Time/Place
This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join...here's the info:
- Time: 11:00am Eastern Daylight Time US (UTC-4)
- Dial-in Number: (712) 775-7035
- IRC:
Attendees
Agenda
Update from "Alignment to Spec" sprint
Progress so far
Versioning and Auth design subgroup: Versioning/Authorization Design
Big question: Get rid of snapshot versioning paradigm? (versioning a container would version only that resource, not children)
- Big question: Get rid of modeshape node references (removes referential integrity, may have good or bad implications on performance)
Engaging Memento spec team
- Wrapping up fcrepo-import-export-verify: https://github.com/fcrepo4-labs/fcrepo-import-export-verify/pull/48
- ..
Ticket Summaries
Please squash a bug!
Tickets resolved this week:
Tickets created this week:
Minutes
- Good progress being made
- migrated key webac and rbacl classes into fcrepo4 core.
- moving towards making the URI of the user available via the core api for messaging and other purposes.
- resolving issues with the one-click jar.
- Discussion about implementation (https://github.com/fcrepo4/fcrepo4/pull/1242) if you have any opinion).
- Versioning and Auth design subgroup needs feedback on the big questions listed above. For details related to the questions see link above.
- Re big question 1: Shall we provide the ability to version children automatically from the get go or shall we push that out?
- Seems to be consensus around abandoning the snapshot versioning paradigm and punting on auto versioning objects in the subtree.
- Performance of reads could be expected to improve.
- Question about whether writes would become significantly slowdown.
- Re big question 2: proposal:
- we could start by replacing node references in mementos with URIs.
- Eventually we could replace all node references in the repo with URIs.
- We must ensure there is a migration path for people who wish to preserve their versions.
- import export tool?
- Or some kind of direct transformation within the repo?
- Stuck thread issue is still a serious problem with ModeShape5.4 for Peter Eichman and Joshua Westgard
- Was not an issue on the dev server, but no is an issue on staging server.
- Still an open question about whether or not it is truly absent from the dev server.
- If dev continues not to show symptom, Joshua Westgard will look for differences in VM configs.
- Suggestions:
- Look closely at the hardware
- Can you shrink the size of the batches?
- Can you slowdown the ingest calls - introduce short pauses between calls?
- Smaller transactions with retries?
- Developing...
Actions
- Ben Pennell of the Versioning and Auth team to write up big questions, solutions and trade-offs and pass that along to the community for feedback.