[mod_cluster-dev] Session temporarily unavailable

Bela Ban bban at redhat.com
Wed Jul 22 11:28:05 EDT 2009


OK, here's what happens:

    * Nodes node1 and node2
    * My session is on node1
    * I kill -9 node1 and access the webapp
    * "Service not available" for ca. 10 secs
    * After ca 10 seconds, the session is failed over to node2
    * However, mod-cluster-manager still lists node1 (forever) !
    * I guess the AJP connection is closed by mod-cluster due to the
      ping (every 10 secs ?)

Can we add functionality to remove a worker from httpd by the singleton
telling it that a worker crashed ?

-- 
Bela Ban
Lead JGroups / Clustering Team
JBoss


More information about the mod_cluster-dev mailing list