Page Comparison

...

• Time: 9:00am PDT / Noon EDT

• Zoom Meeting URL: https://princeton.zoom.us/j/7739591625
  (link will launch Zoom client – if you do not have Zoom, expand the instructions below)
  

...

• Nurax Instances

  • Who needs to be in the conversation for managing Nurax instances?

  • Two conversations need to be held, one today, one at a later day

  • DCE provided the hardware resources for Nurax some years ago (2016?)

    • DCE can provide support for the hardware with funding, however, there are very limited resources for maintaining the database administration

    • Someone with knowledge of Valkyrie needs to be accessible for addressing the database administration (this is a major shift for what was previously Hyrax)

  • DCE also does not have the resources for supporting reviewing and merging changes for the code base (e. g. CSS updates), or, for creating and managing tester administrative user accounts

  • DCE has been running low on disk space on the deployment servers

  • 3 levels of support:

    • pull requests/day-to-day maintenance

    • who is in charge of the content which is ingested into Nurax? this consumes disk space - this is collection administration

    • when the server needs infrastructure changes/systems administration, this is going to require a conversation regarding whether this has been needed before

  • First level of support

    • This should be the Hyrax Tech Lead (perhaps Julie until this person is selected?)

    • Mark: Most experienced contributors can review and merge pull requests for adding/updating test users

    • The list for contributors for the Nurax code base may need to be updated

  • Second level of support

    • There should be a cleanup schedule in order to avoid unnecessary disk space consumption

    • Concept of data seeding - ideally repository should have seed data and content imports for feature testing

  • Third level of support

    • This is the level of support which is going to require a separate, focused conversation

    • nurax-pg labor is going to require significant changes for the server infrastructure and the deployment of Nurax

  • Who is currently is responsible for documenting Valkyrie architecture

    • Trey Pendragon can be contacted

  • Challenges for adopters

    • How do I deploy Hyrax installations or Hyku installations into a production environment?

    • Documentation for proposed processes in lacking, and this is currently not working with legacy Ansible Playbooks for nurax-pg

    • Updating installation steps for production deployment are going to be necessary (e. g. fcrepo no longer needs to be deployed)

  • Docker for referencing the architecture

    • Please note that Dockerfiles and Helm Charts can be used as reference points when documenting this process

• CircleCI and GitHub Actions

  • Are there plans for migrating to GitHub Actions for supporting a release for Hyrax?

  • What would be the benefit?

    • GitLab, GitHub Actions, and CircleCI are actively being used between various community members

    • Where is the existing community knowledge base strongest?

    • This may well be a longer conversation. Migrating from Travis CI to CircleCI was necessary, however, there are really only a limited number of community contributors who have a firm understanding of CircleCI

      • The knowledge of GitHub Actions might also be as limited, and there would be time required to migrate away from CircleCI

    • GitHub Actions may well also require funding for supporting complex, long-running builds for continuous integration

      • Heather has been evaluating the possible price for support in GitHub Actions

      • Collin: CircleCI Orb was required in order to ensure that generic, configurations could be repurposed

• Next Samvera Tech Call

  • Moderator: Chris

  • Notetaker: James

• Hyrax Question

  • FactoryBot seems to require a different setup where ordered_members is undefined within spec/factories/generic_works.rb:59

  • Perhaps the test order is somehow related to the this bug?

  • IiifManifestCachePrewarmJob is where the bug can be reproduced

  • Note that this also does pass in CircleCI

  • Brad clarified that this is not affected by the order in which the test suites are executed

  • This seems to have only started to become problematic with the Rails 6.0.5 upgrade

  • Chris could not identify any immediate possible solutions

• Hyrax PR Review

  • https://github.com/samvera/hyrax/pull/5669

  • This ensures that cases where Collection is not humanized into collection