library cache latch, lock and pin (don't care if I spin)

From: Poras, Henry R. <Henry_Poras_at_dfci.harvard.edu>
Date: Wed, 10 Mar 2004 14:24:30 -0500
Message-ID: <0D3AAB332E29BA4F9D8BE559F244A914558ACB@phsexch27.mgh.harvard.edu>

I just want to make sure I am picturing this correctly. In order to lock an object in the library cache you need to adjust the data structure of the library cache object handle. This data structure is protected by a latch, so first you need the latch to lock the handle. If you cannot aquire the library cache lock
(or subsequently the library cache pin), you wait on a 'library cache lock' wait
event. I assume at this point the latch is released and when you reattempt to aquire the lock you need to reaquire the latch. If there are a number of processes attempting to aquire this lock, there can be a queue (does it queue?) of 'library cach lock' waits, and consequently an increase in latch free
(library cache latch) waits, as each process tries over and over to get the
lock.

Is this about right? Normally I acquaint library cache latch waits with a large library cache and lack of bind variables. Can waits for library cache locks/pins also lead to an increase of these latch waits?

Thanks.

Henry

Please see the official ORACLE-L FAQ: http://www.orafaq.com

To unsubscribe send email to: oracle-l-request_at_freelists.org put 'unsubscribe' in the subject line.

--
Archives are at http://www.freelists.org/archives/oracle-l/
FAQ is at http://www.freelists.org/help/fom-serve/cache/1.html
-----------------------------------------------------------------

Received on Wed Mar 10 2004 - 13:35:12 CST