| opendevreview | Brian Haley proposed openstack/neutron-lib master: pyupgrade changes for Python3.9+ https://review.opendev.org/c/openstack/neutron-lib/+/934095 | 02:20 |
|---|---|---|
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] Check LSP.up status before setting the port host info https://review.opendev.org/c/openstack/neutron/+/933836 | 06:40 |
| opendevreview | liuyulong proposed openstack/neutron master: Add basical functionalities for metadata path extension https://review.opendev.org/c/openstack/neutron/+/881535 | 06:53 |
| opendevreview | liuyulong proposed openstack/neutron master: Add metadata path extension openflows https://review.opendev.org/c/openstack/neutron/+/888097 | 06:53 |
| opendevreview | liuyulong proposed openstack/neutron master: Fullstack case for metadata path https://review.opendev.org/c/openstack/neutron/+/888098 | 06:53 |
| opendevreview | liuyulong proposed openstack/neutron master: Add devstack plugin to enable ovs metadata_path https://review.opendev.org/c/openstack/neutron/+/928586 | 06:53 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron stable/2023.1: [OVN] Fix the revision number retrieval method https://review.opendev.org/c/openstack/neutron/+/934026 | 06:54 |
| ralonsoh | slaweq, hello! no rush, if you have time (once approved, I'll backport both patches) | 06:57 |
| ralonsoh | * https://review.opendev.org/c/openstack/neutron/+/933969 | 06:57 |
| ralonsoh | * https://review.opendev.org/c/openstack/neutron/+/933877 | 06:58 |
| ralonsoh | thanks! | 06:58 |
| opendevreview | liuyulong proposed openstack/neutron master: WIP: Fullstack case for metadata path with real curl https://review.opendev.org/c/openstack/neutron/+/934102 | 07:07 |
| opendevreview | Liushy proposed openstack/neutron-fwaas master: [OVN] Fix the provider error in devstack settings https://review.opendev.org/c/openstack/neutron-fwaas/+/934103 | 07:10 |
| slaweq | ralonsoh: done | 08:29 |
| ralonsoh | slaweq, thanks! | 08:29 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: [OVN] The Port_Group deletion implies the ACLs deletion https://review.opendev.org/c/openstack/neutron/+/933877 | 08:48 |
| yosef | Hi everybody, I wanna know why we cant just use openvswitch metering agent for ovn? Doesn't ovn use ovs? https://bugs.launchpad.net/neutron/+bug/2048773 | 09:33 |
| slaweq | yosef it is explained in the linked bug and in https://docs.openstack.org/neutron/latest/ovn/gaps.html as well - this agent base on what neutron-l3-agent is doing and this agent don't works with ovn backend at all | 10:14 |
| opendevreview | Merged openstack/neutron master: [OVN] Fix the revision number retrieval method https://review.opendev.org/c/openstack/neutron/+/933752 | 10:42 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: DNM - Test "neutron-ovn-tempest-*" for OVN with WSGI https://review.opendev.org/c/openstack/neutron/+/931842 | 10:46 |
| ralonsoh | bcafarel, hello! if you have some minutes: https://review.opendev.org/q/I12079de78773f7409503392d4791848aea90cb7b | 11:07 |
| ralonsoh | thanks in advance! | 11:07 |
| opendevreview | Sahid Orentino Ferdjaoui proposed openstack/neutron master: rpc/dhcp: avoid get_network_info to return segments if not needed https://review.opendev.org/c/openstack/neutron/+/880131 | 11:22 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron master: Allow net owner reader to get subnets https://review.opendev.org/c/openstack/neutron/+/900236 | 12:06 |
| opendevreview | Merged openstack/ovn-octavia-provider master: Remove Python 3.8 support https://review.opendev.org/c/openstack/ovn-octavia-provider/+/933286 | 12:58 |
| bcafarel | ralonsoh: done, thanks for putting the RevisionNumberNotDefined class in neutron first before n-lib for easy backports :) | 13:50 |
| ralonsoh | bcafarel, thanks! | 13:50 |
| * haleyb is making sure it's meeting time, always confused with clock change | 14:00 | |
| mlavalle | haleyb: it is | 14:01 |
| haleyb | #startmeeting networking | 14:01 |
| opendevmeet | Meeting started Tue Nov 5 14:01:15 2024 UTC and is due to finish in 60 minutes. The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot. | 14:01 |
| opendevmeet | Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. | 14:01 |
| opendevmeet | The meeting name has been set to 'networking' | 14:01 |
| ralonsoh | hello | 14:01 |
| elvira | o/ | 14:01 |
| haleyb | mlavalle: thanks for confirming :) | 14:01 |
| ykarel | o/ | 14:01 |
| haleyb | Ping list: bcafarel, elvira, frickler, mlavalle, mtomaska, obondarev, slaweq, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, amotoki, haleyb, ralonsoh | 14:01 |
| obondarev | o/ | 14:01 |
| mlavalle | \o | 14:01 |
| lajoskatona | o/ | 14:01 |
| frickler | \o | 14:01 |
| bcafarel | o/ | 14:02 |
| slaweq | o/ | 14:02 |
| haleyb | #announcements | 14:02 |
| rubasov | o/ | 14:02 |
| cbuggy | o/ | 14:02 |
| haleyb | #link https://releases.openstack.org/epoxy/schedule.html | 14:02 |
| haleyb | hope everyone had a good week last week, and thanks ralonsoh for running the meeting | 14:03 |
| ralonsoh | yw | 14:03 |
| haleyb | i am working on my write-up from PTG, travel and meetings kept me from it | 14:04 |
| haleyb | 2023.1 is transitioning to unmaintained, think we are all set just waiting for release patches | 14:04 |
| haleyb | Reminder: If you have a topic for the drivers meeting on Friday, please add it to the wiki @ https://wiki.openstack.org/wiki/Meetings/NeutronDrivers | 14:05 |
| haleyb | although saying that i might be out on friday, will see what topics we have beforehand | 14:06 |
| haleyb | Let's continue to use the priorities dashboard for patches in the "ready to merge" state, it has worked well so far - see wiki for link (wiki does not allow url shorteners) | 14:07 |
| lajoskatona | +1 | 14:07 |
| haleyb | any other announcements? | 14:07 |
| haleyb | #topic bugs | 14:08 |
| haleyb | bcafarel was deputy last week, his report is at | 14:08 |
| haleyb | #link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/ONVMDOUQ4RDL7GVE2YRJ4AWCPUZCNFKO/ | 14:08 |
| haleyb | thanks bcafarel! | 14:08 |
| bcafarel | np, it was a quiet week overall! | 14:08 |
| haleyb | and it looks like most everything has a patch proposed | 14:09 |
| haleyb | the only one that doesn't was for vpnaas | 14:10 |
| haleyb | #link https://bugs.launchpad.net/neutron/+bug/2086189 | 14:10 |
| haleyb | not sure if anyone from vpnaas is present and can take a look? | 14:11 |
| bcafarel | it would need some vpnaas eyes, the idea is nice but probably not a one-liner fix | 14:12 |
| lajoskatona | I can send a mail out to the ones added OVN driver for vpnaas end of week if nobody checks it | 14:12 |
| lajoskatona | agree with bcafarel | 14:12 |
| haleyb | bcafarel: right, and nothing is propbably seamless with a reboot, but should know limitations | 14:13 |
| haleyb | lajoskatona: thanks, that would work | 14:13 |
| haleyb | i did not have any other bugs to discuss, does anyone else? | 14:14 |
| haleyb | this week elvira is the bug deputy, next week is slaweq - ok with both? | 14:14 |
| ralonsoh | (both are OK) | 14:15 |
| slaweq | yeap | 14:16 |
| haleyb | ok, moving on | 14:16 |
| haleyb | #topic community goals | 14:16 |
| elvira | yes, it's ok :) | 14:16 |
| haleyb | looks like i need to update the wiki a little for this cycle, anyways... | 14:16 |
| haleyb | lajoskatona: any progress on client deprecation? | 14:17 |
| lajoskatona | no | 14:17 |
| haleyb | lajoskatona: ack, thanks | 14:18 |
| haleyb | ralonsoh: and i noticed in last week's notes you were still creating LP bugs for eventlet deprecation? | 14:18 |
| ralonsoh | yes, the main one is the WSGI migration | 14:19 |
| ralonsoh | not what we have done but the ASGI migration | 14:19 |
| ralonsoh | I'm still reading the documentation and checking what is agreed in the community | 14:19 |
| ralonsoh | but there is no opinion on this and I'm afraid of implementing a solution in Neutron that won't be supported by the community | 14:20 |
| ralonsoh | in any case, I'll add this info in the related LP bug | 14:20 |
| haleyb | i will have to catch-up with that thread | 14:20 |
| ralonsoh | that's crazy, IMO | 14:20 |
| ralonsoh | we should agree on a server and a FW | 14:20 |
| lajoskatona | ok, so it inost just my lack of knowledge that I can't see the agreed solution and way forward :-) | 14:20 |
| ralonsoh | in any case, at least for Neutron, WSGI works fine | 14:21 |
| ralonsoh | we don't have the problems that, for example, Nova has with multiple threads during the API call | 14:21 |
| ralonsoh | anyway, I'll document that in a LP bug | 14:21 |
| ralonsoh | the rest of the bugs will be created too, to make work the rest of the agents | 14:21 |
| haleyb | has there been any direction from the TC on the way forward in their meetings? | 14:21 |
| ralonsoh | no | 14:22 |
| slaweq | nothing I would be aware of | 14:22 |
| ralonsoh | but so far, as commented, we work fine with WSGI | 14:22 |
| ralonsoh | we need to "clean" the agents code and make them work without eventlet | 14:22 |
| haleyb | ack, and yes we are fine with WSGI for now | 14:23 |
| ralonsoh | that's all I have | 14:23 |
| haleyb | thanks ralonsoh | 14:23 |
| haleyb | #topic on-demand | 14:24 |
| ralonsoh | one item | 14:24 |
| ralonsoh | please check https://review.opendev.org/c/openstack/neutron/+/933836 | 14:24 |
| ralonsoh | this is fixing a recurrent issue in the CI | 14:24 |
| ralonsoh | thanks! | 14:24 |
| haleyb | ralonsoh: ack, thanks. and is your comment in the wiki old? regarding postgresql jobs, etc? i think it is | 14:25 |
| ralonsoh | ah yes | 14:25 |
| ralonsoh | so should we start deleting the jobs? | 14:25 |
| haleyb | i usually clean before meeting but this is one hour earlier than normal | 14:25 |
| ralonsoh | or removing the code in Neutron (tests, for example) | 14:25 |
| ralonsoh | no no, the comment is for this meeting | 14:26 |
| haleyb | ah | 14:26 |
| slaweq | I think we can remove periodic jobs | 14:26 |
| slaweq | and maybe keep tests in the functional tests for now | 14:26 |
| ihrachys | why would we keep them? | 14:26 |
| ralonsoh | ok to both suggestions | 14:26 |
| slaweq | and add documentation warning that this backend is not tested by us | 14:27 |
| ralonsoh | ihrachys, because FTs don't have such CI impact | 14:27 |
| slaweq | ihrachys I think it would be similar to what e.g. nova is doing | 14:27 |
| ihrachys | I don't know why nova is doing it, but I think in general, if we decide that something is not supported, then it's fair game to clean up code relevant to it (psql, linuxbridge or whatever else) | 14:28 |
| ralonsoh | that's a fair suggestion, we tend to leave leftovers in the code... | 14:30 |
| ihrachys | and if it's a capacity issue, I am happy to help with some of the cleanups (or take them over if you are busy) | 14:30 |
| slaweq | I'm fine either way, we can also remove everything related to postgresql if that's the way team wants to go :) | 14:31 |
| haleyb | i was going to ask if someone was interested in doing this cleanup? and if we should do in a single change if possible | 14:31 |
| haleyb | i would +1 removing it all, instead of leaving some cruft behind we will all forget about until it fails CI | 14:32 |
| ihrachys | imho whatever is most expedient; one piece is preferrable but if splitting simplifies life, also fine | 14:32 |
| ralonsoh | +1 to this then | 14:32 |
| lajoskatona | +1 to clean postgres things | 14:32 |
| ihrachys | ralonsoh: I read you are going to clean up but if not, let us know | 14:32 |
| ihrachys | and thank you | 14:33 |
| ralonsoh | no no, +1 to the idea of removing all | 14:33 |
| ralonsoh | this week I have no time | 14:33 |
| ihrachys | ah ok I will look into it then | 14:33 |
| ralonsoh | thanks! | 14:33 |
| haleyb | ihrachys: thanks! ping for reviews | 14:33 |
| haleyb | any other topics? | 14:34 |
| haleyb | if not it's coffee time | 14:35 |
| mlavalle | e\o/ | 14:35 |
| slaweq | o/ | 14:35 |
| haleyb | thanks for attending everyone and have a good week | 14:35 |
| obondarev | \o | 14:35 |
| haleyb | #endmeeting | 14:35 |
| opendevmeet | Meeting ended Tue Nov 5 14:35:30 2024 UTC. Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4) | 14:35 |
| opendevmeet | Minutes: https://meetings.opendev.org/meetings/networking/2024/networking.2024-11-05-14.01.html | 14:35 |
| opendevmeet | Minutes (text): https://meetings.opendev.org/meetings/networking/2024/networking.2024-11-05-14.01.txt | 14:35 |
| opendevmeet | Log: https://meetings.opendev.org/meetings/networking/2024/networking.2024-11-05-14.01.log.html | 14:35 |
| lajoskatona | o/ | 14:35 |
| ralonsoh | bye | 14:35 |
| bcafarel | o/ | 14:35 |
| ihrachys | ralonsoh: the patch is in gate, could you please check my comment before it perhaps hits the tree? https://review.opendev.org/c/openstack/neutron/+/933836/comments/5e9340c4_0d1499c1 thank you. | 14:59 |
| ralonsoh | ihrachys, sure, right now | 15:02 |
| ralonsoh | ihrachys, I'll update the patch | 15:03 |
| ralonsoh | ihrachys, replied to the question | 15:07 |
| kevko | Hi, I'm really desperate at this point , I'm using Neutron with OVN, and OVS starts spinning at 100% CPU, with the log showing entries https://paste.openstack.org/show/bnleYUdDaH5W7Ohk1LiS/ and the time just keeps increasing. The ovs-appctl command hangs, so I can’t use coverage. While ovs-vsctl show works, any attempt to change anything | 15:07 |
| kevko | causes it to freeze. If I delete OVS (including the volume) and restart the server, when I reconfigure OVS (re-create it), everything is initially fine because there are no rules in OVS. At this stage, if I remove the interface from br-ex and call reconfigure on OVN (which sets the metadata), everything is still OK. However, as soon as I put the | 15:07 |
| kevko | interface back into br-ex, it crashes again | 15:07 |
| kevko | Kolla-ansible, kolla, images ubuntu 22.04, ovs 3.3.0 , 24.03.2 | 15:07 |
| kevko | for me it looks like kind a bug in ovs or something ..anybody to help me ? :( | 15:08 |
| ihrachys | ralonsoh: question - is the order of ovsdb events guaranteed? namely, that the "another update_lsp_host_info call will be done *later*" is always true? (I mean, is it ever possible that the "another call" was already processed? | 15:13 |
| ralonsoh | kevko, I think you'll have more feedback in the OVS mail threads. But you should also provide more info than 3 log lines. Maybe you can also increase the log level and try to debug this exact moment when OVS hangs | 15:13 |
| ralonsoh | ihrachys, no, the order is not guaranteed | 15:14 |
| ralonsoh | ihrachys, the UP and DOWN events can be treated by different workers or nodes | 15:14 |
| ralonsoh | we must sync both calls using the only info that is global, that is the Neutron DB | 15:15 |
| ihrachys | I can be totally confused, but here's my line of reasoning: your set_up handler fires. it sets neutron port=UP, but leaves lsp:host-id unset because nb lsp is down; and 2) your "set_down" handler is already complete, so there won't be another chance to clean host-id up. Now you seem to have neutron port UP with no host-id set. | 15:20 |
| kevko | ralonsoh: The problem is that I'm in serious trouble, I'm running out of time, and I really don’t know what else to do. I just thought I’d give it a try here. :( | 15:22 |
| ralonsoh | ihrachys, if we receive a single event (UP or DOWN), everything is handled correctly | 15:23 |
| ralonsoh | ihrachys, the problem is when we receive two events too close. The Neutron DB update is done synchronously with the API call, but the OVN event method is handled in other thread | 15:24 |
| ralonsoh | so, in the example of two events, UP and DOWN (in this order): if we start treating the lsp host info for UP but at this time the port is already down, the call of "update_lsp_host_info" for UP won't do anything but it will the call for "update_lsp_host_info" for DOWN | 15:25 |
| ralonsoh | the same for the inverse order | 15:26 |
| ihrachys | ralonsoh: if update_lsp_host_info is doing nothing, then shouldn't the set_port_status_down that called it also "do nothing" (namely, not set neutron port to DOWN)? | 15:28 |
| ihrachys | AFAIU it will first call self._plugin.update_port_status, then call update_lsp_host_info (which will do nothing). | 15:29 |
| ralonsoh | ihrachys, we need the transition and to inform about this to Nova | 15:29 |
| ralonsoh | we need to UP->DOWN and DOWN->UP | 15:29 |
| ralonsoh | but the problem when updating the LSP host info is that we not longer have the host info (in the UP event that is processed late) | 15:30 |
| ihrachys | I am not sure in this case we should tell Nova anything; hasn't "another call" already (or about to) set it to DOWN? if so, it should stay DOWN. | 15:31 |
| ralonsoh | nova can handle the "unneeded" vif-events | 15:31 |
| ralonsoh | we also handle the provisioning blocks in the up/down calls | 15:32 |
| ihrachys | I still don't see how this patch addresses the scenario I described above where neutron port may be left UP because "another down call" is already complete (by another worker). | 15:34 |
| ralonsoh | ihrachys, the port is not left up, this is done outside this method | 15:35 |
| ralonsoh | ihrachys, the problem we also had in this method "update_lsp_host_info" is that the port binding get method has a retry decorator | 15:37 |
| ihrachys | the update_lsp_host_info method called from 2 places: set_port_status_up and set_port_status_down. these two will change neutron port status (UP/DOWN) while - in the scenario you are trying to address - leaving host info untouched - and, depending on the order of ovsdb handlers - permanently inconsistent. | 15:37 |
| ralonsoh | and we can wait too much time for this when is not actually needed | 15:37 |
| ralonsoh | ihrachys, the method does the same if "_wait_for_port_bindings_host" fails, it returns | 15:39 |
| ralonsoh | we are doing that but earlier | 15:39 |
| ihrachys | this is true | 15:40 |
| opendevreview | Rodolfo Alonso proposed openstack/neutron-lib master: Define physical and tunnelled network types https://review.opendev.org/c/openstack/neutron-lib/+/934146 | 16:04 |
| opendevreview | Merged openstack/neutron master: [OVN] Delete duplicated OVN resource type constants https://review.opendev.org/c/openstack/neutron/+/933754 | 16:05 |
| opendevreview | Lajos Katona proposed openstack/neutron-dynamic-routing master: pyupgrade changes for Python3.9+ https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/934148 | 16:19 |
| opendevreview | Lajos Katona proposed openstack/neutron-dynamic-routing master: pyupgrade changes for Python3.9+ https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/934148 | 16:28 |
| opendevreview | Lajos Katona proposed openstack/networking-bagpipe master: pyupgrade changes for Python3.9+ https://review.opendev.org/c/openstack/networking-bagpipe/+/933725 | 16:31 |
| opendevreview | Merged openstack/neutron master: [OVN] Check LSP.up status before setting the port host info https://review.opendev.org/c/openstack/neutron/+/933836 | 17:01 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: tests: always raise an explanation when status_int is wrong https://review.opendev.org/c/openstack/neutron/+/934162 | 17:43 |
| ihrachys | slaweq: ^ using your _check_http_response for all requests in the test module | 17:44 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: Remove postgresql code https://review.opendev.org/c/openstack/neutron/+/934171 | 18:40 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron-lib master: nit: Drop mentions of postgresql https://review.opendev.org/c/openstack/neutron-lib/+/934173 | 18:42 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: Remove postgresql code https://review.opendev.org/c/openstack/neutron/+/934171 | 18:49 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron-tempest-plugin master: nit: Drop mention of postgresql https://review.opendev.org/c/openstack/neutron-tempest-plugin/+/934174 | 18:51 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: Remove postgresql code https://review.opendev.org/c/openstack/neutron/+/934171 | 19:01 |
| opendevreview | Jakub Libosvar proposed openstack/neutron master: OVN metadata agent additional_chassis detection https://review.opendev.org/c/openstack/neutron/+/934188 | 21:59 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: Remove postgresql code https://review.opendev.org/c/openstack/neutron/+/934171 | 22:12 |
| opendevreview | Merged openstack/neutron master: refactor: split out code to verify a lb member into a function https://review.opendev.org/c/openstack/neutron/+/929795 | 22:26 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: tests: move test_db_base_plugin_v2.py to common https://review.opendev.org/c/openstack/neutron/+/934190 | 22:33 |
| opendevreview | Merged openstack/neutron master: refactor: construct member dict in one go https://review.opendev.org/c/openstack/neutron/+/929796 | 22:34 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: mypy: enable for most modules https://review.opendev.org/c/openstack/neutron/+/929866 | 22:57 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: gate: squash mypy into pep8 job https://review.opendev.org/c/openstack/neutron/+/934191 | 23:01 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: style: Clarify a check for mixed stateful/stateless SGs https://review.opendev.org/c/openstack/neutron/+/924087 | 23:07 |
| opendevreview | Brian Haley proposed openstack/neutron master: Enable pylint useless-super-delegation check https://review.opendev.org/c/openstack/neutron/+/934192 | 23:11 |
| opendevreview | Brian Haley proposed openstack/neutron master: Start enforcing additional pylint checks https://review.opendev.org/c/openstack/neutron/+/934193 | 23:11 |
| opendevreview | Ihar Hrachyshka proposed openstack/neutron master: DNM: run all ovs connections as leader-only=false https://review.opendev.org/c/openstack/neutron/+/921461 | 23:12 |
| opendevreview | Brian Haley proposed openstack/neutron master: Start enforcing additional pylint checks https://review.opendev.org/c/openstack/neutron/+/934193 | 23:58 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!