7 Adding and Deleting Cluster Nodes

Adding a Cluster Node on Linux and UNIX Systems

This procedure describes how to add a node to your cluster. This procedure assumes that:

• There is an existing cluster with two nodes named node1 and node2

• You are adding a node named node3 using a virtual node name, node3-vip, that resolves to an IP address, if you are not using DHCP and Grid Naming Service (GNS)

• You have successfully installed Oracle Clusterware on node1 and node2 in a local (non-shared) home, where Grid_home represents the successfully installed home

To add a node:

1. Ensure that you have successfully installed Oracle Clusterware on at least one node in your cluster environment. To perform the following procedure, Grid_home must identify your successfully installed Oracle Clusterware home.

   See Also:

   Oracle Grid Infrastructure Installation Guide for Oracle Clusterware installation instructions

2. Verify the integrity of the cluster and node3:

   $ cluvfy stage -pre nodeadd -n node3 [-fixup] [-verbose]
   

   You can specify the -fixup option to attempt to fix the cluster or node if the verification fails.

3. To extend the Oracle Grid Infrastructure home to the node3, navigate to the Grid_home/addnode directory on node1 and run the addnode.sh script as the user that installed Oracle Clusterware.

   To run addnode.sh in interactive mode, run addnode.sh from Grid_home/addnode.

   You can also run addnode.sh in silent mode for both Oracle Clusterware standard Clusters and Oracle Flex Clusters.

   For an Oracle Clusterware standard Cluster:

   ./addnode.sh -silent "CLUSTER_NEW_NODES={node3}" "CLUSTER_NEW_VIRTUAL_
     HOSTNAMES={node3-vip}"
   

   If you are adding node3 to an Oracle Flex Cluster, then you can specify the node role on the command line, as follows:

   ./addnode.sh -silent "CLUSTER_NEW_NODES={node3}" "CLUSTER_NEW_VIRTUAL_
     HOSTNAMES={node3-vip}" "CLUSTER_NEW_NODE_ROLES={hub}"
   

   Note:

   Hub Nodes always have VIPs but Leaf Nodes may not. If you use the preceding syntax to add multiple nodes to the cluster, then you can use syntax similar to the following, where node3 is a Hub Node and node4 is a Leaf Node:

   ./addnode.sh -silent "CLUSTER_NEW_NODES={node3,node4}" "CLUSTER_NEW_VIRTUAL_HOSTNAMES={node3-vip,}" "CLUSTER_NEW_NODE_ROLES={hub,leaf}"
   

4. If prompted, then run the orainstRoot.sh script as root to populate the /etc/oraInst.loc file with the location of the central inventory. For example:

   # /opt/oracle/oraInventory/orainstRoot.sh
   

5. If you have an Oracle RAC or Oracle RAC One Node database configured on the cluster and you have a local Oracle home, then do the following to extend the Oracle database home to node3:

   1. Navigate to the Oracle_home/addnode directory on node1 and run the addnode.sh script as the user that installed Oracle RAC using the following syntax:

      $ ./addnode.sh "CLUSTER_NEW_NODES={node3}"
      

   2. Run the Oracle_home/root.sh script on node3 as root, where Oracle_home is the Oracle RAC home.

   If you have a shared Oracle home that is shared using Oracle Automatic Storage Management Cluster File System (Oracle ACFS), then do the following to extend the Oracle database home to node3:

   1. Run the Grid_home/root.sh script on node3 as root, where Grid_home is the Oracle Grid Infrastructure home.

   2. Run the following command as the user that installed Oracle RAC from the Oracle_home/oui/bin directory on the node you are adding to add the Oracle RAC database home:

      $ ./runInstaller -attachHome ORACLE_HOME="ORACLE_HOME" "CLUSTER_NODES={node3}"
      LOCAL_NODE="node3" ORACLE_HOME_NAME="home_name" -cfs
      

   3. Navigate to the Oracle_home/addnode directory on node1 and run the addnode.sh script as the user that installed Oracle RAC using the following syntax:

      $ ./addnode.sh -noCopy "CLUSTER_NEW_NODES={node3}"
      

      Note:

      Use the -noCopy option because the Oracle home on the destination node is already fully populated with software.

   If you have a shared Oracle home on a shared file system that is not Oracle ACFS, then you must first create a mount point for the Oracle RAC database home on the target node, mount and attach the Oracle RAC database home, and update the Oracle Inventory, as follows:

   1. Run the srvctl config database -db db_name command on an existing node in the cluster to obtain the mount point information.

   2. Run the following command as root on node3 to create the mount point:

      # mkdir -p mount_point_path
      

   3. Mount the file system that hosts the Oracle RAC database home.

   4. Run the following command as the user that installed Oracle RAC from the Oracle_home/oui/bin directory on the node you are adding to add the Oracle RAC database home:

      $ ./runInstaller -attachHome ORACLE_HOME="ORACLE_HOME" "CLUSTER
      _NODES={local_node_name}" LOCAL_NODE="node_name" ORACLE_HOME_NAME="home_name" -cfs
      

   5. Navigate to the Oracle_home/addnode directory on node1 and run the addnode.sh script as the user that installed Oracle RAC using the following syntax:

      $ ./addnode.sh -noCopy "CLUSTER_NEW_NODES={node3}"
      

      Note:

      Use the -noCopy option because the Oracle home on the destination node is already fully populated with software.

6. Run the Grid_home/root.sh script on the node3 as root and run the subsequent script, as instructed.

   Notes:

   • If you ran the root.sh script in the step 5, then you do not need to run it again.

   • If you have a policy-managed database, then you must ensure that the Oracle home is cloned to the new node before you run the root.sh script.

7. Start the Oracle ACFS resource on the new node by running the following command as root from the Grid_home/bin directory:

   # srvctl start filesystem -device volume_device_name -node node3
   

   Note:

   Ensure the Oracle ACFS resources, including Oracle ACFS registry resource and Oracle ACFS file system resource where the Oracle home is located, are online on the newly added node.

8. Run the following CVU command as the user that installed Oracle Clusterware to check cluster integrity. This command verifies that any number of specified nodes has been successfully added to the cluster at the network, shared storage, and clusterware levels:

   $ cluvfy stage -post nodeadd -n node3 [-verbose]
   

   See Also:

   "cluvfy stage [-pre | -post] nodeadd" for more information about this CVU command

Check whether either a policy-managed or administrator-managed Oracle RAC database is configured to run on node3 (the newly added node). If you configured an administrator-managed Oracle RAC database, you may need to use DBCA to add an instance to the database to run on this newly added node.

Deleting a Cluster Node on Linux and UNIX Systems

This section describes the procedure for deleting a node from a cluster.

To delete a node from a cluster:

1. Ensure that Grid_home correctly specifies the full directory path for the Oracle Clusterware home on each node, where Grid_home is the location of the installed Oracle Clusterware software.

2. Run the following command as either root or the user that installed Oracle Clusterware to determine whether the node you want to delete is active and whether it is pinned:

   $ olsnodes -s -t
   

   If the node is pinned, then run the crsctl unpin css command. Otherwise, proceed to the next step.

3. On the node you want to delete, run the following command as the user that installed Oracle Clusterware from the Grid_home/oui/bin directory where node_to_be_deleted is the name of the node that you are deleting:

   $ ./runInstaller -updateNodeList ORACLE_HOME=Grid_home "CLUSTER_NODES=
   {node_to_be_deleted}" CRS=TRUE -silent -local
   

4. On the node that you are deleting, depending on whether you have a shared or local Oracle home, complete one of the following procedures as the user that installed Oracle Clusterware:

   • For a local home, deinstall the Oracle Clusterware home from the node that you want to delete, as follows, by running the following command, where Grid_home is the path defined for the Oracle Clusterware home:

     $ Grid_home/deinstall/deinstall -local
     

     Caution:

     • If you do not specify the -local flag, then the command removes the Oracle Grid Infrastructure home from every node in the cluster.

     • If you cut and paste the preceding command, then paste it into a text editor before pasting it to the command line to remove any formatting this document might contain.

   • If you have a shared home, then run the following commands in the following order on the node you want to delete.

     Run the following command to deconfigure Oracle Clusterware:

     $ Grid_home/perl/bin/perl Grid_home/crs/install/rootcrs.pl -deconfig -force
     

     Run the following command from the Grid_home/oui/bin directory to detach the Grid home:

     $ ./runInstaller -detachHome ORACLE_HOME=Grid_home -silent -local
     

     Manually delete any configuration files, as prompted by the installation utility.

5. On any node other than the node you are deleting (except for a Leaf Node in an Oracle Flex Cluster), run the following command from the Grid_home/oui/bin directory where remaining_nodes_list is a comma-delimited list of the nodes that are going to remain part of your cluster:

   $ ./runInstaller -updateNodeList ORACLE_HOME=Grid_home "CLUSTER_NODES=
   {remaining_nodes_list}" CRS=TRUE -silent
   

   Notes:

   • You must run this command a second time from the Oracle RAC home, where ORACLE_HOME=ORACLE__RAC_HOME and CRS=TRUE -silent is omitted from the syntax, as follows:

     $ ./runInstaller -updateNodeList ORACLE_HOME=ORACLE_HOME
      "CLUSTER_NODES={remaining_nodes_list}"
     

   • Because you do not have to run this command if you are deleting a Leaf Node from an Oracle Flex Cluster, remaining_nodes_list must list only Hub Nodes.

   • If you have a shared Oracle Grid Infrastructure home, then append the -cfs option to the command example in this step and provide a complete path location for the cluster file system.

6. From any node that you are not deleting, run the following command from the Grid_home/bin directory as root to delete the node from the cluster:

   # crsctl delete node -n node_to_be_deleted
   

7. Run the following CVU command to verify that the specified nodes have been successfully deleted from the cluster:

   $ cluvfy stage -post nodedel -n node_list [-verbose]
   

   See Also:

   "cluvfy stage -post nodedel" for more information about this CVU command

8. If you remove a cluster node on which Oracle Clusterware is down, then determine whether the VIP for the deleted node still exists, as follows:

   $ srvctl config vip -node deleted_node_name
   

   If the VIP still exists, then delete it, as follows:

   $ srvctl stop vip -node deleted_node_name
   $ srvctl remove vip -node deleted_node_name
   

Adding a Node to a Cluster on Windows Systems

Ensure that you complete the prerequisites listed in "Prerequisite Steps for Adding Cluster Nodes" before adding nodes.

This procedure describes how to add a node to your cluster. This procedure assumes that:

• There is an existing cluster with two nodes named node1 and node2

• You are adding a node named node3

• You have successfully installed Oracle Clusterware on node1 and node2 in a local home, where Grid_home represents the successfully installed home

To add a node:

1. Verify the integrity of the cluster and node3:

   C:\>cluvfy stage -pre nodeadd -n node3 [-fixup] [-verbose]
   

   You can specify the -fixup option and a directory into which CVU prints instructions to fix the cluster or node if the verification fails.

2. On node1, go to the Grid_home\addnode directory and run the addnode.bat script, as follows:

   C:\>addnode.bat "CLUSTER_NEW_NODES={node3}"
   "CLUSTER_NEW_VIRTUAL_HOSTNAMES={node3-vip}"
   

3. Run the following command on the new node:

   C:\>Grid_home\crs\config\gridconfig.bat
   

4. The following steps are required only if you have database homes configured to use Oracle ACFS:

   1. For each database configured to use Oracle ACFS, run the following command from the Oracle RAC database home:

      C:\> ORACLE_HOME/bin/srvctl stop database -db database_unique_name
      

      Note:

      Run the srvctl config database command to list all of the databases configured with Oracle Clusterware. Use the srvctl config database -db database_unique_name to find the database details. If the ORACLE_HOME path leads to the Oracle ACFS mount path, then the database uses Oracle ACFS. Use the command output to find the database instance name configured to run on the newly added node.

   2. Use Windows Server Manager Control to stop and delete services.

   3. For each of the databases and database homes collected in step a, run the following command:

      C:\> ORACLE_HOME/bin/srvctl start database -db database_unique_name
      

5. Run the following command to verify the integrity of the Oracle Clusterware components on all of the configured nodes, both the preexisting nodes and the nodes that you have added:

   C:\>cluvfy stage -post crsinst -n all [-verbose]
   

After you complete the procedure in this section for adding nodes, you can optionally extend Oracle Database with Oracle RAC components to the new nodes, making them members of an existing Oracle RAC database.

Creating the OraMTS Service for Microsoft Transaction Server

Oracle Services for Microsoft Transaction Server (OraMTS) permit Oracle databases to be used as resource managers in Microsoft application-coordinated transactions. OraMTS acts as a proxy for the Oracle database to the Microsoft Distributed Transaction Coordinator (MSDTC). As a result, OraMTS provides client-side connection pooling and allows client components that leverage Oracle to participate in promotable and distributed transactions. In addition, OraMTS can operate with Oracle databases running on any operating system, given that the services themselves are run on Windows.

On releases earlier than Oracle Database 12c, the OraMTS service was created as part of a software-only installation. Starting with Oracle Database 12c, you must use a configuration tool to create this service.

Create the OraMTS service after adding a node or performing a software-only installation for Oracle RAC, as follows:

1. Open a command window.

2. Change directories to %ORACLE_HOME%\bin.

3. Run the OraMTSCtl utility to create the OraMTS Service, where host_name is a list of nodes on which the service should be created:

   C:\..bin> oramtsctl.exe -new -host host_name
   

Deleting a Cluster Node on Windows Systems

This section describes how to delete a cluster node on Windows systems. This procedure assumes that Oracle Clusterware is installed on node1, node2, and node3, and that you are deleting node3 from the cluster.

To delete a cluster node on Windows systems:

1. Only if you have a local home, on the node you want to delete, run the following command with -local option to update the node list:

   C:\>Grid_home\oui\bin\setup.exe -updateNodeList ORACLE_HOME=Grid_home "CLUSTER_NODES=
   {node_to_be_deleted}" CRS=TRUE -local
   

   Note:

   If you are deleting a Leaf Node from an Oracle Flex Cluster, then you do not have run this command.

2. Run the deinstall tool on the node you want to delete to deinstall and deconfigure the Oracle Clusterware home, as follows:

   C:\Grid_home\deinstall\>deinstall.bat -local
   

   Caution:

   • If you do not specify the -local flag, then the command removes the Oracle Grid Infrastructure home from every node in the cluster.

   • If you cut and paste the preceding command, then paste it into a text editor before pasting it to the command line to remove any formatting this document might contain.

3. On any node that you are not deleting, run the following command from the Grid_home\oui\bin directory where remaining_nodes_list is a comma-delimited list of the nodes that are going to remain part of your cluster:

   C:\>setup.exe –updateNodeList ORACLE_HOME=Grid_home
   "CLUSTER_NODES={remaining_nodes_list}" CRS=TRUE -silent
   

   Notes:

   • You must run this command a second time where ORACLE_HOME=ORACLE_HOME, and CRS=TRUE -silent is omitted from the syntax, as follows:

     C:\>setup.exe -updateNodeList ORACLE_HOME=ORACLE_HOME
      "CLUSTER_NODES={remaining_nodes_list}"
     

   • If you have a shared Oracle Grid Infrastructure home, then append the -cfs option to the command example in this step and provide a complete path location for the cluster file system.

4. On a node that you are not deleting, run the following command:

   C:\>Grid_home\bin\crsctl delete node -n node_to_be_deleted
   

5. Run the following CVU command to verify that the specified nodes have been successfully deleted from the cluster:

   C:\>cluvfy stage -post nodedel -n node_list [-verbose]
   

   See Also:

   "cluvfy stage -post nodedel"

6. If you remove a cluster node on which Oracle Clusterware is down, then determine whether the VIP for the deleted node still exists, as follows:

   C:\> ORACLE_HOME/bin/srvctl config vip -node deleted_node_name
   

   If the VIP still exists, then delete it, as follows:

   C:\> ORACLE_HOME/bin/srvctl srvctl stop vip -node deleted_node_name
   C:\> ORACLE_HOME/bin/srvctl srvctl remove vip -node deleted_node_name