From: Boris Z. (boris@conley.com)
Date: 02/23/99-12:50:41 PM Z
Message-ID: <01BE5F33.81837740.boris@conley.com> From: "Boris Z." <boris@conley.com> Subject: RE: Locking and callbacks Date: Tue, 23 Feb 1999 13:50:41 -0500 On Tuesday, February 23, 1999 12:47 PM, Mike Eisler [SMTP:mre@Eng.Sun.COM] wrote: > > - client A is writing the file, and client B is running a tail -f on > the file. > > - client's A and B memory map the same file, and use > primitives like spin locks to synchronize access to each other. > Note that both NQNFS and Spritely NFS do not solve this either, > because it amounts to client's A and B's being simultaneous > writers to the same file. Again, my impression is that > DCE/DFS would. > Also, we should not forget CIFS. Can two plain vanilla (non-mapped) cases be added to the list? -client A writes to the file and Client B reads from the file - client A flushed its write cache before data returned to client B. or -client A keeps a read cache and client B writes into the cached region - read cache should be discarded.
This archive was generated by hypermail 2.1.2 : 03/04/05-01:46:42 AM Z CST