April 2025 - openstack-discuss - lists.openstack.org

[glance] Flamingo PTG summary
by Cyril Roelandt 15 Apr '25

15 Apr '25

Hi, We have just had our virtual PTG. The full list of topics, along with notes, can be found on our Etherpad[1]. Here is a summary of the topics that will be targeted during this cycle. If you would like to view one of the recordings, please reach out to us so we can figure out a way to share it. # glance-tempest-plugin work There has not been a lot of activity on the Glance Tempest Plugin for a while. We decided to put more effort into it this cycle. The first step is to decide what stale patches to abandon or rebase. We will also try to add tests for newer APIs added in the past cycles (location, quota, cache). # Eventlet removal While Glance can already be deployed in production without eventlet, this library is still required for half of our functional tests, as well as for non-core features (namely the scrubber). We will be focusing on migrating our functional tests away from eventlet during this cycle. # Image encryption Not much progress has been made on this topic in the past cycle. The original authors of this feature are willing to spend time working on the required patches during the Flamingo cycle; the Glance core dev will make it a priority to quickly review the required patches so that we can finally merge the feature. # Glanceclient chunked transfer By default, the Glance client transfers data to the Glance API in chunks. This may force the Cinder backend to resize the volume multiple times, slowing down the upload. Providing the total size of the image will allow us to fix this issue. # Cinder common configuration options Operators often have parameters set to the exact same values for multiple Cinder backends. We plan on making it possible to configure "common configuration options", thus allowing operators not to repeat configuration snippets. # Freeze glanceclient development As we approach feature parity between the Glance client and the OpenStack unified client, we believe it will be best to freeze development of the Glance client. In the future, we will only fix bugs and security issues. During this cycle, we will work on actually reaching feature parity with the unified client. # new location API (Cinder & Nova) Glance has introduced two new Location APIs in the Dalmatian cycle. We can use these APIs to address OSSN-0090 and OSSN-0065. Patches for Nova and Cinder must still be merged, hopefully during the Flamingo cycle. # freeze glance client development Cinder and Nova will need to use the OpenStack SDK instead of the glanceclient. There is no need to complete this work during this cycle, but we should at least have a good idea of what APIs are currently being used, so that we can have a plan for the next cycles. # cache-cleaner/pruner We discussed possible minor improvements to both the cache-pruner and the cache-cleaner. This work is not going to be a priority for this cycle. # Deprecate the filesystem_store_datadirs configuration option This option has not been needed since the introduction of multistore support. Happy Hacking! Cyril Roelandt [1] https://etherpad.opendev.org/p/apr2025-ptg-glance

1 0

[horizon] Launch instance: some port names not displayed
by Eugen Block 15 Apr '25

15 Apr '25

Hi *, I've been searching for an existing bug report but haven't found one yet. OpenStack Zed (Antelope 2023.1 is also affected) openstack-dashboard versions 23.0.0 and also 23.2.0 Both environments have been upgraded from Victoria relatively quick, that's why we haven't noticed it until now, but apparently there have been major workflow changes between Victoria and Wallaby how network ports are displayed in Horizon (Launch Instance). When we open the "Launch Instance" dialog in Horizon, and we want to attach an existing port (created via CLI) to the instance, only some of the ports in the list are displayed with a name, many don't have one, although all ports were created with a name via CLI (it would have failed without providing a name). A colleague of mine helped debugging and he suspects a JavaScript error, I'm translating his statement: > Some sort of cache is built, coming from an array which is supposed > to store all relevant ports. Apparently, not all ports have been > stored yet in that array in time, leading to this inconsistency. If > there's a missing "async" or "await" somewhere, it would explain > this behavior. He then points to these functions: getPortsObj, isPortsObjGenerated which are called by getNameOrId. I found those calls here: /usr/lib/python3/dist-packages/openstack_dashboard/dashboards/project/static/dashboard/project/workflow/launch-instance/networkports/ports.controller.js and here /usr/lib/python3/dist-packages/openstack_dashboard/dashboards/project/static/dashboard/project/workflow/launch-instance/networkports/ports.spec.js Is anyone aware of an existing bug report for this? I can create one, if necessary. Thanks! Eugen

2 2

[tc][all] OpenStack Technical Committee Weekly Summary and Meeting Agenda (2025.2/R-24)
by Goutham Pacha Ravi 15 Apr '25

15 Apr '25

Hello Stackers, Welcome back from a week of design confabulations. In the past couple of weeks, we moved quickly from the coordinated release of OpenStack 2025.1 "Epoxy" to the virtual Project Teams Gathering to plan the next release, OpenStack 2025.2 "Flamingo." We've got 24 weeks to go [1] to implement these plans. Now is a good time for project teams to add any project-specific deadlines to the release schedule. As PTG summaries slowly trickle in, I wanted to send out a brief note about the TC's weekly meeting that happened prior to the PTG, and our plans for the next one. === Weekly Meeting === The last weekly meeting of the OpenStack Technical Committee was held simultaneously on Zoom [2] and IRC [3] on 2025-04-01. The committee discussed security concerns with UI projects, mainly Skyline. Our attempts to independently rebuild the project failed partially because JavaScript libraries included had changed, and there were no build trackers. We didn't get an opportunity to continue this discussion at the PTG. For now, we'll try to keep the conversation on the mailing list and simultaneously explore solutions with the Skyline project team. The rest of the meeting was used to discuss CI updates, updates to the TC's tracker, and scheduling conversations for the vPTG. I'll share the minutes of the TC's PTG proceedings with you soon. In the meanwhile, please feel free to peruse the etherpad [4] or the playlist with recordings of the proceedings [5]. The next meeting of the OpenStack TC is today, 2025-04-15 at 1700 UTC. Please note the change in timing — it is an hour earlier than our past meetings used to be. This meeting will be held on IRC (OFTC's #openstack-tc channel). Please find the agenda on the meeting's wiki page [6]. I hope you'll be able to join us there. === Governance Proposals === ==== Merged ==== - Fix nova PTL IRC nick | https://review.opendev.org/c/openstack/governance/+/946175 ==== Open for Review ==== - [resolution] Extend scope of VMT to cover all project teams | https://review.opendev.org/c/openstack/governance/+/944817 - Fix Dan Smith's email | https://review.opendev.org/c/openstack/governance/+/946960 - Add aetos to telemetry project | https://review.opendev.org/c/openstack/governance/+/946744 === Upcoming Events === - 2025-04-29: StackConf https://stackconf.eu/ - 2025-04-29: SCS Hackathon / OpenStack Birthday: https://events.sovereigncloudstack.org/hackathon5-2025/ - 2025-05-13: OpenInfra Board Meeting: https://board.openinfra.org/ - 2025-05-22: OpenInfra UG Meetup, Sweden: https://www.meetup.com/openinfra-user-group-sweden/events/306139678/ Thank you very much for reading! On behalf of the OpenStack TC, Goutham Pacha Ravi (gouthamr) OpenStack TC Chair [1] 2025.2 "Flamingo" Release Schedule: https://releases.openstack.org/flamingo/schedule.html [2] TC Meeting Video Recording, 2025-04-01: https://youtu.be/HcWJY2DeLYc [3] TC Meeting IRC Log 2025-04-01: https://meetings.opendev.org/meetings/tc/2025/tc.2025-04-01-17.00.html [4] OpenStack TC's 2025.2 PTG Etherpad: https://etherpad.opendev.org/p/r.51cb4e74dd9980932d32e9a9f4e8a0ce [5] OpenStack TC's 2025.2 PTG recordings: https://www.youtube.com/playlist?list=PLhwOhbQKWT7XYzqrx21j1uizxNxnbDf5Q [6] TC Meeting Agenda, 2025-04-15: https://wiki.openstack.org/wiki/Meetings/TechnicalCommittee#Next_Meeting

1 0

[I18n] Broken ko_KR translation for OpenStack-Ansible
by Dmitriy Rabotyagov 15 Apr '25

15 Apr '25

Hey folks, Today it was brought to my attention that OpenStack-Ansible does not get any periodic updates for translations for around 2 years now. Once I've checked the result of periodic job, it seems that a ko_KR translation makes Zuul job to fail: https://zuul.opendev.org/t/openstack/build/fc90cb4be70e45e8bec07e981eaff01e Specifically, following translations are containing the error (/n in translated version, which are not present in original message): * resource-id:3db2c4ee55f9791e84ff3137e05be661 * resource-id:9f3ab9a8642b37abfcdc47804c05a49f * resource-id:8cd037448cc79d28322aeb8ec25dac58 * resource-id:182414aa5a44b5a113335bc17346b853 It would be much appreciated if somebody with access to ko_KR could fix that as soon as possible. Or, alternatively, grant me permissions to translations of the project, as to PTL. Thanks in advance!

2 2

[Magnum] Issues with Magnum (K8s 1.21.11 Clusters)
by Oliver Weinmann 14 Apr '25

14 Apr '25

Hi all, We’re running a Kolla-Ansible 2023.2 deployment with Magnum. While we primarily use CAPI, we also have some older Kubernetes 1.21.11 clusters. In these clusters, we’re encountering two specific issues: Random Crashes of openstack-cloud-controller-manager Pod In one of the older clusters, the openstack-cloud-controller-manager pod is crashing intermittently with the following error: F0410 12:02:45.673777 1 main.go:84] Cloud provider could not be initialized: could not init cloud provider "openstack": Post "http://openstack.xxxxxxx:5000/v3/auth/tokens": dial tcp 172.xxx.xxx.xxx:5000: i/o timeout However, when I manually test connectivity to the Keystone API from the master node, it appears to work fine: while true; do status=$(curl -s -o /dev/null -w "%{http_code}" http://172.xxx.xxx.xxx:5000/v3 2>/dev/null) || status="failed" echo "$(date): Keystone API status: $status" sleep 10 done Sample output: Thu Apr 10 19:19:50 UTC 2025: Keystone API status: 200 Thu Apr 10 19:20:00 UTC 2025: Keystone API status: 200 Thu Apr 10 19:20:10 UTC 2025: Keystone API status: 200 ... Despite the API being reachable, the pod continues to crash sporadically with the timeout error. "Too Many Open Files" Error in Pods Users running pods in these clusters are reporting frequent "Too many open files" errors. Since Kubernetes 1.21.11 relies on CoreOS, I’m looking for a way to permanently increase the maximum number of open file descriptors on CoreOS. Any guidance or suggestions on how to address these issues would be greatly appreciated! Best regards, Oliver

2 2

[openstack-ansible][ptg] Flamingo PTG results
by Dmitriy Rabotyagov 14 Apr '25

14 Apr '25

Hey folks, Last week along with other teams we had a PTG session where a development focus for the project was discussed. I want to thank everyone who attended the session and participated in discussions. Despite the small amount of time scheduled for the session, it was very productive and you might find the summary below. Finalizing Epoxy release As OpenStack-Ansible does have a trailing release model, we still have some time left for releasing the Epoxy, with the deadline being set at June 6, 2025. With that a few topics were highlighted to be worked on for the release to be finished: * Finalize migration for Ansible to use FQCN. While this is not obligatory as of today nor a blocker for further development - we decided to adopt our roles to the new standard introduced with ansible-core 2.10. And this topic is pretty much merged at the moment: https://review.opendev.org/q/topic:%22osa/fqcn%22 * Deprecate openstack-ansible-tests repo to avoid releasing it for Epoxy. At the moment, the only useful data in the repo is the sync tool (proposal bot). We agreed to move the content/functionality we need to the integrated repository. * Finalize migration of services to HTTPD role. At the moment, only Keystone and Horizon are left for the migration (the toughest ones). One of the challenges for the migration was highlighted the need to install/gather list of required modules for HTTPD once, to avoid service restarts needed for loading new modules vs reloads for vhosts updates. Though it still should not be a hard blocker for the migration, as this issue persists today anyway. * Linux Bridge Removal. While main patches for removal were merged: https://review.opendev.org/q/topic:%22osa/lxb_cleanup%22 we still have a Zun relying on LXB tests, which is obviously broken now. Some love needs to be given for tests to pass. * Fix compressed upload of CoreOS for Magnum Heat driver tests. As CoreOS is compressed out of the box, and Glance has no support for uploads of compressed files anymore, we need to implement decompression logic to the openstack_resources roles. Topic https://review.opendev.org/q/topic:%22osa/decompress_image%22 is created to track progress. * Document ceph source & version constraints for Noble, as there are no community repositories, only "native" one with ceph reef. Scoped for the Flamingo (2025.2): * PKI * driver for hashi-vault to be scoped for Flamingo as it requires quite a refactoring and we're too late into the release now. * Refactor standalone to create CSR and private key on the service side only * Issue certificate on the deploy host * Upgrade script to remove secret data from the deploy host * Reload libvirt config on certificates change to address VNC isuses * Shorter lifetimes * Drop Ubuntu 22.04 early in the cycle, as it is a blocker now for ansible-core 2.18 * ceph jobs are still running jammy * Add support for CentOS Stream 10 / Rocky 10 * We need to handle / create nodeset with infra to target only supported providers for tests * modular libvirt support * Use smaller nodes for molecule tests * Import Freezer role from https://github.com/noonedeadpunk/openstack-ansible-os_freezer once central scheduler support is merged to Freezer itself. * Continue coverage with Molecule tests * HAProxy role is smth we need to cover next * Complexity for TLS backends logic * IPv6/Multiple frontends * Coverage for individual roles inside of the plugins repository

1 0

[openstack][kolla-ansible][neutron] router ha was split brain when restarting l3 agent docker
by Nguyễn Hữu Khôi 14 Apr '25

14 Apr '25

Hello., I deploy a multi node cluster with 2 network nodes separately with l3_ha=true and max_l3_agents_per_router=2. It is ok but when I restart neutron_l3_agent on a host which has an active router then 2 routers become standby-standby or active-active. I feel 2 network nodes very easily to split brain. Could I have some ideas on this problem. Thank you. openstack 2025.1-master kolla-ansible2025.1 master ubuntu 22.04 Nguyen Huu Khoi

2 5

[Swift] S3/Swift compatibility: getBucketPolicy
by coolnodje＠gmail.com 14 Apr '25

14 Apr '25

I'm meeting some issues providing S3 compatible storage on OpenStack with editor solutions requiring getBucketPolicy in their S3 usage. I'm checking what can be done from the editor side, but I'd like to know if there exists a roadmap where getBucketPolicy would become supported. Thanks

1 0

OpenInfra Days Korea 2025 발표자 모집 안내
by Seongsoo Cho 14 Apr '25

14 Apr '25

안녕하세요 오픈인프라 한국 사용자모임 대표 조성수입니다. OpenInfra Days Korea 2025 발표자 모집이 시작되어 공지합니다. 미래를 여는 열쇠, 오픈인프라 ( Open Infrastructure, the Key to the Future ) 최근 클라우드 시장의 다양한 변화와 함께 오픈스택과 오픈인프라의 열기가 다시 뜨거워지고 있습니다. 특히 AI, HPC, 5G와 같은 떠오르는 첨단 기술들의 뒤에는 그 모든 가능성을 현실로 만드는 기반, 오픈인프라가 있습니다. 유연하고 확장 가능한 오픈인프라는 변화하는 기술 흐름에 민첩하게 대응하며 AI와 같은 고도화된 기술이 안정적으로 작동할 수 있도록 뒷받침합니다. 화려한 기술의 이면에서, 미래를 향한 문을 열고 있는 진짜 열쇠는 오픈인프라입니다. 지금 이 순간에도, 오픈인프라는 기술의 가능성을 넓히고, 세상을 연결하며, 더 나은 내일을 만들어가고 있습니다. OpenInfra Days Korea 2025에서 미래를 여는 열쇠를 함께 제안하고 함께 만들어보시기 바랍니다. 우리 손으로 기술의 내일을 열어가는 여정에 발표자 여러분들을 초대합니다 행사 일시 : 2025년 8월 26일 (화) 장소 : 서울 aT센터 발표자 모집 일정 : 4월 14일 (월) ~ 6월 15일 (일) 발표자 선정 안내 : 6월 30일 (월) 발표 시간 : 30분 문의 : program(a)openinfradays.kr 지원하러 가기 : https://2025.openinfradays.kr 이후 모든 공지는 https://forum.openinfra-kr.org 의 공지 게시판을 참고해주시기 바랍니다. -- Seongsoo Cho OpenStack Korea User Group / Community Leader IRC #seongsoocho

1 1

[ironic][networking][ops] PTG - Networking cases in Ironic
by Iury Gregory 11 Apr '25

11 Apr '25

Greetings everyone, During this week's PTG, we discussed networking aspects and the challenges we've encountered getting everyone on the same page. Ultimately, our biggest challenge is competing needs *and* requirements. The best way to describe the challenge is as a giant venn diagram where some interests overlap such that there are logical relationships and paths forward which can be drawn. But then there are also broadly useful themes to discuss, and requirements visit which then can also cloud the discussion. At the end of the day, the challenge is not to boil the ocean, but to focus on the distinct cases while also maybe draw the appropriate boundaries to defer specific items or have the necessary plan or idea of a plan in our back pocket before proceeding forward. So we tried bucketing in the form of hands raised during the PTG[0] to begin to give us some initial data points in an unscientific model of +1s. - Standalone operators who need non-neutron networking (mvp: tag/untag pots): +6 - Neutron Integrated Operators who are just using switches and need better existing ML2 plugin support: +3 - 3rd party SDN integration (w/neutron): +2 - Ironic additional capabilities at the front-end (dynamic portgroup assembly/weightable scheduling of specific ports): +7 If you have specific interest in any of these buckets, please reach out to Ironic contributors by replying to this mailing list post, messaging us on IRC, or even leaving some additional comments on our etherpad[0] from the PTG. That being said, the challenge is sort of how to get started and begin to make velocity. Consensus at the PTG was to sort of take a "two specific focus areas" approach and allow sub-teams to self organize. The key aspect is to get both subteams to talk to each other on a regular basis to ensure there is any coss-collaboration, reviews, and other overall awareness is highlighted for the purposes of collaboration. Discussion has also yielded that there is some consensus in trying to keep that sync as a higher bandwidth video call. Thoughts? - The Ironic Team [0]: https://etherpad.opendev.org/p/ironic-ptg-april-2025 -- *Att[]'s* *Iury Gregory Melo Ferreira * *MSc in Computer Science at UFCG* *Ironic Core* *Senior Software Engineer at Red Hat Brazil* *Social*: https://www.linkedin.com/in/iurygregory *E-mail: iurygregory(a)gmail.com <iurygregory(a)gmail.com>*

1 0