Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: High CPU usage

Re: High CPU usage

From: shiva <sshivayogi_at_gmail.com>
Date: 11 Oct 2006 21:32:12 -0700
Message-ID: <1160627532.134172.168760@h48g2000cwc.googlegroups.com>


> Are you running in shared server mode or dedicated? Dedicated

> Is this config the same as the old system? YES

> Could you post following please?
> 1. your init file
> 2. the output of a sqlplus "show sga"
node2.__db_cache_size=2919235584
node1.__db_cache_size=3154116608
node2.__java_pool_size=134217728
node1.__java_pool_size=134217728
node2.__large_pool_size=33554432
node1.__large_pool_size=33554432
node2.__shared_pool_size=788529152
node1.__shared_pool_size=553648128
node2.__streams_pool_size=402653184

node1.__streams_pool_size=402653184
*._job_queue_interval=1
*.cluster_database_instances=2
*.cluster_database=true
*.compatible='10.2.0.2.0'
*.db_block_size=8192
*.db_cache_size=536870912
*.db_file_multiblock_read_count=16
*.dispatchers='(PROTOCOL=TCP) (SERVICE=pcpdbXDB)'
*.fast_start_mttr_target=300
*.java_pool_size=134217728
*.job_queue_processes=10
*.large_pool_size=33554432
*.open_cursors=600
*.pga_aggregate_target=2147483648
*.processes=2000
*.sessions=2400
*.sga_target=4293918720
*.shared_pool_size=536870912
*.streams_pool_size=402653184

node2.thread=2
node1.thread=1

SQL> show sga

Total System Global Area 4294967296 bytes

Fixed Size                  2034408 bytes
Variable Size            3204453656 bytes
Database Buffers         1073741824 bytes
Redo Buffers               14737408 bytes

>
> shiva wrote:
> > > Did you test the performance of your applications into RAC before
> > > migrating?
> > We did load tests but admit not with same application.
> >
> > > Is the "cpu utilization" showing up in IO wait?
> > IO Wait is not significant. It raises up to 30% but comes down righ
> > away. It doesn't stay there. On average it is about 5%.
> >
> > >  Which oracle processes exactly are using more cpu than they used to?
> >  Each client connection used to consume about 0.01% CPU before and
> > there were about 80 connections. Now each connection is consuming about
> > 0.25% CPU and there are about 1000 connections. We definately see more
> > connections then we saw before. (PGA was set to 256MB before and now is
> > set to 2GB)
> >
> > > What happens to the cpu utilization if you temporarily turn off RAC (
> > > run on one node of the cluster ONLY with cluster_database set to false
> > > for the instance you are running )? > > We haven't done this yet. Thanks for quick response. Received on Wed Oct 11 2006 - 23:32:12 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US