Time/Place

This meeting is a hybrid teleconference and IRC chat. Anyone is welcome to join...here's the info:

• Time: 11 am Eastern Daylight Time US (UTC-4)
• Voice
  • Via Skype+: Dial from anywhere: \+9900827047086940 (callto://+9900827047086940)
  • Via Phone+: US Number: +1 (201) 793-9022, Room #7086940
• IRC:
  • Join the #duraspace chat room via Freenode Web IRC (enter a unique nick)
  • Or point your IRC client to #duraspace on irc.freenode.net

Agenda

1. GSoC Update
   1. Progress so far
   2. Next steps?
2. 3.5 has been released!
   1. Any comments or last minute additions to the Release Notes?
3. Thoughts on future 3.5.1, 3.6, or 4.0 release?
   1. In-progress issue: FCREPO-951
   2. In-progress issue: FCREPO-238
4. Do we want to hold Git special topic meeting?
   1. What have we liked or disliked so far?
   2. Do we want to solicit opinions for improved procedure?
      1. Chris puts forth http://nvie.com/posts/a-successful-git-branching-model/ for ideas
5. Currently open issues
6. Recently submitted issues

Notes

• GSOC ends today. Code is in, works. Still need to work on tests
  • We shall add an issue to Jira for integration into the branch that shall become 3.6
  • Asger: put out a beta release for projects such as GSearch to adapt?
  • Eddie: Let's only do so if we need to
  • Asger: makes sense
  • Action item: Code review. Rebase/merge.
  • Action item: Add issue in Jira for integration to CXF

• Steve: Should we add University of York in acknowledgement section?
  • Yes, definitely

• Thinking about 3.5.1 or 3.6 release
  • FCREPO-238 - Will put in maintenance branch
  • FCREPO-951 - Makes sense put in maintenance branch
  • FCREPO-990 - Actually considered a bugfix, so can/should go in maintenance branch
  • FCREPO-991 - Ultimately, it is a bug, and can go in maintenance branch.

• What is the essence of 3.6?
  • Steve: Performance
  • Eddie: profiling, benchmarking
  • Asger: noticed some obvious performance sore points
  • Eddie: what are we aiming for re: benchmarking?
  • Aaron: who is considering performance issues currently?
    • Frank: FIZ has been investigating ingest/delete performance under high concurrency. Have found issues
  • Eddie: Who has done serious profiling? Would be nice to incorporate into our testing
    • Dan: Performance is variable on AMI instances, though can normalize by checking performance of machine or storage
    • Frank: JMeter maven plugin. Have some work up here: https://github.com/astromatthias/JMeter-Testplan
    • Dan: Grinder
  • Special topics meeting to investigate what and how we wish to profile for fedora 3.6
  • Action item gather literature and experiences, approaches, etc. Finalize details for special topics meeting next week.

• Issues discovered in FIZ testing:
  • Ingest: 200 OK messages sent, but some of these objects were not actually ingested
  • Delete: some objects not removed from DORegistry