Siebel Server Installation Guide for UNIX > Clustering Your Siebel Deployment for Failover >

Installing Siebel eBusiness Applications with IBM HACMP

High Availability Cluster Multi-Processing (HACMP) Cluster Server lets you monitor systems and application services, and restart services on a different system when hardware or software fails. A HACMP cluster consists of multiple systems connected in various combinations to a shared storage device.

All systems within a cluster are connected by redundant private networks over which they communicate by heartbeats, or signals sent periodically from one system to another.

Applications can be configured to run on specific nodes within the cluster. Storage is configured to provide access to shared data for nodes hosting the application.

For detailed instructions on how to install and configure a HACMP Cluster, refer to the High Availability Cluster Multi-Processing for AIX - Enhanced Scalability Installation and Administration Guide (HACMP Guide). For examples on how to deploy a Siebel application in a HACMP environment, refer to your IBM documentation.

NOTE:  The cluster services should be started on all nodes and all application resource groups (RG's) have to be online before any Siebel software installation.

For detailed Siebel software installation steps, see Installing the Siebel Gateway and Installing the Siebel Server.

Configuring HACMP Cluster Resource Groups

A Resource Group (RG) is a filesystem located on an external disk (SSA disk) that can be mounted on nodes in a cluster. It also has an associated IP address (known as Service Address) which moves with the RG to the node where it is mounted. When an RG is initially created, the filesystem in the RG is empty. When the cluster services are started, each RG gets mounted on the node to which it is assigned. At this point, the filesystem and the service address become available for software installation. You need to determine the contents of each RG; for example, each RG can contain only one module like the Siebel Server or the IHS Web Server. Alternatively, each RG can contain multiple software such as the Siebel Gateway Name Server, Siebel Server, and the IHS Web Server.

NOTE:  When the RG moves, all the software in the RG moves with it.

Selecting User Groups in a HACMP Clustered Environment

There are no UNIX User group restrictions within HACMP. It does not matter who owns what on each node. However, you must use root to install HACMP on each node, and therefore root owns HACMP. You can install the software in the RG's using any UNIX user. However, when an RG moves to another node, you must make sure that the UNIX user used to install and run the software on the original node has the same UID and GID on the second node; otherwise you will encounter file access permission problems within the RG.

Configuring Login Profiles in a HACMP Clustered Environment

The $HOME directories of each user is local to each node and not in the RG's. This means the .profile for each user is local. Nothing specific to Siebel software is set in the users .profile. This is because in a dynamic HACMP environment, any RG can be connected to any node and each RG may have different environment requirements. The HAMCP scripts should set up the appropriate environment before launching the application. Since the siebenv.sh script is part of the installed Siebel application, it should be in the RG's.

Installing Additional Language Packs in a HACMP Clustered Environment

Before installing additional language packs or patches, you must make sure all RGs containing Siebel products are moved back to their primary nodes.

Synchronize System Information in a HACMP Clustered Environment

Files like /etc/hosts, /etc/resol.conf, /etc/services and others are files that the user needs to synchronize manually as part of the HACMP installation and configuration. HACMP relies on these files to be accurate; otherwise, you may encounter unexpected behavior from the cluster.

Assign Resources for Siebel Servers in a HACMP Clustered Environment

Assign appropriate resources, including User IDs virtual IP addresses, for each of the logical Siebel Servers. Validate that the directory in $SIEBEL_ROOT is owned by the corresponding Siebel User ID.

Use Clear Naming Conventions in a HACMP Clustered Environment

If you will be operating a heterogeneous server environment, use UNIX naming conventions for servers; use names that are no longer than 12 characters and do not contain spaces or special characters.

Use machine names for the Siebel Enterprises, Siebel Gateway and Siebel Servers that are cluster-independent and node-independent.

Clustering the Siebel File System in a HACMP Clustered Environment

If the Siebel File System is on an NFS server, refer to the IBM HACMP guide for clustering NFS filesystems.

Clustering the Database Server in a HACMP Clustered Environment

If you intend to cluster your Siebel Database Server, follow the procedures provided by your database vendor.

Cluster Event Scripts

Once you have installed the Siebel software and have started the cluster, you can use custom HACMP or ES Cluster Event scripts to start, stop, and monitor the Siebel components. For more information, see your IBM documentation.


 Siebel Server Installation Guide for UNIX 
 Published: 24 June 2003