Re: New feature

Mark C. Stock wrote:

>
> "Hans Forbrich" <forbrich_at_yahoo.net> wrote in message
> news:N7Xac.4414$hH3.3538_at_edtnps84...
> | Mark C. Stock wrote:
> |
> | >
> | > "Glen A Stromquist" <glen_stromquist_at_no_spam_yahoo.com> wrote in
> | > message news:JwWac.4061$hH3.3709_at_edtnps84...
> | > | Daniel Morgan wrote:
> | > |
> | > | > Jonathan Lewis wrote:
> | > | >
> | > | >> This only works if you set hidden parameter
> | > | >> _unbreakable = false
> | > | >
> | > | >
> | > | > But to get that to work don't you also have to set
> | > | >
> | > | > _dba = expendable
> | > | >
> | > | I just tried another undocumented feature,
> _dba_auto_pay_increase=TRUE,
> | > | on our financial system.
> | > |
> | > | I'll keep you all posted....
> | > |
> | > |
> | >
> | > there is a posted bug on that feature (available only via the MetaLink
> | > Classic interface, and only for 12z beta testers) that indicates it
> | > does not handle NULLs properly
> | >
> |
> | When I checked, the reference was that _dba_auto_pay_incr_rate had to be
> set
> | and could not be left null. Otherwise the increase rate was
> | indeterminate (leading to the black hole of tri-valued logic).
> |
> | BTW ... From what I can tell, it's best to set _dba_auto_pay_incr_rate
> | to
> a
> | multiple of DB_CACHE_SIZE to get optimal DBa performance.
> |
> |
>
> you need to do a little more testing, i'm pretty sure the recommendation
> is an exponent of DBA_CASH_SIZE
>
> ;-{ mcs

It surely seemed like a typo. Your analysis makes more sense ... and ties in closely with the (undocumented) definition of the _dba_cash_size_target parameter

/HF Received on Thu Apr 01 2004 - 10:31:13 CST