[jboss-jira] [JBoss JIRA] Created: (JBAS-6733) Avoid re-reading expired sessions from distributed cache

[Brian Stansberry (JIRA)]

Avoid re-reading expired sessions from distributed cache
--------------------------------------------------------

                 Key: JBAS-6733
                 URL: https://jira.jboss.org/jira/browse/JBAS-6733
             Project: JBoss Application Server
          Issue Type: Sub-task
      Security Level: Public (Everyone can see)
          Components: Clustering, Web (Tomcat) service
            Reporter: Brian Stansberry
            Assignee: Brian Stansberry
            Priority: Minor
             Fix For: JBossAS-5.1.0.CR1


Add logic such that if we don't find a session locally, only check the distributed cache if we haven't previously handled this session id on this request. If we handled it previously but it's no longer local, that means it's been invalidated. If we request an invalidated session from the distributed cache, it will be missing from the local cache but may still exist on other nodes (i.e. if the invalidation hasn't replicated yet because we are running in a tx or with async repl). With buddy replication, asking the local cache for the session will cause the out-of-date session from the other nodes to be gravitated, thus resuscitating the session.

Previously releases had logic to deal with this, but only covered requests for the session that come before the code returns from ClusteredSessionValve. Such calls can happen after that, in particular StandardHostValve makes such a call.

Fix is to store info on invalidated sessions in a thread local, as was done previously, but to defer clearing the thread local until the Session.recycle method rather than doing it in the ClusteredSessionValve.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: https://jira.jboss.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira