Richard Sims wrote:
>
> On Apr 10, 2007, at 7:57 PM, David Bronder wrote:
>
> > ... However, the
> > automated migrations seem to not be very sensitive to the LOWMIG value
> > (I've been moving it closer to HIGHMIG but the migrations still keep
> > on running). ...
>
> David -
>
> See "Migration" in the TSM Concepts redbook, and "LOwmig"
> in http://people.bu.edu/rbs/ADSM.QuickFacts .

Richard, Steve,

Yeah, I'm aware of how migration works. :)  I'm pretty sure none of
my nodes are consuming 40% or so of any one of my disk pools most
nights, though.

That said, it's also a subjective observation, and with multiple
migration processes running that impression would be further enhanced
(thus my lament about the lack of a MIGPROCESS override).  It certainly
could be that the times I notice it happen to be the times where the
top 3-4 consumers of the pool do in fact consume 40%+ of the pool.

On the other hand, I did notice in TSM 5.3 that a different threshold,
copy pool reclamation I think, became _more_ sensitive to the value.
Where previous versions would continue the reclamation processing to
conclusion even if the threshold changed, TSM 5.3 cancels the process
almost immediately.


--
Hello World.                                    David Bronder - Systems Admin
Segmentation Fault                                     ITS-SPA, Univ. of Iowa
Core dumped, disk trashed, quota filled, soda warm.   david-bronder AT uiowa 
DOT edu