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)
- U.S.A/Canada toll free: 866-740-1260, participant code: 2257295
- International toll free: http://www.readytalk.com/intl
- Use the above link and input 2257295 and the country you are calling from to get your country's toll-free dial-in number
- Once on the call, enter participant code 2257295
- IRC:
- Join the #fcrepo chat room via Freenode Web IRC (enter a unique nick)
- Or point your IRC client to #fcrepo on irc.freenode.net
Attendees
- Unknown User (acoburn)
- Bethany Seeger
- Jared Whiklo
- Nick Ruest
- A. Soroka
- James R. Griffin III
- David Wilcox
- Yinlin Chen
- Aaron Birkland
- Osman Din
- Andy Wagner
- Andrew Woods
Agenda
Camel component issue - priority for next week's workshop
Next release planning
- Switching to a new conference line
- Notes from the field: HydraConnect
- Issues raised:
- Constructive interop discussions: best practices and common us of PCDM
- F3 to F4 migration interest group
- Issues raised:
- Scale/performance opportunities – brainstorm testing scenarios
- Vagrant refactoring: new base-box?
Landing in-flight issues - revisited
- ...
Ticket Summaries
Please squash a bug!
Tickets resolved this week:
Tickets created this week:
Minutes
Vagrant/Apache Camel Component Issue
Next week, D. Wilcox and A. Woods are to be at the Fedora Users Group meeting in Washington D. C.
They shall not be available for the next meeting
This features a Fedora 4 workshop (requires the spinning up of a Vagrant Box)
Recently, this Vagrant stopped working
When 4 commits are rolled back, the Box is fine
Determining which of these commits is the perpetrator is necessary
Unless an easy fix can be determined, the rollback will suffice for the time being
Apache Camel Components
Vagrant depends upon snapshot versions of Camel components
Without previously spinning up Vagrant...if a commit went into toolbox...and Jenkins pushed it...it could potentially be pulled into the Vagrant box
Tying Vagrant box to released vs. bleeding-edge components
Confirmed as reasonable by acoburn
Next Release
Community has gone longer than preferred to go between releases
Call for a new release
Contributors: should we be waiting on any outstanding issues?
Not everything works...clustering still doesn't work (since 4.3)
awoods prefers to have frequent releases, rather than wait to introduce fixes
acoburn: fixes in review preferred to make it into a release (Issues and )
Issue 1706
Discussed between ajs6f and acoburn
They have a temporary patch; wait to fix this in Modeshape
Implement a patch (which works fine), or update the Visitor Class (an interface in Jena; this requires more work)
A. Soroka (ajs6f): Good with pull request as it stands
Osman Din: confirms that he shall create a ticket in Modeshape for this
Create another ticket for the Visitor approach? Or will Modeshape fix handle this?
acoburn: Modeshape ticket will handle it; Visitor implementation would still be useful in other contexts
acoburn: confirms, another ticket for Visitor within Jena would be good; created a related ticket here:
GitHub Shuffle
Need to sort out maintainers on a number of GitHub repositories
A Technical subtask is outstanding ()
acoburn: Outstanding issues for these shouldn't hold up a release
awoods: These 6 repositories appear to have been moved out of fcrepo-labs, and into fcrepo-exts where appropriate
awoods: Just a matter of updating the README file for each of them
Identifying the GitHub Repo. Maintainers
Finding the most commits, assigning the maintainers based upon this
Coburn and Ruest coordinating who shall be maintainers
A. Woods:
- Call for maintainers for any of the following GitHub repositories
- webapp-plus: awoods
- transform: whikloj, acoburn
- audit is assumed to be Esmé Cowles
- This still leaves audit, message-consumer, and webapp-plus with 0 or 1 maintainers
Without owners, note that these modules are to be put on the fence, staying where they are
Still outstanding: selecting which of the modules to be released...and which of these lie with the maintainers
Response: create a matrix as a part of this process
Let's say we do that...which modules to be released (not a huge deal)
It was suggested in a previous release that release candidate branches be created
Let these float out there for testing (as integration tests alone may not be sufficient)
Are people in the position to do sanity testing (if candidate branches created over the next week)?
Ideally, all developers in Fedora who run into issues while testing would create a ticketFor example, the most Vagrant issues (tests are passing, but not working)
whikloj, acoburn, and N. Ruest suggest the usage of release-candidate branches with a code freeze
awoods: Prefers to get in some more tickets in review before Monday
Following this, a code freeze begins on Monday and throughout next week
Week of 10/12/15: Can perform a release
The code freeze message, hence, shall be announced to the community
New Conference Line
Current line...costs money...about 4 cents per minute per person
Try to switch to free line...
Current conference line is quite expensive
D. Wilcox: New, free line is almost exactly the same...except without offering "1-800" toll-free numbers
With numbers from every country...long distance is likely to be a potential issue
Call for parties who may have difficulties with this
Canada offers free calls if it's a Canadian number
Administrative Dashboard is also available for D. Wilcox and A. Woods
Dedicated for Fedora (no conflicts with DuraSpace)
Try it next week....
Birkland: DuraSpace previously offered free conference call HD, accompanied with a Skype back-channel for conference calls...is this similar?
Birkland: This could mitigate issues by offering access over Skype if this is offered
Wilcox: Uncertain if this is offered
HydraConnect 2015
Unknown User (daniel-dgi) can speak to many of the exchangesRepository Interoperability
Two sessions...one was a preconference event on the night of 10/21/15 (Monday)
Not much on what Islandora is doing wrong for Fedora 4
Hydra focused upon what they are improving (e. g. binding their Views)
Discussion landing on PCDM as least common denominator
Islandora over Fedora 4...Hydra over same Fedora 4...ideally, be able to have an admin view of Fedora resources
Lot of interest among the group in moving Hydra forward
Abstraction in Fedora 4...a standardized API is currently being established
PCDM conversation in particular is moving forward within the community
Addressing issues of value to the Hydra community
E. Cowles is progressing with a branch is moving towards resolving E-Tags
E-Tags are generated solely based upon last modified date (managed by ModeShape)
Doesn't always work...issue with inbound references (i. e. cases for referencing between repository resources)
Updating B so that it references A...updates timestamp for resource A
Some parties find this behavior to be valid and appropriate...because the one potential representation has been changed for A
Others which develop applications that rely on E-Tags find that changing the timestamps often can trigger failures
Strong E-Tags should be Weak E-Tags
Reasoning: Strong E-Tags are byte-for-byte identical with regards to response
With content negotiation, responses aren't byte-for-byte identical, but the E-Tags are same
Further, for a higher-level application (e. g. Hydra) weak E-Tags have additional benefits
Fedora 3 to 4 Migration Interest Group
Workshop at HydraConnect (all day on Monday, three parts)
Woods: Core features and hands-on exercises
Benjamin Armintor: Session on migration using fedora-migrate; ActiveFedora 8 connecting to fcrepo 3; ActiveFedora 9 connecting to fcrepo 4 and migrating the content
Lamb: Integrating with external components via Apache Camel...just used fcrepo-camel (didnt use reindexer, solr, triplestoreupdater); How can you integrate with Fedora using Camel (using routes defined in XML)
Re-ran migration utility...using Camel routes for image resources, filtered for just JPEG images, executed ImageMagick, and re-ingesting the thumbnail as a child of the resource with the image
Session later in week: Hydra, and migrating into fcrepo4 (A. Woods, B. Armintor are in the interest group)
Hydra community is in the position: no active development for fcrepo3 stack; how do I get on to Fedora 4 then?
Hydra is strictly focused upon fcrepo 4; Hence, the IG was born (define practices, documentation around migration)
(Within Hydra context, a consistent way of migrating into fcrepo4)
Exercising the limits of the Fedora stack
People are being drawn to the fcrepo4 projectCurrent and potential scalability appeals to these new parties
Concrete descriptions of tests are desired
These render it easier for individuals coming into project...leading to the creation of infrastructure and the addressing of testing
Infinispan and the potential of the fcrepo4 stack (particularly focusing upon scalability targets)
Create isolated test scenarios
Technical WG (Technical Working Group) from last year has some of this already
Interested Parties
- whikloj
- osmandin
- acoburn
- awagner
- jrgriffiniii
- ajs6f (Volunteering to Help)
Call to starting a Google Doc
Over the next couple of weeks, collectively come up with a test plan
Please include everyone in the community if this is initiated; otherwise, A. Woods will address after traveling
This should be an outline of a test plan...again, refer to the work of the Technical WG
Examples of Test Cases:
- Scaling for resources
- Scaling for a number of bytes
- scaling for a large repository collection
- iterating over reads
- performing an operation offline and pushing back updates
- at each resource, POSTing back an update
Include references for components at the Infinispan layer
Also discuss deployment into the cloud (particularly AWS)
Cleaning up Vagrant
N. Ruest has progressed much with the Islandora ecosystem
Vagrant Box with main components installed in the base box
Pulls down pertinent versions of Fedora components...installing all of the infrastructure for the Box
Question of Packer/Vagrant for Docker?
Deep-dive for next installation...sorting out the Vagrant/Packer discussion
Working with Yinlin's Docker image https://hub.docker.com/r/yinlinchen/fcrepo4-docker/
Still need expertise for Packer
N. Ruest: Each institution requires its own Vagrant
ksclarke: has a number of features which he has to offer
