After a cluster resource is enabled, the load and unload scripts take care of starting and stopping the services or mounting and dismounting the volumes that are configured in the resource. You start services and mount devices by onlining the resource. You stop services and unmount devices by offlining the resource.
Onlining a resource runs the load script, which loads the resource on its primary preferred node, or on an alternate preferred node if possible, according to the order in its Preferred Nodes list. No action is taken if the preferred nodes are not active in the cluster, or if there are Resource Mutual Exclusion conflicts on its active preferred nodes.
Offlining a resource runs the unload script and unloads the resource from the server. The resource cannot be loaded on any other servers in the cluster and remains unloaded until you load it again.
If monitoring is enabled for a resource and you start, stop, or restart services outside the cluster framework while the cluster resource is online, a monitoring failure occurs, which triggers a failover of the cluster resource.
If you edit the scripts for a resource that is online, the changes you made do not take effect until the resource is taken offline and brought online again.
If a resource is in a comatose state, you must take the resource offline before it can be brought online again.
There are maintenance situations where you might need to control services by using the standard interfaces. In these rare cases, you must first offline the cluster resource, then use non-cluster methods to start and stop a service or mount and dismount a volume outside the cluster framework. When you are done, online the cluster resource to resume normal cluster handling.
In iManager, select Clusters > My Clusters.
Select the cluster that you want to manage.
If the cluster does not appear in your list, add the cluster to your list as described in Section 8.2, Setting Up a Personalized List of Clusters to Manage.
On the Cluster Manager page, select the check box next to the resource you want to manage, then click Online or click Offline.
To bring a resource online on its most preferred node that is currently active:
As the root user, enter the following at the command prompt of a cluster node:
cluster online resource_nameTo bring a resource online on a specific active node:
As the root user, enter the following at the command prompt of a cluster node:
cluster online resource_name node_name
To take a resource offline:
As the root user, enter the following at the command prompt of a cluster node:
cluster offline resource_nameThe cluster migrate command can be used to move a specified resource from the node where it is currently running to a another node. If a node name is not provided, the resource is brought online according to its Preferred Nodes list if it is possible. You can alternatively specify a destination node from its Preferred Nodes list that is running in the cluster. It migrates the resource to the specified node if it is possible. No action is taken if the specified node is not in the resource’s preferred nodes list, there is a Resource Mutual Exclusion conflict on the specified node, or the specified node is not currently active in the cluster.
As the root user, enter the following at the command prompt of a cluster node:
cluster migrate resource_name [node_name]