RE: data buffer access confusion


With 11g,  not just PARALLEL  but also SERIAL  Direct Path Read operations.
See <http://oracledoug.com/serendipity/index.php?/archives/1321-11g-and-direct-path-reads.html>http://oracledoug.com/serendipity/index.php?/archives/1321-11g-and-direct-path-reads.html http://afatkulin.blogspot.com/2009/01/11g-adaptive-direct-path-reads-what-is.html 
and
<http://dioncho.wordpress.com/2009/07/21/disabling-direct-path-read-for-the-serial-full-table-scan-11g/>http://dioncho.wordpress.com/2009/07/21/disabling-direct-path-read-for-the-serial-full-table-scan-11g/
I do agree with Mark that "I think it's safe to say that Oracle will *never* read data from disk into PGA, and then move from PGA to SGA" (I think !)
The Buffer Cache is for sharing buffers. Blocks read into the PGA are not shared with other sessions. 

Hemant K Chitale

At 03:29 PM Monday, Bobak, Mark wrote:

Aman,
Actually, it depends. Physical reads can be into cache, or directly into PGA. Examples of reads directly to PGA and bypassing buffer cache would be read of temp data from TEMP tablespace or parallel direct reads when parallel query is involved.
However, I think it's safe to say that Oracle will *never* read data from disk into PGA, and then move from PGA to SGA. What would be the point? So, reads can go from disk to PGA or to SGA, but should never "hop" from disk to PGA to SGA. That just doesn't make sense, in any scenario I can imagine. 

Hope that helps,

-




Hemant K Chitale

http://hemantoracledba.blogspot.com





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

Other related posts: