What is this?

This knowledgebase contains questions and answers about PRTG Network Monitor and network monitoring in general.

Learn more

PRTG Network Monitor

Intuitive to Use. Easy to manage.
More than 500,000 users rely on Paessler PRTG every day. Find out how you can reduce cost, increase QoS and ease planning, as well.

Free Download

Top Tags


View all Tags

Pacemaker/Corosync cluster monitoring

Votes:

0

Hi guys,

We are runing a 2 node Pacemaker/Corosync HA cluster and as such the services (cluster resources) are only started on 1 node which cause PRTG to report on the down service on the passive node. Is there a way to configure PRTG to monitor the cluster but only report on the started services on the active node an not report on the stopped services on the passive node. In an ideal world it would be perfect to flag a down service if it's down on all nodes but fine as long as 1 node has the service running.

We are using Debian Jessie with the lastes Pacemaker/Corosync and standard LSB (SysV) daemons.

Thanks, Jason

cluster linux pacemaker

Created on Sep 11, 2017 7:33:53 AM



1 Reply

Votes:

0

Hello Jason,
thank you for your KB-Post.

There's no easy approach for this. PRTG doesn't implement any sort of "cluster-awareness", but some things are possible depending mainly of the sensor types that you're using.

First of all. Does the cluster have a DNS or IP that always points to the active node? If this exists, you should be able to use this address for monitoring, so that the queries would always be done against the "active" node. The problem is, you may not be notified if the failover is suddenly dead, so this alone doesn't solve the issue completely. You should still monitor the nodes in a way that allows you to see that they still 'exist'.

I'm not familiar with the "Pacemaker/Corosync HA cluster", but some clusters will implement API's (or SNMP MIB's) that report the status of all nodes. This usually requires some investigation and trial and error, but if such an API exists for your cluster service, is the most reliable way of knowing the cluster's status (all nodes), because the cluster already needs to "monitor itself" to be able to work.

Lastly, the PRTG implementation will strongly depend on the sensor types that you're using. Many sensor us "internal" error statuses for alerting, in most cases you won't be able to customize these. For some sensor types (that use lookups or limits for alerting), you may be able to customize their behavior/operation.

If you're able to share more concrete details about how you're monitoring this (what sensors) I may be able to provide further input/recommendations.

Best Regards,
Luciano Lingnau [Paessler Support]

Created on Sep 13, 2017 7:59:17 AM by  Luciano Lingnau [Paessler]




Disclaimer: The information in the Paessler Knowledge Base comes without warranty of any kind. Use at your own risk. Before applying any instructions please exercise proper system administrator housekeeping. You must make sure that a proper backup of all your data is available.