Could someone address (or point me to documentation that addresses) the
expectations I should have of the UserApp Cluster? If I don't know how it
SHOULD behave, it becomes tough for me to assess whether or not my failover
scenarios are showing me "Working As Designed" or "Issue to address"
My cluster is working, through a load balancer.
I'm looking to answers for things like the following:
1. User Logs in and is taken to the "main" page. Before strking another
key, the node to which the user is attached goes down.
The user then chooses, say, the Requests and Approvals tab.
Should I see:
a) seamless execution?
b) reversion to the UserApp login page, with a session timeout error?
c) 404 error from the JBOSS/Tomcat server on the node where the userapp has
been taken down?
d) Repeated browser refreshes, bringing the eventual, proper execution,
after the failover timeouts of my load-balancer have been satisfied?
I've seen all of the above.
As it stands, my load balancer works...perhaps with too long of a delay in
The database works, meaning that if I create a workflow on one node, then
fail it, I can access the workflow on the second node.
The themes are shared and matching. All seems well, except I was expecting
some kind of instantaneous, transparent session maintenance that would keep
the whole cluster failover hidden to the end user.
Am I missing a UserApp setup? A load balancer best practice?
Again, if there is a white paper with expected behaviors and best practices,
that would be great.