If you reduce the diskpool to 0 utilization before the beginning of the
window and keep the hi/lo very low for the first half of the window, with
a maximum practicable migpr set, you should be fine.

It's a juggling act we have every night. I know we have set it right when
I get in the morning and no nodes are queueing for tapes and the diskpool util
is greater than 75%.

Cheers, Suad
--
On Wed, Oct 31, 2001 at 12:46:53PM -0500, Selva, Perpetua wrote:
On Wed, Oct 31, 2001 at 12:46:53PM -0500, Selva, Perpetua wrote:
> >       Hi,
> >
> >       Does anyone know what might be the cause of this problem?
> >
> >       I have a diskpool of 40G, but my daily backups on a nightly basis
> > exceeds 100G.
> >       We are in the process of purchasing more disks.
> >
> >       However, this problem has not occured till last nite.
> >
> >       Is it fair to assume that the backup from clients to diskpool is
> > faster than the rate
> >       of migration onto tape depending on the threshold set for the
> > diskpool.
> >
> >       For the interim period, i have set it lower than the what it was set
> > to.. Would this
> >       be sufficient? or should I look at other ways to improve this? or
> > occuring again?
> >
> >         [Selva, Perpetua]  Thanks in Advance