We have 2 identical servers with WildFly-8.2.0.Final installed that we use to run our warehouse-management-system. At any time only 1 server is "active" while the other serves as backup.
Our application uses hornetq/jms for logging messages for statistics (1 topic and 2 subscribers, default configuration). After running fine for a few days hornetq starts throwing XAExceptions caused by the following error:
Caused by: HornetQConnectionTimedOutException[errorType=CONNECTION_TIMEDOUT message=HQ119014: Timed out waiting for response when sending packet 63] at org.hornetq.core.protocol.core.impl.ChannelImpl.sendBlocking(ChannelImpl.java:390) at org.hornetq.core.client.impl.ClientSessionImpl.setTransactionTimeout(ClientSessionImpl.java:1857)
It seems it can recover from this the first few times, but after a while the messaging system stops working altogether (e.g. no more messages get through). Switching to the other server is then the only option.
The strange thing is that only one of the servers is having this problem. The only difference is that the problematic server also runs a tomcat-process that houses the WMS for our Chinese partners.
Could it be that this extra load causes these issues ? Is there anything that we can do to avoid these problems?