Below is the procedure we follow to shutdown the 10g 2 node RAC
database on the Sun solaris. If you are not using the srvctl utility
then you will need to shutdown manually everything on each server.
Recomended method is to use srvctl utility as given below. This should
help you.
$ $ORA_CRS_HOME/bin/crs_stat -t -v
$ srvctl stop service -d <database_name> -s <service_name>
$ srvctl stop database -d <database_name>
$ srvctl stop asm -n <node1_hostname>
$ srvctl stop asm -n <node2_hostname>
$ srvctl stop nodeapps -n <node1_hostname>
$ srvctl stop nodeapps -n <node2_hostname>
$ $ORA_CRS_HOME/bin/crs_stat -t -v
Scott wrote:
> I would say no that is not expected behaviour.
>
> Normally if you shut down an instance gracefully it should stay down. You
> did shut it down with srvctl stop instance -d xxx -i xxx (correct?) The
> only time I have seen crs start something up automatically is when there has
> been a node eviction or a crash and crs will attempt to restart the
> instance.
>
> I have not tried this but maybe you can srvctl disable instance -d xxx -i
> xxx
> to stop it from starting unexpectedly. Or you may want to stop crs on node
> b
> to ensure nothing starts ( /pathto startup scripts/init.d/init.crs stop ),
> when you start crs again it will start the
> database instance registered in crs. (make sure the db, asm, nodeapps are
> down first )
>
> Hope that helps
> Scott.
>
>
> "steve" <stevenoyle1_at_yahoo.com> wrote in message
> news:1158779912.612195.120190_at_h48g2000cwc.googlegroups.com...
> > Hi,
> > Can someone explain if the following is expected in a RAC environment
> > (10gR2 HPUX)?
> >
> > scenario
> > 1. shuts down instance A on node A
> > 2. shuts down node B
> > 3. starts up node B, starts up instance B
> >
> > After instance B was started, instance A was brought up automatically
> > by Oracle CRS. Is this an expected behavior? Is there a way of
> > preventing CRS from bringing up instance A automatically?
> >
> > thanks.
> >
Received on Wed Sep 20 2006 - 18:22:46 CDT
