Sun Cluster Data Service for PostgreSQL Guide for Solaris OS

Planning the Sun Cluster HA for PostgreSQL Installation and Configuration

This section contains the information you need to plan your Sun Cluster HA for PostgreSQL installation and configuration.

PostgreSQL and Solaris Containers

Sun Cluster HA for PostgreSQL is supported in Solaris Containers, Sun Cluster is offering two concepts for Solaris Containers.

Configuration Restrictions

The configuration restrictions in the subsections that follow apply only to Sun Cluster HA for PostgreSQL.


Caution – Caution –

Your data service configuration might not be supported if you do not observe these restrictions.


Restriction for the Location of the Database Cluster

The PostgreSQL database cluster is where the database files and the configuration files are stored. The database cluster, represented by the configuration variable PGDATA, needs to be placed on the shared storage.

Restriction for the Listening Policy of the PostgreSQL Database Server

Sun Cluster HA for PostgreSQL requires that the PostgreSQL listens at the localhost. Otherwise the monitoring of your data service will not work. For more information, see Prepare Your PostgreSQL Installation for Cluster Control.

Restriction for the PostgreSQL postgresql.conf File

The postgresql.conf file is one of the central configuration files for a specific PostgreSQL database cluster.

The postgresql.conf file must be stored in the PGDATA path. You cannot register Sun Cluster HA for PostgreSQL if the file postgresql.conf is not in the directory referenced in the PGDATA variable. The other configuration files can be kept elsewhere. For more information about registration, see Registering and Configuring Sun Cluster HA for PostgreSQL.

Restriction for the Password Policy for the Sun Cluster HA for PostgreSQL Monitoring Database

Sun Cluster HA for PostgreSQL requires a database to which it can connect and where it can manipulate a table for monitoring purposes. The password policy of this database for access from the localhost must be either trust or password. All other password policies can be whatever is applicable. For more information about setting the password policy, see Registering and Configuring Sun Cluster HA for PostgreSQL. For more information about the password policy, go to http://www.postgresql.org.

Restriction for the PostgreSQL smf Service Name in a Failover Zone

The PostgreSQL configuration in a failover zone uses the smf component of Sun Cluster HA for Solaris Containers. The registration of the Sun Cluster HA for PostgreSQL data service in a failover zone defines an smf service to control the PostgreSQL database. The name of this smf service is generated in this naming scheme: svc:/application/sczone-agents:resource-name. No other smf service with exactly this name can exist.

The associated smf manifest is automatically created during the registration process in this location and naming scheme: /var/svc/manifest/application/sczone-agents/resource-name.xml. No other manifest can coexist with this name.

Configuration Requirements

The configuration requirements in this section apply only to Sun Cluster HA for PostgreSQL.


Caution – Caution –

If your data service configuration does not conform to these requirements, the data service configuration might not be supported.


Dependencies Between Sun Cluster HA for PostgreSQL Components

The dependencies between the Sun Cluster HA for PostgreSQL components are described in the following table.

Table 2 Dependencies Between Sun Cluster HA for PostgreSQL Components

Component 

Dependency 

PostgreSQL resource in a Solaris 10 global zone,zone or in Solaris 9 

SUNW.HAStoragePlus This dependency is required only if the configuration uses a failover file system, of file systems in a zone.

SUNW.LogicalHostName

PostgreSQL resource in a Solaris 10 failover zone. 

Sun Cluster HA for the Solaris Container boot resource. 

SUNW.HAStoragePlus

SUNW.LogicalHostName — This dependency is required only if the zones boot resource does not manage the zone's IP address.

You set these dependencies, when you register and configure Sun Cluster HA for PostgreSQL. For more information, see Registering and Configuring Sun Cluster HA for PostgreSQL.

If more elaborate dependencies are required, see the r_properties(5) and rg_properties(5) man pages for further dependencies and affinities settings.

Parameter File for Sun Cluster HA for PostgreSQL

Sun Cluster HA for PostgreSQL requires a parameter file to pass configuration information to the data service. You must create a directory for this file. Because the directory must be available on each node that is to host the PostgreSQL database, place the directory on the shared storage. If Sun Cluster HA for PostgreSQL is configured for a failover zone, this file must be available in this zone. The parameter file is created automatically when the resource is registered.