Platform - Oracle 8.0.4.3.1 running on AIX 4.2.1

I setup the DB_BLOCK_LRU_EXTENDED_STATISTICS=20 parameter in my init.ora
file and recycled my database.  There is a 3rd party application server
running on this box that connected to the database, and almost
immediately 2 of it's connections pegged the CPU usage to 100%.  I could
not spawn any SQL*Plus sessions, they just hung following my password,
and the only way out was shutdown abort.  Shutdown immediate attempts
hung, even though there were no open connections to the database.  I
called Oracle support, and their solution was not to use this parameter.

There were no trace files generated, and no alert log entries made.

Anybody have any experience/thoughts/insight into this?

Sent via Deja.com http://www.***.com/
Share what you know. Learn what you don't.

All I can offer is that I've used this parameter on 8.0.5.1 on AIX 4.3.2
with no major problems (besides the performance hit that it
encounters)..

Sorry I cannot assist further..
--
===========================================
Connor McDonald
"These views mine, no-one elses etc etc"

"Some days you're the pigeon, and some days you're the statue."

Hi

Read your doc's on tuning memory in the online guides. This collects
stats in the v$recent_bucket view which helps tell you the added benefit
(cache hits) of adding for db_block_buffers.

A similar parameter is db_block_lru_statistics which gathers info that
is accessed via the v$current_bucket view. This view tells you about the
effect of removing block buffers - ie how much less hits you will incur.

I may have the uses mixed around, but that the long and short of it. It
does have a performance penalty, but I cannot say how much.

Check the html pages that come with your server install cd.

Good luck
Kugendran Naidoo
South Africa