uninstallsfrac, modify the configuration files on the existing remaining nodes to remove references to the deleted node(s).
On the each of the existing nodes, using
vi or another editor, edit the file
/etc/llthosts, removing lines corresponding to the removed nodes. For example, if
saturn is the node being removed from the cluster, remove the line "3 saturn" from the file:
The file should now resemble:
In the file
/etc/gabtab, change the command contained in the file to reflect the number of systems after the node is removed:
/sbin/gabconfig -c -nN
where N is the number of nodes remaining. For example, with two nodes remaining, the file resembles:
/sbin/gabconfig -c -n2
You can modify the VCS configuration using one of three possible methods. You can edit
/etc/VRTSvcs/conf/config/main.cf (the VCS configuration file) directly, you can use the VCS GUI (Cluster Manager), or you can use the command line, as illustrated in the following example. Please refer to the Veritas Cluster Server User's Guide for details about how to configure VCS.
At this point in the process, all Oracle binaries have been removed from the system to be deleted. The instance has been removed from the database, that is, the thread disabled, and the
spfile<SID>.ora edited by Oracle to remove any references to this instance. The next step is to remove all references in the
main.cf to the deleted node(s).
As root user execute the following on the CVM master node only.
To modify the CVM group in the main.cf file
vxdctl -c mode
cp main.cf main.cf.3node.bak
hagrp -modify cvm SystemList -delete saturn
hares -modify cvm_clus CVMNodeId -delete saturn
hasys -delete saturn
haconf -dump -makero
Example of main.cf file: see Sample main.cf for Removing an Oracle 10g Node.
hacf -verify .
The main.cf file now should not contain entries for system saturn.
hastop -all -force
main.cfto each system in the cluster.
rcp (or scp) main.cf galaxy:/etc/VRTSvcs/conf/config
rcp (or scp) main.cf nebula:/etc/VRTSvcs/conf/config