e07@nikhefh.nikhef.nl (Eric Wassenaar) (12/18/89)
In non-apollo NFS implementations it is possible to mount remote NFS filesystems on top of each order, like /etc/mount rhost:/ /mountpoint /etc/mount rhost:/usr /mountpoint/usr so that the whole remote filesystem tree can be transparently accessed using the same pathname syntax, starting at /mountpoint, as on the remote host itself. The mount points must be existing directories. In the Apollo SR10 NFS implementation, this is still impossible. Mount points must be non-existing. They are created as objects of type 'nfs_gate' by the mount operation, and obviously cannot reside on already mounted remote filesystems. It is very unfortunate that Apollo NFS breaks this transparency. It breaks sophisticated symbolic link structures on such remote systems. Eric Wassenaar -- Organization: NIKHEF-H, National Institute for Nuclear and High-Energy Physics Address: Kruislaan 409, P.O. Box 41882, 1009 DB Amsterdam, the Netherlands Phone: +31 20 592 0412, Home: +31 20 909449, Telefax: +31 20 592 5155 Internet: e07@nikhef.nl