Hi,
I just tested Seafile pro 6.2.0 with a ceph cluster (10.2.10) as storage backend under Scientific Linux 7.4 (CentOS/RedHat 7). I followed the instructions in the server manual. The migration script was running fine, but after logging into seafile, an error occured and the libraries couldn’t be accessed.
As mentioned under “Notes for Ubuntu 16.04”, I had to delete the bundled libraries (librados.so.2, libstdc++.so.6, libnspr4.so) from seafile. After that seafile was running fine with the ceph backend. Maybe you should mentioned this in the server manual. And maybe this problem will no longer exist, if you compile seafile unter CentOS 7.