Transaction Reaper Timeout causes all nodes to stop responding on EAP 6.
Issue
Each env is a set of identical JBoss deployments (not clustered) that use remote EJBs to share load via the JBoss EJBClient.
At roughly 08:38, the transaction timeout is reached on the 01 server and a number of transactions are reaped. All servers appear to be using the 01 server for their remote EJB call and so all of their transactions fail. I'm told all servers become unusable and must be restarted.
I've attached the server logs along with the configuration used. After initial analysis, I've suggested the server ordering for EJBClient setup be changed so that the servers would prefer to talk to themselves first (previously, the order was 01, 02, 03, 04,...).
Environment
- Red Hat JBoss Enterprise Application Platform
- 6.x
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.