RE: Running utlrp creates too many Pnnn processes
Date: Thu, 18 Dec 2008 21:26:27 +0100
I have seen this too. Although I have never tested it, the solution seems to use $ORACLE_HOME/rdbms/admin/utlprp.sql instead, see below for the relevant snippet from the header of that file (BTW ultrp is implemented as a simple call to utlprp with parameter 0)
Rem The degree of parallelism for recompilation can be controlled by
Rem providing a parameter to this script. If this parameter is 0 or Rem NULL, UTL_RECOMP will automatically determine the appropriate Rem level of parallelism based on Oracle parameters cpu_count and Rem parallel_threads_per_cpu. If the parameter is 1, sequential Rem recompilation is used. Please see the documentation for package Rem UTL_RECOMP for more details.
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Maureen English Sent: Thursday, December 18, 2008 8:32 PM To: ORACLE-L
Subject: Running utlrp creates too many Pnnn processes
We recently created some databases on Solaris 10 machines and I'm having a problem running $ORACLE_HOME/rdbms/admin/utlrp.sql.
To the best of my knowledge, I downloaded the correct files for the installation on a SPARC, 64-bit machine. The installation completed successfully and patching to 10.2.0.3.0 also completed successfully. Database creation completed successfully, too.
My problem, however, is that when I run utlrp, Pnnn processes start up uncontrollably. If I have the processes parameter set to 100, so many Pnnn processes start up that it reaches the maximum number of processes allowed. If I increase that value to 200, the same thing happens. I also tried increasing the sga_max_size and sga_target values, but that also made no difference.
Has anyone ever seen this problem before? I've searched Metalink, and opened a Service Request with Oracle, but have not gotten any information back, yet.
Any thoughts/suggestions are welcome.
http://www.freelists.org/webpage/oracle-l Received on Thu Dec 18 2008 - 14:26:27 CST