This article applies as of PRTG 22
If a primary master node fails permanently
If a primary master node fails, a failover node becomes a failover master node until the primary master node is back in the cluster again. Sometimes a server crashes and a primary master node installation cannot be restored.
We recommend that you recover the primary master node from a proper backup.
However, if a primary master node cannot be recovered, there is a possibility to convert a failover master node so that it becomes the new primary master node of a cluster.
How to convert a failover master node
To convert a (temporary) failover master node to be the new primary master node of the cluster, follow the steps below.
Note: We only recommend this procedure in cases where you cannot restore the primary master node of your cluster. In this article, we assume there are at least two running cluster nodes left in your cluster: the current failover master node and a failover node.
Step 1: Revert the primary master node to standalone mode
On the failing primary master node (if still accessible):
- Open the PRTG Administration Tool.
- Select the Service Start/Stop tab.
- Stop the PRTG core server service by clicking the Stop Service button.
- Select the Cluster tab.
- Click the Revert to Standalone button.
- Make sure you do not (re)start the PRTG core server and keep the server stopped permanently.
Standalone mode prevents the primary master node from rejoining the cluster (and overwriting the new cluster configuration) in case it should "accidentally" be started some time later.
Step 2: Stop the PRTG core server services on failover nodes
On all failover nodes:
- Open the PRTG Administration Tool.
- Select the Service Start/Stop tab.
- Stop the PRTG core server service by clicking the Stop Service button.
Step 3: Set the failover master node to be the new primary master node
On the current failover master node:
- Open the PRTG Administration Tool.
- Select the Cluster tab.
- Click the Create a Cluster button.
- When asked if you want to convert this PRTG installation to the master node of a cluster, confirm with Yes.
- Follow the procedure to create a new master node by confirming the cluster port and cluster access key. Leave all settings as they are.
- Write down the settings for cluster port, cluster access key, and the IP address of the machine that is running the master node, as you will need this information later.
- Close all windows by clicking the Ok/Save & Close buttons.
- When asked if you want to (re)start the PRTG core server service, confirm the restart.
This node is now your new primary master node. If you have made changes to the configuration on the failover master node, they will now be saved permanently and deployed to the cluster.
Step 4: Correct the existing cluster configuration
For the new primary master node you created in the previous step:
- Log in to the PRTG web interface and from the main menu bar, select Setup | System Administration | Cluster.
- In the Cluster Node Setup, copy the data (the Node Name) of the primary master node to line number 1 and clear the IP address if necessary. Node ID is updated automatically, check if it is correct.
- Delete the entries in all other lines (lines two to five).
Step 5: Rejoin failover nodes
On all failover nodes:
- Open the PRTG Administration Tool.
- Select the Cluster tab.
- Click the Join a Cluster button.
- Follow the procedure to create a new failover node and use the information you wrote down earlier to insert the IP address of the new primary master node, the cluster port, as well as the correct cluster access key.
- Close all windows by clicking the Ok/Save & Close buttons.
- When asked if you want to (re)start the PRTG core server service, confirm the restart.
This node is now a failover node. It will connect to the new primary master node and receive its configuration. Repeat the procedure in this step for each of your failover nodes.
Step 6: Clean up on the new primary master node
On your new primary master node, hardware-device specific sensors on the local probe device may not work anymore. For example, the sensor for the network card adapter may show a Network card not found message, because the hardware has changed. In this case, delete the affected sensors and add them again.
More
Add comment