| rm_work | <clarkb>rm_work: you need to use non legacy devstack/tempest for that to work. Otherwise its using the old devstack-gate based system which requires the env var for d-g to write into the localrc | 00:02 |
|---|---|---|
| rm_work | <clarkb>rm_work: ya you have to set the devstack-gate option to set that value if there is one or otherwise set it in the append to localrc var | 00:03 |
| *** yamamoto has joined #openstack-lbaas | 00:06 | |
| *** keithmnemonic[m] has quit IRC | 00:09 | |
| *** kbyrne has quit IRC | 00:10 | |
| *** PagliaccisCloud has quit IRC | 00:11 | |
| *** yamamoto has quit IRC | 00:11 | |
| *** PagliaccisCloud has joined #openstack-lbaas | 00:13 | |
| *** kbyrne has joined #openstack-lbaas | 00:14 | |
| *** slaweq has joined #openstack-lbaas | 00:17 | |
| johnsom | So, I'm going to put the amphora agent in a venv | 00:21 |
| *** slaweq has quit IRC | 00:21 | |
| johnsom | Do we care what on the file system I put that venv? | 00:22 |
| rm_work | johnsom: you mean where? | 00:26 |
| rm_work | do we already have /var/lib/octavia ? | 00:26 |
| rm_work | if so... /var/lib/octavia/venv IMO | 00:27 |
| johnsom | /opt/amphora-agent-venv | 00:27 |
| rm_work | ah | 00:27 |
| johnsom | Put it next to the checkout? | 00:27 |
| rm_work | that's fine with me too I guess | 00:27 |
| johnsom | I think opt makes sense | 00:27 |
| rm_work | I just like keeping everything in the same place so i don't have to think about it so hard | 00:27 |
| rm_work | :/ | 00:27 |
| rm_work | and python itself goes in /var/lib so i figure it can't be too bad | 00:27 |
| johnsom | yeah, this is actually *easier* to find that digging in /usr/local/lib land | 00:28 |
| rm_work | yeah but, inside an amp... | 00:28 |
| rm_work | we already have to direct people to /var/lib/octavia | 00:28 |
| rm_work | so we could keep it to one place | 00:28 |
| johnsom | I point people to /opt to get the SHA they are running. | 00:28 |
| rm_work | oh is octavia already installed there? i thought it ended up in /var/lib/python27/ or something | 00:29 |
| rm_work | unless we do system-lib sharing, wouldn't that actually end up increasing the size of our amp? :/ | 00:29 |
| rm_work | as there may be some overlap | 00:30 |
| rm_work | i guess maybe it won't be too bad | 00:30 |
| johnsom | It's cloned there, then currently installed in to /usr/local/lib/pythonX/site- whatever | 00:30 |
| rm_work | ah usr/lib | 00:30 |
| johnsom | Yeah, that is the downside, but, this solves the war between the python-X debs and pip | 00:30 |
| johnsom | Would be nice to have a requirements.txt just for the amphora-agent. It would save a bunch of space too. | 00:33 |
| johnsom | Hmmm, darn it. yeah, 2gb might not work with the venv | 00:33 |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 00:40 |
| rm_work | yes, we could make an amp-requirements.txt | 00:40 |
| rm_work | and honestly... | 00:40 |
| rm_work | it doesn't need to co-install | 00:40 |
| johnsom | We would just have to keep it in sync | 00:41 |
| rm_work | I don't know if i feel like we NEED to gate against g-r / l-c | 00:41 |
| rm_work | which was the main reason we didn't do it before | 00:41 |
| rm_work | well... the other issue being that we basically have to load up all of the other code anyway tho ... | 00:42 |
| rm_work | so maybe it doesn't help us a lot? | 00:42 |
| rm_work | unless i'm wrong and it doesn't | 00:42 |
| johnsom | It doesn't run the other stuff, just installs it | 00:43 |
| rm_work | well | 00:43 |
| rm_work | we load up a lot of shared code | 00:43 |
| rm_work | which imports other stuff | 00:43 |
| rm_work | and i am kinda thinking it might end up loading up most of our codebase :/ | 00:44 |
| rm_work | depending on what we use | 00:44 |
| johnsom | Yeah, darn, venv adds 70MB to the qcow2 | 00:44 |
| johnsom | Oh, typo, 7MB | 00:46 |
| rm_work | oh | 00:46 |
| rm_work | lol k meh | 00:46 |
| rm_work | though we can probably get BACK like 100mb | 00:46 |
| rm_work | if we trim it down | 00:46 |
| johnsom | We will when we switch to minimal | 00:46 |
| rm_work | we'll do an amp-requirements.txt and trim it down a lot when we switch to minimal? | 00:47 |
| rm_work | (the virtualenv) | 00:47 |
| johnsom | Then there is the whole build environment thing for the one python package that doesn't come as a wheel | 00:47 |
| rm_work | T_T which one is that | 00:47 |
| rm_work | we could build a wheel locally | 00:47 |
| johnsom | ubuntu-minimal drops about 100MB off the qcow just itself | 00:47 |
| johnsom | Do we get into cross platform issues building it outside the image chroot? Like if you made an ubuntu amp on suse? | 00:48 |
| rm_work | i mean we can drop another 100 in the venv | 00:49 |
| rm_work | since just our venv is 200-300 isn't it? | 00:49 |
| johnsom | This is the build environment I wish we could drop: http://logs.openstack.org/60/559460/5/check/octavia-v1-dsvm-scenario/e354ff1/logs/devstacklog.txt.gz#_2018-04-10_00_39_08_297 | 00:53 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Switch to ubuntu-minimal for default amphora image https://review.openstack.org/559416 | 01:07 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Switch to ubuntu-minimal for default amphora image https://review.openstack.org/559416 | 01:07 |
| *** yamamoto has joined #openstack-lbaas | 01:07 | |
| johnsom | Ok, that should work. I stacked it on the minimal patch so we don't have to increase the default image size to 3gb | 01:08 |
| *** yamamoto has quit IRC | 01:13 | |
| rm_work | gain space! -> lose space again | 01:18 |
| *** dlundquist has quit IRC | 01:20 | |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 02:07 |
| *** yamamoto has joined #openstack-lbaas | 02:09 | |
| *** yamamoto has quit IRC | 02:15 | |
| johnsom | So this looks good for the gate fix: https://review.openstack.org/559416 | 03:06 |
| johnsom | CentOS is still broken on vitrualenv, but I think Ian is on top of that fix for DIB | 03:06 |
| *** yamamoto has joined #openstack-lbaas | 03:11 | |
| *** yamamoto has quit IRC | 03:16 | |
| *** annp has joined #openstack-lbaas | 03:28 | |
| *** gans has joined #openstack-lbaas | 03:55 | |
| *** amitry has quit IRC | 04:03 | |
| openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow detail pages to auto refresh upon action https://review.openstack.org/561458 | 04:06 |
| *** yamamoto has joined #openstack-lbaas | 04:13 | |
| openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Replace unicode punctation with ascii https://review.openstack.org/561780 | 04:13 |
| *** yamamoto has quit IRC | 04:18 | |
| *** annp has quit IRC | 04:21 | |
| *** annp has joined #openstack-lbaas | 04:21 | |
| *** sanfern has joined #openstack-lbaas | 04:31 | |
| openstackgerrit | Ganpat Agarwal proposed openstack/octavia master: Active-Active: ExaBGP amphora L3 distributor driver https://review.openstack.org/537842 | 04:50 |
| openstackgerrit | Ganpat Agarwal proposed openstack/octavia master: Active-Active: ExaBGP amphora L3 distributor driver https://review.openstack.org/537842 | 04:51 |
| openstackgerrit | Ganpat Agarwal proposed openstack/octavia master: ACTIVE-ACTIVE: Initial distributor data model https://review.openstack.org/528850 | 04:52 |
| *** gans has quit IRC | 05:04 | |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 05:06 |
| *** sapd_ has quit IRC | 05:07 | |
| *** sapd_ has joined #openstack-lbaas | 05:08 | |
| *** links has joined #openstack-lbaas | 05:23 | |
| *** keithmnemonic[m] has joined #openstack-lbaas | 05:30 | |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 05:32 |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 06:00 |
| *** yamamoto has joined #openstack-lbaas | 06:16 | |
| *** pcaruana has joined #openstack-lbaas | 06:16 | |
| *** yamamoto has quit IRC | 06:22 | |
| *** ianychoi has joined #openstack-lbaas | 06:38 | |
| *** gokhan has quit IRC | 06:39 | |
| *** yboaron has joined #openstack-lbaas | 06:50 | |
| *** slaweq has joined #openstack-lbaas | 06:50 | |
| *** rcernin has quit IRC | 06:57 | |
| *** numans has quit IRC | 07:08 | |
| *** numans has joined #openstack-lbaas | 07:08 | |
| *** velizarx has joined #openstack-lbaas | 07:09 | |
| *** yamamoto has joined #openstack-lbaas | 07:10 | |
| *** yamamoto has quit IRC | 07:14 | |
| *** numans has quit IRC | 07:14 | |
| *** yamamoto has joined #openstack-lbaas | 07:14 | |
| *** tesseract has joined #openstack-lbaas | 07:14 | |
| *** numans has joined #openstack-lbaas | 07:14 | |
| *** velizarx has quit IRC | 07:21 | |
| *** numans has quit IRC | 07:21 | |
| *** numans has joined #openstack-lbaas | 07:21 | |
| *** velizarx has joined #openstack-lbaas | 07:24 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 07:31 | |
| *** cgoncalves has joined #openstack-lbaas | 07:34 | |
| *** cgoncalves has left #openstack-lbaas | 07:35 | |
| *** cgoncalves has joined #openstack-lbaas | 07:38 | |
| *** yamamoto_ has joined #openstack-lbaas | 07:39 | |
| *** yamamoto has quit IRC | 07:42 | |
| *** ianychoi has quit IRC | 08:04 | |
| *** ianychoi has joined #openstack-lbaas | 08:10 | |
| *** yboaron_ has joined #openstack-lbaas | 08:26 | |
| *** yboaron has quit IRC | 08:29 | |
| *** yamamoto_ has quit IRC | 08:36 | |
| *** yamamoto has joined #openstack-lbaas | 08:37 | |
| openstackgerrit | Nir Magnezi proposed openstack/octavia master: Revert "Rename q- to neutron- services" https://review.openstack.org/558194 | 08:43 |
| *** yboaron_ has quit IRC | 08:52 | |
| dulek | Hey, kind of dumb question, but where are Octavia logs in http://logs.openstack.org/63/549263/7/check/octavia-v1-dsvm-scenario-multinode/4b4d75b/logs/ ? | 09:00 |
| *** velizarx has quit IRC | 09:46 | |
| *** velizarx has joined #openstack-lbaas | 09:52 | |
| openstackgerrit | Carlos Goncalves proposed openstack/octavia master: Add sos element to Red Hat family images https://review.openstack.org/561860 | 09:59 |
| dmellado | bcafarel: nmagnezi cgoncalves any chance you're getting this kind of issue when deploying dsvm? | 10:22 |
| dmellado | 135412 2018-04-17 09:55:02.560 | Found existing installation: PyYAML 3.11 | 10:22 |
| dmellado | 135413 2018-04-17 09:55:02.753 | Cannot uninstall 'PyYAML'. It is a distutils installed project and thus we cannot accurately determine which files belong to it which would lead to only a pa rtial uninstall. | 10:22 |
| dmellado | looks like it's on the process of creating the amphora | 10:22 |
| nmagnezi | dmellado, yes | 10:40 |
| nmagnezi | dmellado, diskimage-builder got broken because of pip10 | 10:41 |
| dmellado | nmagnezi: oh, awesome | 10:41 |
| dmellado | I'll retry with the amphora qcow2 | 10:41 |
| nmagnezi | dmellado, i tried this https://review.openstack.org/#/c/561740/ but it didn't work for me locally and I couldn't figure out why (the logs don't say much) | 10:42 |
| nmagnezi | dmellado, also i know https://review.openstack.org/#/c/561479/ is in progress but last I heard from johnsom it didn't work that well for him | 10:42 |
| dmellado | nmagnezi: ack, thanks for the pointers | 10:46 |
| nmagnezi | happy to help | 10:46 |
| dmellado | I guess for my tests I can live with the qcow2 for now | 10:46 |
| nmagnezi | yup | 10:46 |
| *** yboaron_ has joined #openstack-lbaas | 10:59 | |
| *** AlexeyAbashkin has quit IRC | 11:07 | |
| *** sanfern has quit IRC | 11:19 | |
| bcafarel | yay more pip fun :/ | 11:23 |
| *** atoth has joined #openstack-lbaas | 11:27 | |
| *** AlexeyAbashkin has joined #openstack-lbaas | 11:28 | |
| *** velizarx has quit IRC | 11:48 | |
| *** velizarx has joined #openstack-lbaas | 11:51 | |
| *** samccann has joined #openstack-lbaas | 11:54 | |
| *** slaweq_ has joined #openstack-lbaas | 12:03 | |
| *** slaweq has quit IRC | 12:06 | |
| *** slaweq has joined #openstack-lbaas | 12:07 | |
| *** slaweq_ has quit IRC | 12:08 | |
| openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow members to be set as "backup" https://review.openstack.org/561387 | 12:16 |
| *** velizarx has quit IRC | 12:17 | |
| openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Add timeout options for listener https://review.openstack.org/561402 | 12:18 |
| openstackgerrit | Jacky Hu proposed openstack/octavia-dashboard master: Allow detail pages to auto refresh upon action https://review.openstack.org/561458 | 12:19 |
| *** velizarx has joined #openstack-lbaas | 12:25 | |
| *** yamamoto has quit IRC | 12:25 | |
| *** sanfern has joined #openstack-lbaas | 12:25 | |
| *** yamamoto has joined #openstack-lbaas | 12:25 | |
| *** yamamoto has quit IRC | 12:26 | |
| *** yamamoto has joined #openstack-lbaas | 12:26 | |
| *** sanfern has quit IRC | 12:36 | |
| openstackgerrit | boden proposed openstack/neutron-lbaas master: use rpc Connection rather than create_connection https://review.openstack.org/561028 | 12:43 |
| *** atoth has quit IRC | 12:47 | |
| *** atoth has joined #openstack-lbaas | 13:03 | |
| *** ianychoi_ has joined #openstack-lbaas | 13:07 | |
| *** ianychoi has quit IRC | 13:10 | |
| *** andreykurilin has joined #openstack-lbaas | 13:20 | |
| *** links has quit IRC | 13:38 | |
| *** yamamoto has quit IRC | 13:39 | |
| *** yamamoto has joined #openstack-lbaas | 13:39 | |
| *** yamamoto has quit IRC | 13:44 | |
| *** yamamoto has joined #openstack-lbaas | 14:01 | |
| *** johnsom has quit IRC | 14:11 | |
| *** xgerman_ has quit IRC | 14:11 | |
| *** dayou has quit IRC | 14:12 | |
| *** johnsom has joined #openstack-lbaas | 14:12 | |
| *** xgerman_ has joined #openstack-lbaas | 14:12 | |
| *** dayou has joined #openstack-lbaas | 14:13 | |
| *** cristicalin has joined #openstack-lbaas | 14:16 | |
| *** slaweq_ has joined #openstack-lbaas | 14:18 | |
| *** slaweq has quit IRC | 14:21 | |
| dmellado | bcafarel: you can always say yeee-pip | 14:39 |
| dmellado | xD | 14:40 |
| * dmellado flees | 14:40 | |
| * bcafarel throws pypi angry users at dmellado | 14:40 | |
| dmellado | xDDDD | 14:40 |
| *** cristicalin has quit IRC | 14:48 | |
| *** cristicalin has joined #openstack-lbaas | 14:49 | |
| *** yboaron_ has quit IRC | 14:49 | |
| johnsom | Would be nice to get some reviews on the ubuntu gate fix: https://review.openstack.org/#/c/559416/ | 14:54 |
| cgoncalves | johnsom, hi. how can users create TERMINATED_HTTPS listeners in a cloud without barbican or castellan? I'm failing to see how to create with the local cert manager | 14:58 |
| cgoncalves | do users have to create and publish TLS containers themselves somewhere and pass it in --default-tls-container-ref? | 14:59 |
| johnsom | cgoncalves I will start with "It's a bad idea" since they will be stored insecurely on the filesystem. | 15:00 |
| cgoncalves | I'm well aware of that :) | 15:00 |
| johnsom | Yes, with the https://github.com/openstack/octavia/blob/master/etc/octavia.conf#L116 local_cert_manager you have to pre-drop them in the folder | 15:01 |
| dmellado | johnsom: I'll make some of my patches depends on that one | 15:01 |
| dmellado | we're currently having some issues with octavia so I'll update later on | 15:01 |
| johnsom | Cool, it' fixes Ubuntu, but Ian has a fix in flight for CentOS | 15:02 |
| cgoncalves | johnsom, uh, most likely only operators can do that | 15:02 |
| johnsom | Right, local cert manager was really only intended for testing.... | 15:02 |
| johnsom | It's a subtle, "don't do this" | 15:03 |
| cgoncalves | I understand that. I am trying to verify the fix for the cryptography+pyOpenSSL issue in a cloud without barbican or castellan | 15:05 |
| cgoncalves | I'm not sure creating a TERMINATED_HTTPS listener without dropping any certs in the local cert manager is enough to verify. I guess not | 15:06 |
| johnsom | Yeah, only think I can offer is using the local cert manager. See the test for info on how to use it: http://git.openstack.org/cgit/openstack/neutron-lbaas/tree/neutron_lbaas/tests/unit/common/cert_manager/test_local.py | 15:06 |
| johnsom | Probably not a good test, no | 15:06 |
| *** yamamoto has quit IRC | 15:17 | |
| *** yamamoto has joined #openstack-lbaas | 15:17 | |
| cgoncalves | actually I can test with barbican as it's gonna be supported downstream, just learned that \o/ | 15:21 |
| johnsom | Nice | 15:21 |
| *** pcaruana has quit IRC | 15:28 | |
| *** velizarx has quit IRC | 15:40 | |
| *** cristicalin has quit IRC | 15:50 | |
| xgerman_ | +1 | 16:03 |
| *** links has joined #openstack-lbaas | 16:15 | |
| *** pcaruana has joined #openstack-lbaas | 16:18 | |
| *** harlowja has joined #openstack-lbaas | 16:23 | |
| *** yamamoto_ has joined #openstack-lbaas | 16:30 | |
| *** yamamoto has quit IRC | 16:33 | |
| dulek | cgoncalves: Did you overcome this issue already: http://logs.openstack.org/74/561974/1/check/kuryr-kubernetes-tempest-octavia/c264aa1/logs/devstacklog.txt.gz#_2018-04-17_16_14_13_629 ? | 16:37 |
| dulek | Or are we hitting some strange ones in Kuryr? | 16:37 |
| johnsom | dulek https://review.openstack.org/#/c/559416/ | 16:37 |
| dulek | johnsom: That's the fix? | 16:38 |
| johnsom | dulek Yes, pip 10 came out and broke a bunch of stuff. That patch fixes the ubuntu DIB image builds | 16:38 |
| dulek | johnsom: Are you planning to backport this to stable/queens? | 16:39 |
| johnsom | dulek Yes, we will have to do something. The challenge and why it's bundled in this patch is the venv solution makes the image build fail with a 2GB filesystem. We will have to figure out if we just pin pip to 9.0.3 for the older branches or if we bump the default filesystem size. | 16:40 |
| dulek | johnsom: Hm, so our stable/queens is broken until that's fixed. | 16:41 |
| dulek | BTW pip's not in https://github.com/openstack/requirements/blob/stable/queens/upper-constraints.txt ? | 16:41 |
| johnsom | nope | 16:41 |
| johnsom | In this case, it's installed by disk-image-builder | 16:42 |
| dulek | johnsom: Okay, thanks for help. There's no bug to track the progress on this fix? | 16:42 |
| johnsom | dulek Hmm, no, it appears DIB does not have one open | 16:43 |
| johnsom | The parallel DIB fix is here: https://review.openstack.org/#/c/561479/ | 16:44 |
| johnsom | I'm just not sure how successful that path is going to be, so I did the venv thing for Octavia master | 16:45 |
| dulek | Okay, thank you for clarifying this stuff! :) | 16:46 |
| johnsom | Yeah, pip 10 landed over the weekend and broke devstack and diskimage-builder | 16:47 |
| *** harlowja has quit IRC | 16:52 | |
| *** harlowja has joined #openstack-lbaas | 16:57 | |
| *** fnaval has joined #openstack-lbaas | 16:58 | |
| *** yamamoto_ has quit IRC | 17:15 | |
| *** yamamoto has joined #openstack-lbaas | 17:16 | |
| *** harlowja has quit IRC | 17:20 | |
| *** links has quit IRC | 17:35 | |
| *** links has joined #openstack-lbaas | 17:47 | |
| rm_work | johnsom: i imagine pinning it in stable branches to 9.x makes sense | 17:54 |
| johnsom | Yeah, but I just realized DIB is not stable, so... | 17:55 |
| johnsom | Maybe *we* override DIB and pin it ourselves | 17:55 |
| rm_work | mayhaps | 17:55 |
| *** AlexeyAbashkin has quit IRC | 18:02 | |
| *** tesseract has quit IRC | 18:06 | |
| johnsom | rm_work If you have a minute the ubuntu gate fix: https://review.openstack.org/#/c/559416/ | 18:08 |
| *** AlexeyAbashkin has joined #openstack-lbaas | 19:25 | |
| *** AlexeyAbashkin has quit IRC | 19:34 | |
| *** imacdonn has quit IRC | 19:35 | |
| *** imacdonn has joined #openstack-lbaas | 19:35 | |
| rm_work | i feel like jsonpath-rw should like... be read-write | 19:38 |
| rm_work | but i think they are read only, no tools to update stuff? :/ | 19:39 |
| rm_work | rw must mean something else in this context | 19:39 |
| rm_work | so i guess figuring out how to combine jsonpath-rw and jsonpatch | 19:39 |
| rm_work | but their formats are a little out of sync :/ | 19:39 |
| openstackgerrit | German Eichberger proposed openstack/octavia master: Updates the docs with new admin tips https://review.openstack.org/562035 | 19:40 |
| rm_work | ah i see, it is, but he ... didn't document it | 19:40 |
| *** harlowja has joined #openstack-lbaas | 19:50 | |
| *** yamamoto has quit IRC | 20:29 | |
| *** yamamoto has joined #openstack-lbaas | 20:30 | |
| *** yamamoto has quit IRC | 20:34 | |
| openstackgerrit | German Eichberger proposed openstack/octavia master: Updates the docs with new admin tips https://review.openstack.org/562035 | 20:35 |
| *** sapd__ has joined #openstack-lbaas | 20:40 | |
| *** links has quit IRC | 20:42 | |
| *** pcaruana has quit IRC | 20:42 | |
| *** sapd_ has quit IRC | 20:43 | |
| openstackgerrit | Merged openstack/octavia master: Switch to ubuntu-minimal for default amphora image https://review.openstack.org/559416 | 21:10 |
| johnsom | Ok, https://review.openstack.org/562055 saves another 50MB out of the image | 21:27 |
| rm_work | ugh testing pecan hook is annoying | 21:34 |
| rm_work | can't seem to get the pecan app to reload with a changed config var | 21:34 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Add Octavia v2 tempest plugin jobs to Octavia https://review.openstack.org/541002 | 21:36 |
| johnsom | rm_work you can HUP it | 21:36 |
| johnsom | I think | 21:36 |
| openstackgerrit | Michael Johnson proposed openstack/octavia master: Make scenario gates use devstack-tempest parent https://review.openstack.org/541039 | 21:37 |
| johnsom | Ok, those are rebased now that the gates are functional | 21:37 |
| rm_work | johnsom: i mean in functional tests | 21:38 |
| rm_work | i got it working | 21:38 |
| rm_work | but i want testing to work | 21:38 |
| rm_work | to write the functional test i had to just hardcode the hook on | 21:38 |
| rm_work | because using oslo config fixture and trying to reload the app... it doesn't see the updated config | 21:38 |
| rm_work | ahh because the make_app thing is too late in the process | 21:40 |
| rm_work | it doesn't cause api/config.py to do its thing again | 21:40 |
| rm_work | yeah i need to change this if i want to test it | 21:42 |
| rm_work | oooooor | 21:43 |
| rm_work | ok | 21:43 |
| *** slaweq_ has quit IRC | 21:48 | |
| *** slaweq has joined #openstack-lbaas | 21:48 | |
| *** yamamoto has joined #openstack-lbaas | 21:52 | |
| *** slaweq has quit IRC | 21:53 | |
| rm_work | johnsom: when i add a requirement do i have to also change lower-constraints? | 21:55 |
| johnsom | rm_work Yes | 21:55 |
| johnsom | Oldest version that works for you and is inside G-R | 21:56 |
| rm_work | so should I like ... | 21:56 |
| rm_work | go back to the first version in pypi | 21:56 |
| rm_work | and try a binary search | 21:56 |
| rm_work | or can i just like... | 21:57 |
| johnsom | Um, sure, if you feel like it. I personally would just take the lowest valid value in G-R | 21:57 |
| rm_work | pick a version | 21:57 |
| rm_work | erg k | 21:57 |
| rm_work | g-r is < | 21:57 |
| rm_work | so like | 21:57 |
| johnsom | What is it? | 21:58 |
| rm_work | jsonpatch!=1.20 # BSD | 21:58 |
| rm_work | jsonpath-rw<2.0 # Apache-2.0 | 21:58 |
| rm_work | those two | 21:58 |
| *** yamamoto has quit IRC | 21:58 | |
| rm_work | the second is weird because the highest version is 1.4.0 | 21:59 |
| rm_work | so i don't get what the <2.0 is for | 21:59 |
| johnsom | Maybe they know something is coming??? no idea | 22:00 |
| rm_work | lol | 22:00 |
| rm_work | yeah last release in 2015 | 22:00 |
| rm_work | maybe they know there is a new thing | 22:00 |
| johnsom | jsonpath-rw===1.2.0 and jsonpatch===1.16 | 22:00 |
| johnsom | I see they still have a lower constraints file in G-R. | 22:01 |
| rm_work | ah so i should just use theirs? | 22:01 |
| rm_work | i guess | 22:01 |
| johnsom | Though I'm pretty sure I saw a patch that removes it | 22:01 |
| johnsom | Sure, why not. | 22:01 |
| rm_work | oh wut we already have jsonpatch in our l-c | 22:01 |
| rm_work | but not in our req | 22:01 |
| rm_work | >_> | 22:01 |
| johnsom | Yeah, they imported a base one with everything people were using. | 22:02 |
| johnsom | Should match the G-R version. Though, that is a bit annoying having all of that in there when we don't use it... hmmm | 22:02 |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Add pecan middleware for neutron-lbaas compat https://review.openstack.org/562063 | 22:08 |
| rm_work | ok, one down | 22:08 |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 22:09 |
| rm_work | lets see what happens now | 22:09 |
| rm_work | johnsom: what other gotchas have you and xgerman_ run into with the compatability? i remember the 404 thing | 22:14 |
| *** yamamoto has joined #openstack-lbaas | 22:15 | |
| rm_work | johnsom: is that something that could go in this compat hook? | 22:15 |
| xgerman_ | so you have no faith in my proxy | 22:15 |
| johnsom | Well, that is a bug more than a gotcha | 22:15 |
| rm_work | xgerman_: i don't care whose it is, i have never thought it is the right approach | 22:15 |
| rm_work | i've said for multiple years now: we should just be able to l7-forward the traffic | 22:16 |
| rm_work | anything heavier weight than that seems kinda crazy | 22:16 |
| rm_work | so i'm seeing if i can get a gate to test the theory | 22:16 |
| rm_work | i just forgot that we got ahead of things a little bit with the tenant_id deprecation T_T | 22:17 |
| rm_work | like ... the proxy requires neutron-lbaas to be installed and configured, and i don't even think we should be requiring THAT | 22:18 |
| johnsom | Well, technically neutron should have removed it too by now | 22:18 |
| *** yamamoto has quit IRC | 22:20 | |
| xgerman_ | rm_work: the proxy takes neutron quotas into account and returns things 100% compatible with neutron — how useful that is for a specific installation I don’t know | 22:24 |
| rm_work | well, octavia SHOULD be returning things 100% compatible with neutron too | 22:24 |
| rm_work | or we have failed to make an API to the spec | 22:24 |
| xgerman_ | there are quirks | 22:25 |
| rm_work | and we should fix those | 22:25 |
| rm_work | that is why i'm working on this | 22:25 |
| rm_work | ideally your proxy plugin could also just rely on this compat-hook thing, and it would simplify it | 22:25 |
| rm_work | by putting the changes there | 22:25 |
| rm_work | my assertion is that the *octavia API* should be able to make a provable claim to be 100% compatible, as we promised it would be repeatedly | 22:26 |
| xgerman_ | we don’t need to replicate every error neutron makes… | 22:27 |
| johnsom | I agree | 22:27 |
| rm_work | johnsom: to which | 22:27 |
| johnsom | rm_work to you. | 22:27 |
| rm_work | i wish we didn't have to replicate neutron dumbness, but | 22:27 |
| rm_work | i think we legitimately need to | 22:27 |
| johnsom | Well, yeah, the neutron oddness is... like returning 404 instead of 403. That one bugs me | 22:27 |
| rm_work | though i am trying to make it so the compat layer is configurable | 22:27 |
| rm_work | oh whoops, speaking of... i forgot to configure that, damnit | 22:29 |
| xgerman_ | personally I would leave all the neutron-lbaas stuff behind in that repo and have us do our own thing in octavia. After all they are frozen in time whereas we evolve so there will be differences over time and the deprecation cycle ensure that the lbaas goes away - if we do a compat layer deprecation will be more complex… | 22:32 |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 22:34 |
| rm_work | xgerman_: as long as we're on lbaas-v2.0 spec, we can't change things in a backwards-incompatible way | 22:34 |
| rm_work | the tenant_id thing actually is a little sketchy as it's a field being removed | 22:34 |
| rm_work | i'm not 100% sure how that is valid to remove without an API version change | 22:35 |
| rm_work | as much as I hated it | 22:35 |
| rm_work | and having it at this middleware layer actually will make it super easy to lift off | 22:35 |
| *** yamamoto has joined #openstack-lbaas | 22:35 | |
| xgerman_ | Being 100% backward compatible in the API and aping neutron lbaas responses are two distinct things… so I am skeptical abiuyt adding more layers to Octavia | 22:38 |
| xgerman_ | but we have more cores and I don't need to approve everything ;-) | 22:39 |
| rm_work | again, if you want to continue with the proxy that's fine too, it's just really heavyweight for what we need I think | 22:39 |
| rm_work | but if this compat layer merges, you can simplify the proxy a lot by using it | 22:40 |
| rm_work | (if you want) | 22:40 |
| *** yamamoto has quit IRC | 22:42 | |
| *** rcernin has joined #openstack-lbaas | 22:42 | |
| *** fnaval has quit IRC | 22:56 | |
| *** yamamoto has joined #openstack-lbaas | 22:58 | |
| *** yamamoto has quit IRC | 23:03 | |
| johnsom | Man, the stable/pike(ocata) gates for neutron-lbaas-dashboard hate me | 23:14 |
| *** yamamoto has joined #openstack-lbaas | 23:19 | |
| openstackgerrit | Adam Harwell proposed openstack/octavia master: Add pecan middleware for neutron-lbaas compat https://review.openstack.org/562063 | 23:23 |
| openstackgerrit | Adam Harwell proposed openstack/neutron-lbaas master: WIP: Test l7 proxy to octavia https://review.openstack.org/561049 | 23:23 |
| *** yamamoto has quit IRC | 23:24 | |
| *** atoth has quit IRC | 23:25 | |
| rm_work | forgot that the error was actually a *request* issue and not a response issue, lol | 23:25 |
| rm_work | afk a min while test runs | 23:25 |
| *** yamamoto has joined #openstack-lbaas | 23:40 | |
| *** yamamoto has quit IRC | 23:46 | |
| *** slaweq has joined #openstack-lbaas | 23:49 | |
| *** slaweq has quit IRC | 23:54 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!