Background: This is a multi-node Qlik Sense environment. There are total 6 nodes: Central node, Consumer Node 1, Consumer Node 2, Consumer Node 3, Consumer Node 3A, Scheduler Node. There are load balancing rules for the consumption with stream based, where Central node is for Proxy use, Node 1 and 2 are only for consumption, Node 3 and 3A are for consumption and scheduler, Scheduler node is for scheduling. The consumption and scheduling are landing on the correct node.
Scenario of issue hit: The client has implemented ODAG and encounter the issue, which the app generation is using random node instead of following the load balancing rule. Looking in Qlik Help portal, the reload will follow the rule set on the template app. They have the selection app in stream A (lb to Node 2), and template app in stream B (lb to Node3A). When they generate the app, the relevant ODAG service log is created in random node (eg. Node2). Furthermore, the log has indicated the app is open in random node (eg. Node 3.) By right it should follow the load balancing rule, to open and reload the app in Node 3A.
Looking forward for any advise as randomness of the reload will cause usage issue to the servers.
