Like all highly available services supported by IRIS FailSafe software options (interfaces, volumes, filesystems, nfs, and so on), the configuration of the application you are making highly available must be described in the IRIS FailSafe configuration file /var/ha/ha.conf.

For example, if you are adding named as a highly available application, the blocks and parameters for named in the configuration file might be as follows (shown with line numbers):

 1 application-class named
 2 {
 3     server-node = ha1        # active node's hostname
 4 }
 5 
 6 action named
 7 {
 8     local-monitor = /var/ha/actions/ha_named_lmon
 9 }
10 
11 action-timer named
12 {
13     start-monitor-time = 60 
14     lmon-probe-time = 120
15     lmon-timeout = 30
16 }
17
18 named named_block
19 {
20     server-node = ha1        # primary node's hostname
21     backup-node = ha2        # backup node's hostname
22     process-name = named
23 }

named is an application that needs one instance running in an IRIS FailSafe cluster. Thus, it is run in an active/backup configuration and just one server-node parameter is needed in the application-class block (line 3). The node ha1 is the active (or primary) node, the node that runs named when the nodes are in normal state.

Each application class needs one action block (lines 6 to 9) and one action-timer block (lines 11 to 16). These blocks have standard parameters, which are described in Chapter 4 of the IRIS FailSafe Administrator's Guide .

Because there is just one instance of named running in the cluster, there is just one named block (lines 18 to 23). The block's label is named_block (line 18). The block specifies the active node as ha1 (the server-node parameter on line 20) and the backup node as ha1 (the backup-node parameter on line 21). The process-name parameter (line 22) will be used by the local monitoring script /var/ha/actions/ha_named_lmon. (See Example 3-1.)

You must develop the configuration file contents for a new highly available service in several phases as you proceed with developing and testing the scripts for the new highly available services. The phases are:

The parameters in the IRIS FailSafe configuration file specify many details for each highly available service, including the number of instances and which node is their primary node, which monitoring script is to be used, and various monitoring frequency values. These parameters are used by monitoring scripts and failover scripts. Placing parameters in the configuration file, rather than simply hardcoding values in the scripts, simplifies maintenance when values are used in more than one script and simplifies tuning by putting all values that are likely to change into just one file.

For each parameter you add, you must choose a name. You have two choices:

Consider the structure of your blocks when adding parameters. Simple parameter specification is structured like this:

block-name {
    parameter-name = value
    ...
}

Parameters can take several values, known as a list:

block-name {
    parameter-name = ( value1 value2 ... )
    ...
}

If necessary, you can include sections as follows:

block-name {
    section-name label {
        parameter-name = value
    }
    ...
}

The formal syntax definition of the configuration file is given in Chapter 4 of the IRIS FailSafe Administrator's Guide .

You can check the syntax and to some extent the content of the configuration file that you create by entering this ha_cfgverify command:

# /usr/etc/ha_cfgverify configuration_file 

configuration_file is the pathname of your configuration file under development. The messages produced by ha_cfgverify are described in Appendix A of the IRIS FailSafe Administrator's Guide .