Versions Compared


  • This line was added.
  • This line was removed.
  • Formatting was changed.


  1. Announcements:
    1. Community Conversation re: Archival Groups vs. Atomic Resources - May 3 @12pm Eastern
    2. Presentations from Fedora, Dan  and Arran in the coming weeks at TCDL, Code4Lib and Open Repositories
  2. Pop-up/Other Topics:
    1. Samvera flyway error megathread
      1. Currently awaiting an update from Steve.  
    2. New OCFL-Java release v2.1.0: Performance Improvements for s3 Backend (Thomas Bernhart)
    3. PostgreSQL 15/16 - anyone using it? (Thomas Bernhart)
      1. Yes. James is using Postgres 16 after a recent upgrade and all seems fine. 
    4. PUT over Tombstone question from slack by Tom Wrobel
        1. No update
    5. Tech Listserv
      1. - F4 to 6 migration last email on April 16
        1. No update
    6. Camel toolbox issues
      1. High CPU utilization -
        1. No update
      2. No containment triples available in Solr indexing -
        1. Nobody had any answers as to why the predicates were removed from the response
      3. Doron brought up the issue that his co-worker had been having with calling additional fedora resources from within the camel toolbox sold xssalts, however this is a limitation of the document() function of XSLT. It was mentioned that the camel-toolbox was seen as more of an example reference implementation of the event messaging system and not designed to be an ent to end solution. Recommendations for a scripted solution, which could request multiple sources and build a single flat solr document for indexing, were brought forward. 
    7. Fcrepo 5.1.2 docker
      1. Thomas noted that this may need to be a manual process and was missed during the last release cycle
    8. Josh brought statistics from their work at AWS to migrate Fedora 4 to 6 for a client. They have managed to run the OCFL indexing over 69 million resources in 6 days of processing. He noted that although they attempted to make the process parallel across 8 fedora instances, the processing time did not decrease by more than 1.5x. This appeared to be due to the DB being in wait state the majority of the time