Before describing the steps to configure the cluster, the hierarchical
structure of cluster members and resources needs to be considered.
Cluster members and resources can be thought of in terms of a
hierarchical structure, as shown in Figure 3-5.
The cluster is comprised of cluster nodes. The cluster nodes are
connected to a fencing device. Nodes can be separated by failover
domains to a cluster service. The services are comprised of managed
resources, such as NFS exports, IP addresses, and shared GFS partitions.
The structure is ultimately reflected in the
/etc/cluster/cluster.conf XML structure.
The steps to configure the cluster software consist of the
following:
Run the Cluster Configuration Tool on one
cluster member.
Click on the Cluster property and click the
Edit Cluster Properties button. Enter a name
for the cluster in the Name field. The name
should be descriptive enough to distinguish it from other clusters
and systems on your network (for example,
nfs_cluster or
httpd_cluster). The Config
Version field should be set to 1
by default and is automatically incremented each time you save your
cluster configuration.
To add a cluster node, click the property, then click the Add a Cluster
Node button. Enter a cluster name that you can
remember (such as the hostname assigned to the node, if any). Then
click the amount of Quorum Votes you want to
assign to the system.
Click on the Fence Device property and
click the Add a Fence Device button to
configure a fence device. Choose the device from the supported list
in the drop-down menu. Each device has different configurations that
you must enter (such as Name, IP
Address, Login name and
Password, if applicable.
Click the triangle next to Managed
Resources, and click the Failover
Domains property to set up one or more failover
domains, if needed. Failover domains restrict the members on which
a cluster service can run or restrict the order of members
followed when a cluster service fails over from one failover
domain member to another. Refer to Section 3.6 Configuring a Failover Domain for details.
Configure the Resources to be managed by
the cluster, specifying any IP addresses, file systems, NFS exports,
or user scripts to be managed by the cluster. Refer to Section 3.7 Adding Cluster Resources for more information.
Configure a failover domain for the cluster
service. Additionally, configure the cluster service using the
available cluster resources, or create private resources used
exclusively for the that service. Refer to Section 3.8 Adding a Cluster Service to the Cluster for more information.
Save cluster configuration changes by selecting
=> . When you
save the cluster configuration, the command service
rgmanager reload command is executed to cause the cluster
software to load the changed configuration file.
Quit the application by selecting =>
.
Running the Cluster Configuration Tool for the
first time causes the cluster configuration file
/etc/cluster/cluster.conf to be created
automatically.
Copy the /etc/cluster/cluster.conf file to
the other nodes using the scp command.
