RE: Solaris 10 shmmax
- From: "Bobak, Mark" <Mark.Bobak@xxxxxxxxxxxxxxx>
- To: <kevinc@xxxxxxxxxxxxx>, "Hameed, Amir" <Amir.Hameed@xxxxxxxxx>, <ax.mount@xxxxxxxxx>, <oracle-l@xxxxxxxxxxxxx>
- Date: Mon, 26 Mar 2007 14:05:08 -0400
Does it really matter if your shared memory is in one chunk or not? I
seem to remember a bit of misinformation that used to float around was
that if the SGA was not in a single shared memory chunk, then ISM was
disabled. Turns out that isn't true. I can't think of any reason why
it would be important for the SGA to be in one contiguous memory chunk.
-Mark
--
Mark J. Bobak
Senior Oracle Architect
ProQuest/CSA
"There are 10 types of people in the world: Those who understand
binary, and those who don't."
________________________________
From: oracle-l-bounce@xxxxxxxxxxxxx
[mailto:oracle-l-bounce@xxxxxxxxxxxxx] On Behalf Of Kevin Closson
Sent: Monday, March 26, 2007 1:09 PM
To: Hameed, Amir; ax.mount@xxxxxxxxx; oracle-l@xxxxxxxxxxxxx
Subject: RE: Solaris 10 shmmax
The 490 is a flat memory box and the Solaris port is smart enough to
behave correctly.
On the other hand, I have a lab server( Sun v490) running 10.2.0.2 on
Solaris9. The shmmax is set to 20GB. My SGA is 12GB and I am seeing only
one segment from ipcs. I have not tinkered with the NUMA optimization
parameter. Per Metalink note 39926.1, multiple shared memory segments in
10gR2 are there for performance reason. Why is the behaviors of shared
memory segments in 10gR2 is different on the same platform (v490 versus
6900)?
Other related posts:
