| *** tkajinam is now known as Guest299 | 04:46 | |
| jm1 | gtema: hey :) openstacksdk's remove_floating_ip_from_server() is failing randomly in aoc ci tests. often a simple recheck solves the issue. looking at nova's api doc it looks like the action being called is deprecated, so maybe we shall mark that sdk function as deprecated as well? https://docs.openstack.org/api-ref/compute/#remove-disassociate-floating-ip-removefloatingip-action-deprecated | 07:59 |
|---|---|---|
| gtema | I will have a deeper look. Can you see any dependency whether this fails when running on certain cloud? | 08:01 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Use Neutron instead of Nova when detaching floating ips if available https://review.opendev.org/c/openstack/ansible-collections-openstack/+/869284 | 08:09 |
| jm1 | gtema: we run tests only against devstack https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_921/869167/1/check/ansible-collections-openstack-functional-devstack/9219698/job-output.txt | 08:11 |
| jm1 | gtema: error at 2023-01-04 21:23:49.821932 | 08:11 |
| opendevreview | Felix Huettner proposed openstack/openstacksdk master: Support passing a subnetpool for create_subnet https://review.opendev.org/c/openstack/openstacksdk/+/868104 | 08:12 |
| jm1 | gtema: ah, its failing now because of this: https://review.opendev.org/c/openstack/openstacksdk/+/867890 | 08:13 |
| gtema | Ok, will check. Even it is devstack things sometimes run differently depending on whether they run in ovh or rax or vex (timeouts and co) | 08:13 |
| jm1 | gtema: stephenfin's patch is fine, its just our code which is using deprecated code. submitted a patch to aoc already | 08:13 |
| gtema | But why then recheck helps, this is what disturbs me | 08:14 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Use Neutron instead of Nova when detaching floating ips if available https://review.opendev.org/c/openstack/ansible-collections-openstack/+/869284 | 08:15 |
| jm1 | gtema: forget about my recheck comment. that was about the other floating ip error we are chasing in cloud layer | 08:16 |
| jm1 | gtema: when checking ci output i was first assuming its the same random floating ip error which happens from time to time. but its a different issue | 08:18 |
| gtema | Ok | 08:19 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored security_group{,_info} modules https://review.opendev.org/c/openstack/ansible-collections-openstack/+/867740 | 09:21 |
| opendevreview | Arnaud Morin proposed openstack/ansible-collections-openstack master: Rename reboot into reboot_soft https://review.opendev.org/c/openstack/ansible-collections-openstack/+/869343 | 10:36 |
| sshnaidm | dtantsur, hi, probably https://review.opendev.org/c/openstack/openstacksdk/+/868099 broke baremetal node module and SDK,it fails in bifrost job: https://storage.gra.cloud.ovh.net/v1/AUTH_dcaab5e32b234d56b626f72581e3644c/zuul_opendev_logs_edf/869284/2/check/bifrost-keystone-collections-src/edf27b6/job-output.txt | 11:12 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored security_group_rule{,_info} modules https://review.opendev.org/c/openstack/ansible-collections-openstack/+/869355 | 12:10 |
| opendevreview | Dr. Jens Harbott proposed openstack/openstacksdk master: Revert "Add 'details' parameter to various 'find' proxy methods" https://review.opendev.org/c/openstack/openstacksdk/+/869315 | 12:46 |
| frickler | gtema: stephenfin: ^^ see sshnaidm's comment, this was shown in the non-voting jobs on that patch already, I guess we should make at least the ironic job voting? | 12:47 |
| frickler | also take better care of hard broken n-v jobs like legacy so that people do not get used to simply ignoring them | 12:48 |
| gtema | ack | 12:48 |
| frickler | if you can fix it instead of reverting, that would of course be fine, too | 12:48 |
| stephenfin | Yeah, better to fix. I'll take a look now | 12:49 |
| opendevreview | Merged openstack/ansible-collections-openstack master: Use Neutron instead of Nova when detaching floating ips if available https://review.opendev.org/c/openstack/ansible-collections-openstack/+/869284 | 12:52 |
| opendevreview | Taehyun Park proposed openstack/openstacksdk master: Add image metadef object operations https://review.opendev.org/c/openstack/openstacksdk/+/858350 | 13:04 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored compute_service_info module https://review.opendev.org/c/openstack/ansible-collections-openstack/+/869167 | 13:09 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored project{,_info} modules https://review.opendev.org/c/openstack/ansible-collections-openstack/+/869254 | 13:10 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored loadbalancer module breaking backward compatibility https://review.opendev.org/c/openstack/ansible-collections-openstack/+/867558 | 13:10 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored lb_{health_monitor,listener,member,pool} modules https://review.opendev.org/c/openstack/ansible-collections-openstack/+/867560 | 13:11 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored coe_cluster{,_template} modules https://review.opendev.org/c/openstack/ansible-collections-openstack/+/866796 | 13:11 |
| opendevreview | Stephen Finucane proposed openstack/openstacksdk master: Rework 'details' parameter for baremetal proxy https://review.opendev.org/c/openstack/openstacksdk/+/869370 | 13:45 |
| stephenfin | frickler: ^ | 13:45 |
| frickler | stephenfin: thx, I'll mostly have to defer to what the CI says though | 14:23 |
| opendevreview | Artem Goncharov proposed openstack/openstacksdk master: Move _normalize_coe_* into _coe class https://review.opendev.org/c/openstack/openstacksdk/+/865043 | 15:10 |
| lajoskatona | stephenfin, frickler: Hi, a question to think about regarding networking stadium projects CLI and python binding code | 15:50 |
| lajoskatona | stephenfin, frickler: back in December we discussed it on team meeting: https://meetings.opendev.org/meetings/networking/2022/networking.2022-12-20-14.01.log.html#l-108 | 15:50 |
| lajoskatona | stephenfin, frickler: frickler already shared his opinion but I think it is better to have the discussion here also perhaps with more people | 15:51 |
| gtema | can you please reiterate what is stadium and what you want to achieve? | 15:52 |
| lajoskatona | gtema: stadium projects for Neutron are projects that implement some networking feature/functionality with extra API that is not in core Neutron (like neutron-fwaas for Firewalls, neutron-dynamic-routing for BGP etc...) | 15:53 |
| gtema | vpnaas as well? | 15:54 |
| lajoskatona | gtema: so these are not core networking features, but now we have the CLI code (both old neutron commands and OSC plugin) in most cases in python-neutronclient repo | 15:54 |
| lajoskatona | gtema yes vpnaas is also a stadium project | 15:54 |
| gtema | okay | 15:54 |
| lajoskatona | gtema: and we already deprecated the CLI code in python-neutronclient but the python bining code is also there, and even some OSC plugins call into the python binding of neutronclient instead of SDK | 15:56 |
| gtema | ok, and so we should add those into the sdk and cli | 15:58 |
| lajoskatona | gtema: and the question is that can we put these SDK methods (like add_bgp_speaker_to_dragent) to openstacksdk or keep them in neutronclient, and the same for the CLI part, shall we add these to osc or keep them as plugins in neutronclient? | 15:59 |
| gtema | I would really like to see us starting adding those to sdk so that later also ansible modules can make use of that | 15:59 |
| gtema | and once sdk is done also osc | 15:59 |
| lajoskatona | gtema: if you can check the meeting log from December which I linked, the argument was to move only openstack network * commands, but that was just a first glance idea at that time | 16:00 |
| gtema | or alternatively keep those in neutronclient (cli part), but use sdk as base | 16:00 |
| lajoskatona | gtema: ok | 16:00 |
| gtema | from my pov it is really important to have sdk coverage | 16:01 |
| gtema | sooner or later somebody would want to use those things with ansible (as is now the case for vpn) | 16:01 |
| gtema | and having sdk support is crucial here | 16:01 |
| lajoskatona | gtema: I pushed a first wip for tap-as-a-service (https://review.opendev.org/c/openstack/tap-as-a-service/+/867786 ) its CLI code is in tap-as-a-service repo so not in neutronclient but perhaps good to see what it means | 16:02 |
| lajoskatona | if we have to do everythin outside SDK, so if we can add these methods to openstacksdk that is a big help I think | 16:02 |
| gtema | definitely move to sdk | 16:02 |
| gtema | I see no reason not to do this | 16:02 |
| gtema | cli part may be separate OSC plugin, but sdk from my pov should be having support for everything what exists in openstack | 16:03 |
| lajoskatona | gteam: sounds cool | 16:04 |
| lajoskatona | gtema: do we need to bring this to some formal meeting, or open some spec or bug or similar? | 16:04 |
| gtema | nope, just feel free to start adding directly to sdk | 16:05 |
| lajoskatona | gtema: ok, thanks, we have a neutron bug to track this effort, so I can use that for tag and in commit msg to have the context for everybody | 16:06 |
| gtema | sounds great | 16:06 |
| opendevreview | Merged openstack/openstacksdk master: Move normalize security group to sg class https://review.opendev.org/c/openstack/openstacksdk/+/865025 | 16:25 |
| opendevreview | Jakob Meng proposed openstack/ansible-collections-openstack master: Refactored coe_cluster{,_template} modules https://review.opendev.org/c/openstack/ansible-collections-openstack/+/866796 | 16:33 |
| opendevreview | Merged openstack/openstacksdk master: Move normalize_server to compute mixin https://review.opendev.org/c/openstack/openstacksdk/+/866567 | 16:47 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!