Hi there,
First, any time you encounter a 500 error, the first thing we recommend is to get more information by looking for the full error message in the web server logs. If you don't have access to the server backend, you may need to ask a system administrator to get this. See:
Once we have the full error message, we can provide more detailed suggestions. That said....
Choosing "Ignore matches and create new records on import" is not an update to existing records - it's explicitly telling AtoM the opposite. You're telling it that even if all matching criteria are met, ignore the match and treat these as new records. If you want to update existing records in AtoM via a CSV update, then you should likely be using the "Update matches ignoring blank fields in the CSV" option. See:
That said, making updates in a single system (i.e. roundtripping - exporting a CSV, making some changes, then importing it as an update) is unfortunately poorly supported at the moment in AtoM. For more background on how the matching logic works currently and why, as well as some workaround options for improving local roundtripping efforts, please see:
Hopefully that helps future attempts, and with more information we can help you resolve the current 500 error!
Cheers,