Re: Shared Pool causing packages to automatically go invalid!

  • From: "Alex Gorbachev" <gorbyx@xxxxxxxxx>
  • To: paulastankus@xxxxxxxxx
  • Date: Mon, 6 Nov 2006 20:03:19 -0500


Just an idea... package status is stored somewhere in the dictionary
so invalidation should cause a datablock change in the system
tablespace and obviously generate some redo. Now from there on - Log
Miner is your friend (if my memory serves me ok, it comes with 8i and
can analyze redos from 8.0 even). If you are really desperate - redo
block dumps but that's more tricky.
Of some relevance might be this note I posted recently -


On 11/1/06, Paula Stankus <paulastankus@xxxxxxxxx> wrote:

Version:  Oracle 8.1.7 - sigh

I have been trying to get production owner accounts locked down to the dba
group and not opened to vendor with turn-over and staffing issues.  Today I
spent a good deal of time (less time next time) tracing a performance
emergency to one invalid package body.  I have some questions:

-Is the last_ddl_time in user_objects the definitive time in which the
package could have become invalid?

-Is there anyway for sure I can tell who, how, when a package became

-The developer is saying that it relates to the size of the shared pool.
Can the shared pool being "too" small cause a database object to just "go"
invalid.  I never heard of such a thing.


Low, Low, Low Rates! Check out Yahoo! Messenger's cheap PC-to-Phone call

Best regards,
Alex Gorbachev

The Pythian Group
Sr. Oracle DBA

Other related posts: