NetXMS Support Forum

We are using Cluster object on small OpenVPN AS cluster with two nodes.
They have a floating IP switching between the nodes with UCARP detecting wich node should be primary.

Latly when doing apt upgrade for some reason primary detach the IP without secondary taking over so there is no master of the cluster. Probly some bug in UCARP setup, nothing todo with NetXMS..

But NetXMS does not detect that the floating IP is not active on any node, even not pingable..
First when old primary network is manually restarted alarm with "Cluster resource "X" moved from node A to node "
In Cluster log there is a event that floating/cluster IP is deleted too..
Some version back im pritty sure this situation caused this untouch netxms setup to trigger a "Cluster down" alarm.

Did anything change latly version of NetXMS for Cluster Objects?

In old NetXMS v2. days we didnt use this way of monitoring such cluster, but added a 3th node with the floating IP as PrimaryHostname.
But thats not possible anymore with the detection and merging of nodes with multi IPs to one Object..

Something in cluster operation seems to be broken. We will check the code.