Metacat

For unexpected errors, we need to track and retry the indexing task before giving up and notifying the administrator.
This may happen because the object type is unknown and cannot be parsed correctly, or because there is some error when submitting to the solr engine.
We do not want these errors to accumulate and bog down the system for weeks and weeks (until, say, support for the new object type is added) but we do want to add them back to the indexing queue once the reason for the error has been resolved.