Comment 6 for bug 815753

Its a fastdowntime issue. These processes where all connecting as the same user (archivepublisher), but only some of them need to be shutdown manually before a fastdowntime deployment.