Tuesday, 2023-11-14

opendevreviewLucas Alvares Gomes proposed openstack/neutron master: [OVN] Enhanced external port scheduling  https://review.opendev.org/c/openstack/neutron/+/89476708:55
opendevreviewLucas Alvares Gomes proposed openstack/neutron master: [OVN] Update the External Ports documentation  https://review.opendev.org/c/openstack/neutron/+/90003008:55
opendevreviewMerged openstack/neutron stable/yoga: [DHCP agent] Add route to OVN metadata port if exists  https://review.opendev.org/c/openstack/neutron/+/90019808:56
mnederlofHi, would someone be able to review this change? https://review.opendev.org/c/openstack/neutron/+/89688309:00
opendevreviewMerged openstack/neutron master: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90064710:58
opendevreviewBartosz Bezak proposed openstack/neutron stable/2023.2: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90088111:02
opendevreviewBartosz Bezak proposed openstack/neutron stable/2023.1: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90088211:03
opendevreviewBartosz Bezak proposed openstack/neutron stable/zed: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90088311:13
opendevreviewBartosz Bezak proposed openstack/neutron stable/zed: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90088311:25
opendevreviewBartosz Bezak proposed openstack/neutron stable/yoga: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90088511:27
opendevreviewBartosz Bezak proposed openstack/neutron stable/wallaby: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90088611:29
opendevreviewRoberto Acosta proposed openstack/neutron-dynamic-routing master: WIP: Allow to configure the BGP peer connect mode  https://review.opendev.org/c/openstack/neutron-dynamic-routing/+/89950511:46
opendevreviewJayce Houtman proposed openstack/neutron master: Change exception messages to error log messages for DNS integration.  https://review.opendev.org/c/openstack/neutron/+/90021211:52
lajoskatonahaleyb: (anybody): Hi, shall I ask 2 quick reviews for taas: https://review.opendev.org/c/openstack/tap-as-a-service/+/884239 &  https://review.opendev.org/c/openstack/tap-as-a-service/+/89651512:28
lajoskatonahaleyb: I tested both, and it looks really useful12:29
haleyblajoskatona: i'll try and take a look13:13
haleybzigo: so the amd64 regression is gone, i see the arm64 one fails, but it's not using the new code, should i worry?13:14
zigohaleyb: No, you should just be patient ! :)13:15
zigoThe arm64 hosts are usually a lot slower to process new packages.13:15
haleybmy slide rule is faster :)13:15
* haleyb just wants to merge that and release a new neutron-lib13:17
zigohaleyb: It's been 20 days I'm waiting for Ceph 18.2.0 to build in mips64el: https://buildd.debian.org/status/package.php?p=ceph&suite=experimental :/13:18
zigoHopefully the new buildd machines will speed this up.13:18
haleyb8-o13:19
lajoskatonahaleyb: thanks13:44
opendevreviewJakub Libosvar proposed openstack/neutron-lib master: api-ref: Document only unbound ports vnic_type updates  https://review.opendev.org/c/openstack/neutron-lib/+/90082613:55
haleybping bcafarel, elvira, frickler, mlavalle, mtomaska, obondarev, slaweq, tobias-urdin, ykarel, lajoskatona, jlibosva, averdagu, amotoki14:00
jlibosvao/14:00
haleyb#startmeeting networking14:00
opendevmeetMeeting started Tue Nov 14 14:00:47 2023 UTC and is due to finish in 60 minutes.  The chair is haleyb. Information about MeetBot at http://wiki.debian.org/MeetBot.14:00
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.14:00
opendevmeetThe meeting name has been set to 'networking'14:00
mlavalleo/14:00
rubasovo/14:01
obondarevo/14:01
bcafarelo/14:01
haleybhi everybody14:01
lajoskatonao/14:01
frickler\o14:01
haleyb#topic announcements14:01
haleybThis week is the C-1 milestone, and library releases have started to happen14:02
haleybos-ken was released and python-neutronclient is ready to do14:02
haleybovsdbapp had a few things ready to merge and neutron-lib has some more14:03
haleyb#link https://review.opendev.org/q/project:openstack/neutron-lib14:03
haleybif there is anything you want merged for C-1 there please update it and ping for reviews, i'll be updating the release patch either tonight or tomorrow to move it along14:04
haleybi think the only thing i had was fixing some unit tests to support sqlalchemy 2 - https://review.opendev.org/c/openstack/neutron-lib/+/90055714:05
haleybit's at least passed the debian amd64 testing14:05
haleybnext announcement14:06
slaweqo/14:06
slaweqsorry for being late14:06
haleybi was going to cancel the drivers meeting this friday as i'm on pto, and next friday since it's a US holiday, unless someone wants to run them14:06
haleybslaweq: o/14:07
mlavalledo we have any proposals that my warrant a meeting?14:08
mlavalleor can we wait until december?14:08
haleybthere is nothing at https://wiki.openstack.org/wiki/Meetings/NeutronCI at the moment14:08
mlavalleIf needed I can run the meeting this coming Friday14:08
haleyband no new bugs up to today14:08
mlavallenext week's I have the same holiday as you \o/14:09
haleybmlavalle: sure, we can decide thursday based on need, thanks!14:09
mlavalleyeap14:09
lajoskatona+1 for deciding on thursday14:09
haleybthe other announcement is just with this being C-1 release week, I'm guessing there will be a neutron release proposed as well, so if there is anything you need there work on getting reviews14:12
slaweqalso, regarding releases - yoga is going to be unmaintained soon14:12
slaweqwe have many not released patches there - maybe it's worth to make one more release before this "final" step14:12
lajoskatonaelodilles sent out a mail with useful links: https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/SYPTRTPUNWZFJNSL2PEYK4NX67RQ5LLW/14:13
haleybslaweq: ack, thanks14:13
lajoskatonathe neutron etherpad from the mail: https://etherpad.opendev.org/p/neutron-stable-yoga-eom14:14
haleybi know we had gone through the stable branches last week trying to merge things, lets take a look again14:14
haleybfwiw i know someone here at canonical was going to propose releases for all the neutron stable branches, i will ping him to get that started14:15
haleybso since i'm dense, yoga will be deleted but u/v/w/x will still be around?14:18
haleybor maybe the *deleted* doesn't mean what i think14:19
lajoskatonaPerhaps ask on the release channel as I know things are fluid14:20
lajoskatonabut I just hve rumors from elodilles :-)14:20
haleybjust yoga-em ?14:20
haleybsure i'll ask14:21
haleybany other announcements?14:21
haleyb#action (haleyb) ask what *deleted* means in yoga release email14:21
haleyb#topic bugs14:22
haleyblucasagomes was the deputy last week, his report is at14:22
haleyb#link https://lists.openstack.org/archives/list/openstack-discuss@lists.openstack.org/thread/PXVY4XBS32G5RDDU5MZJTQ5PL5LIID7H/14:22
haleybnot too busy, and as of this morning only one was unassigned14:23
haleyb#link https://bugs.launchpad.net/neutron/+bug/204294114:23
haleybi think yatin is off today, but that might be related to the oslo.utils issue we saw last week14:24
haleyboh, i misread the bug14:24
haleybthey are not installing later versions?14:25
haleybis anyone more familiar with the job syntax on that to confirm?14:27
lajoskatonaI can check it14:28
haleyblajoskatona: thanks14:28
haleybmy weekly bug query shows we're down 2 bugs to 765 so moving in the right direction14:29
haleybhopefully more will get auto-closed as patches merge14:29
haleybthis week's deputy is jlibosva - are you ok doing it?14:30
jlibosvayes :)14:30
haleyband next week is obondarev14:31
obondarevo/14:31
haleybjlibosva: ack, and i see zero bugs so far :)14:31
haleybany other bugs anyone wants to discuss?14:32
haleyb#topic specs14:33
haleyb#link https://review.opendev.org/q/project:openstack%252Fneutron-specs+status:open14:33
haleybcurrently 3 open, the two BGP ones are close to merging, i just had some nits :-o14:34
haleybotherwiseguy had raised a question on the OVN interconnect one but hasn't had any follow-up yet14:36
haleyb#link https://review.opendev.org/c/openstack/neutron-specs/+/89120414:36
haleybso i'll just ask if you have an opinion to comment there14:37
haleyb#topic community_goals14:39
haleyblajoskatona: any updates?14:40
lajoskatonanothing, sorry14:41
haleyback14:41
haleybmoving on14:41
haleyb#topic on_demand14:41
haleybi had some items, but will let others talk first if they have something14:42
haleybok, in no particular order14:43
haleybOVN support for VPNaas. Does anyone have cycles to take a look?14:43
haleyb#link https://review.opendev.org/c/openstack/neutron-vpnaas/+/76535314:43
haleybit has one +2 from slaweq and a +1 from lajoskatona, it's just an XXL patch14:44
haleybsomeone on the ML asked about this and it came up at our PTG14:44
mlavallewow 4k+ lines14:44
lajoskatonaI checked and stated it in devstack, so I am fine with it basically, but not deep testing as I am not vpn master14:45
opendevreviewLuis Tomas Bolivar proposed openstack/ovn-bgp-agent master: Add initial support for local OVN cluster instead of kernel-networking  https://review.opendev.org/c/openstack/ovn-bgp-agent/+/88177914:46
haleybright, who are the contributors working on vpnaas? besides bodo and dmitriy i guess14:46
haleybwould be good to have another +1 from someone familiar with vpn14:47
mlavalleyeap14:47
lajoskatonamnaser perhaps came earlier with questions or bugs14:47
haleyb"This looks quite fair to me and it's confirmed to work by multiple operators. This also passes tempest test, so might worth having a shot and do a follow-up work whenever needed." - last comment from dmitriy14:48
mlavalleok, I'll take a stab at it over the next 2 weeks. I'll go slowly and deploy in devstack14:49
haleybif i have time i'll look through it, i guess as long as the db migration is sane, etc we can always fix the bugs later14:49
mlavalleI might ping lajoskatona for some guidance14:49
haleybmlavalle: ack, thanks14:49
lajoskatona+114:50
haleybok, my second topic is...14:50
mlavalleif by next Tuesday I feel I'm not making progress, I'll let you know during this meeting14:50
haleybIs anyone familiar with the neutron-lib API reporting tool? It has not been maintained and does not currently run14:50
haleyb#link https://review.opendev.org/c/openstack/neutron-lib/+/89900914:51
haleybhas anyone run 'tox -e api-report' in neutron-lib in the past 7 years?14:51
lajoskatonaI checked first when saw a patch around it, and now it is not working 14:52
slaweqhaleyb what is this "api-report"?14:52
slaweqit's first time I see that :)14:52
haleybslaweq: you answered my question, i had no idea it was there either14:52
haleyband it doesn't run14:53
haleyb#link https://review.opendev.org/c/openstack/neutron-lib/+/33857114:54
haleybintroduced there ^^14:54
haleybwe can either remove it, or fix the issue with the shebang and let someone try and get it working14:55
haleybbut it seems we have enough on our plates right now14:55
slaweqjust a note: it was removed from pep8 job with https://review.opendev.org/c/openstack/neutron-lib/+/35556314:56
mlavalleI say remove it14:58
haleybmlavalle: ack, just wanted a second opinion, someone can always revert and fix it14:58
lajoskatona+14:58
mlavalleyeap14:58
lajoskatona+114:58
haleybok, that patch just needs an update to remove the rest of it, will ping submittor14:58
slaweq++14:58
haleyblast topic with one minute left14:59
haleybi had mentioned review days at the PTG, is everyone good at trying it for a day or two?14:59
mlavalleyes15:00
lajoskatonayes15:00
mlavallewell, I am15:00
mlavalleI'll let others speak for themselves15:00
slaweq++15:00
haleybok, i guess that makes four of us at least :)15:01
slaweqjust please let me know when it will be so I will organize other stuff in the way to prioritize this :)15:01
haleybi was going to propose next week, but begin a short week maybe the week after?  does november 30th/dec 1st work?  or can do later15:02
mlavallehow about sending a doodle. Let's target second week of december15:02
mlavallethe week after Thanksgiving is also good15:03
haleybmlavalle: that's a good idea, will send one15:03
haleyb#action (haleyb) send a doodle about review days15:03
mlavalle+115:03
haleybok, we are over time15:03
haleybthere is no CI meeting today, will be next week i believe15:04
slaweqok, thx for the info :)15:04
mlavallenope, Yatin if off today15:04
lajoskatonathanks, just wanted to ask15:04
haleyblajoskatona: sorry, did you have a question?15:05
mlavalleno, he was asking about the CI meeting15:05
lajoskatonano, just thanks for the info that no CI meeting today15:05
mlavalleanswered15:05
haleybah, ok15:05
lajoskatonaexactly15:05
haleyb#endmeeting15:06
opendevmeetMeeting ended Tue Nov 14 15:06:04 2023 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)15:06
opendevmeetMinutes:        https://meetings.opendev.org/meetings/networking/2023/networking.2023-11-14-14.00.html15:06
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/networking/2023/networking.2023-11-14-14.00.txt15:06
opendevmeetLog:            https://meetings.opendev.org/meetings/networking/2023/networking.2023-11-14-14.00.log.html15:06
mlavalleo/15:06
haleybhave a good week o/15:06
lajoskatonao/15:06
opendevreviewMerged openstack/ovn-bgp-agent master: Add initial support for local OVN cluster instead of kernel-networking  https://review.opendev.org/c/openstack/ovn-bgp-agent/+/88177915:06
opendevreviewBrian Haley proposed openstack/neutron master: Enable liberal TCP connection tracking for SNAT namespaces  https://review.opendev.org/c/openstack/neutron/+/61820815:10
opendevreviewTakashi Kajinami proposed openstack/neutron-lib master: Remove unmaintained API report tooling  https://review.opendev.org/c/openstack/neutron-lib/+/89900915:10
opendevreviewBrian Haley proposed openstack/neutron-lib master: api-ref: Document only unbound ports vnic_type updates  https://review.opendev.org/c/openstack/neutron-lib/+/90082615:14
haleybslaweq: can you take a look at https://review.opendev.org/c/openstack/neutron-lib/+/900557 ? it seems to fix the issue based on the debian build logs15:22
opendevreviewTakashi Kajinami proposed openstack/neutron-lib master: Remove unmaintained API report tooling  https://review.opendev.org/c/openstack/neutron-lib/+/89900915:26
slaweqhaleyb done15:26
otherwiseguy@haleyb: on the ovn interconnect db-sync issue, I just can't think of any reason for the sync code to assume, unless otherwise instructed with a special flag on each potential object type, that it is free to delete it. Seems better to use the flags we already set and only touch neutron-owned stuff.15:28
otherwiseguyhistorically we've treated the ovn db like it is just ours, but back in the day there was no reason not to.15:29
otherwiseguyNow that there is the possibility of sharing the db with others, it just seems better to only touch our stuff instead of enumerating all of the possible things we might share with and looking for their signatures and creating an "ignore-this!" flag for things that don't have obvious signatures.15:30
opendevreviewJakub Libosvar proposed openstack/neutron-lib master: api-ref: Document only unbound ports vnic_type updates  https://review.opendev.org/c/openstack/neutron-lib/+/90082615:39
haleybotherwiseguy: right, it somehow seems more correct to look for things with our fingerprints whenever possible15:41
elodilleshaleyb: the unmaintained process will be something like this: 1) tag the HEAD of stable/$series ($series-eom == end-of-maintenance); 2) delete stable/$series; 3) cut unstable/$series from $series-eom tag15:42
elodilles(and tag $series-eol at some point at the HEAD of unstable/$series branch)15:46
haleybelodilles: ok, so yoga-em is created and stable/yoga goes away. i don't understand the unstable part there15:48
haleyboh, and i guess this means we should tag a 20.5.0 or such before this15:48
elodilleshaleyb: when we had EM then the last release was tagged with $series-em, now we tag the HEAD of stable/$series instead. so the release is necessary if you decide it to be important15:50
haleybelodilles: i think it's important for us downstream, let me ask our build team, just makes for easier work i think15:51
fricklerelodilles: s/unstable/unmaintained/15:51
frickler(for the branch names)15:52
elodillesfrickler: /o\ thanks15:59
haleybfrickler: i don't see any neutron branches with 'unmaintained' in their names, just tags with -em and -eol, so i'm still confused on that part16:03
opendevreviewJakub Libosvar proposed openstack/neutron master: Forbid updating vnic type on a bound port  https://review.opendev.org/c/openstack/neutron/+/89281516:04
opendevreviewJakub Libosvar proposed openstack/neutron master: Forbid updating vnic type on a bound port  https://review.opendev.org/c/openstack/neutron/+/89281516:05
haleybjlibosva: hey, can you look at this one too? https://review.opendev.org/c/openstack/neutron-lib/+/895940 - thanks in advance16:50
haleybbcafarel: can you look at https://review.opendev.org/c/openstack/neutron/+/900196 and related changes? i see the yoga one got merged but not some of the others17:19
bcafarelhaleyb: ah yes I remember looking at that chain, some backports had failing CI and missing conflicts - now looks good, let me review/approve them17:20
haleybrecheck; recheck; recheck worked :)17:20
opendevreviewMerged openstack/neutron-lib master: Change test_sqlalchemytypes.py to support sqlalchemy 2.0  https://review.opendev.org/c/openstack/neutron-lib/+/90055717:24
haleybzigo: your build looks clean and patch has merged, does that need to be backported? otherwise i'll just update the currently planned release to include that17:37
opendevreviewTakashi Kajinami proposed openstack/neutron-lib master: Remove unmaintained API report tooling  https://review.opendev.org/c/openstack/neutron-lib/+/89900917:38
haleybotherwiseguy: when you get a chance can you look at my comment in https://review.opendev.org/c/openstack/ovsdbapp/+/896686 ? trying to tag the ovsdbapp release this week, thanks!17:44
fricklerhaleyb: the switch hasn't happened yet, and yoga will be the first series which is transitioned. how to handle older branches in EM status is still to be discussed17:48
opendevreviewMerged openstack/neutron master: Remove obsolete PID files before start  https://review.opendev.org/c/openstack/neutron/+/89583218:02
opendevreviewMerged openstack/tap-as-a-service master: Support for ovs-dpdk  https://review.opendev.org/c/openstack/tap-as-a-service/+/88423919:07
opendevreviewMerged openstack/neutron stable/2023.1: [DHCP agent] Add route to OVN metadata port if exists  https://review.opendev.org/c/openstack/neutron/+/90019619:22
opendevreviewMerged openstack/neutron stable/zed: [DHCP agent] Add route to OVN metadata port if exists  https://review.opendev.org/c/openstack/neutron/+/90019719:22
opendevreviewMerged openstack/neutron stable/xena: [DHCP agent] Add route to OVN metadata port if exists  https://review.opendev.org/c/openstack/neutron/+/90019919:22
opendevreviewMerged openstack/neutron stable/wallaby: [DHCP agent] Add route to OVN metadata port if exists  https://review.opendev.org/c/openstack/neutron/+/90020019:22
opendevreviewMiro Tomaska proposed openstack/neutron master: Metadata: handle process exceptions  https://review.opendev.org/c/openstack/neutron/+/89098619:40
opendevreviewJakub Libosvar proposed openstack/neutron stable/wallaby: Revert "[OVN][Trunk] Add port binding info on subport when parent is bound"  https://review.opendev.org/c/openstack/neutron/+/90089621:05
opendevreviewJakub Libosvar proposed openstack/neutron stable/wallaby: Revert "[OVN][Trunk] Add port binding info on subport when parent is bound"  https://review.opendev.org/c/openstack/neutron/+/90089621:23
opendevreviewMerged openstack/neutron stable/wallaby: Partially revert "Do not link up HA router gateway in backup node"  https://review.opendev.org/c/openstack/neutron/+/84540623:03
opendevreviewMerged openstack/neutron stable/victoria: Partially revert "Do not link up HA router gateway in backup node"  https://review.opendev.org/c/openstack/neutron/+/84763323:03
opendevreviewMerged openstack/neutron stable/2023.1: Ensure ovn loadbalancer FIPs are centralized upon neutron restarts  https://review.opendev.org/c/openstack/neutron/+/90088223:48
haleybotherwiseguy: you still around? didn't know if you saw my comment on https://review.opendev.org/c/openstack/ovsdbapp/+/89668623:54
otherwiseguyI hadn't seen it. I'll investigate.23:55
haleybjust a question on the func test, i hate adding code without tests :)23:56

Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!