Curtis Preston wrote:

So Sklyar, you are requesting to create an archive that's tracked in
TSM, able to be specified at any level (e.g. filespace, directory,
file), AND to make that within the server (instead of it being required
to create essentially a full backup of the files being archived).

Do I have it now? ;)

Yup.

Although you didn't ask for free advice, I can't help myself, as this
falls onto one of my hot buttons.  ;)

I have no problems with free advice. :)

I believe that this falls into the category of "using backup software as
archive software," which I strongly discourage, almost regardless of
which software package we're talking about.  The biggest use of archives
e-discovery requests. An e-discovery request says, "give me all the
files/emails that have the phrase ABC in them, or that were created by
FRED SMITH."  You cannot meet such requests with backup software, as the
first question will be "what server/filespace were those files on?"  The
e-discovery request doesn't know and you don't know.  (An actual
archiving product would know.)

Unless you're required to keep them by regulation, actually having
backups that are years old is a big liability when you get an
e-discovery request, as the FRCP guidelines specify that if you have it,
you have to give it up.  And if you have really old backups, getting the
data for an e-discovery request can be brutal, incredibly
time-consuming, and costly.

So, unless your backup software can meet these requirements, you're
doing yourself more harm than good making archives.

Some of the labs I support get grant money from the Howard Hughes
Medical Institute, which requires that source data be kept for seven
years. Source data is defined such that it's any information used to
back up research publications, so even processed data has retention
requirements. The biggest use of these archives would be for HHMI to
verify claims made in the labs' publications. The only data subject to
archival retention is the source data; everything else is backed up with
TSM, with the inactive copies lasting a year.

--
-- Skylar Thompson (skylar2 AT u.washington DOT edu)
-- Genome Sciences Department, System Administrator
-- Foege Building S048, (206)-685-7354
-- University of Washington School of Medicine