On Mon, Sep 26, 2005 at 01:16:14PM +0300, Hakan Avci wrote:
> Hi Jan,
> 
> OK.I did as what you say.First open file "hakan" ,than open same file on 
> second client.But there is no kernel error or in-use error??is this a 
> bug?or Am I do something wrong??

I guess I didn't understand your question.

Coda has no file locking, we're optimistically assuming that write
sharing is not very common.

An evaluation of this statement can be found in J.J. Kistlers thesis [1]
at page 221. Here he examines the write-sharing patterns over a 12 month
period on an AFS server cluster at CMU with about 300 users.

[1] Disconnected Operation in a Distributed File System
    Kistler, J.J.
    School of Computer Science, Carnegie Mellon University
    May 1993, CMU-CS-93-156 

You can find links to a postscript and an acrobat copy of his thesis
around the bottom of this page,
http://www.cs.cmu.edu/afs/cs/project/coda-www/ResearchWebPages/docs-coda.html

His thesis also has references to several other file access studies.

    Short-Term File Reference Patterns in a UNIX Environment
    Floyd, R
    Tech. Rep. 177 University of Rochester Computer Science Department, 1986

    Directory Reference Patterns in a UNIX Environment
    Floyd, R.
    Tech. Rep. 179 University of Rochester Computer Science Department, 1986

    A Trace-Driven Analysis of the Unix 4.2 BSD File System
    Ousterhout, J., Da Costa, H., Harrison, D., Kunze, J., Kupfer, M.,
    and Thompson, J.
    Proceedings of the Tenth ACM Symposium on Operating System
    Principles (December 1985)

    A Study of File Sizes and Functional Lifetimes
    Satyanarayanan, M.
    Proceedings of the Eighth ACM Symposium on Operating System
    Principles (December 1981)

    Long Term File Reference Patterns and their Applications to File
    Migration Algorithms
    Smith, A.
    IEEE Transactions on Software Engineering 4, 7 (July 1981)

Jan