Running JBossAS 4.x, and JBossMQ on RH-Linux. Our application is a multi-process app, where processes interact over JMS topics and Queues.
I am seeing the following exceptions thrown from the JBoss process in some of our installations, very occasionally:
org.jboss.mq.SpyJMSException: Connection Failed; - nested throwable: (java.io.IOException: ping timeout.)
org.jboss.mq.SpyJMSException: Cannot authenticate user; - nested throwable: (java.net.NoRouteToHostException: No route to host)
I have seen other postings describing similar situations, but never describing a specific cause/solution. One posting stated that they could reproduce the "NoRouteToHostException" if they disconnected their server from the network. Why would connectivity to the network cause JMS to throw exceptions when supporting inter-process comm within the same host?
Does JBossMQ get impacted by the server-OS DHCP/DNS operation/configuration?
The "java.io.IOException: ping timeout" is described in http://wiki.jboss.org/wiki/Wiki.jsp?page=IGetIOExceptionPingTimeoutHowDoIFixIt
Here is says that large JMS messages and/or heavy paging could cause this. I have been able to create a ping timeout by changing the Linux date/time via the shell!
Any insights would be greatly appreciated.
This is a TCP/IP question about your network configuration and OS
(not a JMS question). You need to ask your network admin.
i.e. why is the host/ip address unreachable when you disconnect the cable,
most likely because it has to do a remote dns lookup to resolve its own ip address
from its own host name? :-)