Hi Ujang Gee you get all the good ones to work out. http://jonathanlewis.wordpress.com/2006/12/04/resizing-the-sga/ has some comment about the use of SGA_TARGET The following metalink notes may assist you in identifying if you are able to get it to work or if you need to find a suitable value for a fixed SGA Note:396940.1, Note:62143.1, and Note:146599.1 might all help you to find the cause of this, whether it is a problem with the application or 4GB is too small. I am suspecting that it is an application design issue that you will find, probably not reusing code effectively, may not use binds effectively HTH Cheers Peter -----Original Message----- From: Ujang Jaenudin [mailto:ujang.jaenudin@xxxxxxxxx] Sent: Tuesday, 10 July 2007 02:08 PM To: oracle-l@xxxxxxxxxxxxx Subject: Re: ORA-4031 (10.1.0.5 on solaris) folks, found at metalink to see free of this query, when > 8400 we could flush the shared pool. SQL> SELECT KSMCHCLS CLASS, COUNT(KSMCHCLS) NUM, SUM(KSMCHSIZ) SIZ FROM X$KSMSP WHERE KSMCHSIZ > 8400 GROUP BY KSMCHCLS; 2 CLASS NUM SIZ -------- ---------- ---------- R-free 293 1053020760 R-freea 283 11540472 R-perm 9 11824008 free 13942 1762175536 freeabl 1115 25464840 perm 812 473364592 recr 11 1168544 according metalink bug. 4238390 , this could be gone by flushing shared_pool. so, what is the root cause of this case? regards ujang On 7/10/07, Ujang Jaenudin <ujang.jaenudin@xxxxxxxxx> wrote: > folks, > > shared pool is 4GB > sga_target activated > > but so many ORA-4031 occured. > the effect are some packages were invalidated..... > > i found on the trace files that so many vary requests of heap, such as > > kglsim heap > 380097728 > > kglsim object batch > 380097728 > > kggfaAllocSeg > > kgghteInit > 1505964060 > > thanks b4 for your share.... > > > -- > regards > ujang > -- regards ujang -- //www.freelists.org/webpage/oracle-l -- //www.freelists.org/webpage/oracle-l