There is something diabolical, and honestly I really do not understand, in the
way that IBM indicates to quoting the Spectrum Protect Suite consumer-metric
TB/FrontEnd (FE) model.
IBM spread a manual "IBM Spectrum Protect Suite - FE Licensing Guide", 64
pages(!) in Ver.7.1.6, procedures that almost certainly lead to the execution
of Perl or Powershell scripts (based on the sysop)
data collection utility (dsmfecc - CRT) etc .... and all this to only know the
occupation of the active data for one or more nodes.
Moreover those scripts executes SQL queries to the Spectrum server, widely
known as "custodian" of node's data. I think it is more simple to share these
SQL SELECT and document that, specifying application field, execution and a
matrix of compatibility. Isnt' it?
Since it is increasingly being leveraged on the OC portal, which is widely
advertised reporting, it was no longer appropriate to complete that portal so
that there is a single point of
aggregation and visibility of these occupations?
Nowadays, given the availability of backup software, given the "zero tolerance"
average, to any complexity or constriction of constructs, still it makes sense
to produce such procedures
so complex, and that let me say, close to madness? (It is unthinkable to run
scripts on farms made of hundreds or thousands of nodes with different types or
login credentials!)
Why this? Why is the FE data in most of our installations the most convenient
for a quotation?
I believe that the future of this product is still in waters made of complexity
and difficulty in management and...Error is only human, but to persist in it,
is diabolical!
Sorry for my poor english. Here speak a disappointed SP (TSM) fan.
________________________________________________________
Tommaso Bollini
Spectrum Protect Specialist
|