Discussion items:

AWS credentials must be entered each time.  we want a way to get credentials to accept automatically. Colleagues at EPA brought this to our attention.  A collaborator in Slack was asked to test this configuration for us to make sure it works.

Rob notes that Mike found two bugs on failed upload for incorrectly decrement + byte counts.  If you do 2 uploads in a row, transfer manager, you need to do a shutdown the S3 file between uploads. This has been fixed.

Fix python builds Jsonform to version 2.25 and do a bug release.

Maxwell Burnette can review the PR's by Monday.

Discussion of service model - getting collaborators to pay for DOI minting by NCSA.

Good discussion from Steering Committee - how to sell Clowder to specific communities. Make is user friendly.

Elizabeth Yanello send out Steering Committee notes to committee members today.

Luigi will create code for Mongo DB and its problems with tight reliance in Clowder.

Clowder 2.0 release - we still need to do bug fixes.  Delegate and divide workload between FastAPI and React

Let's breakdown what aspects we are going to work in a prioritization.  Start with Authentication.  Most people want to share datasets, not spaces.

Rob wants to talk about the structure of the spaces as they seem to be disorganized.  Projects should have sub-projects, rather than collections.  It would be more organized.

Max suggests that we do a design of how to structure these spaces before we try to implement it.

Let's have discussions on HOW we are going to implement with mock-ups.

DOI minting capabilities - part of the CSSI project.  Researchers want to show NSF, and other publishers that they actually have a DOI.  The DOI points to a URL, but that URL needs to be available in perpetuity.

Rob notes that CEED has an archived read only file to link to the DOI.  Do we want to do this with Clowder? Create a static webpage with metadata that the DOI points to using bagit app.

Is there documentation of CEED of how it used to work with staging etc?  Luigi will try to pull this up.

We can do bogus DOI for testing. DataCite has a testing site.

Federation keeps coming up.  Instead of several individual instances that one needs to go to the other instance page, how to we see datasets across all instances? 

Let's work on 2.0 rather than a webinar.

We are part of the github issues data under Clowder Projects.

  • Introduce Kastan Day
  • PRs and patch release - small but important.
  • Github pages
  • Clowder Service Model (Richelle working on it, Discussion in the Steering Committee)
  • FastAPI tests and architecture
  • Clowder 2.0 plan
    • Max and Todd on FastAPI
    • Chen and Mike on React
    • Still need to do bug fixes
  • DOI minting capabilities
  • Federation keeps coming up
Mike L.
  • PR to update jsonschema to v2.2.5
    • Adds support for navigational tabs and other new interesting templates
  • PR to support more options from DefaultAWSCredentialsProviderChain
    • Fixed a bug where subsequent uploads would fail without refreshing the view or navigating
      • Apparently shutdownNow takes an optional parameter to keep the underlying S3 client open
    • Fixed a bug where failed uploads would decrement file + byte counts
      • We shouldn't decrement in this case, since counts were never incremented to begin with (a completed upload performs the increment)
Mike B.
Michael Jabsent

Newsletter and Webinar

Max suggests the workbench would be a good webinar. This would end CDDR. Permafrost is also linked to CDDR.

Action Items/To Dos: