Cisco Cisco Prime Central 1.2 Installation Guide

To set up and manage a cluster, you can use the CLI or the GUI. This section explains how to use the CLI. To use the GUI, see 
the 

, sections 3 and 4.

To set up the Prime Central cluster service, complete the following steps in parallel on both nodes, except where noted:

Modify the /etc/cluster/cluster.conf file by setting unique values for the parameters listed in 

.

Copy the edited cluster.conf file to the /etc/cluster/ directory.

Whenever you modify the cluster.conf file, increment the config_version value so the cluster.conf file propagates 
correctly to the nodes. To propagate the cluster.conf file manually, you must stop the cluster, copy the file, and then 
start the cluster.

Start the cluster services:

# service cman start

# service rgmanager start

Enter each command on one node and then immediately enter the same command on the other node.

For example, when cman starts on a node, it waits for the other node to start cman. If the other node takes too long to 
start cman, cman times out on the first node.

(For the RHCS luci GUI only) Using the username admin, start the RHCS ricci service:

# service ricci start

(For the RHCS luci GUI only) On the first node only, start the RHCS luci services:

# luci_admin init

# service luci start

To test failover, relocate the service to another node:

# clusvcadm -r vmpcservice -m prime-ha-node2.cisco.com

After the Prime Central service is running in an HA cluster, you cannot restart its components (such as the portal, 
integration layer, and database) without first freezing the cluster. After you restart the component, you can unfreeze the 
cluster.

For example, attaching or detaching an application to or from Prime Central requires an integration layer restart. On 
the active node, freeze the HA cluster, restart the integration layer, and unfreeze the cluster:

# clusvcadm -Z Prime-Central-service-name

# su - primeusr

# itgctl stop

# itgctl start

# exit

# clusvcadm -U Prime-Central-service-name

The following table lists the parameters in the /etc/cluster/cluster.conf file for which you must set unique values.

Cluster name

bm1cluster

—

Multicast address

224.0.0.251

The multicast address must be unique per subnet and must be 
working before you start your cluster. For a tool to verify that your 
multicast address is correct, see 

Service IP address

192.168.1.130

—