Re: strange behavior in 9.2.0.4 - no blocks in keep buffer (KEEP broken?)

From: Paul Drake <discgolfdba_at_yahoo.com>
Date: Wed, 9 Jun 2004 22:06:15 -0700 (PDT)
Message-ID: <20040610050615.23141.qmail@web20418.mail.yahoo.com>

• Connor McDonald <hamcdc_at_yahoo.co.uk> wrote:
  > Note that the FAQ item assumes that if you define an
  > object as KEEP then it will indeed go into
  > the KEEP pool...Sounds fair enough but it seems to
  > suggest that KEEP (and I think RECYCLE as well)
  > may be broken in v9. Segment header blocks from the
  > candidate table seem to go into the correct
  > pool but that seems to be all.
  >
  > Here's my demo - I'll be very glad if someone proves
  > me or the queries below wrong here - I'd like
  > to use a KEEP pool on v9 but have so far resisted
  > because of what follows:
  >
  

Connor,

All that I did to make it work was, give it some time. I was distracted by other things.
By the time I recompiled the view with Jonathan's fix, the keep pool was populated.
the environment was not so tightly controlled, as some developers were testing code there, but none with dba access (sys_privs, role) of course. I wish I knew what kicked in the 'normal' behavior.

this one took a back seat to a test case that ran fine in 8.1.7, but blew thru 3.5 GB of temp in 9.2 when optimizer_features_enabled was set forward to 9.2.

does one *have* to open an iTAR to use
_complex_view_merging in production?

Paul                          

---

Do you Yahoo!?
Friends. Fun. Try the all-new Yahoo! Messenger. http://messenger.yahoo.com/

---

Please see the official ORACLE-L FAQ: http://www.orafaq.com

---

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.

--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Received on Thu Jun 10 2004 - 00:03:01 CDT