Installing and upgrading HP Insight Management 7.3 on Windows Server 2012 Failover Clusters

Technical white paper | HP Insight Management 7.3

• Two private disks, one for each ProLiant system’s boot drive, disk C:

• One shared disk for the cluster witness, disk W:

• One shared disk for Microsoft Distributed Transaction Coordinator, disk T:

• One shared disk for Microsoft SQL Server, disk S:

• One shared disk for the Insight Management program files, disk V:

HP recommends that SAN storage be used for boot disks as well as shared disks. See the document HP Boot from SAN

Configuration Guide, and also see Boot from SAN.

In testing for this white paper, the disks for the cluster witness and Distributed Transaction Coordinator were sized at 1 GB

each.

Tip:

Create the shared disks W:, T:, S:, and V: on the SAN and present them to the cluster nodes, but do not “online,” “initialize,” or

“create” a volume for any disk until instructed to do so in Creating the Failover Cluster

Supporting Systems

It is assumed that an existing Active Directory domain infrastructure and DNS services suitable for hosting the cluster are

available. Multi-site clusters are not supported.

If the Insight Management SQL instance is to be hosted in another cluster, that cluster must be in the same domain as the

Insight Management cluster.

Network

Microsoft Failover Clusters require that each virtual server in a cluster have a unique IP address and network name

registered in the DNS. Starting Insight Management 7.3 release, the Central Management Server is supported on both, IPv4

only and the dual stacked (IPv4 and IPv6, both) highly available environment. With the limitation in IMA (via SNMP) that does

not have complete support for IPv6 Cluster, it becomes mandatory to have DNS names of all the Cluster IP addresses to be

same on the network. The HP Systems Insight Manager (HP SIM) relies on the FQDN of the Cluster for discovery. Hence, if

the Cluster and Cluster resources have multiple addresses (say a combination of IPv6 and IPv4 address), all of the addresses

should be set to the same host name in the DNS. For the configuration described in this white paper, six static IP addresses

(four for IPv4 and two for IPv6 addresses, if configured in dual stack mode) are required for the following four virtual

servers. An additional two static IP addresses are needed for the physical nodes.

1. Failover Cluster virtual server and the client access point

Function: To act as the virtual server IP address and name for the Failover Cluster software

Suggested action: Windows management tools use this address to get the status and manage the Failover Cluster

software. In addition, Systems Insight Manager uses this address to discover the Failover Cluster. For the examples

discussed in this white paper, the IPv4 and IPv6 addresses are 15.199.192.144 and

2620:0000:0a04:f115:0000:0000:0000:65, and the name is CMS-C17-CLUSTER.

Note:

The order of IP address allocation for the Cluster nodes and Cluster virtual server in DNS is important. The Insight

Management Advisor forward and reverse DNS lookup test displays an error if the last octet value for Cluster virtual

server has a lower number compared to the last octet in the IP address of the Cluster nodes.

2. Systems Insight Manager and Insight Management virtual server and the client access point

Function: To act as the virtual server IP address and name for Systems Insight Manager and Insight Management

You specify this name and address to connect to Systems Insight Manager through your browser.

Suggested action: All SNMP devices that report their events to Systems Insight Manager should use this IP address and

name for the location of the Event Consolidator, also known as the SNMP Trap Destination. The host defined for the

WMI Mapper proxy should use this name and IP address (unless the WMI Mapper service runs on a system that is not a

cluster member and the WMI Mapper is not installed as part of Insight Management). For the examples discussed in this