Depending on whether the node application can be restarted on the same node or not, the following measures may be necessary to prevent data loss:
If the node application can be restarted on the same node after the failure, it is possible to continue working with the previous data without any problems. A failure script can, for instance, initiate an automatic restart of the node application.
The following alternatives are available if it is not possible to restart the application on the same node, for instance if the computer has failed:
Move the node application to a spare computer with the same host name / IP address. It is then possible to restart the node application on this new computer without the need to take any further measures.
Move the node application to a spare computer with an identical virtual host name/IP address. Before the node application can be started on this new computer, the administration functions must be used to change the host name of the failed node in the cluster configuration file to the host name of the spare computer. After this has been done, it is possible to restart the node application on this new computer.
Perform a node recovery, see section “Node recovery”.