To gracefully shut down one or more nodes, use the following rhino-console commands or related MBean operations.

Note When using the pool clustering mode, only the node the management command is invoked on may be shutdown or rebooted. Other cluster nodes can only be shutdown or rebooted by connecting a management client directly to them.

Console command: shutdown

Command

 
shutdown [-nodes node1,node2,...] [-timeout timeout] [-restart]
  Description
    Gracefully shutdown and terminate the cluster (or the specified nodes). If the
    SLEE is running in any namespace on any target node, event routing functions are
    allowed to complete before termination without affecting existing desired state.
    The optional timeout is specified in seconds. Optionally restart the node(s)
    after shutdown

Examples

To shut down the entire cluster (when using Savanna clustering):

$ ./rhino-console shutdown
Shutting down the SLEE
Shutdown successful

To shut down only node 102:

$ ./rhino-console shutdown -nodes 102
Shutting down node(s) [102]
Shutdown successful
Note Since Rhino 3.0.0 the shutdown console command will shut down the specified nodes regardless of the desired SLEE state. If the SLEE is running in any namespace on any target node, event routing functions are allowed to complete before termination without affecting existing desired state.
Note When using the pool clustering mode, the unadorned shutdown console command will only shut down the node that the console is connected to, as if the command had been given a -nodes <node-id-of-connected-node> argument, and not the entire cluster of pool nodes. It is not possible to shut down all pool cluster nodes in one operation using rhino-console.

MBean operation: shutdownCluster

MBean

SLEE Management

Rhino extension
Shut down all nodes
public void shutdownCluster(boolean restart)
  throws InvalidStateException, ManagementException;

When using the Savanna clustering mode, the shutdownCluster operation terminates every node in the cluster. When using the pool clustering mode, the shutdownCluster operation will ony terminate the node the operation is executed on. If the restart flag is set, terminated nodes will be restarted to the currently configured desired state.

Rhino extension
Shut down all nodes with a timeout
public void shutdownCluster(boolean restart, long timeout)
  throws InvalidStateException, ManagementException;

When using the Savanna clustering mode, the shutdownCluster operation terminates every node in the cluster. When using the pool clustering mode, the shutdownCluster operation will ony terminate the node the operation is executed on. If the restart flag is set, terminated nodes will be restarted to the currently configured desired state. If the timeout argument is greater than zero, any nodes that still have live activities will be shutdown anyway. This may result in call failures.

MBean operation: shutdownNodes

MBean

SLEE Management

Rhino extension
Shut down specific nodes
public void shutdownNodes(int[] nodeIDs, boolean restart)
  throws InvalidStateException, ManagementException;

The shutdownNodes operation terminates the specified set of nodes. When using the pool clustering mode, the only node ID that may be legally specified is the node ID of the node the command is executed on. If the restart flag is set, terminated nodes will be restarted to the currently configured desired state.

Rhino extension
Shut down specific nodes with a timeout
public void shutdownNodes(int[] nodeIDs, boolean restart, long timeout)
  throws NullPointerException, InvalidArgumentException,
    InvalidStateException, ManagementException;

The shutdownNodes operation terminates the specified set of nodes. When using the pool clustering mode, the only node ID that may be legally specified for this operation is the node ID of the node the command is executed on. If the restart flag is set, terminated nodes will be restarted to the currently configured desired state. If the timeout argument is greater than zero, any nodes that still have live activities will be shutdown anyway. This may result in call failures.

MBean operation: reboot

MBean

SLEE Management

Rhino extension
Reboot all nodes
public void reboot(SleeState[] states)
  throws InvalidArgumentException, InvalidStateException, ManagementException;

When using the Savanna clustering mode, this operation reboots every node in the cluster to the state specified. When using the pool clustering mode, this operation will only reboot the node the operation is executed on.

Rhino extension
Reboot specific nodes
public void reboot(int[] nodeIDs, SleeState[])
  throws NullPointerException, InvalidArgumentException, InvalidStateException, ManagementException;

Extension to reboot that adds an argument which lets you control which nodes to shut down (by specifying node IDs). When using the pool clustering mode, the only node ID that may be legally specified is the node ID of the node the command is executed on.
Note Event-router nodes can restart to either the RUNNING state or the STOPPED state. Quorum nodes must have a state provided but do not use this in operation.
Previous page Next page
Rhino Version 3.2