You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, DOI creation happens after the published dandiset is created on our side and committed to the database (see https://github.com/dandi/dandi-archive/blob/master/dandiapi/api/services/publish/__init__.py#L184). This means that if the call to the datacite API to create a new DOI fails, the dandiset will remain published but without a DOI. The publish workflow should be updated to instead fail gracefully and roll back the publish if the call to the datacite API fails.
The text was updated successfully, but these errors were encountered:
Even better would be to first create a draft DOI, and only continue on to perform the publish if that succeeds, and then finally promote the draft DOI to a prod DOI (which is an operation that can be retried until it succeeds).
(This is based on my tentative understanding of how Datacite works after hearing @danlamanna talk about this once.)
Currently, DOI creation happens after the published dandiset is created on our side and committed to the database (see https://github.com/dandi/dandi-archive/blob/master/dandiapi/api/services/publish/__init__.py#L184). This means that if the call to the datacite API to create a new DOI fails, the dandiset will remain published but without a DOI. The publish workflow should be updated to instead fail gracefully and roll back the publish if the call to the datacite API fails.
The text was updated successfully, but these errors were encountered: