From: Robert Thurlow (Robert.Thurlow@eng.sun.com)
Date: 10/10/02-01:53:20 PM Z
Date: Thu, 10 Oct 2002 12:53:20 -0600 (MDT) From: Robert Thurlow <Robert.Thurlow@eng.sun.com> Subject: RE: Replication/Migration issues list Message-ID: <Roam.SIMC.2.0.6.1034276000.14788.thurlow@jurassic.eng> Ted wrote: > > > A filesystem version number helps the NFS client provide better > > > semantics in several ways. Dave wrote: > First, I think this is an interesting facility even within the confines > of fairly standard NFS implementations. To see that, just turn Ted's > statement around ... [to] > "When the client checks the number and sees that it > hasn't changed, it knows that the filesystem has not been updated > and so it does not have to do any revalidation of individual cached > objects." Thanks, that helped - and I agree this is interesting. We'd need to work out lots of semantics, but I will keep this on the table as something to work on (I guess mostly as an NFSv4 core protocol feature for a future minor version). Rob T
This archive was generated by hypermail 2.1.2 : 03/04/05-01:50:25 AM Z CST