> How else could you explain a lag of several minutes before system B can
(in
> some cases) see changes made by system A? It makes no sense to me unless
I
> view it in terms of two distinct caches (or I am particularly near-sighted
> and cannot see the obvious answer, which is, of course, quite possible).
>
> In the particular observed case, system A is a desktop machine and system
B
> is a thin client attached to a W2k server running Metaframe.
I don't doubt that you're seeing the problem you report only that, in
theory, a disk cache shouldn't be causing it. Obviously something is. I
don't know whether it's possible for a program to "fool" the disk cache so
that somehow the data is read directly from disk rather than from the cache,
but perhaps something like that is going on.
Also, I don't know anything at all about Metaframe, but is it possible that
it is keeping its own cache, separate from the operating system? If so,
that cache could still contain the old disk based pages while the operating
system cache had not yet written the data to disk. The two caches could be
on the same machine, just not coordinated.
--
Larry
_________________________________________________________
Do You Yahoo!?
Get your free @yahoo.com address at http://mail.yahoo.com
