| opendevreview | Dong Ma proposed openstack/ovn-bgp-agent master: WIP: Add the support of create kubernetes resource https://review.opendev.org/c/openstack/ovn-bgp-agent/+/937457 | 08:24 |
|---|---|---|
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == [eventlet-deprecation] Change the state change WSGI server https://review.opendev.org/c/openstack/neutron/+/938404 | 08:38 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [eventlet-deprecation] Replace ``eventlet.spawn_n`` usage https://review.opendev.org/c/openstack/neutron/+/938541 | 08:42 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == [eventlet-deprecation] Change the state change WSGI server https://review.opendev.org/c/openstack/neutron/+/938404 | 09:36 |
| sahid_ | o/ | 09:39 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-ken master: hub: remove eventlet patching https://review.opendev.org/c/openstack/os-ken/+/938337 | 09:44 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: common: fix wait_until_true to support native thread https://review.opendev.org/c/openstack/neutron/+/937843 | 09:46 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: remove the usage of eventlet in the OVS agent https://review.opendev.org/c/openstack/neutron/+/937765 | 09:46 |
| opendevreview | Slawek Kaplonski proposed openstack/neutron master: Make sure that policy enforcer is initialized before use https://review.opendev.org/c/openstack/neutron/+/938640 | 10:08 |
| opendevreview | yatin proposed openstack/neutron master: [DNM] functional pyroute2 master https://review.opendev.org/c/openstack/neutron/+/938641 | 10:19 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Synchronize the network segment range initialization https://review.opendev.org/c/openstack/neutron/+/938319 | 10:34 |
| ralonsoh | folks, please check https://review.opendev.org/c/openstack/neutron/+/938319 | 10:35 |
| ralonsoh | slaweq, https://review.opendev.org/c/openstack/neutron/+/938640, some nits | 10:41 |
| opendevreview | Slawek Kaplonski proposed openstack/neutron master: Make sure that policy enforcer is initialized before use https://review.opendev.org/c/openstack/neutron/+/938640 | 11:28 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [eventlet-deprecation] Remove eventlet usage in ``BatchNotifier`` https://review.opendev.org/c/openstack/neutron/+/938648 | 11:40 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [eventlet-deprecation] Nova notifier retrieves local thread https://review.opendev.org/c/openstack/neutron/+/938651 | 11:53 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [eventlet-deprecation] Nova notifier retrieves local thread https://review.opendev.org/c/openstack/neutron/+/938651 | 11:53 |
| opendevreview | Bence Romsics proposed openstack/os-vif master: Do not add taps in trunk bridges to the dead vlan https://review.opendev.org/c/openstack/os-vif/+/923036 | 12:00 |
| slaweq | <ralonsoh> "folks, please check https://..." <- done | 13:12 |
| ralonsoh | slaweq, thanks! | 13:30 |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Install DVR src to mac flow for AAPs https://review.opendev.org/c/openstack/neutron/+/938657 | 13:35 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: WIP == [eventlet-deprecation] Remove the usage of eventlet in the Neutron API https://review.opendev.org/c/openstack/neutron/+/938659 | 13:43 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [eventlet-deprecation] Nova notifier retrieves local thread https://review.opendev.org/c/openstack/neutron/+/938651 | 13:45 |
| ralonsoh | folks, new patches related to the eventlet deprecation | 13:45 |
| ralonsoh | very small and the CI is passing | 13:45 |
| ralonsoh | * https://review.opendev.org/c/openstack/neutron/+/938648 | 13:45 |
| ralonsoh | * https://review.opendev.org/c/openstack/neutron/+/938651 | 13:46 |
| ralonsoh | thanks in advance | 13:46 |
| ralonsoh | bcafarel, hello! if you have 1 min: https://review.opendev.org/c/openstack/neutron/+/937036 | 13:48 |
| ralonsoh | thanks! | 13:48 |
| bcafarel | ralonsoh: sure, looking :) | 14:19 |
| ralonsoh | thanks | 14:19 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Create a separate file for the conntrackd HA script https://review.opendev.org/c/openstack/neutron/+/938669 | 14:59 |
| ralonsoh | frickler, slaweq ykarel_ ^ | 15:00 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Create a separate file for the conntrackd HA script https://review.opendev.org/c/openstack/neutron/+/938669 | 15:01 |
| frickler | ralonsoh: did you mean to ping fungi? because he is the one who answered on the ML. also ianal, but is including a standalone GPL file really better? | 15:05 |
| ralonsoh | frickler, sorry, my bad | 15:06 |
| fungi | yeah, i'm in the process of replying on the ml, but please see https://governance.openstack.org/tc/reference/licensing.html | 15:06 |
| ralonsoh | fungi, but this new file is not part of the Neutron source code | 15:08 |
| fungi | oh, if it's not being put in the neutron git repository then it's probably fine | 15:08 |
| ralonsoh | hmmmm no, it is inside the repo | 15:08 |
| fungi | how is something that's in the git repository not part of the source code then? | 15:08 |
| fungi | i guess i'm missing something | 15:09 |
| ralonsoh | because is not part of the python code | 15:09 |
| ralonsoh | at least this is what I understand from this: | 15:09 |
| ralonsoh | https://review.opendev.org/c/openstack/neutron/+/917430/comment/009f9f98_40027b0c/ | 15:09 |
| ralonsoh | from this reply: | 15:09 |
| ralonsoh | " Keep the Template Separate: Instead of embedding the template directly in the Python code, you could store it as a separate file (e.g., a .tpl file) and have the Python code read the file at runtime. This way, the GPL license applies only to the template file, and your Python code can remain under the Apache 2.0 license, as long as there's no direct "linking" between the Python code and the GPL code." | 15:09 |
| fungi | did that quote come from the foundation's legal counsel? | 15:10 |
| ralonsoh | no, from chatgpt | 15:10 |
| fungi | bwahahahahahahaha | 15:10 |
| ralonsoh | I know... | 15:10 |
| ralonsoh | but that's all I have! | 15:10 |
| fungi | sorry, just seems like a meme ;) | 15:11 |
| ralonsoh | In any case, I'll use https://lists.openstack.org/mailman3/lists/legal-discuss.lists.openstack.org/ to raise this question | 15:11 |
| fungi | okay, so anyway, this is why i was asking if it could be reverted and then discussed more broadly, as the tc members at a minimum are going to want to discuss the possibility of neutron shipping files in its git repository which aren't under an approved license | 15:12 |
| fungi | (which is also why i added the tc topic tag in the mail subject for that subthread) | 15:13 |
| ralonsoh | fungi, ok, I'll send a revert right now before implementing any other solution | 15:13 |
| fungi | if there's a solution that doesn't require openstack to include/distribute this file anywhere, then it side-steps the whole problem of course | 15:14 |
| ralonsoh | I'm checking this now | 15:16 |
| ralonsoh | maybe it could be possible to install the source code that contains this file, as a dependency | 15:16 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Revert "Add conntrackd support to HA routers in L3 agent" https://review.opendev.org/c/openstack/neutron/+/938670 | 15:18 |
| frickler | iiuc that governance doc also disallows GPLed external dependencies :-/ so likely someone would have to reinvent the same thing under a different license? | 15:19 |
| ralonsoh | frickler, but this is a script used by conntrackd, we can format it differently but basically needs to be as is now | 15:20 |
| slaweq | we can also potentially just add config option with path to the script which should be used and it will be up to the user to install it and provide for neutron | 15:22 |
| ralonsoh | we can also re-implement it in python, as an independent python script that makes the needed conntrackd calls | 15:24 |
| fungi | out of curiosity, why isn't that script installed by conntrackd distro packages? | 15:25 |
| ralonsoh | fungi, let me check | 15:26 |
| ralonsoh | fungi, no, this is part of the documentation | 15:28 |
| ralonsoh | https://github.com/Distrotech/conntrack-tools/blob/94ed290b11266d50f4008333fd5141193a9e4afb/doc/sync/primary-backup.sh#L29 | 15:28 |
| ihrachys | fungi: I understand it's funny but this chatgpt quote matches my understanding of how gpl works that I built before llms :p (but of course IANAL) | 15:28 |
| ihrachys | I was just too lazy / not feeling like writing in legalese so I asked my legal assistant :p | 15:29 |
| fungi | yeah, i mean, what chatgpt came up with is surprisingly on point for general gpl(v3)/apache v2 license compatibility, but misses that openstack has more stringent requirements beyond just that licenses be compatible with one another | 15:29 |
| ihrachys | yeah I haven't thought about governance aspect when I raised a flag | 15:30 |
| fungi | as an example, the zuul project (which is an openinfra project with its own governance and not part of openstack) does ship some gpl files copied from ansible in a repo alongside apache licensed code, but that situation ended up getting discussed by the foundation board of directors as to whether it was an acceptable choice (beyond mere license compatibiltiy concerns) | 15:31 |
| ihrachys | "disallows gpled external dependencies" - this can't be true, can it? I mean, maybe for python runtime linkage, but not as external process? | 15:31 |
| iurygregory | ralonsoh, slaweq, hey o/ I have a friend that would like to contribute and he saw https://bugs.launchpad.net/neutron/+bug/2092174 would you be able to provide guidance for him ? from what I see in the bug is just a matter of marking the option as deprecated for removal (but I might be wrong) =) | 15:33 |
| fungi | ihrachys: yes, that part has generally been taken to mean python dependencies where the module is imported by the project's python code, or similar linking where non-python projects are concerned | 15:33 |
| ihrachys | so in theory, we could tell users to take the script there and put it here and let distros deal with the whole package if they need to. | 15:34 |
| ralonsoh | that's slaweq's proposal | 15:34 |
| fungi | right, the main concern is that if neutron ships a copy of that file then it could violate the requirement that "The project must be licensed under Apache License, Version 2.0 (ASLv2)." | 15:34 |
| ihrachys | the script doesn't seem too complex to reimplement though | 15:35 |
| ralonsoh | I don't think so, but needs further testing | 15:35 |
| ralonsoh | in any case, right now I don't have too much time for this, maybe the author of the patch | 15:36 |
| ihrachys | is the author aware of the revert / of the issue? maybe they could do the final push. | 15:36 |
| ralonsoh | ihrachys, it is as long as I've created the revert | 15:37 |
| ihrachys | gaudenz_: ^ I assume you did conntrackd? fyi it's about to get reverted because of gpl violation. | 15:37 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-ken master: hub: prepare support for native python libraries https://review.opendev.org/c/openstack/os-ken/+/938326 | 15:37 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-ken master: hub: refactor move Event and implement native Timeout https://review.opendev.org/c/openstack/os-ken/+/938327 | 15:37 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-ken master: hub: implement threads related functions https://review.opendev.org/c/openstack/os-ken/+/938673 | 15:37 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/os-ken master: hub: use native as default implementation https://review.opendev.org/c/openstack/os-ken/+/938337 | 15:42 |
| opendevreview | Merged openstack/os-vif master: Integrate pre-commit https://review.opendev.org/c/openstack/os-vif/+/934240 | 15:45 |
| ihrachys | ok I pushed the revert to merge queue. I hope we have the original feature back in time before next release. | 15:47 |
| ralonsoh | ihrachys, thanks, I'll add this topic for the next neutron meeting | 15:48 |
| ralonsoh | in order to have more traction (== people working on this) | 15:48 |
| ihrachys | manual instructions are probably the easiest if no capacity for more this cycle. | 15:49 |
| fungi | https://packages.debian.org/sid/amd64/conntrackd/filelist indicates that at least some distros do package that script | 15:52 |
| fungi | in debian's case it's installed at /usr/share/doc/conntrackd/examples/sync/primary-backup.sh | 15:53 |
| ralonsoh | same in ubuntu 22 | 15:56 |
| ralonsoh | and in fedora in /usr/share/doc/conntrack-tools/doc/sync | 15:58 |
| greatgatsby | not sure if anyone noticed my QueuePool question yesterday. I think it's related to eventlet and SQLA. I found this commit in SQLA that to fix connection leakage with eventlet: | 16:01 |
| greatgatsby | https://github.com/sqlalchemy/sqlalchemy/commit/6221c53 | 16:01 |
| fungi | ralonsoh: ihrachys: yeah, so manual instructions could work, if you need to exercise it in ci then have the right package in bindep.txt if it's not already and copy the file to where neutron wants it. deployment projects can of course install the package and copy the script to the expected location as well | 16:01 |
| greatgatsby | this seems to be SQLA 2+ from what I can tell. I noticed SQLAlchemy>=1.4.23 in the neutron requirements.txt. Does neutron actually support SQLA 2+? Installing neutron via pip (just as a test) does in fact pull in SQLA 2+ | 16:02 |
| ralonsoh | greatgatsby, Neutron installs the latest version supported in requirements | 16:03 |
| ihrachys | greatgatsby: pretty sure we install sqla2 | 16:03 |
| ralonsoh | yes, for sure | 16:03 |
| ihrachys | there was a whole mini-project to make it work :) | 16:03 |
| ralonsoh | and I'm sure that is tested (I did that) | 16:03 |
| greatgatsby | ah, ok, thanks a lot. | 16:05 |
| ralonsoh | folks, I need to leave now. But I'll leave this here: https://review.opendev.org/c/openstack/neutron/+/938659 | 16:05 |
| ralonsoh | so far, the CI is passing! | 16:05 |
| ralonsoh | (not pep8, of course) | 16:05 |
| ralonsoh | and maybe that could improve the problems we have right now with ml2/ovn and the API | 16:06 |
| ralonsoh | I'll test that tomorrow with more APi workers (using Yatin patches) | 16:07 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: Remove dibbler code from l3 agent and elsewhere https://review.opendev.org/c/openstack/neutron/+/934283 | 16:10 |
| greatgatsby | ihrachys: sorry, quick question. I just looked in the 2024.1 neutron-server kolla image and in the /var/lib/kolla/venv SQLA is version 1.4.51. When you say you install SQLA2, where do you mean? Is there a minimum neutron version that supports SQLA2? I just want to make sure I have all my ducks in a row if I suggest upgrading SQLA in our neutron-server docker image. | 16:10 |
| ihrachys | I mean in upstream neutron tree / in neutron project CI. kolla may make their own decision that is still compatible with neutron requirements. (since we don't list >=2 as a requirement, keeping 1.4 is probably fine though I don't think we test master patches against it anymore) | 16:14 |
| ihrachys | greatgatsby: there were different fixes across different projects: https://review.opendev.org/q/topic:%22sqlalchemy-20%22 | 16:15 |
| greatgatsby | ihrachys: great, thanks again | 16:16 |
| ihrachys | greatgatsby: I don't have an exact answer to which version is sqla2+ compat but maybe this pin https://review.opendev.org/c/openstack/neutron/+/841246 can give an idea | 16:17 |
| ihrachys | pin -> bump | 16:17 |
| greatgatsby | excellent, thanks again. Just trying to determine if I should backport that specific SQLA commit I posted, or if it's safe to simply upgrade SQLA to a a recent version. This is great info, much appreciated. | 16:19 |
| opendevreview | Merged openstack/neutron stable/2024.2: [OVN] Update the LRP gateway_mtu when the external network changes https://review.opendev.org/c/openstack/neutron/+/937036 | 16:20 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: Remove dibbler code from l3 agent and elsewhere https://review.opendev.org/c/openstack/neutron/+/934283 | 16:39 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: common: fix wait_until_true to support native thread https://review.opendev.org/c/openstack/neutron/+/937843 | 16:52 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: ovs: remove the usage of eventlet in the OVS agent https://review.opendev.org/c/openstack/neutron/+/937765 | 16:52 |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Install DVR src to mac flow for AAPs https://review.opendev.org/c/openstack/neutron/+/938657 | 17:07 |
| opendevreview | Merged openstack/neutron master: [eventlet-deprecation] Remove eventlet usage in ``BatchNotifier`` https://review.opendev.org/c/openstack/neutron/+/938648 | 17:18 |
| opendevreview | Merged openstack/neutron master: Read the environment variable EVENTLET_MONKEYPATCH https://review.opendev.org/c/openstack/neutron/+/937535 | 17:18 |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Install DVR src to mac flow for AAPs https://review.opendev.org/c/openstack/neutron/+/938657 | 17:20 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: DNM Remove linuxbridge driver https://review.opendev.org/c/openstack/neutron/+/927216 | 17:22 |
| ihrachys | haleyb: was a discussion about potential removal for linuxbridge ever started? | 17:25 |
| ihrachys | (I mean, on ml; I know we had some discussion in ptg) | 17:26 |
| *** haleyb is now known as haleyb|out | 17:43 | |
| opendevreview | Vasyl Saienko proposed openstack/neutron master: Install DVR src to mac flow for AAPs https://review.opendev.org/c/openstack/neutron/+/938657 | 18:09 |
| greatgatsby | ok, I figured out why the low SQLA version ends up in the image. It's taken right from here: https://github.com/openstack/requirements/blob/stable/2024.1/upper-constraints.txt | 18:24 |
| greatgatsby | SQLA isn't bumped to 2+ until 2024.2 | 18:24 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: DNM Remove linuxbridge driver https://review.opendev.org/c/openstack/neutron/+/927216 | 18:31 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: DNM Remove linuxbridge driver https://review.opendev.org/c/openstack/neutron/+/927216 | 19:00 |
| *** jamesdenton_ is now known as jamesdenton | 19:24 | |
| opendevreview | Arnaud Morin proposed openstack/neutron master: [wip] Add test to compare time on process loop https://review.opendev.org/c/openstack/neutron/+/938694 | 19:47 |
| opendevreview | Arnaud Morin proposed openstack/neutron master: [wip] Add test to compare time on process loop https://review.opendev.org/c/openstack/neutron/+/938696 | 19:49 |
| gaudenz_ | Regarding the revert of the conntrackd support and the GPL violation. I got aware during a review in the week before christmas and was in the process of replacing this code with Python code in keepalived_state_change.py. But did tno get around to finish it before I went on vacation. I was caught a bit by surprise when it got merged before this was fixed. | 20:30 |
| ihrachys | gaudenz_: thanks for the update. so you think you'll have a fix that gets rid of the gpl code completely before EOW? | 20:31 |
| gaudenz_ | I have a fixed version more or less ready, just want to verify it actually works. I can submit it later today or tomorrow. So maybe we can still stopp the revert. | 20:31 |
| ihrachys | gaudenz_: I just removed my +2/+W from the patch | 20:32 |
| ihrachys | thanks for working on a more proper fix | 20:32 |
| jamesdenton | I've seen mention of VTEP support w/ OVN, but curious if anyone has implemented a compute or gateway node into an existing VXLAN fabric (Cisco/Arista/etc) with good results. I haven't come across any implementation details for what might be supported | 20:43 |
| opendevreview | Jakub Libosvar proposed openstack/ovn-bgp-agent master: WIP: Introduce multinode tempest job https://review.opendev.org/c/openstack/ovn-bgp-agent/+/936968 | 20:55 |
| opendevreview | Jakub Libosvar proposed openstack/ovn-bgp-agent master: WIP: Introduce multinode tempest job https://review.opendev.org/c/openstack/ovn-bgp-agent/+/936968 | 21:01 |
| opendevreview | Merged openstack/os-vif master: Do not add taps in trunk bridges to the dead vlan https://review.opendev.org/c/openstack/os-vif/+/923036 | 21:21 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!