Re: Open by inode? (was Re: knfsd)

Brandon S. Allbery KF8NH (allbery@kf8nh.apk.net)
Mon, 11 Jan 1999 19:41:21 -0500


In message <77c9q9$a8r$1@palladium.transmeta.com>, Linus Torvalds writes:
+-----
| any well-thought-out network filesystem (and no, I don't consider NFS
| to be well-thought-out) will use pathnames, and even local
+--->8

I think I'm channeling the AFS developers when I boggle at this... files are
identified by 4-tuples in AFS (and Coda, which is based on AFS). "By path"
is rather difficult when a remote cell's file server can't possibly know
*where* its volume is mounted in my local cell's AFS tree (nothing but a
desire to easily identify what's going on just by looking at the path stops
me from doing "fs mkmount ~/.private/pafs/foo remote.cell:volume.name" where
"volume.name" is not "root.cell" --- and even the latter is just a
convention). The local client resolves the pathname to the appropriate
4-tuple and then hands that off to the remote cell's servers. The only
fixed requirement is that there is exactly one mount point for the AFS
client's root cell on a local filesystem (by convention,
$(</usr/vice/etc/ThisCell):root.afs on /afs).

Or I could use "arlad -d", which behaves much like an FTP client for
browsing AFS volumes (compare to Samba's Unix SMB client...). The path is
irrelevant, as I can tell it to mount as a root cell any volume I can access
in any cell I can access.

Pathnames are less than usable on a global filesystem.

-- 
brandon s. allbery	[os/2][linux][solaris][japh]	 allbery@kf8nh.apk.net
system administrator	     [WAY too many hats]	   allbery@ece.cmu.edu
carnegie mellon / electrical and computer engineering			 KF8NH
     We are Linux. Resistance is an indication that you missed the point.

- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/