First, are you using NFS version 3 or 4? If version 3, will both the server and the clients support the NFSACL extension?
If you are using NFSv3 + NFSACL, or NFSv4, and the actual filesystem on the NFS server supports ACLs, it should be possible to do something like this on the NFS server (adjust permissions to suit your needs):
setfacl -m d:u::rwx,d:g::rwx,d:o::rwx /nfs_anon
If the share already has sub-directories in it, you'll probably want to apply the default ACL to the top directory of the share and all existing sub-directories:
find /nfs_anon -type d -exec setfacl -m d:u::rwx,d:g::rwx,d:o::rwx {} \+
After the default mask ACL has been applied to the share's top directory, any new sub-directories created within that directory should automatically get the same default mask ACL.
Second, note that even with ACLs, this just applies a new default mask. The resulting permissions will still be based on what the program creating the file requests (typically permissions 0666 for regular files, and 0777 for executables/directories).
As far as I've understood, even the ACL will not be able to set the x
permission bits if the program did not request them in the first place - it can only subtract (mask) bits from the program's original request.