Unconfiguring Oracle Clusterware Without Removing Binaries

Running the rootcrs.sh command flags -deconfig -force enables you to unconfigure Oracle Clusterware on one or more nodes without removing installed binaries.

This feature is useful if you encounter an error on one or more cluster nodes during installation when running the root.sh command, such as a missing operating system package on one node. By running rootcrs.sh -deconfig -force on nodes where you encounter an installation error, you can unconfigure Oracle Clusterware on those nodes, correct the cause of the error, and then run root.sh again.


Stop any databases, services, and listeners that may be installed and running before deconfiguring Oracle Clusterware. In addition, dismount Oracle Automatic Storage Management Cluster File System (Oracle ACFS) and disable Oracle Automatic Storage Management Dynamic Volume Manager (Oracle ADVM) volumes.


Commands used in this section remove the Oracle Grid infrastructure installation for the entire cluster. If you want to remove the installation from an individual node, then see Oracle Clusterware Administration and Deployment Guide.

To unconfigure Oracle Clusterware:

  1. Log in as the root user on a node where you encountered an error.

  2. Change directory to Grid_home/crs/install. For example:

    # cd /u01/app/12.2.0/grid/crs/install
  3. Run rootcrs.sh with the -deconfig and -force flags. For example:

    # rootcrs.sh -deconfig -force

    Repeat on other nodes as required.

  4. If you are deconfiguring Oracle Clusterware on all nodes in the cluster, then on the last node, enter the following command:

    # rootcrs.sh -deconfig -force -lastnode

    The -lastnode flag completes deconfiguration of the cluster, including the OCR and voting files.


    Run the rootcrs.sh -deconfig -force -lastnode command on a Hub Node. Deconfigure all Leaf Nodes before you run the command with the -lastnode flag.