OKD WG - Community Development Meetings

The Community Development Meeting has been folded back into the main OKD Working Group Meeting. Please see that agenda.

September 3, 2024

Attendees

  • Jaime Magiera (Emergent Community Systems)
  • Brian Innes (IBM)
  • Bruce Link (BCIT)
  • Dennis Gilmore (Red Hat)
  • Owen Howard (Ziax)
  • Andrew Kiselev (Red Hat)
  • Tyrone Wilson (IONOS)
  • Christophe Bonnaud
  • Owen Howard (Ziax)

Agenda

  • Issue/Discussion Review (All)

  • Needed Webpages/Communications

    • Move community testing page to Website (Jaime)
    • Testing FAQ (Jaime)
    • Testing Known Issues (Jaime)
    • General FAQ (Owen)
    • Road Map/Hot Topics (Brian)

  • For operators we should use the upstream branding, not the OCP branding, e.g. Tekton not Openshift (or OKD) pipelines (Dennis)

August 20, 2024

Attendees

  • Jaime Magiera (Emergent Community Systems)
  • Dennis Gilmore (Red Hat)
  • Bruce Link (BCIT)
  • Brian Innes (IBM)
  • Zed Spencer-Milnes (Ziax)
  • Michael Burke (Red Hat)
  • Andrew Kiselev (Red Hat)
  • Owen Howard (Ziax)

Agenda

  • Roles & Responsibilities

    • It's a good idea to have points of contact for various community tasks. In addition to helping communication, it will provide points for community members to congregate around.
    • Possible Roles
      • Web (Zed)
      • Testing Coordiantor (Tyrone)
      • Documentation (Zed)
        • Project Docs (TBC)
        • Product Docs (Michael B)
      • OKDerator (Owen)
      • Release Management (Jaime)
      • Triage (TBC)

  • OKDerators

    • Release criteria
      *

August 6, 2024

Attendees

  • Jaime Magiera (Emergent Community Systems)
  • Dennis Gilmore (Red Hat)
  • Tyrone Wilson (IONOS)

Agenda

July 23, 2024

Attendees

  • Jaime Magiera (Emergent Community Systems)
  • Bruce Link (BCIT)
  • Tyrone Wilson (IONOS)
  • Sumedh Deepak Patkar (NC State)
  • Zed Spencer-Milnes (Ziax)
  • Brian Innes (IBM)

Agenda

Item Description
XXXX XXXX
  • Prioritization
    • FCOS 15 - SCOS 16

July 9, 2024

Attendees

  • Jaime Magiera (Emergent Community Systems)
  • Brian innes (IBM)
  • Bruce Link (BCIT)
  • Zed Spencer-Milnes (Ziax)
  • Tyrone Wilson
  • Sumedh Deepak Patkar (NC State)
  • Owen Howard (Ziax)
  • Michael Burke (Red Hat)
  • Michael elmiko McCune (Red Hat)

Agenda

  • Community feeding back into OKD Development
    • Ideally, each row should map to a component of the OKD development process.
Item Description Subgroup Home
Centos Stream 9 Builds the Centos Stream Distro ? ?
SCOS Builds the Centos Stream CoreOS ? ?
OKD Additions to SCOS Adds layer ? ?
OKD Builds OKD ? ?
Tooling cli tools for OKD ? ?
Operators Builds the operators for OKD catalog ? ?
Documentation Write and deploy the documentation ? ?

Notes:
* suss out different SCOS uses (delivery, node)
* initial OS
* tooling
* Feeding back into SCOS
* Bug fixes
* Security patches
*https://www.reddit.com/r/CentOS/comments/1dy55hj/centos_stream_case_study_openssh_exploit/
* https://kojihub.stream.centos.org/koji/index
* CI/CD Communications/Notifications
* Can the WG receive emails from the PROW system?
* How to ultimately submit tickets?
* Initial bug is git issue
* triaged to discussion or kept as issue
*

  • Website
    • July 16th

June 25, 2024

Attendees

  • Jaime Magiera (Emergent Community Systems)
  • Dennis Gilmore (Red Hat)
  • Bruce Link (BCIT)
  • Sumedh Deepak Patkar (NC State)
  • Brian Innes (IBM)
  • Michael Burke (Red Hat)
  • Andrew Kiselev (Red Hat)
  • Zed Spencer-Milnes (Ziax)
  • Owen Howard (Ziax)

Agenda

  • Organizing Community Testing

    • New document outlining the process?

      • If yes, website or repo?

      • Testing

        • Samples
        • Community Operators
          • Generate list, break up testing
        • Drivers
        • Branding
          • Generate list of locations

      • Expected differences between OCP and OKD

      • Document lack of support for secure boot

      • Add link to operators for testing

        • Think about bugs/feature requests to providers; build relationships.
    • Where are all the places to submit feedback?
    • What else can we do to make submitting feedback as simple as possible?

  • Release cadence

    • Based on just current internal tests?
    • External real-time testing?

  • Making some nightlies available via quay.io

  • In order to move from the fcos stream to the scos stream you will need to force it. One example: oc adm upgrade –force –allow-explicit-upgrade=true –to-image=quay.io/okd/scos-release@sha256:4910fbfd72d68345846299830babf86c3ceffbd689a25d3fb67d80459632934a

  • This should result in: The cluster is updating from 4.15.0-0.okd-2024-03-10-010116 to 4.15.0-0.okd-scos-2024-01-18-223523
    (succeeded)

June 11, 2024

Attendees

  • Jaime Magiera (ICPSR - University of Michigan)
  • Brian Innes (IBM)
  • Zed Spencer-Milnes (Ziax)
  • Owen Howard (Ziax)
  • Bruce Link (BCIT)
  • Blaise Pabon (self)

Agenda

  • Video version of statement?
  • Showcasing the operator work
  • Organizing Community Testing

May 28, 2024

Attendees

  • Jaime Magiera (ICPSR - University of Michigan)
  • Bruce Link (BCIT)
  • Michael elmiko McCune (Red Hat)
  • Brian Innes (IBM)
  • Owen Howard (Ziax)
  • Blaise Pabon (self)

Agenda

  • Press release

    • https://hackmd.io/3GvNPW0sR2W1mRqxuMnqew
    • do we have enough material to warrant a regular cadence message?
    • what is the community wanting to know currently?
    • what is the state of the operator situation?
    • [bruce] have had issues doing a successful install recently, but know that red hat is also having this issue
    • [dennis] have been deep diving into coreos build process and issues that are blocking builds, would like to work out how we can move the pipeline to outiside of red hat so that it is more community oriented. have had internal discussions about what can be done to make okd successful.
      • think we need to have fresh information if we will have a regular cadence for our messaging
    • [jaime] difference between project updates and community updates, do we want more regular community updates?

  • Blocker

  • Clarifying OS components

May 14, 2024

Attendees

  • Zed Spencer-Milnes (Ziax)
  • Bruce Link (BCIT)
  • Jaime Magiera (ICPSR - University of Michigan)
  • Michael elmiko McCune (Red Hat)
  • Brian Innes (IBM)
  • Andrew Kiselev (Red Hat)
  • Owen Howard (Ziax)

Agenda

  • Meeting time
    • Feedback from Christian was that an earlier time would make it easier
    • Moving meeting time earlier would cause some existing members to have to drop out
    • Conversation around why RedHat involvement in community meetings is important
      • Helps for visibility on ongoing projects (e.g. container rebuilds, linking to relevant GH repo and JIRA tickets etc)
      • Insight on internal technical roadmap for shared OKD/Openshift (e.g moving to Konflux for component building)
      • Sign of commitment form RedHat into OKD community
      • Deliver on getting community builds
    • Decision was to not move meeting time
  • Press release
  • SCOS/Cloud SIG - This thursday!
    *
  • Co-chairs
    • There was a vote and Zed was appointed co-chair

April 30, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Bruce Link(BCIT)
  • Tyrone Wilson (IONOS)
  • Andrew Kiselev (Red Hat)
  • Dennis Gilmore (Red Hat)
  • Brian Innes (IBM)
  • Zed Spencer-Milnes (Ziax)
  • Michael elmiko McCune (Red Hat)
  • Michael Burke (Red Hat)

Agenda

  • Open discussion on the way forward for OKD

  • Press Release Bullet Points

    • Why releases have stopped
      *
    • Focusing on SCOS release (initially x86_64 only)
      • All containers will be centOS Stream based
    • Migration paths
      • Will need testing of different senarios
      • We need to know the most common setups to help ensure we test migration
    • Improve experience for independent users (e.g. bare metal)
    • There will be a re-launch
    • Community Engagement Opportunities
    • Red Hat statement of support
    • Will work in community/Hack MD/slack to curate press release ready for sign off next meeting 7th May

  • Meeting Time Survey

    • will moving the meeting forward 1 or 0.5 hours allow more people to join?
    • Will Christian be able to attend if we move?
    • Is there still a need for the 2 meetings, or should we merge?

  • Relaunch

    • Easier to find information about meetings, communication channels, etc.
    • Web site relaunch - easier to onboard new community participants: good-first-project search, etc…

  • SCOS/Cloud SIG

  • Product Docs - Remove "Support"

April 16, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Owen Howard (Ziax)
  • Dennis Gilmore (Red Hat)
  • Brian Innes (IBM)
  • Andrew Kiselev (Red Hat)
  • Christophe Bonnaud (Fermilab)
  • Blaise Pabon (Independent)
  • Michael elmiko McCune (Red Hat)
  • Bruce Link (BCIT) late

Agenda

  • Open discussion on the way forward for OKD
    • https://sylvaproject.org/ (Dennis Gilmore)
    • Become a presence at the Centos Cloud SIG 3rd Thursday of the month 15:00 UTC
    • could we find a liaison, or two, who might want to visit the centos cloud sig on behalf of the okd community?
      • this might be a nice topic to bring up in the WG meetings, it's a good way for people to get more invovled in okd and in the wider foss process.
    • Ideally we will get to the point where the community can take the code from RH and build OKD, for other OS if need be.
      • Brian: One problem is that lot gets built in Prow that we don't get to see.
    • how do we strategize around answering questions from our user community related to topics like: upgrading, installing specific versions, technical triage/debug
      • next version will be on a new distro
      • we are in the process of making some big changes, build/release-wise
      • [denis] getting everyone migrated to the scos base seems worthwhile as it will help standardize the user base and make answering questions easier
        • this would mean migrating the community, which could be problematic for some users
        • [brian] will be very disruptive if we have to announce "hey, next version you must be on scos"
          • [jaime] need a clear migration path
        • [bruce] something is vastly superior to nothing, but would be really helpful to have a way to transition a working system from one to the other. would make this a "must have" if were doing such things.
        • [bruce] might need to address the notion that scos is newer and is tested differently than fcos
        • [denis] ostree tech makes it easy to pivot the o/s (with some corner cases), older hardware can cause issues but this might be the biggest concern. from the container perspective, most of the container in scos are rhel-based. we will need to do as much testing as possible to find the corner cases, and anything we hit should be considered a bug and fixed. would expect that people should be able to migrate from fcos to scos withou much problem, aside from corner cases.
        • [jaime] when the time comes, having good documentation around this process, and related tasks, will be helpful. eg. how to do backups, how to run the upgrade, how to migrate workloads, etc.
      • [elmiko] could we plan some blog posts to help describe what is happening and explore the various topics around the migration. eg. start with an intro/high level overview, then use followups to explore details. plan the whole thing to be published over the course of a few months to keep a stream of info going.
  • Co-Chairs
    • Luigi is focusing on some other areas and will reduce his involvement
    • we will be down -1 hatter
    • is Denis a possible candidate?
      • Luigi had reached out, not sure about involvement, depends on the charter and demands
      • does align with some goals about okd
      • [denis] need to think about this, but will get back to the group
    • [jaime] we should revisit the charter and make sure that it is still applicapable and up-to-date
    • [brian] think we should put time into determining what is our relation to the source? in a similar way as fedora and centos,we should be formalizing how we percieve this from our community standpoint, and also we should coordinate with red hat to determine what this means. this is especially relevant if we ever are to think about okd as an "upstream" of ocp.
      • [jaime] think this will benefit, perhaps needs, a person from red hat being a co-chair

April 2, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Luigi Mario Zuccarelli (Red Hat)
  • Bruce Link (BCIT)
  • Dennis Gilmore (Red Hat)
  • Zed Spencer-Milnes (Ziax)
  • Owen Howard (Ziax)
  • Brian Innes (IBM)
  • Bruce Link (BCIT)

Agenda

March 19, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Brian Innes (IBM)
  • Luigi Mario Zuccarelli (Red Hat)
  • Bruce Link (BCIT)
  • Dennis Gilmore (Red Hat)
  • Zed Spencer-Milnes (Ziax)
  • Owen Howard (Ziax)
  • Michael elmiko McCune (Red Hat)

Agenda

  • Open discussion on the way forward for OKD
    • Not recorded to permit an open discussion
    • Need a community statement - ideally joint with RH on the commitment to move OKD forward
    • Short term build resolution (Prow + Doozer). Will be based on Centos Stream CoreOS.
      • Need to determine what a community build should look like
    • Need to update community charter, membership, etc.
    • Need to determine OKD relationship with OCP
    • Operator work continues

March 5, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Andrea Fasano (Red Hat)
  • Bruce Link (BCIT)
  • Michael elmiko McCune (Red Hat)
  • Brian Innes (IBM)
  • Owen Howard (Ziax)
  • Luigi Mario Zuccarelli (Red Hat)
  • Michael Burke (Red Hat)
  • Dennis Gilmore (Red Hat)

Agenda

  • [Dennis] talking about future of okd in relation to Red Hat
    • impetus: some content got shipped in okd that was not supposed to
    • how to get the build back on track with respect to ensuring what is included
    • in the short, we need to continue building okd within the red hat garden (until the issues are resolved)
    • we need to improve the build process if we want an open future
    • [jamie] what next for us, operators?
      • [elmiko] focus on day-2 seems our best option
      • [bruce] could we get some sort of list to help us ensure that we don't break this process again. eg if X is included it breaks thing, how do we ensure that X is not included.
        • [dennis] some of this is around the base images that included in every okd, the licensing/subscription agreements are not correct. this ends up being much of the base content in okd. if we can base off centos stream images then we should be fine, but it depends on the source of the operator. in addition, for operators, building the upstream projects (eg Ceph instead of ODF) would help to ensure that the wrong trademarks are not included.
        • [bruce] have run into issues when building from the upstream projects (eg istio) as the expectations are aimed at kubernetes as opposed to openshift. the differences create toil and duplication for anyone attempting to build from the upstream.
        • [brian] think we still plan to use the upstream repos but with the base images changed to something more acceptable (UBI, centos, etc).
          • [dennis] ubi would be fine, the linked repos have some rhel content but they should be acceptable for these uses.
          • [bruce] centos stream is an upstream of rhel? yes. would focusing on centos stream be preferred to ubi as there might be different (more recent) version in centos?
            • [dennis] depends on what is needed, but some of the tooling (eg RDO) is built from the centos image. there is a possibility to make RPMs available if needed. building on top of centos stream makes the most logical sense to me.
              • [jaime] +1 to building on centos stream, this also seems to align with the notion of being an openshift upstream.
          • [bruce] currently we have scos and fcos bases running on the nodes, is there anything we should know about contianers built with fcos running on an scos openshift?
            • [jaime] Vadim has said this is not a problem, there is also some conversion processes that have been hinted at as well.
            • [dennis] there is a potential issue if you run newer things on older software, eg related to libc usage and the kernel. have encountered issues with running fedora containers on rhel when there is a big mismatch between versions. easier to run older on newer as opposed to newer on older.
      • [jaime] couple key folks aren't here today, but i will reach out to them. would like to come to a consensus about the images that we build with.
        • show of hands poll indicates that centos stream would be the preferred base
        • we should reevaluate this policy as changes are updated from red hat
  • Operators
    • demo by Owen of operator pipeline work
  • Centos Cloud Sig

February 20, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Bruce Link (BCIT)
  • Brian Innes (IBM)
  • Zed Spencer-Milnes (Ziax)
  • Michael elmiko McCune (Red Hat)
  • Dennis Gilmore (Red Hat)

Agenda

February 6, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Luigi Mario Zuccareli (Red Hat)
  • B Dewayne Branch (QSBG)
  • Brian Innes (IBM)
  • Andrea Fasano (Red Hat)
  • Dennis Gilmore (Red Hat)
  • Michael elmiko McCune (Red Hat)
  • Bruce Link (BCIT)
  • Leroy Shirto (IONOS)

Agenda

January 23, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Bruce Link (BCIT)
  • Luigi Mario Zuccarelli (Red Hat)
  • Zed Spencer-Milnes (Ziax)
  • Michael Burke (Red Hat)
  • Blaise Pabon (Self)
  • Brian Innes (IBM)

Agenda

January 9, 2024

Attendees

  • Jaime Magiera (ICPSR @ University of Michigan)
  • Brian Innes (IBM)
  • Michael elmiko McCune (Red Hat)
  • Bruce Link (BCIT)
  • Owen Howard (Ziax)
  • Zed Spencer-Milnes (Ziax)
  • Michael Burke (Red Hat)
  • Blaise Pabon (Self)

Agenda

Select a repo