NFS consistency problem?

[Trond Myklebust]

On Mon, 2007-05-07 at 08:02 -0700, Mike Eisler wrote:
> If the NFSv4 server does not increment the "change" attribute on
> each modification, then comparing the GETATTR results on open() with
> that of the previous close() won't help, even with noac.
> 
> I've heard that ext3 does not support sub-second granularity for the
> mtime and ctime attributes in NFSv3. Assuming the NFSv4 server on
> Linux uses "ctime" to simulate "change", and assuming the "echo"
> commands are done less than a second apart, this might explain the
> problem.

This is true, but it will only affect matters if the two changes are
made within less than 1 second of each other. Using XFS can mitigate the
problem a bit (with XFS the ctime resolution is ~ 1/250 sec) but it is
still not a perfect solution.

We're trying to push for a proper NFSv4-compatible change attribute in
the next revision of the ext filesystem (ext4).

Trond