Skip site navigation (1)Skip section navigation (2)

  
 
  

home | help
DTAR(1) 			  mpiFileUtils				 DTAR(1)

NAME
     dtar - create and extract a tar archive

SYNOPSIS
     dtar [OPTION] -c -f ARCHIVE SOURCE...

     dtar [OPTION] -x -f ARCHIVE

DESCRIPTION
     Parallel MPI application to create and extract tar files.

     dtar  writes archives in pax file format.	In addition to the archive file,
     dtar creates an index to record the number of items and the  starting  byte
     offset  of each entry within the archive.	This index enables faster paral-
     lel extraction.  By default, dtar appends its index as the  last  entry  of
     the archive.  Optionally, the index may be written as a separate file (with
     a	.dtaridx extension) or as an extended attribute (named user.dtar.idx) of
     the archive file.

     dtar can extract archives in various tar formats, including  archive  files
     that  were created by other tools like tar.  dtar can also extract archives
     that have been compressed with gzip, bz2, or compress.  Compressed archives
     are significantly slower to extract than uncompressed archives, because de-
     compression inhibits available parallelism.

     Archives are extracted fastest when a dtar index exists.  If an index  does
     not exist, dtar can create and record an index during extraction to benefit
     subsequent extractions of the same archive file.

     When  extracting  an archive, dtar skips the entry corresponding to its in-
     dex.  If other tools, like tar, are used to extract the archive, the  index
     entry  is extracted as a regular file that is placed in the current working
     directory with a file extension of ".dtaridx" and having the same	basename
     as  the  original	archive  file.	For an archive that was named "file.tar"
     when it was created, the dtar index file is named "file.tar.dtaridx".

LIMITATIONS
     dtar only supports directories, regular files, and symlinks.

     dtar works best on Lustre and GPFS.  There are no	known  restrictions  for
     creating  or extracting archives on these file systems.  These file systems
     also deliver the highest bandwidth and file create rates.

     dtar can be used on NFS, but there is one	key  restriction.   Namely,  one
     should  not  create  an  archive  file in NFS.  To create an archive of NFS
     files, the archive file itself should be written to a directory  in  Lustre
     or  GPFS.	 The dtar tool writes to an archive file from multiple processes
     in parallel, and the algorithms used to write the archive are not valid for
     NFS.

     dtar can be used to extract an archive file into  NFS.   The  archive  file
     that is being extracted may be on any file system.

     The target items to be archived must be under the current working directory
     where dtar is running, so commands like these work.

     dtar -cf foo.tar foo/

     dtar -cf foo.tar dir/foo/

     But commands like the following are not supported:

     dtar -cf foo.tar ../foo/

     dtar -cf foo.tar /some/other/absolute/path/to/foo/

OPTIONS
     -c, --create
	    Create a tar archive.

     -x, --extract
	    Extract a tar archive.

     -f, --file NAME
	    Name of archive file.

     -C, --chdir DIR
	    Change directory to DIR before executing.

     --preserve-owner
	    Apply recorded owner and group to extracted files.	Default uses ef-
	    fective uid/gid of the running process.

     --preserve-times
	    Apply  recorded  atime  and  mtime to extracted files.  Default uses
	    current system times.

     --preserve-perms
	    Apply recorded permissions to extracted  files.   Default  subtracts
	    umask from file permissions.

     --preserve-xattrs
	    Record  extended  attributes  (xattrs) when creating archive.  Apply
	    recorded xattrs to extracted files.  Default does not record or  ex-
	    tract xattrs.

     --open-noatime
	    Open source files with O_NOATIME flag when creating archive.

     --fsync
	    Call fsync before closing files after writing.

     --bufsize SIZE
	    Set  the  I/O buffer to be SIZE bytes.  Units like "MB" and "GB" may
	    immediately follow the number without spaces (e.g. 8MB). The default
	    bufsize is 4MB.

     --chunksize SIZE
	    Multiple processes copy a large file in parallel by dividing it into
	    chunks.  Set chunk to be at minimum SIZE bytes.  Units like "MB" and
	    "GB" can immediately follow the number without spaces  (e.g.  64MB).
	    The default chunksize is 4MB.

     --memsize SIZE
	    Set  the  memory  limit to be SIZE bytes when reading archive files.
	    For some archives, dtar can distribute the file across processes  to
	    store  segments  of  the  archive  in  memory for faster processing.
	    Units like "MB" and "GB" may immediately follow the  number  without
	    spaces (eg. 8MB). The default is 256MB.

     --progress N
	    Print progress message to stdout approximately every N seconds.  The
	    number of seconds must be a non-negative integer.  A value of 0 dis-
	    ables progress messages.

     -v, --verbose
	    Run in verbose mode.

     -q, --quiet
	    Run tool silently. No output is printed.

     -h, --help
	    Print a brief message listing the dtar(1) options and usage.

EXAMPLES
     1. To create an archive of dir named dir.tar:

     mpirun -np 128 dtar -c -f dir.tar dir/

     2. To extract an archive named dir.tar:

     mpirun -np 128 dtar -x -f dir.tar

SEE ALSO
     The  mpiFileUtils	source code and all documentation may be downloaded from
     <https://github.com/hpc/mpifileutils>

AUTHOR
     HPC

COPYRIGHT
     2022, LLNL/LANL/UT-Battelle/DDN

0.12				  Jan 29, 2025				 DTAR(1)

Want to link to this manual page? Use this URL:
<https://man.freebsd.org/cgi/man.cgi?query=dtar&sektion=1&manpath=FreeBSD+Ports+15.1.quarterly>

home | help