Tuesday, 2026-03-24

opendevreviewOpenStack Proposal Bot proposed openstack/openstack-ansible master: Imported Translations from Zanata  https://review.opendev.org/c/openstack/openstack-ansible/+/98185203:47
noonedeadpunkmornings07:39
noonedeadpunkso, I am slightly worried about this part of refactored role: https://github.com/adriacloud/ansible-collection-kubernetes/pull/29/changes#diff-c6995490a1fb3bcd6863e96d37408a61e8f71ff2318b5e40e2ebd8c0c9d0a24907:40
noonedeadpunkI'd really love to use systemd_service in there, but there's no way to make a dependency from collection to a standalone role07:40
noonedeadpunkbut then we'd need to figure finally out how to publish things to galaxy, and make smth like openstack.systemd collection07:41
noonedeadpunkwhere put service, networkd, mount, etc...07:42
noonedeadpunkon other note - https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/981776 seems to be fixing regression in respecting constraints for metal jobs07:43
noonedeadpunkthis one feels way better: https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/97868007:44
jrosserimho it is not the worst thing to have a standalone service setup there08:35
jrosseras you say until with have things on galaxy this is all a bit tricky08:35
noonedeadpunkjrosser: I was not sure about proxy part though... I was kinda thinking of leveraging systemd_service overrides to define the Environment'08:38
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-haproxy_server master: Remove support option httpchk  https://review.opendev.org/c/openstack/openstack-ansible-haproxy_server/+/98184610:20
gokhanhi noonedeadpunk , how can we implement multi nova cell deployment with osa? may be we can benefit from multi az deployment https://docs.openstack.org/openstack-ansible/latest/user/l3pods/multi_az.html10:40
noonedeadpunkgokhan: hey! I was actually looking at it the other day10:44
noonedeadpunkand no, I don;t think it's possible with the current code10:44
noonedeadpunkMain issue was actually logic of configuration of the super-conductor, and then I am also not sure you can do parameters templating with multi-cells, which we use to simplify logic10:45
noonedeadpunkhttps://docs.openstack.org/nova/latest/admin/cells.html#template-urls-in-cell-mappings10:46
noonedeadpunkand we use that in mapping: https://opendev.org/openstack/openstack-ansible-os_nova/src/branch/master/tasks/nova_db_setup.yml#L74-L8310:46
noonedeadpunkso no, in case of cells there should be really a quite explicit support in os_nova role for that10:47
gokhanthanks noonedeadpunk for details, we need to work on os_nova role. Firstly I will try to go with creating cells manually. 10:54
noonedeadpunkgokhan: I know that jamesdenton had a multi-cell deployment, but I don't know how they managed to do that :)10:55
noonedeadpunkgokhan: though what are you trying to solve with multi-cell?10:55
noonedeadpunkas given you're not having OVN - neutron is mainly the bottleneck anyway10:56
gokhannoonedeadpunk, jamesdenton gives information  in https://www.youtube.com/watch?v=PmHsRzNYFkM&t=133s10:59
noonedeadpunkI was there in the room :D11:01
noonedeadpunkBut I can't recall any practical details about how11:01
gokhanThe main reason using cell is to establish fault domains. We want to decouple the message bus and DB per segment to prevent a single point of failure from affecting the entire deployment. Cells are  in different pods in dc. 11:19
gokhanI was also there :) 11:19
noonedeadpunkso what about neutron?:)11:25
noonedeadpunkand then superconductor anyway has to talk to each "regional" conductor11:26
gokhannoonedeadpunk, you're right about the Neutron bottleneck. In multicell deployment I will plan to use ovn with Chassis-per-cell (or per-pod) mapping. This can allow us to localize the network control plane traffic and align it with our physical fault domains11:30
noonedeadpunkbut you can't really isolate traffic or tenant networks in segments today afaik11:31
noonedeadpunkyou can build strucutre of sb/nb DBs to spread them across pods11:32
noonedeadpunkbut still all networks gonna be "available" everywhere... Unless I am missing smth11:32
noonedeadpunkthough OVN puts you in a way better place11:33
noonedeadpunkEventually, we neither have ability to build tiering of nb/sb dbs, but this is another thing I really wanted to do, though -ENOTIME11:33
gokhanThat makes sense. Regarding the network visibility/isolation issue you mentioned, do you think implementing Multi-AZ (Availability Zones) combined with Routed Provider Networks could act as a viable workaround for now? In theory, mapping specific AZs to our physical Pods and using routed segments should prevent a single tenant network from spanning across the entire deployment. Do you see any major pitfalls with this approach in a multi-cell 11:36
gokhancontext?11:36
noonedeadpunkafaik AZ do not isolate networks still inside of the AZ11:37
noonedeadpunkit's more of set of nodes where routers will be binded to11:37
noonedeadpunkbut in no way prevent to bind ports from the network on whatever compute you tell it to bind11:37
opendevreviewDamian DÄ…browski proposed openstack/openstack-ansible-os_nova master: Add hashi_vault pki backend support  https://review.opendev.org/c/openstack/openstack-ansible-os_nova/+/94942611:40
gokhanou're absolutely right. Networking remains a significant challenge, and achieving true fault domain isolation there seems out of reach for now without more advanced new OVN features11:42
gokhannoonedeadpunk, given the networking limitations we discussed, it sounds like you'd find a Multi-region deployment more suitable for achieving true fault domain isolation?11:45
noonedeadpunk[e]Well... It is, but then you have really isolated networks 😅11:46
noonedeadpunk[e]But az was never designed to be fault domain isolation11:46
noonedeadpunkso indeed, I guess only multi-region with shared keystone would provide true isolation and fault-tolerance11:51
gokhanyes, you're right 12:02
opendevreviewDamian DÄ…browski proposed openstack/openstack-ansible-os_aodh master: Add hashi_vault pki backend support  https://review.opendev.org/c/openstack/openstack-ansible-os_aodh/+/98159912:10
opendevreviewourdia sadia proposed openstack/openstack-ansible master: doc: Overriding Python dependencies for services in OpenStack-Ansible  https://review.opendev.org/c/openstack/openstack-ansible/+/98126613:37
noonedeadpunkhuh, interesting issue for Amphora in Debian 1314:29
noonedeadpunkFailure: octavia.amphorae.driver_exceptions.exceptions.AmpConnectionRetry: Could not connect to amphora, exception caught: HTTPSConnectionPool(host='172.29.232.185', port=9443): Max retries exceeded with url: / (Caused by SSLError(SSLCertVerificationError(1, '[SSL: CERTIFICATE_VERIFY_FAILED] certificate verify failed: Missing Authority Key Identifier14:29
noonedeadpunkhttps://zuul.opendev.org/t/openstack/build/adf036018f6f416296d0328158597b47/log/logs/host/octavia-worker.service.journal-09-53-21.log.txt#2913-296414:29
damiandabrowskia lot of POST_FAILURES today in CI :/14:59
damiandabrowskihttps://zuul.opendev.org/t/openstack/builds?project=openstack%2Fopenstack-ansible&result=POST_FAILURE&skip=014:59
noonedeadpunk#startmeeting openstack_ansible_meeting15:02
opendevmeetMeeting started Tue Mar 24 15:02:01 2026 UTC and is due to finish in 60 minutes.  The chair is noonedeadpunk. Information about MeetBot at http://wiki.debian.org/MeetBot.15:02
opendevmeetUseful Commands: #action #agreed #help #info #idea #link #topic #startvote.15:02
opendevmeetThe meeting name has been set to 'openstack_ansible_meeting'15:02
noonedeadpunk#topic rollcall15:02
noonedeadpunko/15:02
anfimovirhello 15:02
chubinidzedr_Hi15:02
-opendevstatus- NOTICE: Recent POST_FAILURE job results with no logs were due to upload errors in one of our providers, which has been temporarily disabled now so rechecking those should be safe15:03
damiandabrowskihi!15:03
noonedeadpunkcourtesy ping jrosser15:04
noonedeadpunk#topic CI status15:04
noonedeadpunkExcept POST_FAILURE which are apparently an infra outage, things not looking great when I look at PKI patches specifically15:05
noonedeadpunkThough I believe, that a root cause is likely in python_venv_build for many of them15:05
noonedeadpunk#link https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/98177615:05
noonedeadpunkBut then telemetry looks completely broken as well15:06
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-os_gnocchi master: Use Direct URL syntax for pip packages  https://review.opendev.org/c/openstack/openstack-ansible-os_gnocchi/+/97987715:07
noonedeadpunkat very least gnocchi does require usage of direct url format instead of egg already15:07
noonedeadpunk#link https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/98114815:08
noonedeadpunkAnd I think we should try to switch to this format sooner then later15:08
noonedeadpunkpatches are quite trivial, thankfully15:08
damiandabrowskicool, thanks for pushing fixes!15:09
noonedeadpunkI'd suggest to use the topic `osa/pep-508` to track things15:09
noonedeadpunkbut some help there would be appreciated :)15:09
noonedeadpunk#topic releasing15:10
noonedeadpunkso we're approaching to the coordinated release date, at which it would be amazing to have a beta release15:10
noonedeadpunkthough realistically, I don't see how we can manage to do that 15:11
noonedeadpunkanyway, I am planning to propose at very least switch to track 2026.1 branch for all projects15:11
noonedeadpunkand "fix" the tempest plugins and versions15:11
noonedeadpunkand try our best to get to somewhat stable state15:12
noonedeadpunkregarding release for stable branches - we have a bunch of backports to review15:12
noonedeadpunkit's been a while since our last release15:12
noonedeadpunkand also 2025.2 still does not have X.1 release...15:12
noonedeadpunkso I'd really try to do this first15:13
noonedeadpunk#link https://review.opendev.org/q/parentproject:openstack/openstack-ansible+branch:%5Estable/.*+status:open+15:13
anfimovir++ need finish with 2025.1/2025.215:13
noonedeadpunkanfimovir: do you know if all "important" backports are at least done?15:14
noonedeadpunks/done/proposed/?\15:14
anfimovirNo, one moment..15:14
anfimovirhttps://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/981720 / https://review.opendev.org/c/openstack/openstack-ansible-os_cinder/+/980168 / https://review.opendev.org/c/openstack/ansible-role-systemd_mount/+/979485 / https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/98163515:15
anfimovirafter this we can create new release for 2025.115:16
anfimovirFor 2025.2 same...15:16
noonedeadpunkok, but we are not expecting anything to merge on master?15:16
anfimovirThis one - https://review.opendev.org/c/openstack/ansible-role-systemd_mount/+/971356?usp=cherry-pick15:17
anfimovirNeed back W after backport15:17
anfimovirrebase*15:17
noonedeadpunkaha, ok, set it15:17
noonedeadpunksorry, I am missing the gerrit pings quite frequently 15:18
anfimovirnp :)15:18
anfimovir@damiandabrowski may be you can find time for it15:18
damiandabrowskisure, i'll have a look15:19
anfimovirthank you15:19
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: Remove deprecated HSTS Policies  https://review.opendev.org/c/openstack/openstack-ansible/+/98123215:19
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: Remove openstack-ansible from exclude_project  https://review.opendev.org/c/openstack/openstack-ansible/+/98099215:20
NeilHanlonah crap... i got distracted 👀15:20
NeilHanlonhello :) i'm here lol15:20
noonedeadpunkAlso I have recalled about old patch regarding contributors docs refactoring:15:20
noonedeadpunk#link https://review.opendev.org/c/openstack/openstack-ansible/+/96893215:20
noonedeadpunkNeilHanlon: heya! Nice to see you around!15:20
NeilHanlon:) hoiwzit15:22
anfimovirOps, guys, sorry, forgot about... https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/98110015:23
noonedeadpunkeventually, there's an interesing "backport candidate"15:23
noonedeadpunk#link https://review.opendev.org/c/openstack/openstack-ansible-os_skyline/+/98114915:23
noonedeadpunkI think we do a minor release without it ofc15:23
noonedeadpunkbut eventually skyline yarn builds are broken without it for sure15:23
noonedeadpunkI am not sure if they fully work with it though either...15:24
anfimovirYes, in octavia same fix available too ;))15:24
noonedeadpunkAt least it worked for me in multi-node AIO15:24
noonedeadpunkanfimovir: which one do you mean?15:24
noonedeadpunk(octavia seems to be utterly broken right now)15:25
anfimovirhttps://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/981135?usp=search15:25
noonedeadpunkSo, I looked at the related bug you created, and it kinda feels like that except tons of errors in logs, it's "fine"....15:26
noonedeadpunkI think we can wait with this one as well for the next point release15:26
noonedeadpunkotherwise we never make it :D15:26
noonedeadpunk#topic office hours15:27
noonedeadpunkI have one proposal to discuss15:27
noonedeadpunkAlready multiple teams have added an excalated privilege group in Gerrit, which allows users within this group to do Code Review + 2, but not set +Workflow15:28
noonedeadpunkAnd I'd like to propose to add such group for OSA as well15:28
noonedeadpunkWe can name it Power Reviewers or smth....15:28
noonedeadpunkThe idea behind it - so we could have another step between regular reviewers and core team15:29
noonedeadpunkand this way "recognize" contributions and direction of where they are going15:30
jrossero/ hello15:30
noonedeadpunkto keep ppl motivated and engaged15:30
noonedeadpunkat the same time keeping the "safeguard" of merging not ready things by absent +W15:31
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible master: Migrate k8s cluster from osa ops into main tree  https://review.opendev.org/c/openstack/openstack-ansible/+/97527715:32
opendevreviewDmitriy Rabotyagov proposed openstack/openstack-ansible-os_magnum master: Migrate capi driver installation from osa-ops  https://review.opendev.org/c/openstack/openstack-ansible-os_magnum/+/97626115:34
jrosseragree with the extra reviewer tier15:34
jrossersounds good and will help us move stuff forward15:35
noonedeadpunkNeilHanlon: damiandabrowski any opinions on that?15:35
* noonedeadpunk want to engage all cores on the meeting to this topic :)15:35
damiandabrowskifine for me15:35
jrosserit is still though the case that we have a lot of subtlety about why things are like they are, and this takes quite the understanding15:36
jrosserso the concept is good15:36
noonedeadpunkyeah, we have plenty of history, and $reasons indeed15:37
noonedeadpunkhow would we name it? I see ironic had `approvers` and kolla `reviewers`15:37
noonedeadpunkI don't like reviewers, as each reviewer is important15:37
noonedeadpunkand I do look and value +1, so don't want to downplay that part15:38
jrosserhardest computing problem - names :/15:39
noonedeadpunkright :D15:39
noonedeadpunkwell... I wonder if this problem is solvable with LLMs....15:40
noonedeadpunkMaybe they're good at least for smth...15:40
noonedeadpunkit come up with either approvers or senior reviewers... meh15:41
noonedeadpunkmaybe approvers not that bad and will align with ironic at least...15:42
jrosserit is descriptive too15:43
jrosseras anyone can review15:43
noonedeadpunkright15:43
noonedeadpunkok, agreed?15:44
jrosser`approvers` is ok with me15:44
noonedeadpunk#agreed create a new group named openstack-ansible-approvers with +2 code review privileges on top of regular reviewers role15:45
noonedeadpunkbtw15:47
noonedeadpunkdo we want to add also Backport Candidate label?15:47
noonedeadpunkI think it would be reasonable thing to do?15:47
noonedeadpunk*add privileges for backport candidate label to approvers?15:48
damiandabrowskiyeah, makes sense15:53
noonedeadpunkI'm looking not in more details on Ironic ACL, and it seems that approvers there prety much == cores15:53
noonedeadpunkand they have also `ironic-reviewers` which is seemingly what we wanna name as approvers15:53
noonedeadpunkBut I don't like naming it as reviewers...15:55
noonedeadpunkso yeah, sticking with the decision :)15:55
noonedeadpunk(and they do have backport candidate as well)15:55
noonedeadpunkgiven that I will be asking for permissions changes15:56
noonedeadpunkShould I also get ability to remove WIP state for Cores?15:56
damiandabrowskiyou as PTL wants to remove WIP state for patches uploaded by Cores?16:00
damiandabrowskiI thought it's already possible :D don't have any problem with that16:01
noonedeadpunkno, it's not possible16:02
noonedeadpunkalso it's not only about cores, but about any other user16:02
noonedeadpunkso if somebody uploads a patch in a WIP state and never removes the label - nobody except of this user (or gerrit admins) have ability to remove the WIP16:02
noonedeadpunkand I wanna grant cores ability to do that16:03
anfimovirYes, for cores this need.16:03
noonedeadpunk++16:04
noonedeadpunk#endmeeting16:04
opendevmeetMeeting ended Tue Mar 24 16:04:28 2026 UTC.  Information about MeetBot at http://wiki.debian.org/MeetBot . (v 0.1.4)16:04
opendevmeetMinutes:        https://meetings.opendev.org/meetings/openstack_ansible_meeting/2026/openstack_ansible_meeting.2026-03-24-15.02.html16:04
opendevmeetMinutes (text): https://meetings.opendev.org/meetings/openstack_ansible_meeting/2026/openstack_ansible_meeting.2026-03-24-15.02.txt16:04
opendevmeetLog:            https://meetings.opendev.org/meetings/openstack_ansible_meeting/2026/openstack_ansible_meeting.2026-03-24-15.02.log.html16:04
noonedeadpunkI've pushed patches16:08
noonedeadpunk#link https://review.opendev.org/c/openstack/project-config/+/98192416:08
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-plugins stable/2025.1: Pin setuptools via global-requirement-pins.txt  https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/97611516:50
opendevreviewMerged openstack/ansible-role-systemd_mount master: Do not try to reload mount on absent state  https://review.opendev.org/c/openstack/ansible-role-systemd_mount/+/97135617:08
opendevreviewMerged openstack/ansible-role-systemd_mount master: Remove unused Makefile  https://review.opendev.org/c/openstack/ansible-role-systemd_mount/+/97871317:08
opendevreviewMerged openstack/openstack-ansible-os_neutron stable/2025.2: Ensure OVN logs to the journald  https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/98163417:08
opendevreviewMerged openstack/openstack-ansible-os_neutron stable/2025.1: Ensure OVN logs to the journald  https://review.opendev.org/c/openstack/openstack-ansible-os_neutron/+/98163517:11
opendevreviewMerged openstack/ansible-role-python_venv_build stable/2025.1: Cleanup environment on wheel build failure  https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/98110018:12
opendevreviewMerged openstack/ansible-role-python_venv_build stable/2025.2: Cleanup environment on wheel build failure  https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/98109918:12
opendevreviewMerged openstack/openstack-ansible-os_horizon stable/2025.1: Ensure translation handlers run after pull_catalog  https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/98172018:23
opendevreviewMerged openstack/openstack-ansible-os_horizon stable/2025.2: Ensure translation handlers run after pull_catalog  https://review.opendev.org/c/openstack/openstack-ansible-os_horizon/+/98171918:45
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-plugins stable/2025.1: DNM: CI status  https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/98198818:53
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-plugins stable/2025.1: DNM: CI status  https://review.opendev.org/c/openstack/openstack-ansible-plugins/+/98198818:54
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning stable/2025.1: CI check status  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199219:33
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning master: Remove unused Makefile  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199319:35
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning master: Remove unused Makefile  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199319:35
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning master: Update outdate information in bindep.txt  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199419:36
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning master: Update outdate information in bindep.txt  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199419:37
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning stable/2025.1: Pin setuptools via global-requirement-pins.txt  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199219:50
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning stable/2025.1: Pin setuptools via global-requirement-pins.txt  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199219:51
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning stable/2025.1: Pin setuptools via global-requirement-pins.txt  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98199220:06
opendevreviewMerged openstack/ansible-role-python_venv_build master: Stop removing first/last lines in constraints.txt  https://review.opendev.org/c/openstack/ansible-role-python_venv_build/+/98177620:16
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning master: Symlink the role into tests  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98200620:24
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning master: Symlink the role into tests  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98200620:24
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-apt_package_pinning master: Symlink the role into tests  https://review.opendev.org/c/openstack/openstack-ansible-apt_package_pinning/+/98200620:26
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-os_manila master: CI: start use AIO for prevent problems with Ceph-Ansible  https://review.opendev.org/c/openstack/openstack-ansible-os_manila/+/97187820:27
opendevreviewTakashi Kajinami proposed openstack/openstack-ansible-os_manila master: Avoid leaking internal url for authentication error  https://review.opendev.org/c/openstack/openstack-ansible-os_manila/+/97602720:28
opendevreviewIvan Anfimov proposed openstack/openstack-ansible-os_octavia master: Rename Octavia groups to remove dashes  https://review.opendev.org/c/openstack/openstack-ansible-os_octavia/+/97868020:31

Generated by irclog2html.py 4.1.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!