RE: X$ksmsp (OSEE 10.2.0.2 on Solaris 8)

Sorry, let me clarify. We have thousands of entries in ksmsp as well, but I was under the impression that 1 row = 1 contiguous chunk. Sorry for the confusion.

From: Hallas, John, Tech Dev [mailto:John.Hallas_at_gb.vodafone.co.uk] Sent: Monday, June 26, 2006 9:57 AM
To: Schultz, Charles; oracle-l_at_freelists.org Subject: RE: X$ksmsp (OSEE 10.2.0.2 on Solaris 8)

Whilst I was looking at x$ksmsp I came across the following link which provides a useful summary of the x$tables and some interesting queries against some of them.  

http://www.stormloader.com/yonghuang/computer/x$table.html  

Charles, in your database is there only 1 entry in total for that table as I see many thousands of entries (same Oracle version but 2.9 Solaris)  

From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Schultz, Charles Sent: 26 June 2006 15:27
To: oracle-l_at_freelists.org
Subject: X$ksmsp (OSEE 10.2.0.2 on Solaris 8)  

Granted that x$ksmsp is not documented and therefore subject to changes from version to version (or patch to patch), I have an Oracle Support Engineer telling me that one entry in ksmsp might actually show fragmented memory in the case of a memory leak. My understanding, from what I have gleaned from others much smarter than I, was that each row of ksmsp showed one contiguous chunk of memory, be it small or large, FREE or PERM (or something in between). The whole idea of a memory leak causing ksmsp to report a fragmented chunk as one piece is a little disturbing.

Can anyone corroborate or refute this?

Oracle Server Enterprise Edition 10.2.0.2 Solaris 8 (SunOS 5.8) 64-bit

charles schultz
oracle dba
aits - adsd
university of illinois  

--
http://www.freelists.org/webpage/oracle-l