Skip to main content

Clustering Core on a server with an External IP Address

Overview:
Essentials 4.6 and Core 3.0 offer clustering support out of the box, however if you are using a server with an external IP address (for example on an AWS VM) you need to take some extra steps to whitelist the IP address on all the nodes in the cluster.  By default only link local IP addresses are allowed, IE:
  • 10.0.0.0/8
  • 192.168.0.0/16
  • 172.16.0.0/12
If you're not sure if you need to do this, have a look at the DocumentStore.log file for entries like this:
  
[2016-11-22 13:06:01,072][INFO ][13  ][DocumentStore][Geocortex.Platform.Roles.DocumentStore.Watcher.BasicProxy.OnGetContext] Rejected DocumentStore request http://96.50.10.123:19201/ from 96.50.10.124

If you see these alerts, you should apply the fix below.  Another symptom you may see is the Java.exe process, or the w3wp.exe process using 100% of the CPU on the server.  

Solution:
  1. Follow the installation guide to set up your first node, and any additional nodes.  Make sure they are all members of the same cluster.
  2. Stop the Geocortex Core service on all the nodes
  3. Edit the DocumentStore role.config file on the first node (here by default: C:\Program Files\Latitude Geographics\Geocortex Core\NSRoot\Geocortex\Core\Roles\DocumentStore\3.0.3.1485\role.config) 
Note: The version number in the Document Store folder may be slightly different depending on the version of Essentials and Core you have installed
 
Note: It's good practice to make a backup of a config file before you edit it, however make sure that the folder does not contain more than one .config file, or multiple versions of the role will be loaded.  For example role - Copy.config is not OK, but role.config.bak is fine
  1. Add the following to the bottom of the file (just before the </StartupSettings> tag:
<!-- Enables a whitelist of hosts to accept document-store requests from -->
<StartupSetting Key="WhiteList" Value="96.50.10.123;96.50.10.124;96.50.10.125" />

The value is a semicolon separated list of the IP addresses of all the nodes in your cluster.
  1. Save the file
  2. Start the Geocortex Core service on all the nodes in your cluster
  3. Once all the roles have started on all the nodes, your cluster should be communicating properly.  

 

Related articles

Comments

0 comments

Article is closed for comments.