Splitting up sets of files for archiving (e.g. to tape, optical media)

Ronald F. Guilmette rfg at tristatelogic.com
Fri Jan 19 02:15:57 UTC 2018


This isn't really FreeBSD specific, but in my experience the folks on
this list have a lot of knowledge about a lot nice, useful free software
tools, so I hope nobody will begrudgd me for asking this question here.

I'm looking for a pre-existing software tool, which may or may not already
exist, and which will do the following job...

Problem statement:

Imagine that you have a big set of files that you would like to archive
to some sort of archiving media, such as tapes, or optical media, where
each unit of said archiving media has a capacity considerably less than
the total aggregate size of all of the files you want to archive.

Imagine further that you would like your set of input files to be spread
across the units of the output (archive) media such that no single input
file is ever split across more than one unit of the output media, in order
to simplify recovery/restore of individual files.

Lastly, assume that it is desired to minimize, as much as reasonably
possible, the total number of output (archive) media units used to
archive the entire set of input files.  (And to further this goal,
it is acceptable for files from any single input subdirectory to be
scattered among the various output media units.

+_+_+_+_+_+_+_+_+_+_

In my case, I want to archive several hundred gigabytes onto a set of
blank BD-R disks.

I plan to use ImgBurn to actually write the BD-R disks.

So basically, I just need a tool to analyze the input file set, applying
some sort of bin packing algorithm, and then spit out a list of which
specific files should go into each specific archive volume, e.g. #01, #02,
#03... etc.  Each such set of files will then, in turn, be hard-linked
into a temporary directory, and then, one by one, ImghBurn will be told
to write each of these temp directories to a single output BD-R disk.

I have written a small software tool to do the above "splitting" job,
and I am currently improving upon it, but it occured to me that I
should at least ask if someone else has perhaps already perfected this
exact wheel that I am busy re-inventing.


Regards,
rfg


P.S.  It seems unlikely that I'm the first and only person to have ever
written a tool to do this specific job, but on the off chance that I am,
I am more than willing to contribute my little tool to the ever-expanding
ports tree.


More information about the freebsd-questions mailing list