I have JBoss in a 2-node cluster. After a few days of uptime, I see
the following message show up on the logs.
Once this message starts to appear, the requests start favoring the
node that does not see the messages, but some requests still make it
over to the server instance that is having the issue.
ERROR [ProtocolStack] no down protocol available !
ERROR [org.jgroups.stack.ProtocolStack] no down protocol available !
The JBoss cluster seems to be working fine for a while as I don't see any issue in the cluster communication. Eventually however, both nodes require a restart, as it leads to cluster members failing to communicate. The message keeps being logged in the files every other second, filling up the files. There is no full stack trace. Just the message as shown.
We have an application that uses the Flex framework and communicates with JBoss to get data from a MySQL database.
The JBoss cache is communicating over UDP. We are looking into making a change soon to communicate over TCP instead.
JBoss: 4.0.5GA built on October 16, 2006
JDK: 1.5.0_09-b03 (Sun)
OS: Linux Red Hat (2.6.9-55.0.2.ELsmp)
Hardware: Intel Quad-core
2 physically identical machines each with one node, and running 1.5GB of heap.
Thank you in advance for any pointers.