...
- Slack channel for breaking changes during as Fedora 6 is being developed?
- Move most development conversations into a "bleeding edge" channel (which will replace the "sprint" channel), and then post to the "tech" channel if a backwards breaking change is made
- Demo Server
- Auto-updating on AWS every time a new image is pushed to Dockerhub (in theory)
- REST API is working, but UI isn't available. Andrew Woods and Danny Bernstein will circle up so they both have access and can work on it.
- "Next Demo" topic pushed back until David Wilcox gets back from vacay
- Committers / Leaders Call Debrief
- Getting everyone on Fedora 6 is top priority
- For 4/5
- Export content using Import/Export tooling (this exists)
- Import the exported content (does not exist yet)
- How best to manage the fact that you've got to double the data? Unavoidable, most likely just needs to be clearly explained to the user in documentation.
- Is there a chance to share common code between the other tooling (migration utils, import/export, etc...) when moving to Fedora 6?
- Migration tooling must be there before 6 can be an alpha
- Migration tool strategy:
- extract abstraction layer for writing fedora objects to ocfl from core into a common library
- extract read fcrepo4/5 export functionality from fcrepo-import export tool into another common library
- Use above libraries in the fcrepo-upgrade-util for transforming an export to Fedora OCFL.
- Danny Bernstein to create JIRAs to organize this work.
- Calvin Xu is experiencing slowness when starting up Fedora 6 on 3 million object OCFL repo (migrated from Fedora 3)
- Do we need progress meter for index rebuild activity? Is this even feasible?
- How about "resume" index build functionality?
- We would have to be able to assume that the OCFL repo had not changed
- We also do not know at this time how long it takes to iterate over the repo (without doing any index updates).
- How about "resume" index build functionality?
- Next steps:
- Calvin Xu will try using a separate database server (mysql, mariadb or postgres) on a small subset of items (1000) to measure performance improvement and will report back on next week's call.
- Do we need progress meter for index rebuild activity? Is this even feasible?
Actions
- Danny Bernstein to ensure that a JIRA exists for scenario of transaction does not succeed completely, but rollback does succeed (see: 2020-07-23- Fedora Tech Meeting)
- Peter Winckles to create a JIRA to investigate support for optimistic locking and mutable head (see: 2020-07-23- Fedora Tech Meeting)
- Who: Clarify in documentation that multiple simultaneous writes to OCFL are not supported
- Peter Winckles Create a Jira to update the locations files stored in OCFL
- migration-utils - translating RELS-EXT/INT into RDF in Fedora 6: is the functionality there? If not, add it.
...