Hello,

I configured the cluster with two nodes.

However, we note that during shutdown of the master server and slave activation, many failures of connectivity are generated for the devices being monitored by the remote probes.

Is this normal?

Do we need to configure dependencies?

Thank you.

Article Comments

Hello Sandro,
thank you for your inquiry.

I will need you to be more specific about the errors. What precise errors/messages do you see/get? We also need to know what sort of sensor is generating these errors. For instance, it is expected that the cluster health sensors will go down if one of the cluster nodes is unavailable.

Best Regards,
Luciano Lingnau [Paessler Support]


Oct, 2018 - Permalink

Hello,

We realized this scenario when we enabled the PRTG to open the ticket on an external ticket system.

While disabling the MASTER node, alarms are generated for all devices being monitored by remote probes. Connection loss errors with network devices.

According to what we observe the alarms become unknown during the deactivation of the MASTER nodes and then they return to their normal state of monitoring. But during this process, several tickets are opened in our external ticket system with alarms of communication failure of the devices monitored by the remote probes.

We have a Java API that was querying the PRTG alarms API and all the new alarms created, are opened in our external ticket system.

The cluster sensor alarms, this makes sense to generate alarms.

Thank you.


Oct, 2018 - Permalink

Hello Sandro,
thank you for your reply.

While disabling the MASTER node, alarms are generated for all devices being monitored by remote probes. Connection loss errors with network devices. - According to what we observe the alarms become unknown during the deactivation of the MASTER nodes and then they return to their normal state of monitoring. But during this process, several tickets are opened in our external ticket system with alarms of communication failure of the devices monitored by the remote probes.

Please share a couple of examples of the error messages (or notifications) that you're receiving. Because remote probes should definitively not produce alerts when the master node is disabled or restarted, since the remote probe should continue polling normally even with the Master node unavailable.

Best Regards,
Luciano Lingnau [Paessler Support]


Oct, 2018 - Permalink