|Applies To||RSA Product Set: Archer|
RSA Product/Service Type: Advanced Workflow
RSA Version/Condition: 6.0
|Issue||Suppose you have a load balanced configuration with multiple web nodes. Each web host has an instance of IIS hosting Archer and the Wildfly application server hosting Workpoint. The communication between Archer in IIS and Wildfly on any given web host is over localhost.|
Suppose one instance of Wildfly goes down. Because the load balancer is against IIS, it will not be able to detect that Wildfly is down and will continue to route requests to the affected web host.
For end users executing workflow actions, this means unpredictable behavior depending on whether the web host with Wildfly down is hit by the load balancer or not. For instance, they could get a failure the first time and then retry the action and have it succeed because it gets routed to a different node. Then it could fail again afterwards.
If data feeds are running for records where workflow is enabled, then the creation/update of records will be unpredictable - it will work for some records and fail for others, depending on whether the load balancer routes to the web host with Wildfly down. It may be difficult for the end user to figure out which records did not persisted properly and to clean things up accordingly.
|Resolution||This issue has been resolved in Archer 6.0.1.|