Hi, I'm assuming this isn't done very often (for obvious reasons), but I'm attempting to NFS export a mounted EncFS directory. The NFS clients would have access to the unencrypted data, without running EncFS. The data on the server would be encrypted for off-line protection. Here's an example of the setup:
-- Server -- [/etc/fstab] encfs#/h/crypt-raw /h/crypt fuse rw,user,noauto,allow_other 0 0 [/etc/exports] /h/crypt/foo host(fsid=1,ro,async,all_squash,anonuid=1000,anongid=1000,no_subtree_check) -- Client 1 -- [/etc/fstab] server:/h/crypt/foo /h/foo nfs intr,async,nfsvers=3,bg,actimeo=0,tcp This *mostly* works. I have a variety of exports under "/h/crypt". With one exception, they all behave as hoped. The exception is an export of my music to a media host and a Mac. After a bit of use, an "ls" in the mount on the Linux media client starts spitting out "Stale NFS file handle" errors for some of the subdirectories. iTunes on the Mac will start complaining about a duplicate library (which is located on the mount). The other mounts have never exhibited this behavior. The big difference that jumps out to me is that there are many more directories/files in the music export. FWIW, if I NFS export the raw encrypted data and NFS mount + encfs mount it on the client, the stale link-type errors don't seem to occur. Any suggestions for what might be causing these issues? I realize this is a non-conventional use, but I'm curious if it should work in theory. Thanks! Joe ------------------------------------------------------------------------------ The modern datacenter depends on network connectivity to access resources and provide services. The best practices for maximizing a physical server's connectivity to a physical network are well understood - see how these rules translate into the virtual world? http://p.sf.net/sfu/oracle-sfdevnlfb _______________________________________________ Encfs-users mailing list Encfs-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/encfs-users
