2025-06-27 Hyrax Fedora 6 Working Group

2025-06-27 Hyrax Fedora 6 Working Group

 Date

Jun 27, 2025@9am Eastern

 Participants

  • @Arran Griffith

  • @Ben Pennell

  • @Bradley Watson

  • @Collin Brittle

  • @Dan Field

  • @Daniel Pierce

  • @Emily Porter

  • @Heather Greer Klein

  • @Jon Dunn

  • @Juliet Hardesty

  • @Kate Dohe

  • @Nicholas Mark Homenda

  • @Scott Prater

  • @Randall Floyd

  • @Rebekah Kati

  • @Tom Wrobel

indicates note taker

 Discussion topics

Item

Item

Welcome

  • Volunteer notetaker: @Nicholas Mark Homenda

Updates:

  • Community Performance Testing Update?

    • Community Performance Testing - f6.nurax

    • @Nicholas Mark Homenda has given access to ~5 people to the tester role

    • @Arran Griffith will send out a reminder

    • Folks from this group will do testing, too

    • Close to having Bulkrax in the Nurax instances, Julie had it partway up locally

      • Then, we can launch Phase 2 of testing

    • Frigg adapter is being worked on (additional connection parameters beyond what was needed for PostGres)

      • Are there institutions poised to test this?

        • IU!

    • Working on 5.2 release- mostly bug fixes

  • Migration updates?

    • Finished migration at Emory at beginning of June

    • 2.9 million Fedora objects ~ 34000 Hyrax works

      • Is this number unexpected/unmanageable?

        • No

      • Is this vastly different than number of objects in F4?

        • Unsure, also some different collections available that weren’t in F4

        • F4 Solr could give part of the picture, unsure if it shows proxies, etc.

        • FCR stats call with 1mil+ objects is really slow

          • Query is the one provided in UI

          • Oxford noted it’s still snappy for them

        • @Ben Pennell notes in their Hyrax 4 instance Solr:

          • 2.9 million objects

          • 152114 works

          • 383576 AccessControl

          • 34168 Proxy

          • 230391 FileSet

          • 170924 IndirectContainer

      • Scale of OCFL manifests a concern?

        • If we had our choice, we would’ve structured these as Archival Groups

      • Also a problem for Oxford. Application level knows to ignore, but file system interaction doesn’t

      • Emory attempting to find orphaned objects/resources created and not associated back to parents during large ingest process

    • Tomcat update prompted Fedora reindex, which took 70 hours

      • no idea why Fedora would require that- perhaps a flag is specifying that

      • good to check that fcrepo.rebuild.on.start is turned off, and maybe fcrepo.rebuild.continue

    • Getting ready to launch new app July 8, also upgrading Hyrax 5.0 (pinned to a specific commit) to 5.1

    • Overall, so far, so good

  • OCFL export for Fedora Team

    • At Open Repositories 2025 @Dan Field asked if we can get an export from OCFL after ingesting works into Hyrax

    • Probably easier to obtain from local Hyrax installation rather than from OCFL

    • Easier with a single deposit with multiple files attached

What work remains?

  • Phase 2 (after Bulkrax)

  • Frigg adapter testing

Wrap Up

Next meeting:

  • July 25th @ 9am Eastern

Next note taker:

Notes