-----David Longo wrote: -----

>The majority of the restore time is spent in formatting the DB and
>logs. They are actually restoring from a fileclass backup of the
>TSM DB. I know when they told me these times I found them hard to
>believe as well.

Our DR plan is similar to the one described in earlier postings. We
have been able to shorten the process significantly by running
dsmfmt processes in parallel. At our last DR test we started all
35 dsmfmt commands at the same time, each responsible for a file
a little over 2 GB in size. The processes finished in rapid
succession about 15 minutes later. This was on a mainframe Linux
system.