Configuration of SSL Load Balancer for Universal Messaging


We have 3 UM instances, each UM instance has 4 realms - Persistent_small, Persistent_large, Non-persistent, admin which are in cluster.
We have a load balancer configured between UM cluster and application.
We are using TestHarness provided by SAG to test the UM.
We are able to do ‘nsp’ test successfully with loadbalancer [nsp://:].
But when we do ‘nsps’ test we are facing issues. We are unable to publish message to UM via LoadBalancer using TestHarness.
We are in version 9.10
Are we missing anything. Do we need to do any settings at our end.
LoadBalancer and Error snapshot is also attached.


Your setup is not clear to me from your description, for me, an instance and a realm is the same, a cluster consistes of different realms. Inside a cluster the nodes should be identically, having different configured nodes/reals inside a cluster does not work imho.

Why do you have a loadbalancer between applications and the UM server? The connecitvity should be done by the client libraries as well as the automatic failover.


Instances here refers to machines. We have UM installed on 3 machines [ 3 diff IP ]. Each UM machine has 4 realms created - say A, B, C, D which are in cluster.
We have identical nodes inside cluster.
Cluster A - is reserved to take Persistant_small messages…Cluster B - is for persistent_large and so on.

Application will send message to load balancer and load balancer will take responsibility of forwarding it to UM.


How do the application connect to Universal Messaging and why do you use a loadbalancer in between?

The message says it tries to create a native broker, is the test ahrness you are using aware of UM or does it try to do broker specific things?