mkcodabf will take an existing large file and produce
a directory tree rooted at new-dir of much smaller files,
called hunks. Each hunk, except the last hunk, will be an integral
number of megabytes as controlled by the -s flag. The number
of hunk files and subdirectories in each directory is controlled by
the -f flag. Also, in the
directory new-dir, a meta-data
file, named _Coda_BigFile_, will be created so that new-dir
will appear as a large, read only, regular file after being written to
the Coda Distributed File System.
The reason for these big files is to allow one to write a very
large file, one that is larger than the venus cache, to Coda
and to be able to read it back. These are primarily expected to
be some kind of media files which are written once and read many
times, often in a sequential fashion by a media player. For this
reason, file is expected to not be stored in the Coda file
tree and new-dir is expected to be in the Coda file tree.
mkcodabf supports the following options:
The number of hunk files or subdirectories in each directory
in the big file directory tree. The default number is 100
entries per directory.
The size of each hunk file in megabytes. The default hunk
size is one megabyte.
Print verbose output as each directory and file is created.
The default is to quietly create the directory tree.