Create a cluster system using the following procedure:
1) Determine the system configuration
Determine which operation mode to use (1:1 active/standby, N:1 active/standby, dual node mutual standby, or cascading), considering the strengths and weaknesses of each mode. Also determine how the Systemwalker Operation Manager server system is to be configured as a cluster system.
2) Prepare resources
Prepare the following resources, which are required to install Systemwalker Operation Manager on cluster system:
Shared disk
Logical IP address
The terminology of "shared disk" and "logical IP address" varies from cluster system to cluster system. So, they should be interpreted as explained below:
Oracle Solaris Cluster
Shared disk: global file system
HACMP
Logical IP address: service address
Estimate the amount of space required on the shared disk that will be shared between the nodes in the cluster system. The following Systemwalker Operation Manager directories are placed on the shared disk, so estimate the amount of space that these directories will require.
The directory for calendar control information
The database directory for the Jobscheduler
The spool directory for Job Execution Control
The database directory for Master Schedule Management
The amount of space required for the Jobscheduler, Job Execution Control, and Master Schedule Management directories is the same as for normal system. Refer to the Systemwalker Operation Manager Technical Guide for details on how to estimate the amount of space required.
The amount of space required for the calendar control information directory can be calculated using the following formula:
Amount of space (bytes) on the shared disk required for calendar control information = 50 (bytes) x the number of registered calendars x the number of subsystems being used |
Prepare a logical IP address for the cluster system. The logical IP address is the virtual IP address used in common between the active node and the standby node.
Prepare logical IP address, separate from the node-specific physical IP address. (Prepare one logical IP address for 1:1 active/standby or cascading configuration, "N" logical IP addresses for N:1 active/standby configuration and two logical IP addresses for dual node mutual standby configuration.) Make sure that these IP addresses are not used by any other nodes on the network.
3) Resolve the physical IP address for the primary host name
If the IP address cannot be resolved for the primary NIC host name (in default, it is a local host name displayed by the hostname command), enter the primary host name and physical IP address in the /etc/hosts file so that they are associated each other.
4) Install the cluster system
Install and create the cluster system. Refer to the cluster system manual for details on how to create cluster system.
