(curation request) Investigate the possibility of recovering some VERY OLD backups for some legacy datasets
landreev opened this issue · comments
(placeholder)
from @sbarbosadataverse :
Think it's possible to see this collection in the old production site? the entire collection?
https://dataverse.harvard.edu/dataverse/ndsa
This dataset is particular would be great to see: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/27562
Email from the collection admin describing the nature of the extra data that may be missing:
Reviewing the five studies currently listed in a dataverse I admin (NDSA)
it appears that somewhere along the line, these terms were simply dropped,
and replaced with a CC-0 license. This appears to have
been performed at the dataverse network level, is not documented in the
metadata itself, and is not visible in the version history.
To comment on the "see this collection in the old production site? the entire collection?" part:
If we are talking about bringing up a working DVN 3* instance, that would be a lot of work, and I would strongly prefer to avoid having to do that. I hope that simply recovering any missing license data - if at all possible! - would be sufficient.
The timeline that was specified - perhaps somewhere between the 2014 and 2015 releases? - makes it highly likely that the missing info was lost during the DVN-to-Dataverse migration.
Unfortunately, nothing useful has been found in the most easily-accessible, but incomplete backups from the migration, the _dvn3_* database content that we still have around. (The DVN3 tables that are likely to contain the missing terms of use are not preserved there). So more, and deeper digging will need to be done.