2.5. Preparing for Installation on Windows

2.5.1. Setting up the JDK

The JBoss ON server requires Java 6 or Java 7 JDK.
If necessary, configure Windows to use the appropriate Java version.
  1. Download and install the appropriate version of Java, if necessary.
  2. Set the JAVA_HOME environment variable to the installation directory.
    C:\>set JAVA_HOME=C:\Program Files (x86)\Java
    The default Java service wrapper included with JBoss ON requires a 32-bit JVM, so the Java preference set for the server must be a 32-bit JDK. The JBoss ON server must use a 32-bit JVM even on 64-bit systems.
  3. Set the system to use the bin directory of the correct version of the JDK.
    C:>set path C:\Program Files\Java\jdk1.6.0_29\bin
  4. Set the classpath to the lib directory of the correct version of the JRE distribution.
    C:>set classpath C:\Program Files (x86)\Java\jre1.6.0_03\lib

2.5.2. Configuring the JVM to Run as a Service

JBoss ON includes Tanuki Software's Java service wrapper so that the JBoss ON server can be configured to run as a Windows service. That default Java service wrapper included with JBoss ON requires a 32-bit JVM, so the Java preference set for the server must be a 32-bit JDK.
RHQ_JAVA_HOME=C:\Program Files\Java\jdk1.6.0_29
The JBoss ON server must use a 32-bit JVM even on 64-bit systems.
Running the server or agent with a 32-bit JVM does not in any way affect how JBoss ON manages other resources which may run with a 64-bit JVM. JBoss ON can still manage those resources and those resources can still use the 64-bit Java libraries for their own processes.
To run the server in a 64-bit JVM, an alternative Java wrapper service must be obtained independently from Tanuki Software or from another Java service vendor.

2.5.3. Configuring NTP

Synchronize machine clocks. All JBoss ON servers and agents must have synchronized clocks. Clock variations cause issues in availability reporting, metric measurements, graphing, and even identifying and importing resources into inventory. The Network Time Protocol project, http://www.ntp.org/, has information on installing and configuring NTP to ensure your clocks are synchronized.

2.5.4. Configuring DNS

Both forward and reverse DNS mapping entries must be present for all systems for which host servers, storage nodes, and agents.
Every IP address must have a corresponding entry in the DNS server or must be explicitly defined in every /etc/hosts file for each system which is managed by JBoss ON or hosts a server or storage node..

2.5.5. Configuring Ports

Configure the firewall to allow communication over the server, agent, and storage node ports. If the required ports are blocked, then individual components will be unable to communicate with each other.
Using the default configuration, JBoss ON uses the ports listed in Table 2.3, “Default JBoss ON Ports”.

Table 2.3. Default JBoss ON Ports

Port Purpose
7080 Standard HTTP port for server-client communication
7443 HTTPS port for secure server-client communication
16163 For agent communication from the server
9142 For storage cluster communication
7299 For storage node JMX communication
7100 For the storage node gossip (node-to-node) communication

2.5.6. Selecting Path Names

Make sure that the complete path name for the server installation directory is relatively short. Path names longer than 19 characters can cause problems with executing some server tasks. Use a location such as C:\jon rather than C:\Documents and Settings\myusername\jon-server.
Also be careful when using the extract all command. Expanding the archive automatically creates a directory called jon-server-VER.RELEASE/, which is about 20 characters. Using extract all, instead of specifying the directory to which to extract the archive, can double the directory name by extracting to the archive name and then to a subdirectory — for example, C:\example\jon-server-3.2.0.GA\jon-server-3.2.0.GA. Using other tools may install it to a downloads directory such as C:\Users\Administrator\Downloads.
It is recommended that you extract the archive to a short, top-level directory such as C:\jon, which creates an installation directory of C:\jon\jon-server-3.2.0.GA.
Windows' handling of file and path names is covered in the "Naming Files, Paths, and Namespaces" in the Windows Data Access and Storage API.

2.5.7. Utilities to Use with JBoss ON

The only utilities used to manage the JBoss ON server are a ZIP utility to install the binaries and, possibly, a text editor to view and edit configuration files.
The recommended ZIP utility is WinZip. Examples in this guide usually use the Windows command prompt, so, optionally, install the WinZip CLI utility add-on. WinZip downloads are available at http://www.winzip.com.

2.5.8. Configuring Internet Explorer

Some Internet Explorer settings can prevent the JBoss ON login page from loading properly. By default, Internet Explorer is in stealth mode, which disables some JavaScript access for websites. To allow the login page to load, add the IP address of the JBoss ON server to the whitelist for Internet Explorer.
  1. In Internet Explorer, click the gear icon in the upper right corner and select Internet options.
  2. Open the Security tab, and select the Local intranet icon.
  3. Click the Sites button.
  4. Click the Advanced button at the bottom of the pop-up window.
  5. Enter the JBoss ON server hostname or IP address in the Add this webiste to the zone: field, and click the Add.
  6. Close out the options windows.

2.5.9. Planning the Storage Nodes

There must be at least one backend storage database to store metrics data. This storage node is installed using the rhqctl script (the same as the server). The metrics storage database works as a cluster, so it is possible to have multiple nodes and to add and remove nodes as necessary.
There are some guidelines for planning the storage nodes:
  • The storage node and server do not have to be located on the same machine.
  • Installing a storage node does not require installing a server.
  • At least one storage node must be installed before the server. (If the rhqctl script is run with just the install command, then it automatically installs a storage node first, then the server, then the agent.)
  • Multiple storage nodes can be installed before installing the server. There are several benefits to installing multiple nodes:
    • For upgraded environments, it speeds data migration.
    • It can minimize the administrative and resource requirements that are incurred by deploying nodes after the server is running.