solidDB Help : solidDB Grid : Failure handling in the grid : Node controller
  
Node controller
Note Check the solidDB Release Notes for any limitations that are associated with using a grid in the current release.
The node controller is a separate software component that monitors the health of node processes within the same host.
The node controller must run on the same physical or virtual machine as the grid node that it is monitoring; a node controller cannot monitor a grid node that is running on another physical or virtual machine.
Diagram that shows an overview of the operation of the node controller
Depending on the requirements, you can configure every grid node in your host to have its own node controller or you can configure one node controller to monitor several grid nodes (as long as the nodes are all on the same physical or virtual machine).
When a grid node fails, the node controller attempts to restart the node to minimize the downtime of the grid node. The node controller tests the availability of a node by periodically checking the connection and responsiveness of the node. If a node does not respond, the response takes too long, or the node is no longer active on the system, then the node controller determines that the node has failed and restarts the node.
Note Using the node controller is not mandatory. If there are other means of monitoring and restarting processes in the system that are more practical to use than the node controller, these can be used instead.
See
Using the node controller
Optional node controller scripts
Controller.ini file
Go up to
Failure handling in the grid