Start the blob-archiver process from '1' - genesis, it will rapidly index blobs, however if the process is stopped and started again (without any changes) it will not carry on from where it left off... if will just add the most recent blobs as they become available.
Start the blob-archiver process from '1' - genesis, it will rapidly index blobs, however if the process is stopped and started again (without any changes) it will not carry on from where it left off... if will just add the most recent blobs as they become available.