[jboss-user] [JBossCache] - Re: Memory leak when using JbossCache and JOTM

[atijms]

"bstansberry at jboss.com" wrote : Does the reference chain from the JOTM object lead through a javax.transaction.Synchronization implementation?  For example, org.jboss.cache.interceptors.TxInterceptor$RemoteSynchronizationHandler (or TxInterceptor$LocalSynchronizationHandler)?
  | 

Yes it does. Just for clarity I'll post the whole reference chain anyway:


  | [sticky class] (jprofiler term.)
  | |
  | class com.sun.corba.se.impl.util.Utility
  | |
  | com.sun.corba.se.impl.util.IdentityHashtable
  | |
  | <class>[]
  | |
  | com.sun.corba.se.impl.util.IdentityHashTableEntry
  | |
  | org.objectweb.jotm.SubCoordinator
  | |
  | java.util.Vector
  | |
  | <class>[]
  | |
  | org.jboss.cache.interceptors.OrderedSynchronisationHandler
  | |
  | ...
  | |
  | org.jboss.cache.interceptors.TxInterceptor$LocalSynchronizationHandler
  | |
  | org.jboss.cache.TransactionEntry
  | |
  | ...
  | |
  | org.jboss.cache.marshall.JBCMethodCall
  | |
  | <class>[]
  | |
  | [the object I put into JBossCache]
  | 


anonymous wrote : 
  | 2) JOTM is holding a ref to the Synchronization and won't ever release it.  That would be a bug in JOTM.
  | 

This is what is currently happening, as I don't think "for a little while" means atleast a few hours ;)

I'm going to read the link you posted. Meanwhile I'm also reading through JOTM's sourcecode to see if it's anything trivial to fix (but I'm afraid it isn't).


View the original post : http://www.jboss.com/index.html?module=bb&op=viewtopic&p=4028072#4028072

Reply to the post : http://www.jboss.com/index.html?module=bb&op=posting&mode=reply&p=4028072