Steve,

80% is a good rule of thumb.  The idea being that for master sets, the
hashing of the indexes starts to become much less efficient.  This is
because the location that the key value hashes to may be already in use by a
previous entry, so another location is used.  Then, a find on that key value
takes extra disk seeks.  Similar structural degradations can occur with
other types of indexes and sets  Also, the 80% rule is often used to make
sure the dreaded "dataset full" error does not occur.

For an explanation of the internal indexing mechanisms, see the
TurboImage/XL manual under "Internal Techniques."

For a better way to prevent full datasets, see the section on Dynamic Data
Set Expansion

-Dave

> -----Original Message-----
> From: STEPHEN HUNTER [mailto:[log in to unmask]]
> Sent: Monday, November 27, 2000 8:17 AM
> To: [log in to unmask]
> Subject: Adager
>
>
> Hello all, I would like some help with some info on the
> Database Tool called
> ADAGER..  I am familar with a product called Flexibase, but
> this is my 1st
> exposure to ADAGER.  My question is this:  Currently, the
> operators that I
> have just taken over, will run an Adager on ANY dataset that
> shows up as 76
> to 80 % full(They get these figures from running a capacity
> program that
> shows the percentage of each dataset)
>
> This HP system is running MANMAN as a front-end..  I find
> this very curious
> that one would run Adager on ANY Dataset that is 76 to 80%
> full. Is this
> comment practice??  If not, what is a good measuring tool as
> to when to run
> the Adagers on these datasets??
>
> Any help is greatly appreciated...
>
> Thanks...
> ______________________________________________________________
> _______________________
> Get more from the Web.  FREE MSN Explorer download :
http://explorer.msn.com