On 4/11/07, Don Seiler <don@xxxxxxxxx> wrote:
AUTO SGA: Disabling background sga tuning due to err 604 Also, is the sga auto tuning still disabled? Where can one check?
After reading Jonathan Lewis' article at http://jonathanlewis.wordpress.com/2007/04/16/sga-resizing/, I ran the query against v$sga_resize_ops and sure enough there hasn't been any operations since 11-APR-07 02:30:38, when this error happened. v$sga_resize_ops also shows an INSANE amount of activity, each time shrinking the buffer cache (db_cache_size) and growing the shared pool (shared_pool_size). What confuses me is that the INITIAL_SIZE and FINAL_SIZE never change during this flurry. Here is a small sample of the output, there is lots more just like it: START_TIME OPER_TYPE COMPONENT PARAMETER OPER_MODE INITIAL_SIZE FINAL_SIZE ------------------ ---------- -------------------- -------------------- --------- ------------ ---------- 11-apr-07 02:30:29 SHRINK DEFAULT buffer cache db_cache_size IMMEDIATE 838860800 838860800 GROW shared pool shared_pool_size IMMEDIATE 587202560 587202560 11-apr-07 02:30:30 SHRINK DEFAULT buffer cache db_cache_size IMMEDIATE 838860800 838860800 GROW shared pool shared_pool_size IMMEDIATE 587202560 587202560 11-apr-07 02:30:33 SHRINK DEFAULT buffer cache db_cache_size IMMEDIATE 838860800 838860800 GROW shared pool shared_pool_size IMMEDIATE 587202560 587202560 I'm considering the suggestion of setting decent minimums for the pool sizes, but I don't see how it would address this particular issue. Perhaps it is just that bug, and perhaps ASMM isn't as fantastic as I thought. -- Don Seiler oracle: http://seilerwerks.blogspot.com ultimate: http://www.mufc.us -- //www.freelists.org/webpage/oracle-l