[ https://issues.jboss.org/browse/MODCLUSTER-397?page=com.atlassian.jira.pl... ] Aaron Ogburn commented on MODCLUSTER-397: ----------------------------------------- It looks like this is in part related to the fix from MODCLUSTER-268. Looks like the find_node_context_host check there causes us to not find a node for the route. I tested with MODCLUSTER-268 backed out. That improves the situation since we can then see that the route is being used from find_session_route's debug message. But the request 503s because another find_node_context_host check doesn't find a node for the request again. Some extra debug logging helps clarify the issue: [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(2951): proxy_cluster_trans for 0 (null) (null) uri: / args: (null) unparsed_uri: / [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(2772): cluster: balancer://mycluster Found value Value.4e6189af-0502-3305-8ff3-fad7fee8b516 for stickysession JSESSIONID|jsessionid [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(1943): find_node_context_host: / node: 1 vhost: 1 context: /foo [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(2772): cluster: balancer://mycluster Found value Value.4e6189af-0502-3305-8ff3-fad7fee8b516 for stickysession JSESSIONID|jsessionid [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(1943): find_node_context_host: / node: 1 vhost: 1 context: /foo [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(1943): find_node_context_host: / node: 1 vhost: 1 context: /foo [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(3015): proxy_cluster_trans DECLINED (null) uri: (null) unparsed_uri: / [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(3040): proxy_cluster_canon url: //mycluster/foo/ [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(3480): proxy_cluster_pre_request: url balancer://mycluster/foo/ [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(3217): cluster:No route found [Mon Mar 31 18:03:46 2014] [debug] mod_proxy_cluster.c(2144): proxy: Entering byrequests for CLUSTER (balancer://mycluster) failoverdomain:0 So find_node_context_host is checking for the requested / context and not the /foo context that we proxy the request to. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira