Re: [RFC] inode generation numbers.

Craig Milo Rogers (rogers@ISI.EDU)
Tue, 23 Feb 1999 11:11:30 -0800

>Having all of your filehandles go south on you just because your Linux
>NFS server was upgraded should happen as rarely as possible (never is
>best, but I agree that we need to add the generation number, and I
>don't know if we can add it without breaking filehandles once).

Suppose that the absence of a generation number (or timestamp)
(internally represented as, say, the value 0) acted as a wildcard?
Couldn't you then upgrade the server without disrupting service on the
clients?

You'd want the wildcard feature to be configurable, of course,
so you could turn it off when all the clients have been upgraded.

Straightening out kernel NFS sounds like an important service
improvement... perhaps later in 2.2.x?

Craig Milo Rogers

-
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/