Big Brother tells me that Alex Paschal wrote:
> Hi, Jack. There's a FAQ at Richard Sims's site,
>
> http://people.bu.edu/rbs/ADSM.funcdir
>
> There's a searchable archive at
>
> http://www.adsm.org
>
Thanks. I knew of the searchable archive, but it wouldn't help
much: a search for DISK and RECLAMATION would turn up too many results
to be very useful.
The documentation in TSM didn't help much, either. I have finally
found that TSM does not need to reclaim DISK space (though I still do
not know if it fragments files, or reclaims DISK volumes to themselves).
I found this by reading what _wasn't_ in the documentation. The reference
guide has an option under DEFINE STGPOOL to specify what stgpool to use
to reclaim the one you are defining. However, the specific docs for
DEFINE STGPOOL DISK is missing that option. I can thus infer that it does
not need to do reclamation. The docs should really spell this out, as I
am still interested in knowing if TSM will just fill the holes (in which
case the database could grow significantly, since each file could have
multiple entries to locate all of its fragments) or if it shifts the data
down the volume to recover space at the end. Only time will tell...
PGP signature
