ClusterNFS FAQ

How does ClusterNFS relate to the "transname" patch for linux 2.0.x?

The transname patch for version 2.0.x of the Linux kernel vfs implements essentially the same features as ClusterNFS. Ulf Bartelt (ulf@twc.de) pointed out the transname patch after I announced ClusterNFS.

Advantages of ClusterNFS over the transname patch:

ClusterNFS does not require any kenel changes
ClusterNFS is portable to any *NIX system.
ClusterNFS runs only on the server. Clients run (any) standard NFS client.
ClusterNFS allows matching on user characteristics (UID, GID)

(Current) Advantages of the transname patch over ClusterNFS:

Does name translation on file creation.
Applies to all filesystem types (ie, local filesystems)
Allows matching system CPU type (i486)
Can be restricted to apply only with directories owned by a specific group

Name translation on file creation and restriction of matching to files/directoies owned only by a certain group or user can easily be added to ClusterNFS. Volonteers?