Don't do this. You are archiving twice, and worse, the first part is using async archiving. Remove that whole "cp" part. Once that is fixed, run:
I'm going to see if putting archive_command = 'pgbackrest --stanza=My_Repo archive-push %p' fix the problem. This was haunting me for more than a week.
Thanks for the suggestion.
pgbackrest --stanza=My_Repo check
to verify that WAL is being archived quickly and properly. Then test your backups, checking the file /var/log/pgbackrest/My_Repo-archive-push-async.log and your postgres log if any problems arise.