From Fedora Project Wiki

This is a summary of important goals for the Fedora Engineering team for Red Hat Fiscal Year 2016 (March 2015 - February 2016).

Please note this is not a comprehensive list. Many team members have numerous additional goals for contribution to the project. In addition, all team members work with the community ad-hoc throughout the year to troubleshoot problems, mentor new contributors, and create additional opportunities for community participation.

While Red Hat employees on the team are accountable for the progress of these goals, each one is fully open to the community to participate. Contribution is essential to achieving these goals, and we want and need the community's help! In Fedora, there is no such thing as "someone else's problem." We are all part of the solution. The indicator of a lead person shows accountability for whatever work the Fedora Engineering team has committed to. It may not be the same as a person leading an overall project.

Infrastructure and General

To help with these goals, contact the Infrastructure team -- #fedora-admin[?].

Description Lead Person Target
Migrate old OpenStack cloud to new Icehouse release. This is the new long-term support release for OpenStack. We will migrate the old cloud to the new one. This is the first of a two-part goal (see below). Patrick Uiterwijk Q1
Wiki upgrade Move to the new upstream LTR of mediawiki, switch to openid plugin Patrick Uiterwijk Q1
Migrate remaining hosts to RHEL 7. While there may be a few boxes that must be maintained on RHEL 6, we want to make use of all the innovation in RHEL 7 that comes from Fedora. Migrate remaining hosts that can be moved to RHEL 7. Kevin Fenzi Q2
Migrate from Puppet to Ansible. We believe Ansible is the best new technology for systems deployment and management. So we intend to move all remaining Puppet recipes (78 at start of FY2016) in the infrastructure to Ansible playbooks. If you're interested in helping, start by exploring the Ansible docs and then get in touch with us. Kevin Fenzi Q2
Make new cloudlet ready (with 2-3 nodes) with latest OpenStack and make available for testing Patrick Uiterwijk Q2
Finish storage research. Consult with other consumers and Red Hat owners/users of storage products to ensure in the future we are buying the best solution for our needs, with a preference to fully open source products. This must be completed in advance of the next Red Hat budget cycle. Ebenezer Smooge Q3
Bring secondary arch infra into Fedora infra s390, ppc and arm secondary arches currently manage their own infra in a add hoc and less than ideal manner. Merge them into main Fedora infrastructure for common CM, updates, support and monitoring Kevin Fenzi Q3
Data-driven ansible config make our nagios and fedmsg configs dynamically generated from our inventory and group/host vars. Ralph Bean Q4
Fix current rel-eng staging environment. Identify blocking problems that prevent use of existing infra, and repair. Make sure it's up to date with current tools like ansible. Adam Miller Q1
Drive creation of new release infrastructure. Coordinate with whole Fedora infra + other teams to produce a modern build/compose infrastructure that supports Fedora.next "Rings" model. Adam Miller Q4

Application development

To help with these goals, contact the Infrastructure Applications team -- #fedora-apps[?].

Description Lead Person Alley-oop Target
Deploy HyperKitty. HyperKitty is a web front end to the new Mailman version 3 which allows users to browse topics in a more familiar, forum-like interface. We will complete development of this application and deploy for use with Fedora mailing lists. Aurélien Bompard Q1 Checkmark.png
MirrorManager 2 The code is mostly ready, we just need a few more tests and a deployment plan. Pierre-Yves Chibon Q1 Checkmark.png
python3-fedora We need to get python-fedora ready for python3. Other efforts are waiting on this. Ralph Bean Q1 Checkmark.png
anitya/hotness These currently have about 20 open bugs on them. It would be nice to at least cut that in half to reduce any broken window effect. Ralph Bean Q1 Checkmark.png
pkgdb2 admin actions New releng tools as a part of pkgdb2. This is in staging now, waiting for the next release of Fedora. Pierre-Yves Chibon Q1 Checkmark.png
Bodhi 2 Finishing up rewrite and tests in Q1, look to deploy in Q2. Luke Macken Ralph Bean Q2 Checkmark.png
fedmenu Enable all of our apps with the fedmenu popup. Ralph Bean Q2 Checkmark.png
API keys Implement central API key stuff, and give less messy CLI login. Patrick Uiterwijk Q2
SSO/SLO Add single sign on and single logout to core webapps. Patrick Uiterwijk Ralph Bean Q2
fedora-packages revamp We need to rewrite the backend which is subject to data corruption and race conditions. Ralph Bean Q3 Checkmark.png
FAS 3 Tentative, pending discussion on this plan with the team Q3
composedb This turned into PDC . Maxamillion will be taking lead in consultation with the releng group. Adam Miller Ralph Bean Checkmark.png Q4
Hubs Implementation Not entirely sure what this will entail yet, but we'll know more after collab design process gets underway. Aurélien Bompard Ralph Bean? Q4

Design and content

To help with these goals, contact the Design team -- #fedora-design[?].

Description Category Lead Person Target
Hubs design New Contributor UX Máirín Duffy
Make docs easier to contribute to New Contributor UX Ryan Lerch
<Bootstrap?> everywhere Ryan Lerch
Ask.FPO (curation & redo) User-Facing UX Ryan Lerch
Cookbook-style user guide User-Facing UX Ryan Lerch
App data expansion / curation User-Facing UX Ryan Lerch
Web presence for app data User-Facing UX Ryan Lerch
Anaconda Tweaks User-Facing UX Máirín Duffy
Spins.fpo / labs.fpo Redesign User-Facing UX Máirín Duffy
Real usability testing User-Facing UX

Kernel

To help with these goals, contact the Kernel team -- #fedora-kernel[?].

Overall department goals

  • Modernizing release process and deliverables
  • Create and improve systems, resources, and upstream efforts that benefit general users, and help draw power users into contribution
  • Improve/maintain systems/resources for existing contributors

Team High Level Goals

  • Plan for and deliver high quality kernel for upcoming Fedora releases
  • Improve team efficiency and community collaboration
  • Increase participation in upstream kernel community

Goals

Specific action items will be derived from these goals.

Description Category Lead Person Target
Retrace kernel issue reduction. The retrace server provides a view of the issues that are hitting the broadest subset of Fedora users. We'll use this data to focus our efforts on improving the kernel. Quality/Upstream
Improve kernel task automation. A number of common tasks could be automated. We'll investigate which tasks can leverage tools like fedmsg and fmn. This can range from automated nodebug and kernel-playground repos, to a more automated process for creating the actual kernel package. Efficiency/Automation
Increase upstream kernel reviews and bugfixes. Upstream is continually looking for additional patch reviews. By increasing our participation there we can help avoid bugs. We'll also use this to increase our team's overall knowledge by gaining a deeper understanding of various focus areas within the kernel. Upstream
Publish various communications about the Fedora kernel. The kernel is often a hot topic in various communities. We'll look at continuing the upstream Fedora patch reports, and possibly writing articles for Fedora Magazine. Collaboration
Kernel test system - what's next Quality/Upstream Justin Forbes

Work Items

Description Lead Person Target
Investigate pkg-git creation from exploded git tree jwb Q2
Automated nodebug builds jforbes Q2
Kernel updates in Atomic world jwb Q2/Q3
Linux-next automated testing jforbes
Update/create community focused kernel documentation/wiki labbott Q1/Q2
Focus on retrace issues and signal/noise reduction labbott Q1/Q2
Subsystem focus areas All FY
Upstream patch reviews and bugfixes All FY
Blog/magazine postings on kernel topics All FY
Release maintenance All FY

Problems

  • Retrace is flooded with kernel reports (so is bugzilla but this is always the case)
  • Our team's participation in the upstream kernel community has dwindled
    • This means we're considered outsiders more than members
    • Decreased visibility can lead to decreased responsiveness from upstream
  • The team's integration within the distro is fairly minimal
    • Good and bad
      • Good: We aren't holding anything up. The kernel ships on time and is fairly high quality overall
      • Not so good:
        • We don't necessarily improve the overall release
        • We don't ask for and discuss needs/problems other teams are having (e.g. Desktop)
  • Interaction with Red Hat stakeholders is minimal
    • Normally doesn't impact us day to day, but we could help for future items
  • Team efficiency and community participation is pretty stagnant.
    • Mostly because we haven't had a full team for a while. Let's figure this out.
    • Community members have little insight into how or why we do what we do beyond the wiki pages. Can we make it easier for them to participate?