...
Time | Title | Presenter |
---|---|---|
14:00-14:10 | Welcome and Introductions | |
14:10-14:30 | Migrating Fedora - Staatsbibliothek zu Berlin | Oliver Schöner, Staatsbibliothek zu Berlin |
14:30-14:50 | Ralf Claussnitzer, Saxon State Library Dresden | |
14:50-15:10 | BREAK | |
15:10-15:20 | Danny Lamb, Islandora Foundation | |
15:20-15:40 | Heather Greer Klein, Samvera Community | |
15:40-16:00 | Wrap-Up and discussion | All |
Notes
DAY 1
Andrey - Data Communication Between Fedora, IIIF Server and Madoc
Currently using Fedora 4/5 but started process of migrating to 6
- When might the project be publicly available?
- No concrete plans as of right now
- Wanted to offer insight in to new approach to using data within in Fedora
- What kind of IIIF server are you using in the architecture?
- Own implementation at State Library of Berlin, only for local usage right now
- How do you achieve synchronization between the IIIF server and the image/manifests data from Fedora ?
- No need to sync, it just works
- Reads directly from Fedora, doesn’t get stored in IIIF
- Notice any differences in API from 4 or 5 to 6?
- Using Fedora 4 currently, but don’t anticipate problems migrating
- How much content have you moved from 4/5-6?
- Just started
- Performance bottlenecks?
- Migrated content is larger than source content
- With smaller collections, no problems
- What version of Madoc are you using?
- Version 2
Raman - Institutional Repository; Dealing with Requirements from Different Disciplines
- Have you considered migrating to Fedora 6?
- Would like to start later this year
- Have done some kinds of pre-testing, but also need a strategy because there are over 1 million items in the repository
Felix - Ongoing Work & Plans with Fedora 6
- Ralf discussed their solution for OAI Endpoints
- Could you elaborate on why you want to restructure data?
- Structuring would only be for internally working with repository (nicer environment for working in the backend)
- Would not change how data is presented
Chris - Digital Archiving Using Hyrax & Archivematica
- Where will the metadata be archived in the workflow process?
- Assumption that metadata taking place in Hyrax, but they have been able to do it elsewhere
- If captured in CALM or other places like that it would not exist anywhere for digital preservation purposes
- Chris will share link to Whitepaper once it’s available
- Hull White Paper is available at https://hydra.hull.ac.uk/assets/hull:17757/content1
Fedora Migration Break-Out
- Migration testing so far
- Berlin State Library
- Leipzig University (CLARIN-D)
- External content (Paul - Running Fedora 3.8.1 w/ Islandora 7)
- Fedora API supports this content by storing a reference to the external content
- OCFL does not formally support external content as a first class resource
- From the Fedora perspective, this is just metadata so any updates should not require a migration and would not be a big deal
- Migration tool testing
- If you have lots of versions in Fedora 3 you will get lots of versions in Fedora 6. Can these be squashed? (Ralf)
- There is not currently a public open source tool that can squash OCFL versions
- Migraiton-utils only supports migrating versions, but it would be fairly easy to allow users to ignore versions and only import the latest
- Don’t NEED this, but had noticed versions that weren’t necessary, but not a roadblock to adoption
- Migration tool basically replicates Fedora 3 structure
- You could delete the old versions from the Fedora 3 objects using the API before migrating. If you want to do it on specific objects.
- LDP containment
- Different container types
- Mostly based on metadata, not structural
- Archival groups
- Multiple Fedora objects contained in a single OCFL object
- PCDM in Hyrax
- Indirect containers are used
- Bloat in Fedora 6
- OCFL does use more space than Fedora 3 storage
- Changing SHA-512 to SHA-256 will improve storage space
- May want to schedule a special topics call for OAI
DAY 2
Oliver - Migrating Fedora - Staatsbibliothek zu Berlin
- How hard was migration from 5-6 and how big was the data?
- 5-6 easy, applied upgrade utilities and it went well
- Start at 5.1.1 is recommended
- 10G of data
- Were there changes in the API?
- Didn’t see any (read only repository)
- No major API changes between 5 and 6
Ralf - Centralizing Repository Infrastructure
- Have any problems w/ different workflows needing a different namespace?
- Was kind of an issue - do have different namespaces for publications
- In Fedora they have only 1 namespace for all objects but metadata is different
- How do you manage changes of metadata schema?
- Started by defining metadata schema - made 2 (info and metadata)
Wrap-up and Discussion
Resources
...