<%NUMBERING1%>.<%NUMBERING2%>.<%NUMBERING3%> PRTG Manual: Failover Cluster Step by Step

This section guides you through a step-by-step process to set up a failover cluster. Follow these instructions carefully to successfully integrate two or more PRTG installations into one failover cluster.

icon-i-round-redBefore getting started, make sure that you consider the information in section Failover Cluster Configuration.

icon-prtg-on-demandThis feature is not available in PRTG hosted by Paessler.

Step 1: Install Core Servers

First, you need to set up a single failover cluster consisting of two PRTG core server installations, each running on an individual server. Use your license key twice to install the PRTG core server on two different computers.

If you already run an installation of PRTG, this is your future Master Node. In this case, install a second core server on another computer. Make sure you use the same license key for both installations.

Before you start to set up a cluster, make sure you run exactly the same PRTG version (build number) on all (future) nodes. Install updates for existing installations, if necessary.

icon-i-roundOnce the cluster is established, any updates you install on one node are automatically deployed to all other cluster nodes.

icon-book-arrowsFor details about the installation process, see section Install a PRTG Core Server.

Step 2: Configure Master Node

Decide which of your PRTG core server installations is your future Master Node. If you already have an installation of PRTG in your network running for some time, this should be your master, so your existing monitoring configuration is kept.

On the Master Node, from the Windows start menu, open the PRTG Administration Tool. On the Cluster tab, click the following button:

Create a PRTG Cluster...

  • Click this button to start creating a cluster. The current PRTG core server is then the Master Node of your cluster.
  • Click Yes to confirm the conversion of this installation into a cluster master node.
Converting an Installation into a Cluster Master Node

Converting an Installation into a Cluster Master Node

  • A new dialog box appears.
PRTG Administrator: Creating a Cluster Master

PRTG Administrator: Creating a Cluster Master

  • Enter a Cluster Port. This is the port on which the internal communication between the different cluster nodes is sent. Make sure connections between cluster nodes are possible on the selected port.
  • Enter or paste a Cluster Access Key. This is a unique access key. All nodes in a cluster have to be configured with the same cluster access key to join the cluster. Connection attempts with the wrong access key are rejected.
  • We recommend that you use the default value.
  • Save the Cluster Access Key so you have it at hand when configuring your Failover Nodes.
  • After confirming your settings, you are asked to restart Windows services. Please do so for your changes to take effect.

Step 3: Configure Failover Node

On the server that is your Failover Node, open the PRTG Administration Tool. On the Cluster tab, click the following button:

Join a PRTG Cluster...

  • Click this button to add this installation to an existing cluster that already has a Master Node. The current PRTG core server is then a Failover Node in the cluster.
  • This button is also available if you are currently running your PRTG installation in Cluster Mode: Master Node. This option then changes your master node to a failover node.
  • Click Yes to confirm the conversion of this installation into a failover node.
Converting an Installation into a Cluster Failover Node

Converting an Installation into a Cluster Failover Node

  • A dialog box appears.
PRTG Server Administrator

PRTG Server Administrator

  • Enter a Master IP address/DNS name for your cluster. It must be reachable from the machine running the failover node.
  • Enter the other settings as defined in the settings of your Master Node. Make sure you use the same settings on all nodes in your cluster.
  • Enter a Cluster Port. This is the port on which the internal communication between the different cluster nodes is sent. Make sure connections between cluster nodes are possible on the selected port.
  • Enter or paste a Cluster Access Key. This is a unique access key. All nodes in a cluster have to be configured with the same cluster access key to join the cluster. Connection attempts with the wrong access key are rejected.

After confirming your settings, you are asked to restart Windows services. Please do so for your changes to take effect.

Step 4: Confirm Failover Node

Now you need to confirm the new failover node by setting it to Active in the master node's settings.

In a browser window, log in to the PRTG web interface of the Master Node.

In the System Administration—Cluster settings, you see your Master Node in the first line of the cluster list and your Failover Node below.

icon-i-round-redIf you use remote probes outside your local network, for each core server in the cluster, use a Domain Name System (DNS) name or IP address that the probes can reach from the outside. Enter the entries in the Cluster Node Setup table accordingly (see below). The addresses must be valid for both cluster nodes to reach each other and for remote probes to individually reach all cluster nodes. Remote probes outside your LAN cannot reach private IP addresses or DNS names.

System Administration: Cluster Node Setup

System Administration: Cluster Node Setup

For the Failover Node, set the radio button for Node State to Active and Save the changes. The nodes now connect and exchange configuration data. This may take a few minutes.

Step 5: Check Cluster Connection

In two browser windows, log in to the PRTG web interfaces of both of your PRTG installations. Open the cluster status page in both windows by clicking on the narrow cluster information bars at the top of each window. You should see a cluster status with your two nodes in a Connected state after a few minutes.

PRTG Cluster Status

PRTG Cluster Status

Step 6: Troubleshooting

If your nodes cannot connect, see

In the latest entries of these logs, you can see messages about any errors that might have occurred. These give you hints on where to find a solution.

If you have connection issues with the two cluster nodes, make sure that no software or hardware firewall is blocking communication on the cluster port defined during cluster setup. Communication between the nodes must be possible in both directions for the cluster to work properly.

Step 7: Move Sensors to the Cluster Probe Now

You have successfully set up your failover cluster. All devices that you create or move under the Cluster Probe are monitored by both servers.

icon-i-round-redTo monitor your existing configuration via all cluster nodes, on your master node, move your groups, devices, and sensors from the local probe to the cluster probe. Objects, including their settings, are then automatically transferred to all cluster nodes.

Step 8: Move Custom Content to Failover Nodes

On startup of the cluster master node, maps and custom lookups are automatically transmitted to the failover nodes. While changes to maps are automatically synchronized, you have to manually (re)load lookups on all nodes. Other custom content has to be manually copied from the according folders on the master node to the same folders on the failover nodes:

Add More Failover Nodes

If you want to add an additional failover node to your cluster, you need an additional license key for two and three failover nodes, and two additional license keys to run four failover nodes.

icon-i-roundIn a cluster, only core servers running on the same license type can be combined. To add another failover node to your cluster, set up a new PRTG core server installation on a new server, using an additional license key. Then proceed with step 3 and following. Use a second license key to set up both your second and third failover node, and use a third license key to set up your fourth failover node. Each failover cluster is technically limited to five cluster nodes: as a maximum, you can run one master node and four failover nodes in one cluster.

icon-i-round-redStay below 2,500 sensors per cluster for best performance in a single failover. Clusters with more than 5,000 sensors are not supported. For each additional failover node, divide the number of sensors by two.

More

Knowledge Base: My PRTG Cluster is messed up. How can I start over?

Paessler Website: How to connect PRTG through a firewall in 4 steps

Advanced Topics

Keywords: