due to some network changes we were forced to adjust the interconnect configuration of one our managed grid infrastructure installations today. surprisingly this was a really smooth task, and all that needs to to be done is:
before doing any changes to the network configuration and while the rac databases and grid infrastructure are still running:
backup the current gpng ( grid plug and play ) profile in case you need to restore the old configuration:
cd $GRID_HOME/gpnp/[hostname]/profiles/peer cp profile.xml profile.xml_bak
shutdown the rac database and listener:
srvctl stop database -d db_name srvctl stop listener -l listener_name
make the changes to the interconnect configuration:
oifcfg setif -global eth1/10.167.77.0:cluster_interconnect oifcfg getif
… where eth1 is the interface name for the interconnect and 10.167.77.0 is the new network to use
as root shutdown and disable the cluster stack:
crsctl stop crs crsctl disable crs
shutdown the servers and perform the physical network changes, if any, start them up again and adjust the hosts file and the interface configurations.
as root user enable and start crs:
crsctl enable crs crsctl start crs
wait a few minutes and check the cluster status:
oifcfg getif crsctl status resource -t
that’s it.
Daniel Westermann's Blog 