RE: relation shared pool elements phys mem

  • From: Feighery Raymond <Raymond.Feighery@xxxxxxxxxxxxx>
  • To: "'oracle-l@xxxxxxxxxxxxx'" <oracle-l@xxxxxxxxxxxxx>
  • Date: Mon, 9 Aug 2004 13:35:26 +0100

Find out how much free space is in your large pool. You may wish to monitor
this figure over time to see when it is shrinking.

select  bytes
from    v$sgastat
where   name  = 'free memory'
and     pool = 'large pool';

See which sessions are taking a constantly increasing amount of UGA:

set pages 999
compute sum of kbytes on report
break on report
column username format a15 heading "Username"
column kbytes format 999,999,999.99  heading "Current UGA Memory|(kbytes)"
SELECT  username,
        sess.sid,
        value/1024 kbytes
FROM    v$session sess,
        v$sesstat stat,
        v$statname name
WHERE   sess.sid = stat.sid
AND     stat.statistic# = name.statistic#
AND     name.name = 'session uga memory'
order by value
/

I don't have a HP machine at the moment, so I can't test any HP memory
utilities.

How do your users connect to the database? I know you use MTS, but is there
an application or web server layer? It is possible with connection pooling
that sessions never release the memory they grab.

Ray


-----Original Message-----
From: Jeroen van Sluisdam [mailto:jeroen.van.sluisdam@xxxxxxxxxx]
Sent: Friday, August 06, 2004 6:22 PM
To: 'oracle-l@xxxxxxxxxxxxx'
Subject: RE: relation shared pool elements phys mem


I'm sorry I forgot to mention some detail:
I have 1 Gb available, HP-UX 11.11,
Oracle 9.2.0.4 I already use mts as default because in
Our production environment it is necessary and we want development to
Be as close as possible. Already downsized sga including buffers.
Large pool was set to 5 mb. Runnning 2 databases with 
Total sga around 120 Mb each.

Still top shows around 5 to 8 Mb free memory. Unix shells are acting weird
Also, hanging for several seconds. The errors come with starting a new 
Sqlplus session now and then. Dedicated connections for example
A oem session of my client I don't have sofar

 Didn't see anything strange in syslog. Have these problems for
Months now but were not important up till now several new development
projects are starting. The shmax parameter is set to 1 Gb. I would like to
Understand what the possible relations are between sga and memory usage on
OS level. I know SGA has to fit into 1 memory segment. But is there any
default memory segment size such that my 1Gb is already up with 2 db'es
running? I don't have glace available. Any other trics to measure how much
Memory is used by which process?

Tnx,

Jeroen
-------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------
----------------------------------------------------------------
Please see the official ORACLE-L FAQ: http://www.orafaq.com
----------------------------------------------------------------
To unsubscribe send email to:  oracle-l-request@xxxxxxxxxxxxx
put 'unsubscribe' in the subject line.
--
Archives are at //www.freelists.org/archives/oracle-l/
FAQ is at //www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Other related posts: