JBoss EAP unavailable after Apache restarts when using mod_cluster
Issue
- User needs to restart Apache (but JBoss is untouched)
- Once Apache is started, JBoss is still not accessible for a short time (even though it's up)
-
After Apache is restarted, JBoss logs an error about a STATUS MCMP:
ERROR [org.jboss.modcluster.mcmp.impl.DefaultMCMPHandler] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) Error [MEM: MEM: Can't read node: {4}] sending command STATUS to proxy 127.0.0.1:6666, configuration will be reset
And Apache logs a message like:
[warn] manager_handler STATUS error: MEM: Can't read node
-
mod_cluster drops JBoss nodes when apache is reloaded
- I am using apache with mod_cluster as a load balancer in front of a JBoss Cluster.When I issue apache a reload command, the JBoss nodes are unregistered and it takes approximately 10 seconds for the JBoss nodes to register again.This behaviour would be ok for an apache restart, but for a reload I was expecting no downtime. Is there any solution ?
- httpd restart causes 503 for ~10s with mod_cluster.
- mod_cluster not discovering nodes after restart of httpd.What could be the cause of this?
Environment
- Red Hat JBoss Enterprise Application Platform (EAP) all versions
- Apache or JBoss Enterprise Web Server (EWS) all versions (using mod_cluster)
Subscriber exclusive content
A Red Hat subscription provides unlimited access to our knowledgebase, tools, and much more.