ctime history - Unix

This is a discussion on ctime history - Unix ; The ctime is updated by any operation that changes the inode, with one exception: the change to the atime that results from a read() operation does not update the ctime. My understanding was that the ctime was put into the ...

+ Reply to Thread
Results 1 to 4 of 4

Thread: ctime history

  1. ctime history


    The ctime is updated by any operation that changes the inode, with one
    exception: the change to the atime that results from a read()
    operation does not update the ctime.

    My understanding was that the ctime was put into the inode to provide
    advice to the tape backup system about which inodes needed to be
    backed up. With the atime exception, a file's atime changes will not
    always be backed up, and so could be incorrect after a restore.

    Why was this exception originally put into the Unix semantics?

    Also, the Unix first edition manual, and the 1974 CACM paper by
    Thompson and Ritchie both refer to the inode as containing the "time
    of creation" (!!) of the file. When did the "time of creation" become
    the inode change time, and why? Was it a true semantic change, or was
    it only a change in name?





  2. Re: ctime history

    In article ,
    mjd@red.seas.upenn.edu (Mark Jason Dominus) wrote:

    > The ctime is updated by any operation that changes the inode, with one
    > exception: the change to the atime that results from a read()
    > operation does not update the ctime.
    >
    > My understanding was that the ctime was put into the inode to provide
    > advice to the tape backup system about which inodes needed to be
    > backed up. With the atime exception, a file's atime changes will not
    > always be backed up, and so could be incorrect after a restore.


    True.

    >
    > Why was this exception originally put into the Unix semantics?


    Probably because this would cause too many files to be backed up. For
    example, most of the programs and libraries in /usr/bin and /usr/lib
    would get backed up every time.

    --
    Barry Margolin, barmar@alum.mit.edu
    Arlington, MA
    *** PLEASE post questions in newsgroups, not directly to me ***
    *** PLEASE don't copy me on replies, I'll read them in the group ***

  3. Re: ctime history



    On Jan 24, 8:22 am, Barry Margolin wrote:

    > > Why was this exception originally put into the Unix semantics?

    > Probably because this would cause too many files to be backed up. For
    > example, most of the programs and libraries in /usr/bin and /usr/lib
    > would get backed up every time.


    Sure. But the dump program could have been written to dump only the
    inode, and not the file contents, in such a circumstance.


  4. Re: ctime history



    On Jan 23, 3:08 pm, m...@red.seas.upenn.edu (Mark Jason Dominus) wrote:

    > Also, the Unix first edition manual, and the 1974 CACM paper by
    > Thompson and Ritchie both refer to the inode as containing the "time
    > of creation" (!!) of the file. When did the "time of creation" become
    > the inode change time, and why? Was it a true semantic change, or was
    > it only a change in name?


    In the course of researching this, I made a number of interesting
    discoveries.
    The mtime was not added to Unix until v7. Prior to that, inodes had an
    atime and a ctime. But in v6 and earlier, the ctime was called
    "mtime".

    At least as early as v5, there was no creation time.

    Full details are at http://blog.plover.com/Unix/ctime.html .


+ Reply to Thread