Re: The Cache Hint

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Mon, 12 Mar 2012 12:40:43 -0000
Message-ID: <6A5FEB3F55994587BFE6DCAF54BDB49B_at_Primary>


Too many unknowns.

One thing I can say, though, is that nothing gets loaded into the MRU end of the LRU, everything goes to the mid point. One thought about the cache hint - it's possible that it changes (again) in 11.2 as a side effect of serial direct path reads.

In 11.2 a table over 25% of the cache size (whatever that means exactly) will be read using serial direct path reads (unless specified as parallel) - perhaps if you specify the cache hint it will be read into the cache instead. Conversely, if you specify nocache might this force a serial direct read ?

Regards

Jonathan Lewis
http://jonathanlewis.wordpress.com
Oracle Core (Apress 2011)
http://www.apress.com/9781430239543

  • Original Message ----- From: "Sidney Chen" <huanshengchen_at_gmail.com> To: <oracle-l_at_freelists.org> Sent: Sunday, March 11, 2012 11:56 AM Subject: The Cache Hint

So I wonder the cache hint is just for the small tables, placing the fresh blocks from disk to the MRU end of the LRU list, rathen than the middle point, to let the blocks live longer in buffer pool. I think it makes sense, because if the cache hint can be used to cache any table, it will be a serious potential threat to the buffer pool, anyone will select privelege on big table and do damage to the buffer pool.

--

http://www.freelists.org/webpage/oracle-l Received on Mon Mar 12 2012 - 07:40:43 CDT

Original text of this message