Basic Cluster Installation (single machine)

It is possible to run 2 or more instances of XpoLog on a single machine as a cluster similar to the usage of multiple machines.

The procedure is similar with a few mandatory additional steps as listed below:

Installation

Follow the same steps as described in the cluster installation guide with the following changes during installation:

1. During installation set the installation directory name to be XpoLog_Center_UI (for the UI node) and XpoLog_Center_PROCESSOR (for the PROCESSOR node). At first step you will not be able to run both instances as there will be a port a collision since both instance use the same ports by defaults.

2. Go to PROCESSOR_INSTALL_DIR/ServletContainer/conf/ and rename the file server.xml to be server.xml.orig and the file server.xml.processor to server.xml (this will change the ports of the processor node so that both instances will be able to run simultaneously on the same machine).

3. Start both instances. Processor can be accessed via http://<server-name>:30304 / https://<server-name>:30444 and UI can be accessed on the default ports via  http://<server-name>:30303 / https://<server-name>:30443

4. Please continue with the configuration section at cluster installation guide.

5. If you plan to install more instances make sure all ports are changed in the INSTALL_DIR/ServletContainer/conf/server.xml as each instance must use different ports.

• Note: the HTTP/S ports of all nodes can be changed in the settings section.

Windows Specifics

If you are running multiple nodes on a Windows machine then it is mandatory to give a unique name to each node during installation.
During installation ensure you specify a different installation directory and a different service name to each instance. For example: XpoLogCenterProcessor and XpoLogCenterUI

At the end of the process, you should have a installation directory and a unique service name per each installed instance.

Note: It is also possible to manually remove/create services but recommended to do it directly in the installation wizard:

• Remove service: sc delete [SERVICE_NAME] 

• Create service: sc create [SERVICE_NAME] binpath= "FULL_PATH_TO_INSTALL_DIR\XpoLog.exe -zglaxservice XpoLogCenter" start= auto DisplayName= "[SERVICE_DISPLAY_NAME]"
  For example: sc create XpoLogProcessor binpath= "C:\Program Files\xpologcenter\XpoLog.exe -zglaxservice XpoLogCenter" start= auto DisplayName= "XpoLogProcessor"
  
  

It is recommended to set a service account on each of the services for an optimized connectivity to machines across the network - Windows Post Installation.

Linux Specifics

If you are running multiple nodes on a Linux machine then it is recommended to allocate a specific range of CPU cores to be used by the Processor and a specific range for the UI.

For example, if there are 12 CPU cores available on the machine that the cluster is running on, the recommended configuration should be as follows:

1. Go to PROCESSOR_INSTALL_DIR and edit the file runXpoLog.prop:

   1. For Linux, specify a CPUs range, for example 0-3 for the first 4 CPUs or 4-7 for the second 4 CPUs
      For SunOS, specify a space separated list of processor IDs, for example 0 1 2 3 for the first 4 CPUs or 4 5 6 7 for the second 4 CPUs

      un-comment the  the line #cpus= with the allocated range. For example cpus=0-7

   2. Save and restart the PROCESSOR_NODE and you should see a message after restart indicating that the configured range has be applied.

2. Go to UI_INSTALL_DIR and edit the file runXpoLog.prop:

   1. For Linux, specify a CPUs range, for example 0-3 for the first 4 CPUs or 4-7 for the second 4 CPUs
      For SunOS, specify a space separated list of processor IDs, for example 0 1 2 3 for the first 4 CPUs or 4 5 6 7 for the second 4 CPUs

      un-comment the  the line #cpus= with the allocated range. For example cpus=8-11

   2. Save and restart the UI_NODE and you should see a message after restart indicating that the configured range has be applied.