[jboss-jira] [JBoss JIRA] Created: (EJBTHREE-923) Avoid unnecessary passivation/activation callbacks for beans cached on remote nodes
Brian Stansberry (JIRA)
jira-events at lists.jboss.org
Thu Mar 22 11:42:40 EDT 2007
Avoid unnecessary passivation/activation callbacks for beans cached on remote nodes
-----------------------------------------------------------------------------------
Key: EJBTHREE-923
URL: http://jira.jboss.com/jira/browse/EJBTHREE-923
Project: EJB 3.0
Issue Type: Task
Components: Clustering
Affects Versions: AS 4.2.0 CR1
Reporter: Brian Stansberry
Assigned To: Brian Stansberry
Fix For: AS 4.2.0 CR2
Talking here about backup copy of an SFSB context on a remote node; i.e. not the instance on the node where the client is actively using it. Such a context's bean instance, interceptors and XPCs are still marshalled from replication, as they don't need to be unmarshalled unless used
Two issues:
1) If the context has been passivated to disk, and another replication of the context occurs, the old copy of the context will be read from disk before the new copy overwrites it. That's standard JBC behavior. Currently we are invoking @PostActivate on this old copy, and then immediately replacing it in the cache with the new copy. This @PostActivate call is unnecessary and forces an unneeded unmarshalling of the old context.
2) If the replicated context is in memory, and the local eviction thread decides to passivate it, we are calling @PrePassivate on the bean. This is unnecessary when the bean is fully marshalled; a fully marshalled bean can't be holding any state that needs cleanup.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://jira.jboss.com/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira
More information about the jboss-jira
mailing list