We're about to migrate attachments from a SQL Server to file system. For having a really large database (about 40GB), this procedure takes about 2 days. After it is finished the attachments are stored twice: in home folder of confluence and still inside the database. Confluence is also still storing attachments in database and not on the file system.
Do anybody has exceperience with that large database and SQL Server? We didn't find any exceptions or errors in both log files (tomcat&confluence), so what could be the problem?