We run  on a basis of constant data sent.

So a   backup stg diskpool copypool
is run every half hour or so.
Backups are heaviest (but not exclusive) overnight.
In the morning its:
             backup db type=full    (takes 1 hour)
a   backup stg diskpool copypool usually runs simultaneously
then
            expire inventory       (takes 1-3 hours
migrations are  'pushed along' during expiration
onsite reclamation is given over to the rest of the day
if we are not 'stressed' by extra data causing migrations
to be needed.     another backup db at 7 pm
more onsite reclamations  then
the the first half of the overnight backups start at 9pm.

Weekends are exclusively devoted to offsite reclamations after
getting the diskpool down to a 'no migration needs' state
and an expiration plus db backup are run.

Weekend offsite reclamations (of 15-25 tapes) usually takes from
noon Saturday to sometime late Sunday or the wee hours of Monday AM.

This is a two drive 3494/3590 system.  I plan to re-do this all when I
get one or two more drives.  I use a mixture of perl programs and
the (inadequate) TSM scheduling mixed with scripts.
My migration script is:
         backup stg tapepool copypool wait=yes
         backup stg diskpool copypool wait=yes
         upd stg diskpool migproc=2 hi=59 low=49

Then I usually complement it with a scheduled script that
has   diskpool changed to  hi=80 low=65  to get it to stop.

With a 3rd and maybe 4th drive I hope to eliminate the pre-req processing.
And this works best when you have diskpool that is 1+n times bigger than
what gets sent over a 24 hour cycle. (My diskpool has ranged from 70 gig
down to 45 Gigs)  I get about 25 gigs overnight and 15 gigs during the
day.   The key is to have a more than adequate diskpool size and avoid
tapepool to copypool as much as possible.

         ... joe.f.

Joseph A Faracchio,  Systems Programmer, UC Berkeley
Private mail on any topic should be directed to :
   joef AT socrates.berkeley DOT edu

On Fri, 17 Aug 2001, Maurice van 't Loo wrote:

> Hi,
>
> Can anybode say if i have the housekeeping in the right order......
> Just checkin'....
>
> There are no backups between 06:00 and 12:00 and the startup period is 5 min.
>
> 06:00   DB Backup Incremental (devcl=3Dfile)
> 06:05   Migration Start
> 06:10   Migration Reset Settings
> 08:00   Expire Inventory
> 09:00   Reclamation Start
> 09:50   Reclamation Stop (reset settings)
> 10:00   Backup STGpool
> 11:00   Delete Volhist (db & all)
> 11:00   Audit License
> 11:05   DB Backup Full (devcl=3Dtape)
>
>
> Thanks,
>    Maurice van 't Loo
>    The Netherlands
>