After running the archive job, the records appear in both databases
The archiving of these records did not complete.
The next run of the archiving job will pick up these unfinished records. The job is designed to do this. After the next run these records should no longer appear in the source database.