserver that does concurrency well. | 22:29 |
| johnsom | I know for our C10k tests, simplehttpserver will not work, but it may be an option for a basic test web server. | 22:31 |
| johnsom | The other part of the golang is to be intelligent about the cookie handling, again which might not be needed for your tests | 22:32 |
| cgoncalves | we don't need concurrency for grenade | 22:32 |
| cgoncalves | ok, than are we in agreement that simplehttpserver is good enough for grenade? | 22:33 |
| johnsom | Yeah, as long as it's a once and done, you will be fine with nc. It's just when you repeatedly call it that it can fail | 22:33 |
| johnsom | cgoncalves I am fine with giving it a go. Also, if it is small enough, we might be able to pass it as user data into cirros and not need the SCP work arounds. | 22:34 |
| cgoncalves | johnsom: in total it's 3x GETs only IIRC | 22:35 |
| johnsom | Yeah, if they aren't right after each other, the netcat is probably fine. | 22:36 |
| cgoncalves | johnsom: we would still need to compile it, no? | 22:36 |
| cgoncalves | k, I'll then keep with netcat. if it proves to be troublesome we know why and how to fix it | 22:36 |
| johnsom | The python, no. We can just do shell tricks to get it in there. | 22:36 |
| cgoncalves | yes, for simplehttpserver we don't need tricks | 22:37 |
| johnsom | Yeah. We have merged a pre-compiled golang server for the tempest plugin, so it's there if we want it. | 22:37 |
| johnsom | cgoncalves FYI, https://raymii.org/s/tutorials/Automating_Openstack_with_Cloud_init_run_a_script_on_VMs_first_boot.html | 22:38 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 22:38 | |
| johnsom | I tried base64'ing the golang server, but it's too large for nova to handle, the API bombs | 22:40 |
| cgoncalves | why have I been spending time grenading when you've been bombing?! :D | 22:41 |
| johnsom | Some day you will bomb too.... grin | 22:42 |
| *** AlexeyAbashkin has quit IRC | 22:42 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Add grenade support https://review.openstack.org/549654 | 22:44 |
| cgoncalves | lol | 22:45 |
| *** kobis has joined #openstack-lbaas | 22:49 | |
| *** yamamoto has joined #openstack-lbaas | 22:57 | |
| johnsom | cgoncalves So, the octavia stestr coverage patch, I think the parallel-mode flag is breaking things | 22:58 |
| *** sshank has quit IRC | 22:59 | |
| *** kobis has quit IRC | 23:02 | |
| cgoncalves | hmm | 23:11 |
| cgoncalves | coverage combine {toxinidir}/.coverage-* works locally | 23:11 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Reduce stestr concurrency to 1 for tempest jobs https://review.openstack.org/551801 | 23:11 |
| johnsom | Yeah, I'm explicitly naming the output files | 23:11 |
| cgoncalves | Usage: coverage combine [options] <path1> <path2> ... <pathN> | 23:12 |
| johnsom | cgoncalves You forgot to add the sh whitelist back in | 23:12 |
| cgoncalves | it should take multiple paths | 23:12 |
| cgoncalves | oops! | 23:12 |
| cgoncalves | so sloppy | 23:12 |
| johnsom | Well, I made it explicit because there was some trash coverage files, I don't remember exactly what was happening, but I made it explicit so combine wouldn't pickup something I didn't want. | 23:13 |
| johnsom | probably overkill now | 23:14 |
| johnsom | It was failing because you added parallel mode which just appends ".<something>" to the end of the filename | 23:15 |
| johnsom | Dropping the --parallel-mode would have made the explicit filename way work again | 23:17 |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Reduce stestr concurrency to 1 for tempest jobs https://review.openstack.org/551801 | 23:20 |
| cgoncalves | see you tomorrow | 23:20 |
| johnsom | o/ | 23:20 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Reduce stestr concurrency to 1 for tempest jobs https://review.openstack.org/551801 | 23:23 |
| *** fnaval has quit IRC | 23:32 | |
| *** fnaval has joined #openstack-lbaas | 23:36 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!