Completing Failed or Interrupted Installations and Upgrades

If Oracle Universal Installer (OUI) exits on the node from which you started the upgrade, or the node reboots before you confirm that the rootupgrade.sh script was run on all nodes, then the upgrade remains incomplete.

In an incomplete installation or upgrade, configuration assistants still need to run, and the new Grid home still needs to be marked as active in the central Oracle inventory. You must complete the installation or upgrade on the affected nodes manually.

Completing Failed Installations and Upgrades

Understand how to join nodes to the cluster after installation or upgrade fails on some nodes.

If installation or upgrade of Oracle Grid Infrastructure on some nodes fails, and the installation or upgrade completes with only successful nodes in the cluster, then follow this procedure to add the failed nodes to the cluster.

Remove the Oracle Grid Infrastructure software from the failed nodes:
```
Grid_home/deinstall/deinstall -local
```
As the root user, from a node where Oracle Clusterware is installed, delete the failed nodes using the delete node command:
```
Grid_home/bin/crsctl delete node -n node_name
```
node_name is the node to be deleted.
Run the Oracle Grid Infrastructure installation wizard and follow the steps in the wizard to add the nodes:
```
Grid_home/gridSetup.sh
```

The nodes are added to the cluster.

Continuing Incomplete Upgrade of First Node

Review this information to complete the upgrade, if upgrade of Oracle Grid Infrastructure fails on the first node.

If the root script failure indicated a need to reboot, through the message CLSRSC-400, then reboot the first node (the node where the upgrade was started). Otherwise, manually fix or clear the error condition, as reported in the error output.
If necessary, log in as root to the first node. Change directory to the new Grid home on the first node, and run the rootupgrade.sh script on that node again. For example:
```
[root@node1]# cd Grid_home
[root@node1]# ./rootupgrade.sh
```
Complete the upgrade of all other nodes in the cluster.

Complete the upgrade of all other nodes in the cluster.
```
[root@node2]# ./rootupgrade.sh
```
Configure a response file, and provide passwords for the installation.
To complete the upgrade, log in as the Grid installation owner, and run gridSetup.sh, located in the Grid_home, specifying the response file that you created. For example, where the response file is gridinstall.rsp:
```
[grid@node1]$ Grid_home/gridSetup.sh -executeConfigTools -responseFile Grid_home/install/response/gridinstall.rsp
```
Note:
When you reupgrade Oracle Grid Infrastructure, you must use the -all flag with the executeConfigTools command to run all the configuration tools.

Continuing Incomplete Upgrades on Remote Nodes

Review this information to continue incomplete upgrade on remote nodes.

If the root script failure indicated a need to reboot, through the message CLSRSC-400, then reboot the first node (the node where the upgrade was started). Otherwise, manually fix or clear the error condition, as reported in the error output.
If root automation is being used, click Retry on the OUI instance on the first node.
If root automation is not being used, log into the affected node as root. Change directory to the Grid home, and run the rootupgrade.sh script on that node. For example:
```
[root@node6]# cd Grid_home
[root@node6]# ./rootupgrade.sh
```

Continuing Incomplete Installation on First Node

Review this information to continue an incomplete installation of Oracle Grid Infrastructure, if installation fails on the first node.

If the root script failure indicated a need to reboot, through the message CLSRSC-400, then reboot the first node (the node where the installation was started). Otherwise, manually fix or clear the error condition, as reported in the error output.
If necessary, log in as root to the first node. Run the orainstRoot.sh script on that node again. For example:
```
$ sudo -s
[root@node1]# cd /u01/app/oraInventory
[root@node1]# ./orainstRoot.sh
```
Change directory to the Grid home on the first node, and run the root script on that node again. For example:
```
[root@node1]# cd Grid_home
[root@node1]# ./root.sh
```
Complete the installation on all other nodes.
Configure a response file, and provide passwords for the installation.
To complete the installation, log in as the Grid installation owner, and run gridSetup.sh, located in the Oracle Grid Infrastructure home, specifying the response file that you created. For example, where the response file is gridinstall.rsp:
```
[grid@node1]$ Grid_home/gridSetup.sh -executeConfigTools -responseFile /u01/app/23.0.0/grid/install/response/gridinstall.rsp
```

Continuing Incomplete Installation on Remote Nodes

Review this information to continue incomplete installation on remote nodes.

If the root script failure indicated a need to reboot, through the message CLSRSC-400, then reboot the affected node. Otherwise, manually fix or clear the error condition, as reported in the error output.
If root automation is being used, click Retry on the OUI instance on the first node.
If root automation is not being used, follow these steps:
1. Log into the affected node as root, and run the orainstRoot.sh script on that node. For example:
```
$ sudo -s
[root@node6]# cd /u01/app/oraInventory
[root@node6]# ./orainstRoot.sh
```
2. Change directory to the Grid home, and run the root.sh script on the affected node. For example:
```
[root@node6]# cd Grid_home
[root@node6]# ./root.sh
```
Continue the installation from the OUI instance on the first node.